2014-05-03 23:27:38 +01:00
|
|
|
/*
|
2017-06-30 08:22:17 +01:00
|
|
|
* This file is part of the MicroPython project, http://micropython.org/
|
2014-05-03 23:27:38 +01:00
|
|
|
*
|
|
|
|
* The MIT License (MIT)
|
|
|
|
*
|
|
|
|
* Copyright (c) 2013, 2014 Damien P. George
|
|
|
|
*
|
|
|
|
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
|
|
* of this software and associated documentation files (the "Software"), to deal
|
|
|
|
* in the Software without restriction, including without limitation the rights
|
|
|
|
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
|
|
* copies of the Software, and to permit persons to whom the Software is
|
|
|
|
* furnished to do so, subject to the following conditions:
|
|
|
|
*
|
|
|
|
* The above copyright notice and this permission notice shall be included in
|
|
|
|
* all copies or substantial portions of the Software.
|
|
|
|
*
|
|
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
|
|
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
|
|
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
|
|
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
|
|
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
|
|
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
|
|
|
* THE SOFTWARE.
|
|
|
|
*/
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
// Essentially normal Python has 1 type: Python objects
|
|
|
|
// Viper has more than 1 type, and is just a more complicated (a superset of) Python.
|
|
|
|
// If you declare everything in Viper as a Python object (ie omit type decls) then
|
|
|
|
// it should in principle be exactly the same as Python native.
|
|
|
|
// Having types means having more opcodes, like binary_op_nat_nat, binary_op_nat_obj etc.
|
|
|
|
// In practice we won't have a VM but rather do this in asm which is actually very minimal.
|
|
|
|
|
|
|
|
// Because it breaks strict Python equivalence it should be a completely separate
|
|
|
|
// decorator. It breaks equivalence because overflow on integers wraps around.
|
|
|
|
// It shouldn't break equivalence if you don't use the new types, but since the
|
|
|
|
// type decls might be used in normal Python for other reasons, it's probably safest,
|
|
|
|
// cleanest and clearest to make it a separate decorator.
|
|
|
|
|
|
|
|
// Actually, it does break equivalence because integers default to native integers,
|
|
|
|
// not Python objects.
|
|
|
|
|
|
|
|
// for x in l[0:8]: can be compiled into a native loop if l has pointer type
|
|
|
|
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <string.h>
|
|
|
|
#include <assert.h>
|
|
|
|
|
2015-01-01 20:27:54 +00:00
|
|
|
#include "py/emit.h"
|
2019-10-07 01:56:24 +01:00
|
|
|
#include "py/nativeglue.h"
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
#include "py/objfun.h"
|
2019-09-10 04:47:24 +01:00
|
|
|
#include "py/objstr.h"
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2017-07-24 17:55:14 +01:00
|
|
|
#if MICROPY_DEBUG_VERBOSE // print debugging info
|
2014-04-06 12:58:40 +01:00
|
|
|
#define DEBUG_PRINT (1)
|
|
|
|
#define DEBUG_printf DEBUG_printf
|
|
|
|
#else // don't print debugging info
|
|
|
|
#define DEBUG_printf(...) (void)0
|
|
|
|
#endif
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
// wrapper around everything in this file
|
2019-09-13 04:15:12 +01:00
|
|
|
#if N_X64 || N_X86 || N_THUMB || N_ARM || N_XTENSA || N_XTENSAWIN
|
2016-12-08 06:47:17 +00:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// C stack layout for native functions:
|
2018-10-01 03:36:21 +01:00
|
|
|
// 0: nlr_buf_t [optional]
|
2022-05-20 05:31:56 +01:00
|
|
|
// return_value [optional word]
|
2022-05-19 06:08:45 +01:00
|
|
|
// exc_handler_unwind [optional word]
|
2022-05-15 04:37:06 +01:00
|
|
|
// emit->code_state_start: mp_code_state_native_t
|
2018-10-01 03:36:21 +01:00
|
|
|
// emit->stack_start: Python object stack | emit->n_state
|
|
|
|
// locals (reversed, L0 at end) |
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
//
|
2018-10-01 04:07:04 +01:00
|
|
|
// C stack layout for native generator functions:
|
|
|
|
// 0=emit->stack_start: nlr_buf_t
|
2022-05-20 05:31:56 +01:00
|
|
|
// return_value
|
2022-05-19 06:08:45 +01:00
|
|
|
// exc_handler_unwind [optional word]
|
2018-10-01 04:07:04 +01:00
|
|
|
//
|
|
|
|
// Then REG_GENERATOR_STATE points to:
|
2022-05-15 04:37:06 +01:00
|
|
|
// 0=emit->code_state_start: mp_code_state_native_t
|
2018-10-01 04:07:04 +01:00
|
|
|
// emit->stack_start: Python object stack | emit->n_state
|
|
|
|
// locals (reversed, L0 at end) |
|
|
|
|
//
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// C stack layout for viper functions:
|
2018-10-01 03:36:21 +01:00
|
|
|
// 0: nlr_buf_t [optional]
|
2022-05-20 05:31:56 +01:00
|
|
|
// return_value [optional word]
|
2022-05-19 06:08:45 +01:00
|
|
|
// exc_handler_unwind [optional word]
|
2018-10-01 03:36:21 +01:00
|
|
|
// emit->code_state_start: fun_obj, old_globals [optional]
|
|
|
|
// emit->stack_start: Python object stack | emit->n_state
|
|
|
|
// locals (reversed, L0 at end) |
|
|
|
|
// (L0-L2 may be in regs instead)
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
2019-09-18 04:40:42 +01:00
|
|
|
// Native emitter needs to know the following sizes and offsets of C structs (on the target):
|
2019-09-18 04:45:20 +01:00
|
|
|
#if MICROPY_DYNAMIC_COMPILER
|
|
|
|
#define SIZEOF_NLR_BUF (2 + mp_dynamic_compiler.nlr_buf_num_regs + 1) // the +1 is conservative in case MICROPY_ENABLE_PYSTACK enabled
|
|
|
|
#else
|
2019-09-18 04:40:42 +01:00
|
|
|
#define SIZEOF_NLR_BUF (sizeof(nlr_buf_t) / sizeof(uintptr_t))
|
2019-09-18 04:45:20 +01:00
|
|
|
#endif
|
2022-05-15 04:37:06 +01:00
|
|
|
#define SIZEOF_CODE_STATE (sizeof(mp_code_state_native_t) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_CODE_STATE_STATE (offsetof(mp_code_state_native_t, state) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_CODE_STATE_FUN_BC (offsetof(mp_code_state_native_t, fun_bc) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_CODE_STATE_IP (offsetof(mp_code_state_native_t, ip) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_CODE_STATE_SP (offsetof(mp_code_state_native_t, sp) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_CODE_STATE_N_STATE (offsetof(mp_code_state_native_t, n_state) / sizeof(uintptr_t))
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
#define OFFSETOF_OBJ_FUN_BC_CONTEXT (offsetof(mp_obj_fun_bc_t, context) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_OBJ_FUN_BC_CHILD_TABLE (offsetof(mp_obj_fun_bc_t, child_table) / sizeof(uintptr_t))
|
2019-09-10 04:47:24 +01:00
|
|
|
#define OFFSETOF_OBJ_FUN_BC_BYTECODE (offsetof(mp_obj_fun_bc_t, bytecode) / sizeof(uintptr_t))
|
2022-05-20 05:31:56 +01:00
|
|
|
#define OFFSETOF_MODULE_CONTEXT_QSTR_TABLE (offsetof(mp_module_context_t, constants.qstr_table) / sizeof(uintptr_t))
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
#define OFFSETOF_MODULE_CONTEXT_OBJ_TABLE (offsetof(mp_module_context_t, constants.obj_table) / sizeof(uintptr_t))
|
|
|
|
#define OFFSETOF_MODULE_CONTEXT_GLOBALS (offsetof(mp_module_context_t, module.globals) / sizeof(uintptr_t))
|
2019-09-18 04:40:42 +01:00
|
|
|
|
2019-09-10 04:35:19 +01:00
|
|
|
// If not already defined, set parent args to same as child call registers
|
|
|
|
#ifndef REG_PARENT_RET
|
|
|
|
#define REG_PARENT_RET REG_RET
|
|
|
|
#define REG_PARENT_ARG_1 REG_ARG_1
|
|
|
|
#define REG_PARENT_ARG_2 REG_ARG_2
|
|
|
|
#define REG_PARENT_ARG_3 REG_ARG_3
|
|
|
|
#define REG_PARENT_ARG_4 REG_ARG_4
|
|
|
|
#endif
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Word index of nlr_buf_t.ret_val
|
|
|
|
#define NLR_BUF_IDX_RET_VAL (1)
|
|
|
|
|
2018-09-27 14:27:53 +01:00
|
|
|
// Whether the viper function needs access to fun_obj
|
|
|
|
#define NEED_FUN_OBJ(emit) ((emit)->scope->exc_stack_size > 0 \
|
|
|
|
|| ((emit)->scope->scope_flags & (MP_SCOPE_FLAG_REFGLOBALS | MP_SCOPE_FLAG_HASCONSTS)))
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Whether the native/viper function needs to be wrapped in an exception handler
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
#define NEED_GLOBAL_EXC_HANDLER(emit) ((emit)->scope->exc_stack_size > 0 \
|
2018-10-01 04:07:04 +01:00
|
|
|
|| ((emit)->scope->scope_flags & (MP_SCOPE_FLAG_GENERATOR | MP_SCOPE_FLAG_REFGLOBALS)))
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
2022-05-19 06:08:45 +01:00
|
|
|
// Whether a slot is needed to store LOCAL_IDX_EXC_HANDLER_UNWIND
|
|
|
|
#define NEED_EXC_HANDLER_UNWIND(emit) ((emit)->scope->exc_stack_size > 0)
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Whether registers can be used to store locals (only true if there are no
|
|
|
|
// exception handlers, because otherwise an nlr_jump will restore registers to
|
|
|
|
// their state at the start of the function and updates to locals will be lost)
|
2018-10-01 04:07:04 +01:00
|
|
|
#define CAN_USE_REGS_FOR_LOCALS(emit) ((emit)->scope->exc_stack_size == 0 && !(emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR))
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
// Indices within the local C stack for various variables
|
2018-10-01 03:36:21 +01:00
|
|
|
#define LOCAL_IDX_EXC_VAL(emit) (NLR_BUF_IDX_RET_VAL)
|
|
|
|
#define LOCAL_IDX_EXC_HANDLER_PC(emit) (NLR_BUF_IDX_LOCAL_1)
|
2022-05-20 05:31:56 +01:00
|
|
|
#define LOCAL_IDX_EXC_HANDLER_UNWIND(emit) (SIZEOF_NLR_BUF + 1) // this needs a dedicated variable outside nlr_buf_t
|
|
|
|
#define LOCAL_IDX_RET_VAL(emit) (SIZEOF_NLR_BUF) // needed when NEED_GLOBAL_EXC_HANDLER is true
|
2019-09-18 04:40:42 +01:00
|
|
|
#define LOCAL_IDX_FUN_OBJ(emit) ((emit)->code_state_start + OFFSETOF_CODE_STATE_FUN_BC)
|
|
|
|
#define LOCAL_IDX_OLD_GLOBALS(emit) ((emit)->code_state_start + OFFSETOF_CODE_STATE_IP)
|
|
|
|
#define LOCAL_IDX_GEN_PC(emit) ((emit)->code_state_start + OFFSETOF_CODE_STATE_IP)
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
#define LOCAL_IDX_LOCAL_VAR(emit, local_num) ((emit)->stack_start + (emit)->n_state - 1 - (local_num))
|
|
|
|
|
2022-05-20 05:31:56 +01:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
|
|
|
|
|
|
|
// When building with the ability to save native code to .mpy files:
|
|
|
|
// - Qstrs are indirect via qstr_table, and REG_LOCAL_3 always points to qstr_table.
|
|
|
|
// - In a generator no registers are used to store locals, and REG_LOCAL_2 points to the generator state.
|
|
|
|
// - At most 2 registers hold local variables (see CAN_USE_REGS_FOR_LOCALS for when this is possible).
|
|
|
|
|
|
|
|
#define REG_GENERATOR_STATE (REG_LOCAL_2)
|
|
|
|
#define REG_QSTR_TABLE (REG_LOCAL_3)
|
|
|
|
#define MAX_REGS_FOR_LOCAL_VARS (2)
|
|
|
|
|
|
|
|
STATIC const uint8_t reg_local_table[MAX_REGS_FOR_LOCAL_VARS] = {REG_LOCAL_1, REG_LOCAL_2};
|
|
|
|
|
|
|
|
#else
|
|
|
|
|
|
|
|
// When building without the ability to save native code to .mpy files:
|
|
|
|
// - Qstrs values are written directly into the machine code.
|
|
|
|
// - In a generator no registers are used to store locals, and REG_LOCAL_3 points to the generator state.
|
|
|
|
// - At most 3 registers hold local variables (see CAN_USE_REGS_FOR_LOCALS for when this is possible).
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
#define REG_GENERATOR_STATE (REG_LOCAL_3)
|
2022-05-20 05:31:56 +01:00
|
|
|
#define MAX_REGS_FOR_LOCAL_VARS (3)
|
|
|
|
|
|
|
|
STATIC const uint8_t reg_local_table[MAX_REGS_FOR_LOCAL_VARS] = {REG_LOCAL_1, REG_LOCAL_2, REG_LOCAL_3};
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define REG_LOCAL_LAST (reg_local_table[MAX_REGS_FOR_LOCAL_VARS - 1])
|
2018-10-01 04:07:04 +01:00
|
|
|
|
2015-04-20 14:29:31 +01:00
|
|
|
#define EMIT_NATIVE_VIPER_TYPE_ERROR(emit, ...) do { \
|
|
|
|
*emit->error_slot = mp_obj_new_exception_msg_varg(&mp_type_ViperTypeError, __VA_ARGS__); \
|
|
|
|
} while (0)
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
typedef enum {
|
2013-10-08 22:18:32 +01:00
|
|
|
STACK_VALUE,
|
|
|
|
STACK_REG,
|
|
|
|
STACK_IMM,
|
|
|
|
} stack_info_kind_t;
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2016-02-02 23:16:05 +00:00
|
|
|
// these enums must be distinct and the bottom 4 bits
|
2014-09-29 22:10:41 +01:00
|
|
|
// must correspond to the correct MP_NATIVE_TYPE_xxx value
|
2013-10-08 09:05:10 +01:00
|
|
|
typedef enum {
|
2014-09-29 22:10:41 +01:00
|
|
|
VTYPE_PYOBJ = 0x00 | MP_NATIVE_TYPE_OBJ,
|
|
|
|
VTYPE_BOOL = 0x00 | MP_NATIVE_TYPE_BOOL,
|
|
|
|
VTYPE_INT = 0x00 | MP_NATIVE_TYPE_INT,
|
|
|
|
VTYPE_UINT = 0x00 | MP_NATIVE_TYPE_UINT,
|
2016-02-02 23:16:05 +00:00
|
|
|
VTYPE_PTR = 0x00 | MP_NATIVE_TYPE_PTR,
|
|
|
|
VTYPE_PTR8 = 0x00 | MP_NATIVE_TYPE_PTR8,
|
|
|
|
VTYPE_PTR16 = 0x00 | MP_NATIVE_TYPE_PTR16,
|
|
|
|
VTYPE_PTR32 = 0x00 | MP_NATIVE_TYPE_PTR32,
|
2014-09-29 22:10:41 +01:00
|
|
|
|
2016-02-02 23:16:05 +00:00
|
|
|
VTYPE_PTR_NONE = 0x50 | MP_NATIVE_TYPE_PTR,
|
2014-09-29 22:10:41 +01:00
|
|
|
|
2015-10-13 00:50:17 +01:00
|
|
|
VTYPE_UNBOUND = 0x60 | MP_NATIVE_TYPE_OBJ,
|
|
|
|
VTYPE_BUILTIN_CAST = 0x70 | MP_NATIVE_TYPE_OBJ,
|
2013-10-08 09:05:10 +01:00
|
|
|
} vtype_kind_t;
|
|
|
|
|
2015-04-20 14:29:31 +01:00
|
|
|
STATIC qstr vtype_to_qstr(vtype_kind_t vtype) {
|
|
|
|
switch (vtype) {
|
|
|
|
case VTYPE_PYOBJ:
|
|
|
|
return MP_QSTR_object;
|
|
|
|
case VTYPE_BOOL:
|
|
|
|
return MP_QSTR_bool;
|
|
|
|
case VTYPE_INT:
|
|
|
|
return MP_QSTR_int;
|
|
|
|
case VTYPE_UINT:
|
|
|
|
return MP_QSTR_uint;
|
|
|
|
case VTYPE_PTR:
|
|
|
|
return MP_QSTR_ptr;
|
|
|
|
case VTYPE_PTR8:
|
|
|
|
return MP_QSTR_ptr8;
|
|
|
|
case VTYPE_PTR16:
|
|
|
|
return MP_QSTR_ptr16;
|
2015-10-13 00:50:17 +01:00
|
|
|
case VTYPE_PTR32:
|
|
|
|
return MP_QSTR_ptr32;
|
2015-04-20 14:29:31 +01:00
|
|
|
case VTYPE_PTR_NONE:
|
|
|
|
default:
|
|
|
|
return MP_QSTR_None;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-10-08 22:18:32 +01:00
|
|
|
typedef struct _stack_info_t {
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
stack_info_kind_t kind;
|
|
|
|
union {
|
|
|
|
int u_reg;
|
2014-07-03 13:25:24 +01:00
|
|
|
mp_int_t u_imm;
|
2015-01-24 23:14:12 +00:00
|
|
|
} data;
|
2013-10-08 22:18:32 +01:00
|
|
|
} stack_info_t;
|
|
|
|
|
2018-09-04 05:31:28 +01:00
|
|
|
#define UNWIND_LABEL_UNUSED (0x7fff)
|
|
|
|
#define UNWIND_LABEL_DO_FINAL_UNWIND (0x7ffe)
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
typedef struct _exc_stack_entry_t {
|
|
|
|
uint16_t label : 15;
|
|
|
|
uint16_t is_finally : 1;
|
2018-09-04 05:31:28 +01:00
|
|
|
uint16_t unwind_label : 15;
|
|
|
|
uint16_t is_active : 1;
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
} exc_stack_entry_t;
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
struct _emit_t {
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
mp_emit_common_t *emit_common;
|
2015-04-20 14:29:31 +01:00
|
|
|
mp_obj_t *error_slot;
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
uint *label_slot;
|
|
|
|
uint exit_label;
|
2013-10-08 09:05:10 +01:00
|
|
|
int pass;
|
|
|
|
|
|
|
|
bool do_viper_types;
|
2021-01-17 13:08:59 +00:00
|
|
|
bool prelude_offset_uses_u16_encoding;
|
2013-10-08 22:18:32 +01:00
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
mp_uint_t local_vtype_alloc;
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t *local_vtype;
|
2013-10-08 22:18:32 +01:00
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
mp_uint_t stack_info_alloc;
|
2013-10-08 22:18:32 +01:00
|
|
|
stack_info_t *stack_info;
|
2014-10-19 19:00:51 +01:00
|
|
|
vtype_kind_t saved_stack_vtype;
|
2013-10-08 22:18:32 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
size_t exc_stack_alloc;
|
|
|
|
size_t exc_stack_size;
|
|
|
|
exc_stack_entry_t *exc_stack;
|
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
int prelude_offset;
|
2022-05-10 04:56:24 +01:00
|
|
|
int prelude_ptr_index;
|
2018-10-01 04:07:04 +01:00
|
|
|
int start_offset;
|
2015-04-06 22:38:53 +01:00
|
|
|
int n_state;
|
2018-10-01 03:36:21 +01:00
|
|
|
uint16_t code_state_start;
|
|
|
|
uint16_t stack_start;
|
2013-10-08 09:05:10 +01:00
|
|
|
int stack_size;
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
uint16_t n_info;
|
2019-09-25 06:45:47 +01:00
|
|
|
uint16_t n_cell;
|
2013-10-08 09:05:10 +01:00
|
|
|
|
|
|
|
scope_t *scope;
|
|
|
|
|
2014-09-06 23:06:36 +01:00
|
|
|
ASM_T *as;
|
2013-10-08 09:05:10 +01:00
|
|
|
};
|
|
|
|
|
2022-05-20 05:31:56 +01:00
|
|
|
STATIC void emit_load_reg_with_object(emit_t *emit, int reg, mp_obj_t obj);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
STATIC void emit_native_global_exc_entry(emit_t *emit);
|
|
|
|
STATIC void emit_native_global_exc_exit(emit_t *emit);
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
STATIC void emit_native_load_const_obj(emit_t *emit, mp_obj_t obj);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_t *EXPORT_FUN(new)(mp_emit_common_t * emit_common, mp_obj_t *error_slot, uint *label_slot, mp_uint_t max_num_labels) {
|
2014-05-07 17:24:22 +01:00
|
|
|
emit_t *emit = m_new0(emit_t, 1);
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit->emit_common = emit_common;
|
2015-04-20 14:29:31 +01:00
|
|
|
emit->error_slot = error_slot;
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit->label_slot = label_slot;
|
2018-08-07 07:05:44 +01:00
|
|
|
emit->stack_info_alloc = 8;
|
|
|
|
emit->stack_info = m_new(stack_info_t, emit->stack_info_alloc);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit->exc_stack_alloc = 8;
|
|
|
|
emit->exc_stack = m_new(exc_stack_entry_t, emit->exc_stack_alloc);
|
2016-11-27 22:24:50 +00:00
|
|
|
emit->as = m_new0(ASM_T, 1);
|
|
|
|
mp_asm_base_init(&emit->as->base, max_num_labels);
|
2013-10-08 09:05:10 +01:00
|
|
|
return emit;
|
|
|
|
}
|
|
|
|
|
2014-01-24 22:42:28 +00:00
|
|
|
void EXPORT_FUN(free)(emit_t * emit) {
|
2016-11-27 22:24:50 +00:00
|
|
|
mp_asm_base_deinit(&emit->as->base, false);
|
|
|
|
m_del_obj(ASM_T, emit->as);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
m_del(exc_stack_entry_t, emit->exc_stack, emit->exc_stack_alloc);
|
2014-05-07 17:24:22 +01:00
|
|
|
m_del(vtype_kind_t, emit->local_vtype, emit->local_vtype_alloc);
|
|
|
|
m_del(stack_info_t, emit->stack_info, emit->stack_info_alloc);
|
2014-01-24 14:20:11 +00:00
|
|
|
m_del_obj(emit_t, emit);
|
|
|
|
}
|
|
|
|
|
2018-09-14 08:40:59 +01:00
|
|
|
STATIC void emit_call_with_imm_arg(emit_t *emit, mp_fun_kind_t fun_kind, mp_int_t arg_val, int arg_reg);
|
2015-04-03 16:03:24 +01:00
|
|
|
|
2018-10-14 13:56:46 +01:00
|
|
|
STATIC void emit_native_mov_reg_const(emit_t *emit, int reg_dest, int const_val) {
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, reg_dest, REG_FUN_TABLE, const_val);
|
|
|
|
}
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
STATIC void emit_native_mov_state_reg(emit_t *emit, int local_num, int reg_src) {
|
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
|
|
|
ASM_STORE_REG_REG_OFFSET(emit->as, reg_src, REG_GENERATOR_STATE, local_num);
|
|
|
|
} else {
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, local_num, reg_src);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_mov_reg_state(emit_t *emit, int reg_dest, int local_num) {
|
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, reg_dest, REG_GENERATOR_STATE, local_num);
|
|
|
|
} else {
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, reg_dest, local_num);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_mov_reg_state_addr(emit_t *emit, int reg_dest, int local_num) {
|
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_dest, local_num * ASM_WORD_SIZE);
|
|
|
|
ASM_ADD_REG_REG(emit->as, reg_dest, REG_GENERATOR_STATE);
|
|
|
|
} else {
|
|
|
|
ASM_MOV_REG_LOCAL_ADDR(emit->as, reg_dest, local_num);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-02-03 01:25:50 +00:00
|
|
|
STATIC void emit_native_mov_reg_qstr(emit_t *emit, int arg_reg, qstr qst) {
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
2022-05-20 05:31:56 +01:00
|
|
|
ASM_LOAD16_REG_REG_OFFSET(emit->as, arg_reg, REG_QSTR_TABLE, mp_emit_common_use_qstr(emit->emit_common, qst));
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#else
|
2019-02-03 01:25:50 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, arg_reg, qst);
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#endif
|
2019-02-03 01:25:50 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_mov_reg_qstr_obj(emit_t *emit, int reg_dest, qstr qst) {
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
2022-05-20 05:31:56 +01:00
|
|
|
emit_load_reg_with_object(emit, reg_dest, MP_OBJ_NEW_QSTR(qst));
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#else
|
2019-02-03 01:25:50 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_dest, (mp_uint_t)MP_OBJ_NEW_QSTR(qst));
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#endif
|
2019-02-03 01:25:50 +00:00
|
|
|
}
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
#define emit_native_mov_state_imm_via(emit, local_num, imm, reg_temp) \
|
|
|
|
do { \
|
|
|
|
ASM_MOV_REG_IMM((emit)->as, (reg_temp), (imm)); \
|
|
|
|
emit_native_mov_state_reg((emit), (local_num), (reg_temp)); \
|
|
|
|
} while (false)
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_start_pass(emit_t *emit, pass_kind_t pass, scope_t *scope) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("start_pass(pass=%u, scope=%p)\n", pass, scope);
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
emit->pass = pass;
|
2018-09-15 03:17:14 +01:00
|
|
|
emit->do_viper_types = scope->emit_options == MP_EMIT_OPT_VIPER;
|
2013-10-08 09:05:10 +01:00
|
|
|
emit->stack_size = 0;
|
|
|
|
emit->scope = scope;
|
|
|
|
|
2014-05-07 17:24:22 +01:00
|
|
|
// allocate memory for keeping track of the types of locals
|
|
|
|
if (emit->local_vtype_alloc < scope->num_locals) {
|
|
|
|
emit->local_vtype = m_renew(vtype_kind_t, emit->local_vtype, emit->local_vtype_alloc, scope->num_locals);
|
|
|
|
emit->local_vtype_alloc = scope->num_locals;
|
2013-10-08 22:18:32 +01:00
|
|
|
}
|
2014-05-07 17:24:22 +01:00
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
// set default type for arguments
|
|
|
|
mp_uint_t num_args = emit->scope->num_pos_args + emit->scope->num_kwonly_args;
|
|
|
|
if (scope->scope_flags & MP_SCOPE_FLAG_VARARGS) {
|
|
|
|
num_args += 1;
|
|
|
|
}
|
|
|
|
if (scope->scope_flags & MP_SCOPE_FLAG_VARKEYWORDS) {
|
|
|
|
num_args += 1;
|
|
|
|
}
|
|
|
|
for (mp_uint_t i = 0; i < num_args; i++) {
|
2014-08-15 16:45:41 +01:00
|
|
|
emit->local_vtype[i] = VTYPE_PYOBJ;
|
|
|
|
}
|
2014-08-15 22:39:08 +01:00
|
|
|
|
2018-09-15 04:00:11 +01:00
|
|
|
// Set viper type for arguments
|
|
|
|
if (emit->do_viper_types) {
|
|
|
|
for (int i = 0; i < emit->scope->id_info_len; ++i) {
|
|
|
|
id_info_t *id = &emit->scope->id_info[i];
|
|
|
|
if (id->flags & ID_FLAG_IS_PARAM) {
|
|
|
|
assert(id->local_num < emit->local_vtype_alloc);
|
|
|
|
emit->local_vtype[id->local_num] = id->flags >> ID_FLAG_VIPER_TYPE_POS;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-08-15 22:39:08 +01:00
|
|
|
// local variables begin unbound, and have unknown type
|
2015-04-06 22:38:53 +01:00
|
|
|
for (mp_uint_t i = num_args; i < emit->local_vtype_alloc; i++) {
|
2022-12-15 05:11:27 +00:00
|
|
|
emit->local_vtype[i] = emit->do_viper_types ? VTYPE_UNBOUND : VTYPE_PYOBJ;
|
2014-08-15 22:39:08 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
// values on stack begin unbound
|
|
|
|
for (mp_uint_t i = 0; i < emit->stack_info_alloc; i++) {
|
2014-08-15 16:45:41 +01:00
|
|
|
emit->stack_info[i].kind = STACK_VALUE;
|
2014-08-15 22:39:08 +01:00
|
|
|
emit->stack_info[i].vtype = VTYPE_UNBOUND;
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2016-11-27 22:24:50 +00:00
|
|
|
mp_asm_base_start_pass(&emit->as->base, pass == MP_PASS_EMIT ? MP_ASM_PASS_EMIT : MP_ASM_PASS_COMPUTE);
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
// generate code for entry to function
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2022-05-15 04:37:06 +01:00
|
|
|
// Work out start of code state (mp_code_state_native_t or reduced version for viper)
|
2018-10-01 03:36:21 +01:00
|
|
|
emit->code_state_start = 0;
|
|
|
|
if (NEED_GLOBAL_EXC_HANDLER(emit)) {
|
2022-05-20 05:31:56 +01:00
|
|
|
emit->code_state_start = SIZEOF_NLR_BUF; // for nlr_buf_t
|
|
|
|
emit->code_state_start += 1; // for return_value
|
2022-05-19 06:08:45 +01:00
|
|
|
if (NEED_EXC_HANDLER_UNWIND(emit)) {
|
|
|
|
emit->code_state_start += 1;
|
|
|
|
}
|
2018-10-01 03:36:21 +01:00
|
|
|
}
|
|
|
|
|
2022-05-07 06:51:41 +01:00
|
|
|
size_t fun_table_off = mp_emit_common_use_const_obj(emit->emit_common, MP_OBJ_FROM_PTR(&mp_fun_table));
|
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
if (emit->do_viper_types) {
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Work out size of state (locals plus stack)
|
|
|
|
// n_state counts all stack and locals, even those in registers
|
|
|
|
emit->n_state = scope->num_locals + scope->stack_size;
|
|
|
|
int num_locals_in_regs = 0;
|
|
|
|
if (CAN_USE_REGS_FOR_LOCALS(emit)) {
|
|
|
|
num_locals_in_regs = scope->num_locals;
|
2022-05-20 05:31:56 +01:00
|
|
|
if (num_locals_in_regs > MAX_REGS_FOR_LOCAL_VARS) {
|
|
|
|
num_locals_in_regs = MAX_REGS_FOR_LOCAL_VARS;
|
2015-04-06 22:38:53 +01:00
|
|
|
}
|
2022-05-20 05:31:56 +01:00
|
|
|
// Need a spot for REG_LOCAL_LAST (see below)
|
|
|
|
if (scope->num_pos_args >= MAX_REGS_FOR_LOCAL_VARS + 1) {
|
2018-09-14 08:40:59 +01:00
|
|
|
--num_locals_in_regs;
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
2018-09-15 13:37:07 +01:00
|
|
|
// Work out where the locals and Python stack start within the C stack
|
|
|
|
if (NEED_GLOBAL_EXC_HANDLER(emit)) {
|
|
|
|
// Reserve 2 words for function object and old globals
|
2018-10-01 03:36:21 +01:00
|
|
|
emit->stack_start = emit->code_state_start + 2;
|
2018-09-27 14:27:53 +01:00
|
|
|
} else if (scope->scope_flags & MP_SCOPE_FLAG_HASCONSTS) {
|
|
|
|
// Reserve 1 word for function object, to access const table
|
2018-10-01 03:36:21 +01:00
|
|
|
emit->stack_start = emit->code_state_start + 1;
|
2018-09-15 13:37:07 +01:00
|
|
|
} else {
|
2018-10-01 03:36:21 +01:00
|
|
|
emit->stack_start = emit->code_state_start + 0;
|
2018-09-15 13:37:07 +01:00
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
// Entry to function
|
|
|
|
ASM_ENTRY(emit->as, emit->stack_start + emit->n_state - num_locals_in_regs);
|
2015-04-06 22:38:53 +01:00
|
|
|
|
2018-10-13 04:53:35 +01:00
|
|
|
#if N_X86
|
2019-09-10 04:35:19 +01:00
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 0, REG_PARENT_ARG_1);
|
2015-08-12 23:31:19 +01:00
|
|
|
#endif
|
|
|
|
|
2018-10-13 04:53:35 +01:00
|
|
|
// Load REG_FUN_TABLE with a pointer to mp_fun_table, found in the const_table
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_PARENT_ARG_1, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
2022-05-20 05:31:56 +01:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_QSTR_TABLE, REG_FUN_TABLE, OFFSETOF_MODULE_CONTEXT_QSTR_TABLE);
|
|
|
|
#endif
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_FUN_TABLE, OFFSETOF_MODULE_CONTEXT_OBJ_TABLE);
|
2022-05-07 06:51:41 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_FUN_TABLE, fun_table_off);
|
2018-10-13 04:53:35 +01:00
|
|
|
|
2018-09-15 13:37:07 +01:00
|
|
|
// Store function object (passed as first arg) to stack if needed
|
2018-09-27 14:27:53 +01:00
|
|
|
if (NEED_FUN_OBJ(emit)) {
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_FUN_OBJ(emit), REG_PARENT_ARG_1);
|
2018-09-15 13:37:07 +01:00
|
|
|
}
|
|
|
|
|
2022-05-20 05:31:56 +01:00
|
|
|
// Put n_args in REG_ARG_1, n_kw in REG_ARG_2, args array in REG_LOCAL_LAST
|
2015-04-06 22:38:53 +01:00
|
|
|
#if N_X86
|
2018-09-14 08:40:59 +01:00
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 1, REG_ARG_1);
|
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 2, REG_ARG_2);
|
2022-05-20 05:31:56 +01:00
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 3, REG_LOCAL_LAST);
|
2015-04-06 22:38:53 +01:00
|
|
|
#else
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_1, REG_PARENT_ARG_2);
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_2, REG_PARENT_ARG_3);
|
2022-05-20 05:31:56 +01:00
|
|
|
ASM_MOV_REG_REG(emit->as, REG_LOCAL_LAST, REG_PARENT_ARG_4);
|
2018-09-14 08:40:59 +01:00
|
|
|
#endif
|
|
|
|
|
|
|
|
// Check number of args matches this function, and call mp_arg_check_num_sig if not
|
|
|
|
ASM_JUMP_IF_REG_NONZERO(emit->as, REG_ARG_2, *emit->label_slot + 4, true);
|
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_3, scope->num_pos_args);
|
|
|
|
ASM_JUMP_IF_REG_EQ(emit->as, REG_ARG_1, REG_ARG_3, *emit->label_slot + 5);
|
|
|
|
mp_asm_base_label_assign(&emit->as->base, *emit->label_slot + 4);
|
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_3, MP_OBJ_FUN_MAKE_SIG(scope->num_pos_args, scope->num_pos_args, false));
|
2018-10-13 05:13:55 +01:00
|
|
|
ASM_CALL_IND(emit->as, MP_F_ARG_CHECK_NUM_SIG);
|
2018-09-14 08:40:59 +01:00
|
|
|
mp_asm_base_label_assign(&emit->as->base, *emit->label_slot + 5);
|
|
|
|
|
|
|
|
// Store arguments into locals (reg or stack), converting to native if needed
|
|
|
|
for (int i = 0; i < emit->scope->num_pos_args; i++) {
|
|
|
|
int r = REG_ARG_1;
|
2022-05-20 05:31:56 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_ARG_1, REG_LOCAL_LAST, i);
|
2018-09-14 08:40:59 +01:00
|
|
|
if (emit->local_vtype[i] != VTYPE_PYOBJ) {
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_OBJ_TO_NATIVE, emit->local_vtype[i], REG_ARG_2);
|
|
|
|
r = REG_RET;
|
|
|
|
}
|
2022-05-20 05:31:56 +01:00
|
|
|
// REG_LOCAL_LAST points to the args array so be sure not to overwrite it if it's still needed
|
|
|
|
if (i < MAX_REGS_FOR_LOCAL_VARS && CAN_USE_REGS_FOR_LOCALS(emit) && (i != MAX_REGS_FOR_LOCAL_VARS - 1 || emit->scope->num_pos_args == MAX_REGS_FOR_LOCAL_VARS)) {
|
2018-09-14 08:40:59 +01:00
|
|
|
ASM_MOV_REG_REG(emit->as, reg_local_table[i], r);
|
2015-04-06 22:38:53 +01:00
|
|
|
} else {
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_state_reg(emit, LOCAL_IDX_LOCAL_VAR(emit, i), r);
|
2015-04-06 22:38:53 +01:00
|
|
|
}
|
|
|
|
}
|
2022-05-20 05:31:56 +01:00
|
|
|
// Get local from the stack back into REG_LOCAL_LAST if this reg couldn't be written to above
|
|
|
|
if (emit->scope->num_pos_args >= MAX_REGS_FOR_LOCAL_VARS + 1 && CAN_USE_REGS_FOR_LOCALS(emit)) {
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_LOCAL_LAST, LOCAL_IDX_LOCAL_VAR(emit, MAX_REGS_FOR_LOCAL_VARS - 1));
|
2018-09-14 08:40:59 +01:00
|
|
|
}
|
2015-04-06 22:38:53 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_global_exc_entry(emit);
|
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
} else {
|
|
|
|
// work out size of state (locals plus stack)
|
|
|
|
emit->n_state = scope->num_locals + scope->stack_size;
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
2022-05-10 04:56:24 +01:00
|
|
|
mp_asm_base_data(&emit->as->base, ASM_WORD_SIZE, (uintptr_t)emit->prelude_ptr_index);
|
2018-10-01 04:07:04 +01:00
|
|
|
mp_asm_base_data(&emit->as->base, ASM_WORD_SIZE, (uintptr_t)emit->start_offset);
|
2022-05-19 06:08:45 +01:00
|
|
|
ASM_ENTRY(emit->as, emit->code_state_start);
|
|
|
|
|
|
|
|
// Reset the state size for the state pointed to by REG_GENERATOR_STATE
|
|
|
|
emit->code_state_start = 0;
|
|
|
|
emit->stack_start = SIZEOF_CODE_STATE;
|
2018-10-01 04:07:04 +01:00
|
|
|
|
|
|
|
// Put address of code_state into REG_GENERATOR_STATE
|
|
|
|
#if N_X86
|
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 0, REG_GENERATOR_STATE);
|
|
|
|
#else
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_REG_REG(emit->as, REG_GENERATOR_STATE, REG_PARENT_ARG_1);
|
2018-10-01 04:07:04 +01:00
|
|
|
#endif
|
|
|
|
|
|
|
|
// Put throw value into LOCAL_IDX_EXC_VAL slot, for yield/yield-from
|
|
|
|
#if N_X86
|
2019-09-10 04:35:19 +01:00
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 1, REG_PARENT_ARG_2);
|
2018-10-01 04:07:04 +01:00
|
|
|
#endif
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_VAL(emit), REG_PARENT_ARG_2);
|
2018-10-13 04:53:35 +01:00
|
|
|
|
|
|
|
// Load REG_FUN_TABLE with a pointer to mp_fun_table, found in the const_table
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_GENERATOR_STATE, LOCAL_IDX_FUN_OBJ(emit));
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_TEMP0, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
2022-05-20 05:31:56 +01:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_QSTR_TABLE, REG_TEMP0, OFFSETOF_MODULE_CONTEXT_QSTR_TABLE);
|
|
|
|
#endif
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_TEMP0, OFFSETOF_MODULE_CONTEXT_OBJ_TABLE);
|
2022-05-07 06:51:41 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_TEMP0, fun_table_off);
|
2018-10-01 04:07:04 +01:00
|
|
|
} else {
|
|
|
|
// The locals and stack start after the code_state structure
|
2019-09-18 04:40:42 +01:00
|
|
|
emit->stack_start = emit->code_state_start + SIZEOF_CODE_STATE;
|
2018-08-04 13:41:35 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Allocate space on C-stack for code_state structure, which includes state
|
|
|
|
ASM_ENTRY(emit->as, emit->stack_start + emit->n_state);
|
2015-04-06 22:38:53 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Prepare incoming arguments for call to mp_setup_code_state
|
2017-03-17 03:54:53 +00:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
#if N_X86
|
2019-09-10 04:35:19 +01:00
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 0, REG_PARENT_ARG_1);
|
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 1, REG_PARENT_ARG_2);
|
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 2, REG_PARENT_ARG_3);
|
|
|
|
asm_x86_mov_arg_to_r32(emit->as, 3, REG_PARENT_ARG_4);
|
2018-10-01 04:07:04 +01:00
|
|
|
#endif
|
2015-04-06 22:38:53 +01:00
|
|
|
|
2018-10-13 04:53:35 +01:00
|
|
|
// Load REG_FUN_TABLE with a pointer to mp_fun_table, found in the const_table
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_PARENT_ARG_1, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
2022-05-20 05:31:56 +01:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_QSTR_TABLE, REG_FUN_TABLE, OFFSETOF_MODULE_CONTEXT_QSTR_TABLE);
|
|
|
|
#endif
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_FUN_TABLE, OFFSETOF_MODULE_CONTEXT_OBJ_TABLE);
|
2022-05-07 06:51:41 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_FUN_TABLE, REG_FUN_TABLE, fun_table_off);
|
2018-10-13 04:53:35 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Set code_state.fun_bc
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_FUN_OBJ(emit), REG_PARENT_ARG_1);
|
2017-03-17 03:54:53 +00:00
|
|
|
|
2022-05-10 04:56:24 +01:00
|
|
|
// Set code_state.ip, a pointer to the beginning of the prelude. This pointer is found
|
|
|
|
// either directly in mp_obj_fun_bc_t.child_table (if there are no children), or in
|
|
|
|
// mp_obj_fun_bc_t.child_table[num_children] (if num_children > 0).
|
2022-05-19 06:07:31 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_PARENT_ARG_1, REG_PARENT_ARG_1, OFFSETOF_OBJ_FUN_BC_CHILD_TABLE);
|
2022-05-10 04:56:24 +01:00
|
|
|
if (emit->prelude_ptr_index != 0) {
|
2022-05-19 06:07:31 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_PARENT_ARG_1, REG_PARENT_ARG_1, emit->prelude_ptr_index);
|
2021-01-17 13:08:59 +00:00
|
|
|
}
|
2022-05-19 06:07:31 +01:00
|
|
|
emit_native_mov_state_reg(emit, emit->code_state_start + OFFSETOF_CODE_STATE_IP, REG_PARENT_ARG_1);
|
2015-04-06 22:38:53 +01:00
|
|
|
|
2019-09-24 06:57:08 +01:00
|
|
|
// Set code_state.n_state (only works on little endian targets due to n_state being uint16_t)
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_native_mov_state_imm_via(emit, emit->code_state_start + OFFSETOF_CODE_STATE_N_STATE, emit->n_state, REG_ARG_1);
|
2019-09-24 06:57:08 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Put address of code_state into first arg
|
|
|
|
ASM_MOV_REG_LOCAL_ADDR(emit->as, REG_ARG_1, emit->code_state_start);
|
2015-04-06 22:38:53 +01:00
|
|
|
|
2019-09-10 04:35:19 +01:00
|
|
|
// Copy next 3 args if needed
|
|
|
|
#if REG_ARG_2 != REG_PARENT_ARG_2
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_2, REG_PARENT_ARG_2);
|
|
|
|
#endif
|
|
|
|
#if REG_ARG_3 != REG_PARENT_ARG_3
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_3, REG_PARENT_ARG_3);
|
|
|
|
#endif
|
|
|
|
#if REG_ARG_4 != REG_PARENT_ARG_4
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_4, REG_PARENT_ARG_4);
|
|
|
|
#endif
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Call mp_setup_code_state to prepare code_state structure
|
|
|
|
#if N_THUMB
|
2018-10-13 05:08:31 +01:00
|
|
|
asm_thumb_bl_ind(emit->as, MP_F_SETUP_CODE_STATE, ASM_THUMB_REG_R4);
|
2018-10-01 04:07:04 +01:00
|
|
|
#elif N_ARM
|
2018-10-13 05:03:29 +01:00
|
|
|
asm_arm_bl_ind(emit->as, MP_F_SETUP_CODE_STATE, ASM_ARM_REG_R4);
|
2018-10-01 04:07:04 +01:00
|
|
|
#else
|
2018-10-13 05:13:55 +01:00
|
|
|
ASM_CALL_IND(emit->as, MP_F_SETUP_CODE_STATE);
|
2018-10-01 04:07:04 +01:00
|
|
|
#endif
|
|
|
|
}
|
2015-04-06 22:38:53 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_global_exc_entry(emit);
|
|
|
|
|
|
|
|
// cache some locals in registers, but only if no exception handlers
|
|
|
|
if (CAN_USE_REGS_FOR_LOCALS(emit)) {
|
2022-05-20 05:31:56 +01:00
|
|
|
for (int i = 0; i < MAX_REGS_FOR_LOCAL_VARS && i < scope->num_locals; ++i) {
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
ASM_MOV_REG_LOCAL(emit->as, reg_local_table[i], LOCAL_IDX_LOCAL_VAR(emit, i));
|
|
|
|
}
|
2015-04-06 22:38:53 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
// set the type of closed over variables
|
|
|
|
for (mp_uint_t i = 0; i < scope->id_info_len; i++) {
|
|
|
|
id_info_t *id = &scope->id_info[i];
|
|
|
|
if (id->kind == ID_INFO_KIND_CELL) {
|
|
|
|
emit->local_vtype[id->local_num] = VTYPE_PYOBJ;
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-09-16 13:12:59 +01:00
|
|
|
static inline void emit_native_write_code_info_byte(emit_t *emit, byte val) {
|
|
|
|
mp_asm_base_data(&emit->as->base, 1, val);
|
|
|
|
}
|
|
|
|
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
static inline void emit_native_write_code_info_qstr(emit_t *emit, qstr qst) {
|
|
|
|
mp_encode_uint(&emit->as->base, mp_asm_base_get_cur_to_write_bytes, mp_emit_common_use_qstr(emit->emit_common, qst));
|
|
|
|
}
|
|
|
|
|
2022-03-15 22:37:58 +00:00
|
|
|
STATIC bool emit_native_end_pass(emit_t *emit) {
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_global_exc_exit(emit);
|
2015-04-06 22:38:53 +01:00
|
|
|
|
|
|
|
if (!emit->do_viper_types) {
|
2016-11-27 22:24:50 +00:00
|
|
|
emit->prelude_offset = mp_asm_base_get_code_pos(&emit->as->base);
|
2022-05-10 04:56:24 +01:00
|
|
|
emit->prelude_ptr_index = emit->emit_common->ct_cur_child;
|
2019-09-16 13:12:59 +01:00
|
|
|
|
|
|
|
size_t n_state = emit->n_state;
|
|
|
|
size_t n_exc_stack = 0; // exc-stack not needed for native code
|
|
|
|
MP_BC_PRELUDE_SIG_ENCODE(n_state, n_exc_stack, emit->scope, emit_native_write_code_info_byte, emit);
|
2015-10-23 01:23:11 +01:00
|
|
|
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
size_t n_info = emit->n_info;
|
|
|
|
size_t n_cell = emit->n_cell;
|
|
|
|
MP_BC_PRELUDE_SIZE_ENCODE(n_info, n_cell, emit_native_write_code_info_byte, emit);
|
|
|
|
|
|
|
|
// bytecode prelude: source info (function and argument qstrs)
|
|
|
|
size_t info_start = mp_asm_base_get_code_pos(&emit->as->base);
|
|
|
|
emit_native_write_code_info_qstr(emit, emit->scope->simple_name);
|
|
|
|
for (int i = 0; i < emit->scope->num_pos_args + emit->scope->num_kwonly_args; i++) {
|
|
|
|
qstr qst = MP_QSTR__star_;
|
|
|
|
for (int j = 0; j < emit->scope->id_info_len; ++j) {
|
|
|
|
id_info_t *id = &emit->scope->id_info[j];
|
|
|
|
if ((id->flags & ID_FLAG_IS_PARAM) && id->local_num == i) {
|
|
|
|
qst = id->qst;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
emit_native_write_code_info_qstr(emit, qst);
|
|
|
|
}
|
|
|
|
emit->n_info = mp_asm_base_get_code_pos(&emit->as->base) - info_start;
|
2015-10-23 01:23:11 +01:00
|
|
|
|
|
|
|
// bytecode prelude: initialise closed over variables
|
2019-09-25 06:45:47 +01:00
|
|
|
size_t cell_start = mp_asm_base_get_code_pos(&emit->as->base);
|
2015-10-23 01:23:11 +01:00
|
|
|
for (int i = 0; i < emit->scope->id_info_len; i++) {
|
|
|
|
id_info_t *id = &emit->scope->id_info[i];
|
|
|
|
if (id->kind == ID_INFO_KIND_CELL) {
|
2019-09-25 06:45:47 +01:00
|
|
|
assert(id->local_num <= 255);
|
2016-11-27 22:24:50 +00:00
|
|
|
mp_asm_base_data(&emit->as->base, 1, id->local_num); // write the local which should be converted to a cell
|
2015-10-23 01:23:11 +01:00
|
|
|
}
|
|
|
|
}
|
2019-09-25 06:45:47 +01:00
|
|
|
emit->n_cell = mp_asm_base_get_code_pos(&emit->as->base) - cell_start;
|
2019-09-10 04:47:24 +01:00
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
}
|
|
|
|
|
2014-09-06 23:06:36 +01:00
|
|
|
ASM_END_PASS(emit->as);
|
2013-10-08 09:05:10 +01:00
|
|
|
|
|
|
|
// check stack is back to zero size
|
2017-03-14 12:00:53 +00:00
|
|
|
assert(emit->stack_size == 0);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
assert(emit->exc_stack_size == 0);
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2014-05-07 17:24:22 +01:00
|
|
|
if (emit->pass == MP_PASS_EMIT) {
|
2016-11-27 22:24:50 +00:00
|
|
|
void *f = mp_asm_base_get_code(&emit->as->base);
|
|
|
|
mp_uint_t f_len = mp_asm_base_get_code_size(&emit->as->base);
|
2014-08-15 16:45:41 +01:00
|
|
|
|
2022-05-10 04:56:24 +01:00
|
|
|
mp_raw_code_t **children = emit->emit_common->children;
|
|
|
|
if (!emit->do_viper_types) {
|
|
|
|
#if MICROPY_EMIT_NATIVE_PRELUDE_SEPARATE_FROM_MACHINE_CODE
|
|
|
|
// Executable code cannot be accessed byte-wise on this architecture, so copy
|
|
|
|
// the prelude to a separate memory region that is byte-wise readable.
|
|
|
|
void *buf = emit->as->base.code_base + emit->prelude_offset;
|
|
|
|
size_t n = emit->as->base.code_offset - emit->prelude_offset;
|
|
|
|
const uint8_t *prelude_ptr = memcpy(m_new(uint8_t, n), buf, n);
|
|
|
|
#else
|
|
|
|
// Point to the prelude directly, at the end of the machine code data.
|
|
|
|
const uint8_t *prelude_ptr = (const uint8_t *)f + emit->prelude_offset;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
// Store the pointer to the prelude using the child_table.
|
|
|
|
assert(emit->prelude_ptr_index == emit->emit_common->ct_cur_child);
|
|
|
|
if (emit->prelude_ptr_index == 0) {
|
|
|
|
children = (void *)prelude_ptr;
|
|
|
|
} else {
|
|
|
|
children = m_renew(mp_raw_code_t *, children, emit->prelude_ptr_index, emit->prelude_ptr_index + 1);
|
|
|
|
children[emit->prelude_ptr_index] = (void *)prelude_ptr;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-04-06 22:38:53 +01:00
|
|
|
mp_emit_glue_assign_native(emit->scope->raw_code,
|
|
|
|
emit->do_viper_types ? MP_CODE_NATIVE_VIPER : MP_CODE_NATIVE_PY,
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
f, f_len,
|
2022-05-10 04:56:24 +01:00
|
|
|
children,
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
#if MICROPY_PERSISTENT_CODE_SAVE
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit->emit_common->ct_cur_child,
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
emit->prelude_offset,
|
|
|
|
#endif
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit->scope->scope_flags, 0, 0);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2022-03-15 22:37:58 +00:00
|
|
|
|
|
|
|
return true;
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2018-08-07 07:05:44 +01:00
|
|
|
STATIC void ensure_extra_stack(emit_t *emit, size_t delta) {
|
|
|
|
if (emit->stack_size + delta > emit->stack_info_alloc) {
|
|
|
|
size_t new_alloc = (emit->stack_size + delta + 8) & ~3;
|
|
|
|
emit->stack_info = m_renew(stack_info_t, emit->stack_info, emit->stack_info_alloc, new_alloc);
|
|
|
|
emit->stack_info_alloc = new_alloc;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-09-23 15:10:03 +01:00
|
|
|
STATIC void adjust_stack(emit_t *emit, mp_int_t stack_size_delta) {
|
|
|
|
assert((mp_int_t)emit->stack_size + stack_size_delta >= 0);
|
2018-08-07 07:05:44 +01:00
|
|
|
assert((mp_int_t)emit->stack_size + stack_size_delta <= (mp_int_t)emit->stack_info_alloc);
|
2013-10-08 09:05:10 +01:00
|
|
|
emit->stack_size += stack_size_delta;
|
2014-05-07 17:24:22 +01:00
|
|
|
if (emit->pass > MP_PASS_SCOPE && emit->stack_size > emit->scope->stack_size) {
|
2013-10-08 09:05:10 +01:00
|
|
|
emit->scope->stack_size = emit->stack_size;
|
|
|
|
}
|
2023-09-26 09:40:17 +01:00
|
|
|
#if DEBUG_PRINT
|
2014-10-19 19:00:51 +01:00
|
|
|
DEBUG_printf(" adjust_stack; stack_size=%d+%d; stack now:", emit->stack_size - stack_size_delta, stack_size_delta);
|
|
|
|
for (int i = 0; i < emit->stack_size; i++) {
|
|
|
|
stack_info_t *si = &emit->stack_info[i];
|
2015-01-24 23:14:12 +00:00
|
|
|
DEBUG_printf(" (v=%d k=%d %d)", si->vtype, si->kind, si->data.u_reg);
|
2014-10-19 19:00:51 +01:00
|
|
|
}
|
|
|
|
DEBUG_printf("\n");
|
|
|
|
#endif
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-23 15:10:03 +01:00
|
|
|
STATIC void emit_native_adjust_stack_size(emit_t *emit, mp_int_t delta) {
|
2014-10-19 19:00:51 +01:00
|
|
|
DEBUG_printf("adjust_stack_size(" INT_FMT ")\n", delta);
|
2018-08-07 07:05:44 +01:00
|
|
|
if (delta > 0) {
|
|
|
|
ensure_extra_stack(emit, delta);
|
|
|
|
}
|
2014-09-23 15:10:03 +01:00
|
|
|
// If we are adjusting the stack in a positive direction (pushing) then we
|
|
|
|
// need to fill in values for the stack kind and vtype of the newly-pushed
|
|
|
|
// entries. These should be set to "value" (ie not reg or imm) because we
|
|
|
|
// should only need to adjust the stack due to a jump to this part in the
|
|
|
|
// code (and hence we have settled the stack before the jump).
|
|
|
|
for (mp_int_t i = 0; i < delta; i++) {
|
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size + i];
|
|
|
|
si->kind = STACK_VALUE;
|
2014-10-19 19:00:51 +01:00
|
|
|
// TODO we don't know the vtype to use here. At the moment this is a
|
|
|
|
// hack to get the case of multi comparison working.
|
|
|
|
if (delta == 1) {
|
|
|
|
si->vtype = emit->saved_stack_vtype;
|
|
|
|
} else {
|
|
|
|
si->vtype = VTYPE_PYOBJ;
|
|
|
|
}
|
2014-09-23 15:10:03 +01:00
|
|
|
}
|
|
|
|
adjust_stack(emit, delta);
|
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_set_source_line(emit_t *emit, mp_uint_t source_line) {
|
2015-01-20 12:47:20 +00:00
|
|
|
(void)emit;
|
|
|
|
(void)source_line;
|
2014-09-23 15:10:03 +01:00
|
|
|
}
|
|
|
|
|
2013-10-08 22:18:32 +01:00
|
|
|
// this must be called at start of emit functions
|
2014-03-27 23:30:26 +00:00
|
|
|
STATIC void emit_native_pre(emit_t *emit) {
|
2022-06-17 14:06:24 +01:00
|
|
|
(void)emit;
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-29 18:45:42 +01:00
|
|
|
// depth==0 is top, depth==1 is before top, etc
|
|
|
|
STATIC stack_info_t *peek_stack(emit_t *emit, mp_uint_t depth) {
|
|
|
|
return &emit->stack_info[emit->stack_size - 1 - depth];
|
|
|
|
}
|
|
|
|
|
|
|
|
// depth==0 is top, depth==1 is before top, etc
|
|
|
|
STATIC vtype_kind_t peek_vtype(emit_t *emit, mp_uint_t depth) {
|
2018-10-14 12:58:27 +01:00
|
|
|
if (emit->do_viper_types) {
|
|
|
|
return peek_stack(emit, depth)->vtype;
|
|
|
|
} else {
|
|
|
|
// Type is always PYOBJ even if the intermediate stored value is not
|
|
|
|
return VTYPE_PYOBJ;
|
|
|
|
}
|
2013-10-08 22:18:32 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2013-10-16 23:58:48 +01:00
|
|
|
// pos=1 is TOS, pos=2 is next, etc
|
|
|
|
// use pos=0 for no skipping
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void need_reg_single(emit_t *emit, int reg_needed, int skip_stack_pos) {
|
2013-10-16 23:58:48 +01:00
|
|
|
skip_stack_pos = emit->stack_size - skip_stack_pos;
|
|
|
|
for (int i = 0; i < emit->stack_size; i++) {
|
|
|
|
if (i != skip_stack_pos) {
|
|
|
|
stack_info_t *si = &emit->stack_info[i];
|
2015-01-24 23:14:12 +00:00
|
|
|
if (si->kind == STACK_REG && si->data.u_reg == reg_needed) {
|
2013-10-16 23:58:48 +01:00
|
|
|
si->kind = STACK_VALUE;
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_state_reg(emit, emit->stack_start + i, si->data.u_reg);
|
2013-10-16 23:58:48 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-07-17 14:18:40 +01:00
|
|
|
// Ensures all unsettled registers that hold Python values are copied to the
|
|
|
|
// concrete Python stack. All registers are then free to use.
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void need_reg_all(emit_t *emit) {
|
2013-10-08 22:18:32 +01:00
|
|
|
for (int i = 0; i < emit->stack_size; i++) {
|
|
|
|
stack_info_t *si = &emit->stack_info[i];
|
2013-10-16 23:58:48 +01:00
|
|
|
if (si->kind == STACK_REG) {
|
2021-07-17 14:18:40 +01:00
|
|
|
DEBUG_printf(" reg(%u) to local(%u)\n", si->data.u_reg, emit->stack_start + i);
|
2013-10-08 22:18:32 +01:00
|
|
|
si->kind = STACK_VALUE;
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_state_reg(emit, emit->stack_start + i, si->data.u_reg);
|
2013-10-08 22:18:32 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2018-10-14 12:58:27 +01:00
|
|
|
STATIC vtype_kind_t load_reg_stack_imm(emit_t *emit, int reg_dest, const stack_info_t *si, bool convert_to_pyobj) {
|
|
|
|
if (!convert_to_pyobj && emit->do_viper_types) {
|
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_dest, si->data.u_imm);
|
|
|
|
return si->vtype;
|
|
|
|
} else {
|
|
|
|
if (si->vtype == VTYPE_PYOBJ) {
|
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_dest, si->data.u_imm);
|
|
|
|
} else if (si->vtype == VTYPE_BOOL) {
|
2018-10-14 13:56:46 +01:00
|
|
|
emit_native_mov_reg_const(emit, reg_dest, MP_F_CONST_FALSE_OBJ + si->data.u_imm);
|
2018-10-14 12:58:27 +01:00
|
|
|
} else if (si->vtype == VTYPE_INT || si->vtype == VTYPE_UINT) {
|
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_dest, (uintptr_t)MP_OBJ_NEW_SMALL_INT(si->data.u_imm));
|
|
|
|
} else if (si->vtype == VTYPE_PTR_NONE) {
|
2018-10-14 13:56:46 +01:00
|
|
|
emit_native_mov_reg_const(emit, reg_dest, MP_F_CONST_NONE_OBJ);
|
2018-10-14 12:58:27 +01:00
|
|
|
} else {
|
2020-03-02 11:35:22 +00:00
|
|
|
mp_raise_NotImplementedError(MP_ERROR_TEXT("conversion to object"));
|
2018-10-14 12:58:27 +01:00
|
|
|
}
|
|
|
|
return VTYPE_PYOBJ;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-07-17 14:18:40 +01:00
|
|
|
// Copies all unsettled registers and immediates that are Python values into the
|
|
|
|
// concrete Python stack. This ensures the concrete Python stack holds valid
|
|
|
|
// values for the current stack_size.
|
2021-07-19 04:13:32 +01:00
|
|
|
// This function may clobber REG_TEMP1.
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void need_stack_settled(emit_t *emit) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf(" need_stack_settled; stack_size=%d\n", emit->stack_size);
|
2021-07-17 14:18:40 +01:00
|
|
|
need_reg_all(emit);
|
2013-10-16 23:58:48 +01:00
|
|
|
for (int i = 0; i < emit->stack_size; i++) {
|
|
|
|
stack_info_t *si = &emit->stack_info[i];
|
|
|
|
if (si->kind == STACK_IMM) {
|
2015-01-24 23:14:12 +00:00
|
|
|
DEBUG_printf(" imm(" INT_FMT ") to local(%u)\n", si->data.u_imm, emit->stack_start + i);
|
2014-08-29 20:05:32 +01:00
|
|
|
si->kind = STACK_VALUE;
|
2021-07-19 04:13:32 +01:00
|
|
|
// using REG_TEMP1 to avoid clobbering REG_TEMP0 (aka REG_RET)
|
|
|
|
si->vtype = load_reg_stack_imm(emit, REG_TEMP1, si, false);
|
|
|
|
emit_native_mov_state_reg(emit, emit->stack_start + i, REG_TEMP1);
|
2013-10-16 23:58:48 +01:00
|
|
|
}
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2013-10-16 23:58:48 +01:00
|
|
|
// pos=1 is TOS, pos=2 is next, etc
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_access_stack(emit_t *emit, int pos, vtype_kind_t *vtype, int reg_dest) {
|
2013-10-16 23:58:48 +01:00
|
|
|
need_reg_single(emit, reg_dest, pos);
|
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size - pos];
|
2013-10-08 22:18:32 +01:00
|
|
|
*vtype = si->vtype;
|
|
|
|
switch (si->kind) {
|
|
|
|
case STACK_VALUE:
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state(emit, reg_dest, emit->stack_start + emit->stack_size - pos);
|
2013-10-08 09:05:10 +01:00
|
|
|
break;
|
|
|
|
|
2013-10-08 22:18:32 +01:00
|
|
|
case STACK_REG:
|
2015-01-24 23:14:12 +00:00
|
|
|
if (si->data.u_reg != reg_dest) {
|
|
|
|
ASM_MOV_REG_REG(emit->as, reg_dest, si->data.u_reg);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
2013-10-08 22:18:32 +01:00
|
|
|
case STACK_IMM:
|
2018-10-14 12:58:27 +01:00
|
|
|
*vtype = load_reg_stack_imm(emit, reg_dest, si, false);
|
2013-10-08 09:05:10 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-09-29 22:10:41 +01:00
|
|
|
// does an efficient X=pop(); discard(); push(X)
|
|
|
|
// needs a (non-temp) register in case the popped element was stored in the stack
|
|
|
|
STATIC void emit_fold_stack_top(emit_t *emit, int reg_dest) {
|
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size - 2];
|
|
|
|
si[0] = si[1];
|
|
|
|
if (si->kind == STACK_VALUE) {
|
|
|
|
// if folded element was on the stack we need to put it in a register
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state(emit, reg_dest, emit->stack_start + emit->stack_size - 1);
|
2014-09-29 22:10:41 +01:00
|
|
|
si->kind = STACK_REG;
|
2015-01-24 23:14:12 +00:00
|
|
|
si->data.u_reg = reg_dest;
|
2014-09-29 22:10:41 +01:00
|
|
|
}
|
|
|
|
adjust_stack(emit, -1);
|
|
|
|
}
|
|
|
|
|
|
|
|
// If stacked value is in a register and the register is not r1 or r2, then
|
|
|
|
// *reg_dest is set to that register. Otherwise the value is put in *reg_dest.
|
|
|
|
STATIC void emit_pre_pop_reg_flexible(emit_t *emit, vtype_kind_t *vtype, int *reg_dest, int not_r1, int not_r2) {
|
2014-09-29 18:45:42 +01:00
|
|
|
stack_info_t *si = peek_stack(emit, 0);
|
2015-01-24 23:14:12 +00:00
|
|
|
if (si->kind == STACK_REG && si->data.u_reg != not_r1 && si->data.u_reg != not_r2) {
|
2014-09-29 18:45:42 +01:00
|
|
|
*vtype = si->vtype;
|
2015-01-24 23:14:12 +00:00
|
|
|
*reg_dest = si->data.u_reg;
|
2014-09-29 18:45:42 +01:00
|
|
|
need_reg_single(emit, *reg_dest, 1);
|
|
|
|
} else {
|
|
|
|
emit_access_stack(emit, 1, vtype, *reg_dest);
|
|
|
|
}
|
|
|
|
adjust_stack(emit, -1);
|
|
|
|
}
|
|
|
|
|
2014-09-06 18:38:20 +01:00
|
|
|
STATIC void emit_pre_pop_discard(emit_t *emit) {
|
2014-05-10 13:40:46 +01:00
|
|
|
adjust_stack(emit, -1);
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_pre_pop_reg(emit_t *emit, vtype_kind_t *vtype, int reg_dest) {
|
2013-10-16 23:58:48 +01:00
|
|
|
emit_access_stack(emit, 1, vtype, reg_dest);
|
|
|
|
adjust_stack(emit, -1);
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_pre_pop_reg_reg(emit_t *emit, vtype_kind_t *vtypea, int rega, vtype_kind_t *vtypeb, int regb) {
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_pre_pop_reg(emit, vtypea, rega);
|
2013-10-08 22:18:32 +01:00
|
|
|
emit_pre_pop_reg(emit, vtypeb, regb);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_pre_pop_reg_reg_reg(emit_t *emit, vtype_kind_t *vtypea, int rega, vtype_kind_t *vtypeb, int regb, vtype_kind_t *vtypec, int regc) {
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_pre_pop_reg(emit, vtypea, rega);
|
2013-10-08 22:18:32 +01:00
|
|
|
emit_pre_pop_reg(emit, vtypeb, regb);
|
|
|
|
emit_pre_pop_reg(emit, vtypec, regc);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_post(emit_t *emit) {
|
2015-01-20 12:47:20 +00:00
|
|
|
(void)emit;
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-29 22:10:41 +01:00
|
|
|
STATIC void emit_post_top_set_vtype(emit_t *emit, vtype_kind_t new_vtype) {
|
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size - 1];
|
|
|
|
si->vtype = new_vtype;
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_post_push_reg(emit_t *emit, vtype_kind_t vtype, int reg) {
|
2018-08-07 07:05:44 +01:00
|
|
|
ensure_extra_stack(emit, 1);
|
2013-10-08 22:18:32 +01:00
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size];
|
|
|
|
si->vtype = vtype;
|
|
|
|
si->kind = STACK_REG;
|
2015-01-24 23:14:12 +00:00
|
|
|
si->data.u_reg = reg;
|
2013-10-08 22:18:32 +01:00
|
|
|
adjust_stack(emit, 1);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-07-03 13:25:24 +01:00
|
|
|
STATIC void emit_post_push_imm(emit_t *emit, vtype_kind_t vtype, mp_int_t imm) {
|
2018-08-07 07:05:44 +01:00
|
|
|
ensure_extra_stack(emit, 1);
|
2013-10-08 22:18:32 +01:00
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size];
|
|
|
|
si->vtype = vtype;
|
|
|
|
si->kind = STACK_IMM;
|
2015-01-24 23:14:12 +00:00
|
|
|
si->data.u_imm = imm;
|
2013-10-08 22:18:32 +01:00
|
|
|
adjust_stack(emit, 1);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_post_push_reg_reg(emit_t *emit, vtype_kind_t vtypea, int rega, vtype_kind_t vtypeb, int regb) {
|
2013-10-08 22:18:32 +01:00
|
|
|
emit_post_push_reg(emit, vtypea, rega);
|
|
|
|
emit_post_push_reg(emit, vtypeb, regb);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_post_push_reg_reg_reg(emit_t *emit, vtype_kind_t vtypea, int rega, vtype_kind_t vtypeb, int regb, vtype_kind_t vtypec, int regc) {
|
2013-10-08 22:18:32 +01:00
|
|
|
emit_post_push_reg(emit, vtypea, rega);
|
|
|
|
emit_post_push_reg(emit, vtypeb, regb);
|
|
|
|
emit_post_push_reg(emit, vtypec, regc);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_post_push_reg_reg_reg_reg(emit_t *emit, vtype_kind_t vtypea, int rega, vtype_kind_t vtypeb, int regb, vtype_kind_t vtypec, int regc, vtype_kind_t vtyped, int regd) {
|
2013-10-08 22:18:32 +01:00
|
|
|
emit_post_push_reg(emit, vtypea, rega);
|
|
|
|
emit_post_push_reg(emit, vtypeb, regb);
|
|
|
|
emit_post_push_reg(emit, vtypec, regc);
|
|
|
|
emit_post_push_reg(emit, vtyped, regd);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-08-16 22:31:57 +01:00
|
|
|
STATIC void emit_call(emit_t *emit, mp_fun_kind_t fun_kind) {
|
2013-10-16 23:58:48 +01:00
|
|
|
need_reg_all(emit);
|
2018-10-13 05:13:55 +01:00
|
|
|
ASM_CALL_IND(emit->as, fun_kind);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-08-16 22:31:57 +01:00
|
|
|
STATIC void emit_call_with_imm_arg(emit_t *emit, mp_fun_kind_t fun_kind, mp_int_t arg_val, int arg_reg) {
|
2013-10-10 22:06:54 +01:00
|
|
|
need_reg_all(emit);
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, arg_reg, arg_val);
|
2018-10-13 05:13:55 +01:00
|
|
|
ASM_CALL_IND(emit->as, fun_kind);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-08-16 22:31:57 +01:00
|
|
|
STATIC void emit_call_with_2_imm_args(emit_t *emit, mp_fun_kind_t fun_kind, mp_int_t arg_val1, int arg_reg1, mp_int_t arg_val2, int arg_reg2) {
|
2014-02-02 13:11:48 +00:00
|
|
|
need_reg_all(emit);
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, arg_reg1, arg_val1);
|
|
|
|
ASM_MOV_REG_IMM(emit->as, arg_reg2, arg_val2);
|
2018-10-13 05:13:55 +01:00
|
|
|
ASM_CALL_IND(emit->as, fun_kind);
|
2014-02-02 13:11:48 +00:00
|
|
|
}
|
|
|
|
|
2019-02-03 01:25:50 +00:00
|
|
|
STATIC void emit_call_with_qstr_arg(emit_t *emit, mp_fun_kind_t fun_kind, qstr qst, int arg_reg) {
|
|
|
|
need_reg_all(emit);
|
|
|
|
emit_native_mov_reg_qstr(emit, arg_reg, qst);
|
|
|
|
ASM_CALL_IND(emit->as, fun_kind);
|
|
|
|
}
|
|
|
|
|
2014-08-16 22:06:11 +01:00
|
|
|
// vtype of all n_pop objects is VTYPE_PYOBJ
|
|
|
|
// Will convert any items that are not VTYPE_PYOBJ to this type and put them back on the stack.
|
|
|
|
// If any conversions of non-immediate values are needed, then it uses REG_ARG_1, REG_ARG_2 and REG_RET.
|
|
|
|
// Otherwise, it does not use any temporary registers (but may use reg_dest before loading it with stack pointer).
|
|
|
|
STATIC void emit_get_stack_pointer_to_reg_for_pop(emit_t *emit, mp_uint_t reg_dest, mp_uint_t n_pop) {
|
|
|
|
need_reg_all(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2014-08-16 22:06:11 +01:00
|
|
|
// First, store any immediate values to their respective place on the stack.
|
|
|
|
for (mp_uint_t i = 0; i < n_pop; i++) {
|
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size - 1 - i];
|
|
|
|
// must push any imm's to stack
|
|
|
|
// must convert them to VTYPE_PYOBJ for viper code
|
|
|
|
if (si->kind == STACK_IMM) {
|
|
|
|
si->kind = STACK_VALUE;
|
2018-10-14 12:58:27 +01:00
|
|
|
si->vtype = load_reg_stack_imm(emit, reg_dest, si, true);
|
|
|
|
emit_native_mov_state_reg(emit, emit->stack_start + emit->stack_size - 1 - i, reg_dest);
|
2014-08-16 22:06:11 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
// verify that this value is on the stack
|
|
|
|
assert(si->kind == STACK_VALUE);
|
|
|
|
}
|
|
|
|
|
|
|
|
// Second, convert any non-VTYPE_PYOBJ to that type.
|
|
|
|
for (mp_uint_t i = 0; i < n_pop; i++) {
|
|
|
|
stack_info_t *si = &emit->stack_info[emit->stack_size - 1 - i];
|
|
|
|
if (si->vtype != VTYPE_PYOBJ) {
|
|
|
|
mp_uint_t local_num = emit->stack_start + emit->stack_size - 1 - i;
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state(emit, REG_ARG_1, local_num);
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_NATIVE_TO_OBJ, si->vtype, REG_ARG_2); // arg2 = type
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_state_reg(emit, local_num, REG_RET);
|
2014-08-16 22:06:11 +01:00
|
|
|
si->vtype = VTYPE_PYOBJ;
|
2014-09-29 22:10:41 +01:00
|
|
|
DEBUG_printf(" convert_native_to_obj(local_num=" UINT_FMT ")\n", local_num);
|
2014-08-16 22:06:11 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// Adujust the stack for a pop of n_pop items, and load the stack pointer into reg_dest.
|
|
|
|
adjust_stack(emit, -n_pop);
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state_addr(emit, reg_dest, emit->stack_start + emit->stack_size);
|
2014-08-16 22:06:11 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
// vtype of all n_push objects is VTYPE_PYOBJ
|
|
|
|
STATIC void emit_get_stack_pointer_to_reg_for_push(emit_t *emit, mp_uint_t reg_dest, mp_uint_t n_push) {
|
|
|
|
need_reg_all(emit);
|
2018-08-07 07:05:44 +01:00
|
|
|
ensure_extra_stack(emit, n_push);
|
2014-08-16 22:06:11 +01:00
|
|
|
for (mp_uint_t i = 0; i < n_push; i++) {
|
|
|
|
emit->stack_info[emit->stack_size + i].kind = STACK_VALUE;
|
|
|
|
emit->stack_info[emit->stack_size + i].vtype = VTYPE_PYOBJ;
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state_addr(emit, reg_dest, emit->stack_start + emit->stack_size);
|
2014-08-16 22:06:11 +01:00
|
|
|
adjust_stack(emit, n_push);
|
|
|
|
}
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
STATIC void emit_native_push_exc_stack(emit_t *emit, uint label, bool is_finally) {
|
|
|
|
if (emit->exc_stack_size + 1 > emit->exc_stack_alloc) {
|
|
|
|
size_t new_alloc = emit->exc_stack_alloc + 4;
|
|
|
|
emit->exc_stack = m_renew(exc_stack_entry_t, emit->exc_stack, emit->exc_stack_alloc, new_alloc);
|
|
|
|
emit->exc_stack_alloc = new_alloc;
|
|
|
|
}
|
|
|
|
|
|
|
|
exc_stack_entry_t *e = &emit->exc_stack[emit->exc_stack_size++];
|
|
|
|
e->label = label;
|
|
|
|
e->is_finally = is_finally;
|
2018-09-04 05:31:28 +01:00
|
|
|
e->unwind_label = UNWIND_LABEL_UNUSED;
|
|
|
|
e->is_active = true;
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_RET, label);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_PC(emit), REG_RET);
|
|
|
|
}
|
|
|
|
|
2018-09-04 05:31:28 +01:00
|
|
|
STATIC void emit_native_leave_exc_stack(emit_t *emit, bool start_of_handler) {
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
assert(emit->exc_stack_size > 0);
|
2018-09-04 05:31:28 +01:00
|
|
|
|
|
|
|
// Get current exception handler and deactivate it
|
|
|
|
exc_stack_entry_t *e = &emit->exc_stack[emit->exc_stack_size - 1];
|
|
|
|
e->is_active = false;
|
|
|
|
|
|
|
|
// Find next innermost active exception handler, to restore as current handler
|
|
|
|
for (--e; e >= emit->exc_stack && !e->is_active; --e) {
|
|
|
|
}
|
|
|
|
|
|
|
|
// Update the PC of the new exception handler
|
|
|
|
if (e < emit->exc_stack) {
|
|
|
|
// No active handler, clear handler PC to zero
|
|
|
|
if (start_of_handler) {
|
|
|
|
// Optimisation: PC is already cleared by global exc handler
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
ASM_XOR_REG_REG(emit->as, REG_RET, REG_RET);
|
|
|
|
} else {
|
2018-09-04 05:31:28 +01:00
|
|
|
// Found new active handler, get its PC
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_RET, e->label);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
}
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_PC(emit), REG_RET);
|
2018-09-04 05:31:28 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
STATIC exc_stack_entry_t *emit_native_pop_exc_stack(emit_t *emit) {
|
|
|
|
assert(emit->exc_stack_size > 0);
|
|
|
|
exc_stack_entry_t *e = &emit->exc_stack[--emit->exc_stack_size];
|
|
|
|
assert(e->is_active == false);
|
|
|
|
return e;
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
}
|
|
|
|
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
STATIC void emit_load_reg_with_object(emit_t *emit, int reg, mp_obj_t obj) {
|
2022-05-20 05:31:56 +01:00
|
|
|
emit->scope->scope_flags |= MP_SCOPE_FLAG_HASCONSTS;
|
2022-05-07 06:51:41 +01:00
|
|
|
size_t table_off = mp_emit_common_use_const_obj(emit->emit_common, obj);
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state(emit, REG_TEMP0, LOCAL_IDX_FUN_OBJ(emit));
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_TEMP0, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_TEMP0, OFFSETOF_MODULE_CONTEXT_OBJ_TABLE);
|
2018-09-27 14:27:53 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, reg, REG_TEMP0, table_off);
|
|
|
|
}
|
|
|
|
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
STATIC void emit_load_reg_with_child(emit_t *emit, int reg, mp_raw_code_t *rc) {
|
|
|
|
size_t table_off = mp_emit_common_alloc_const_child(emit->emit_common, rc);
|
|
|
|
emit_native_mov_reg_state(emit, REG_TEMP0, LOCAL_IDX_FUN_OBJ(emit));
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_TEMP0, OFFSETOF_OBJ_FUN_BC_CHILD_TABLE);
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, reg, REG_TEMP0, table_off);
|
2018-09-27 14:27:53 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_label_assign(emit_t *emit, mp_uint_t l) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("label_assign(" UINT_FMT ")\n", l);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
bool is_finally = false;
|
|
|
|
if (emit->exc_stack_size > 0) {
|
|
|
|
exc_stack_entry_t *e = &emit->exc_stack[emit->exc_stack_size - 1];
|
|
|
|
is_finally = e->is_finally && e->label == l;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (is_finally) {
|
|
|
|
// Label is at start of finally handler: store TOS into exception slot
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_TEMP0);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_VAL(emit), REG_TEMP0);
|
|
|
|
}
|
|
|
|
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2013-10-16 23:58:48 +01:00
|
|
|
// need to commit stack because we can jump here from elsewhere
|
|
|
|
need_stack_settled(emit);
|
2016-11-27 22:24:50 +00:00
|
|
|
mp_asm_base_label_assign(&emit->as->base, l);
|
2013-11-02 20:34:54 +00:00
|
|
|
emit_post(emit);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
if (is_finally) {
|
|
|
|
// Label is at start of finally handler: pop exception stack
|
2018-09-11 06:33:25 +01:00
|
|
|
emit_native_leave_exc_stack(emit, false);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_global_exc_entry(emit_t *emit) {
|
|
|
|
// Note: 4 labels are reserved for this function, starting at *emit->label_slot
|
|
|
|
|
|
|
|
emit->exit_label = *emit->label_slot;
|
|
|
|
|
|
|
|
if (NEED_GLOBAL_EXC_HANDLER(emit)) {
|
|
|
|
mp_uint_t nlr_label = *emit->label_slot + 1;
|
|
|
|
mp_uint_t start_label = *emit->label_slot + 2;
|
|
|
|
mp_uint_t global_except_label = *emit->label_slot + 3;
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
if (!(emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR)) {
|
|
|
|
// Set new globals
|
|
|
|
emit_native_mov_reg_state(emit, REG_ARG_1, LOCAL_IDX_FUN_OBJ(emit));
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_ARG_1, REG_ARG_1, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_ARG_1, REG_ARG_1, OFFSETOF_MODULE_CONTEXT_GLOBALS);
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_call(emit, MP_F_NATIVE_SWAP_GLOBALS);
|
2018-09-04 05:31:28 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Save old globals (or NULL if globals didn't change)
|
|
|
|
emit_native_mov_state_reg(emit, LOCAL_IDX_OLD_GLOBALS(emit), REG_RET);
|
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
if (emit->scope->exc_stack_size == 0) {
|
2018-10-01 04:07:04 +01:00
|
|
|
if (!(emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR)) {
|
|
|
|
// Optimisation: if globals didn't change don't push the nlr context
|
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_RET, start_label, false);
|
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
// Wrap everything in an nlr context
|
2018-10-01 03:36:21 +01:00
|
|
|
ASM_MOV_REG_LOCAL_ADDR(emit->as, REG_ARG_1, 0);
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
emit_call(emit, MP_F_NLR_PUSH);
|
2019-09-10 04:44:22 +01:00
|
|
|
#if N_NLR_SETJMP
|
|
|
|
ASM_MOV_REG_LOCAL_ADDR(emit->as, REG_ARG_1, 2);
|
|
|
|
emit_call(emit, MP_F_SETJMP);
|
|
|
|
#endif
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_RET, start_label, true);
|
|
|
|
} else {
|
|
|
|
// Clear the unwind state
|
|
|
|
ASM_XOR_REG_REG(emit->as, REG_TEMP0, REG_TEMP0);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_UNWIND(emit), REG_TEMP0);
|
|
|
|
|
|
|
|
// Put PC of start code block into REG_LOCAL_1
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_LOCAL_1, start_label);
|
|
|
|
|
|
|
|
// Wrap everything in an nlr context
|
|
|
|
emit_native_label_assign(emit, nlr_label);
|
2018-10-01 03:36:21 +01:00
|
|
|
ASM_MOV_REG_LOCAL_ADDR(emit->as, REG_ARG_1, 0);
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
emit_call(emit, MP_F_NLR_PUSH);
|
2019-09-10 04:44:22 +01:00
|
|
|
#if N_NLR_SETJMP
|
|
|
|
ASM_MOV_REG_LOCAL_ADDR(emit->as, REG_ARG_1, 2);
|
|
|
|
emit_call(emit, MP_F_SETJMP);
|
|
|
|
#endif
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
ASM_JUMP_IF_REG_NONZERO(emit->as, REG_RET, global_except_label, true);
|
|
|
|
|
|
|
|
// Clear PC of current code block, and jump there to resume execution
|
|
|
|
ASM_XOR_REG_REG(emit->as, REG_TEMP0, REG_TEMP0);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_PC(emit), REG_TEMP0);
|
|
|
|
ASM_JUMP_REG(emit->as, REG_LOCAL_1);
|
|
|
|
|
|
|
|
// Global exception handler: check for valid exception handler
|
|
|
|
emit_native_label_assign(emit, global_except_label);
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_LOCAL_1, LOCAL_IDX_EXC_HANDLER_PC(emit));
|
|
|
|
ASM_JUMP_IF_REG_NONZERO(emit->as, REG_LOCAL_1, nlr_label, false);
|
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
if (!(emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR)) {
|
|
|
|
// Restore old globals
|
|
|
|
emit_native_mov_reg_state(emit, REG_ARG_1, LOCAL_IDX_OLD_GLOBALS(emit));
|
|
|
|
emit_call(emit, MP_F_NATIVE_SWAP_GLOBALS);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
|
|
|
// Store return value in state[0]
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_TEMP0, LOCAL_IDX_EXC_VAL(emit));
|
2019-09-18 04:40:42 +01:00
|
|
|
ASM_STORE_REG_REG_OFFSET(emit->as, REG_TEMP0, REG_GENERATOR_STATE, OFFSETOF_CODE_STATE_STATE);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Load return kind
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_PARENT_RET, MP_VM_RETURN_EXCEPTION);
|
2018-10-01 04:07:04 +01:00
|
|
|
|
|
|
|
ASM_EXIT(emit->as);
|
|
|
|
} else {
|
|
|
|
// Re-raise exception out to caller
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_ARG_1, LOCAL_IDX_EXC_VAL(emit));
|
|
|
|
emit_call(emit, MP_F_NATIVE_RAISE);
|
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
// Label for start of function
|
|
|
|
emit_native_label_assign(emit, start_label);
|
2018-10-01 04:07:04 +01:00
|
|
|
|
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
|
|
|
emit_native_mov_reg_state(emit, REG_TEMP0, LOCAL_IDX_GEN_PC(emit));
|
|
|
|
ASM_JUMP_REG(emit->as, REG_TEMP0);
|
|
|
|
emit->start_offset = mp_asm_base_get_code_pos(&emit->as->base);
|
|
|
|
|
|
|
|
// This is the first entry of the generator
|
|
|
|
|
|
|
|
// Check LOCAL_IDX_EXC_VAL for any injected value
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_ARG_1, LOCAL_IDX_EXC_VAL(emit));
|
|
|
|
emit_call(emit, MP_F_NATIVE_RAISE);
|
|
|
|
}
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_global_exc_exit(emit_t *emit) {
|
|
|
|
// Label for end of function
|
|
|
|
emit_native_label_assign(emit, emit->exit_label);
|
|
|
|
|
|
|
|
if (NEED_GLOBAL_EXC_HANDLER(emit)) {
|
2018-09-15 13:37:07 +01:00
|
|
|
// Get old globals
|
2018-10-01 04:07:04 +01:00
|
|
|
if (!(emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR)) {
|
|
|
|
emit_native_mov_reg_state(emit, REG_ARG_1, LOCAL_IDX_OLD_GLOBALS(emit));
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
if (emit->scope->exc_stack_size == 0) {
|
|
|
|
// Optimisation: if globals didn't change then don't restore them and don't do nlr_pop
|
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_ARG_1, emit->exit_label + 1, false);
|
|
|
|
}
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
// Restore old globals
|
|
|
|
emit_call(emit, MP_F_NATIVE_SWAP_GLOBALS);
|
|
|
|
}
|
2018-09-15 13:37:07 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Pop the nlr context
|
|
|
|
emit_call(emit, MP_F_NLR_POP);
|
|
|
|
|
2018-10-01 04:07:04 +01:00
|
|
|
if (!(emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR)) {
|
|
|
|
if (emit->scope->exc_stack_size == 0) {
|
|
|
|
// Destination label for above optimisation
|
|
|
|
emit_native_label_assign(emit, emit->exit_label + 1);
|
|
|
|
}
|
py: Fix native functions so they run with their correct globals context.
Prior to this commit a function compiled with the native decorator
@micropython.native would not work correctly when accessing global
variables, because the globals dict was not being set upon function entry.
This commit fixes this problem by, upon function entry, setting as the
current globals dict the globals dict context the function was defined
within, as per normal Python semantics, and as bytecode does. Upon
function exit the original globals dict is restored.
In order to restore the globals dict when an exception is raised the native
function must guard its internals with an nlr_push/nlr_pop pair. Because
this push/pop is relatively expensive, in both C stack usage for the
nlr_buf_t and CPU execution time, the implementation here optimises things
as much as possible. First, the compiler keeps track of whether a function
even needs to access global variables. Using this information the native
emitter then generates three different kinds of code:
1. no globals used, no exception handlers: no nlr handling code and no
setting of the globals dict.
2. globals used, no exception handlers: an nlr_buf_t is allocated on the
C stack but it is not used if the globals dict is unchanged, saving
execution time because nlr_push/nlr_pop don't need to run.
3. function has exception handlers, may use globals: an nlr_buf_t is
allocated and nlr_push/nlr_pop are always called.
In the end, native functions that don't access globals and don't have
exception handlers will run more efficiently than those that do.
Fixes issue #1573.
2018-09-13 13:03:48 +01:00
|
|
|
}
|
|
|
|
|
2018-09-04 05:31:28 +01:00
|
|
|
// Load return value
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_PARENT_RET, LOCAL_IDX_RET_VAL(emit));
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
ASM_EXIT(emit->as);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-04-06 12:58:40 +01:00
|
|
|
STATIC void emit_native_import_name(emit_t *emit, qstr qst) {
|
|
|
|
DEBUG_printf("import_name %s\n", qstr_str(qst));
|
2016-11-10 11:53:04 +00:00
|
|
|
|
|
|
|
// get arguments from stack: arg2 = fromlist, arg3 = level
|
2018-10-14 13:06:09 +01:00
|
|
|
// If using viper types these arguments must be converted to proper objects, and
|
|
|
|
// to accomplish this viper types are turned off for the emit_pre_pop_reg_reg call.
|
|
|
|
bool orig_do_viper_types = emit->do_viper_types;
|
|
|
|
emit->do_viper_types = false;
|
|
|
|
vtype_kind_t vtype_fromlist;
|
|
|
|
vtype_kind_t vtype_level;
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_fromlist, REG_ARG_2, &vtype_level, REG_ARG_3);
|
|
|
|
assert(vtype_fromlist == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_level == VTYPE_PYOBJ);
|
|
|
|
emit->do_viper_types = orig_do_viper_types;
|
2016-11-10 11:53:04 +00:00
|
|
|
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_IMPORT_NAME, qst, REG_ARG_1); // arg1 = import name
|
2014-04-06 12:58:40 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-04-06 12:58:40 +01:00
|
|
|
STATIC void emit_native_import_from(emit_t *emit, qstr qst) {
|
|
|
|
DEBUG_printf("import_from %s\n", qstr_str(qst));
|
|
|
|
emit_native_pre(emit);
|
|
|
|
vtype_kind_t vtype_module;
|
|
|
|
emit_access_stack(emit, 1, &vtype_module, REG_ARG_1); // arg1 = module
|
|
|
|
assert(vtype_module == VTYPE_PYOBJ);
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_IMPORT_FROM, qst, REG_ARG_2); // arg2 = import name
|
2014-04-06 12:58:40 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_import_star(emit_t *emit) {
|
2014-04-06 12:58:40 +01:00
|
|
|
DEBUG_printf("import_star\n");
|
|
|
|
vtype_kind_t vtype_module;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_module, REG_ARG_1); // arg1 = module
|
|
|
|
assert(vtype_module == VTYPE_PYOBJ);
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call(emit, MP_F_IMPORT_ALL);
|
2014-04-06 12:58:40 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2018-05-22 12:58:25 +01:00
|
|
|
STATIC void emit_native_import(emit_t *emit, qstr qst, int kind) {
|
|
|
|
if (kind == MP_EMIT_IMPORT_NAME) {
|
|
|
|
emit_native_import_name(emit, qst);
|
|
|
|
} else if (kind == MP_EMIT_IMPORT_FROM) {
|
|
|
|
emit_native_import_from(emit, qst);
|
|
|
|
} else {
|
|
|
|
emit_native_import_star(emit);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_load_const_tok(emit_t *emit, mp_token_kind_t tok) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("load_const_tok(tok=%u)\n", tok);
|
2018-10-14 12:58:27 +01:00
|
|
|
if (tok == MP_TOKEN_ELLIPSIS) {
|
py: Add support to save native, viper and asm code to .mpy files.
This commit adds support for saving and loading .mpy files that contain
native code (native, viper and inline-asm). A lot of the ground work was
already done for this in the form of removing pointers from generated
native code. The changes here are mainly to link in qstr values to the
native code, and change the format of .mpy files to contain native code
blocks (possibly mixed with bytecode).
A top-level summary:
- @micropython.native, @micropython.viper and @micropython.asm_thumb/
asm_xtensa are now allowed in .py files when compiling to .mpy, and they
work transparently to the user.
- Entire .py files can be compiled to native via mpy-cross -X emit=native
and for the most part the generated .mpy files should work the same as
their bytecode version.
- The .mpy file format is changed to 1) specify in the header if the file
contains native code and if so the architecture (eg x86, ARMV7M, Xtensa);
2) for each function block the kind of code is specified (bytecode,
native, viper, asm).
- When native code is loaded from a .mpy file the native code must be
modified (in place) to link qstr values in, just like bytecode (see
py/persistentcode.c:arch_link_qstr() function).
In addition, this now defines a public, native ABI for dynamically loadable
native code generated by other languages, like C.
2019-02-21 04:18:33 +00:00
|
|
|
emit_native_load_const_obj(emit, MP_OBJ_FROM_PTR(&mp_const_ellipsis_obj));
|
2013-10-08 09:05:10 +01:00
|
|
|
} else {
|
2018-10-14 12:58:27 +01:00
|
|
|
emit_native_pre(emit);
|
|
|
|
if (tok == MP_TOKEN_KW_NONE) {
|
|
|
|
emit_post_push_imm(emit, VTYPE_PTR_NONE, 0);
|
|
|
|
} else {
|
|
|
|
emit_post_push_imm(emit, VTYPE_BOOL, tok == MP_TOKEN_KW_FALSE ? 0 : 1);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-07-03 13:25:24 +01:00
|
|
|
STATIC void emit_native_load_const_small_int(emit_t *emit, mp_int_t arg) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("load_const_small_int(int=" INT_FMT ")\n", arg);
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2018-10-14 12:58:27 +01:00
|
|
|
emit_post_push_imm(emit, VTYPE_INT, arg);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2015-06-25 15:42:13 +01:00
|
|
|
STATIC void emit_native_load_const_str(emit_t *emit, qstr qst) {
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2014-09-12 23:15:06 +01:00
|
|
|
// TODO: Eventually we want to be able to work with raw pointers in viper to
|
|
|
|
// do native array access. For now we just load them as any other object.
|
|
|
|
/*
|
2013-10-08 09:05:10 +01:00
|
|
|
if (emit->do_viper_types) {
|
|
|
|
// load a pointer to the asciiz string?
|
2014-09-08 23:05:16 +01:00
|
|
|
emit_post_push_imm(emit, VTYPE_PTR, (mp_uint_t)qstr_str(qst));
|
2014-09-12 23:15:06 +01:00
|
|
|
} else
|
|
|
|
*/
|
|
|
|
{
|
2019-02-03 01:25:50 +00:00
|
|
|
need_reg_single(emit, REG_TEMP0, 0);
|
|
|
|
emit_native_mov_reg_qstr_obj(emit, REG_TEMP0, qst);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_TEMP0);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-11-27 12:41:25 +00:00
|
|
|
STATIC void emit_native_load_const_obj(emit_t *emit, mp_obj_t obj) {
|
2015-01-13 15:55:54 +00:00
|
|
|
emit_native_pre(emit);
|
2015-01-14 00:11:09 +00:00
|
|
|
need_reg_single(emit, REG_RET, 0);
|
2018-09-27 14:27:53 +01:00
|
|
|
emit_load_reg_with_object(emit, REG_RET, obj);
|
2015-01-13 15:55:54 +00:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
}
|
|
|
|
|
2014-04-20 17:50:40 +01:00
|
|
|
STATIC void emit_native_load_null(emit_t *emit) {
|
|
|
|
emit_native_pre(emit);
|
|
|
|
emit_post_push_imm(emit, VTYPE_PYOBJ, 0);
|
|
|
|
}
|
|
|
|
|
2015-01-16 12:24:49 +00:00
|
|
|
STATIC void emit_native_load_fast(emit_t *emit, qstr qst, mp_uint_t local_num) {
|
|
|
|
DEBUG_printf("load_fast(%s, " UINT_FMT ")\n", qstr_str(qst), local_num);
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype = emit->local_vtype[local_num];
|
|
|
|
if (vtype == VTYPE_UNBOUND) {
|
2020-03-02 11:35:22 +00:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit, MP_ERROR_TEXT("local '%q' used before type known"), qst);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2022-05-20 05:31:56 +01:00
|
|
|
if (local_num < MAX_REGS_FOR_LOCAL_VARS && CAN_USE_REGS_FOR_LOCALS(emit)) {
|
2018-08-15 01:52:38 +01:00
|
|
|
emit_post_push_reg(emit, vtype, reg_local_table[local_num]);
|
2014-08-16 21:55:53 +01:00
|
|
|
} else {
|
2014-09-29 16:25:04 +01:00
|
|
|
need_reg_single(emit, REG_TEMP0, 0);
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_reg_state(emit, REG_TEMP0, LOCAL_IDX_LOCAL_VAR(emit, local_num));
|
2014-09-29 16:25:04 +01:00
|
|
|
emit_post_push_reg(emit, vtype, REG_TEMP0);
|
2014-08-16 21:55:53 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_load_deref(emit_t *emit, qstr qst, mp_uint_t local_num) {
|
2015-01-15 14:41:41 +00:00
|
|
|
DEBUG_printf("load_deref(%s, " UINT_FMT ")\n", qstr_str(qst), local_num);
|
|
|
|
need_reg_single(emit, REG_RET, 0);
|
|
|
|
emit_native_load_fast(emit, qst, local_num);
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
int reg_base = REG_RET;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype, ®_base, -1, -1);
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_RET, reg_base, 1);
|
|
|
|
// closed over vars are always Python objects
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-12-11 00:41:43 +00:00
|
|
|
}
|
|
|
|
|
2018-05-18 15:11:04 +01:00
|
|
|
STATIC void emit_native_load_local(emit_t *emit, qstr qst, mp_uint_t local_num, int kind) {
|
|
|
|
if (kind == MP_EMIT_IDOP_LOCAL_FAST) {
|
|
|
|
emit_native_load_fast(emit, qst, local_num);
|
|
|
|
} else {
|
|
|
|
emit_native_load_deref(emit, qst, local_num);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-05-22 12:16:30 +01:00
|
|
|
STATIC void emit_native_load_global(emit_t *emit, qstr qst, int kind) {
|
|
|
|
MP_STATIC_ASSERT(MP_F_LOAD_NAME + MP_EMIT_IDOP_GLOBAL_NAME == MP_F_LOAD_NAME);
|
|
|
|
MP_STATIC_ASSERT(MP_F_LOAD_NAME + MP_EMIT_IDOP_GLOBAL_GLOBAL == MP_F_LOAD_GLOBAL);
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2018-05-22 12:16:30 +01:00
|
|
|
if (kind == MP_EMIT_IDOP_GLOBAL_NAME) {
|
|
|
|
DEBUG_printf("load_name(%s)\n", qstr_str(qst));
|
2014-09-29 22:10:41 +01:00
|
|
|
} else {
|
2018-05-22 12:16:30 +01:00
|
|
|
DEBUG_printf("load_global(%s)\n", qstr_str(qst));
|
|
|
|
if (emit->do_viper_types) {
|
|
|
|
// check for builtin casting operators
|
2018-09-15 04:50:04 +01:00
|
|
|
int native_type = mp_native_type_from_qstr(qst);
|
2019-05-03 14:18:30 +01:00
|
|
|
if (native_type >= MP_NATIVE_TYPE_BOOL) {
|
2018-09-15 04:50:04 +01:00
|
|
|
emit_post_push_imm(emit, VTYPE_BUILTIN_CAST, native_type);
|
2018-05-22 12:16:30 +01:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
2014-09-29 22:10:41 +01:00
|
|
|
}
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_LOAD_NAME + kind, qst, REG_ARG_1);
|
2018-05-22 12:16:30 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_load_attr(emit_t *emit, qstr qst) {
|
2013-10-08 09:05:10 +01:00
|
|
|
// depends on type of subject:
|
|
|
|
// - integer, function, pointer to integers: error
|
|
|
|
// - pointer to structure: get member, quite easy
|
2014-03-30 13:35:08 +01:00
|
|
|
// - Python object: call mp_load_attr, and needs to be typed to convert result
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype_base;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1); // arg1 = base
|
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_LOAD_ATTR, qst, REG_ARG_2); // arg2 = attribute name
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
}
|
|
|
|
|
2017-04-19 00:45:59 +01:00
|
|
|
STATIC void emit_native_load_method(emit_t *emit, qstr qst, bool is_super) {
|
|
|
|
if (is_super) {
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_2, 3); // arg2 = dest ptr
|
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_2, 2); // arg2 = dest ptr
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_LOAD_SUPER_METHOD, qst, REG_ARG_1); // arg1 = method name
|
2017-04-19 00:45:59 +01:00
|
|
|
} else {
|
|
|
|
vtype_kind_t vtype_base;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1); // arg1 = base
|
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_3, 2); // arg3 = dest ptr
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_LOAD_METHOD, qst, REG_ARG_2); // arg2 = method name
|
2017-04-19 00:45:59 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_load_build_class(emit_t *emit) {
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call(emit, MP_F_LOAD_BUILD_CLASS);
|
2013-10-10 11:24:39 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-04-17 22:10:53 +01:00
|
|
|
STATIC void emit_native_load_subscr(emit_t *emit) {
|
2014-10-12 16:59:29 +01:00
|
|
|
DEBUG_printf("load_subscr\n");
|
|
|
|
// need to compile: base[index]
|
|
|
|
|
|
|
|
// pop: index, base
|
|
|
|
// optimise case where index is an immediate
|
|
|
|
vtype_kind_t vtype_base = peek_vtype(emit, 1);
|
|
|
|
|
|
|
|
if (vtype_base == VTYPE_PYOBJ) {
|
2015-06-04 11:52:16 +01:00
|
|
|
// standard Python subscr
|
|
|
|
// TODO factor this implicit cast code with other uses of it
|
|
|
|
vtype_kind_t vtype_index = peek_vtype(emit, 0);
|
|
|
|
if (vtype_index == VTYPE_PYOBJ) {
|
|
|
|
emit_pre_pop_reg(emit, &vtype_index, REG_ARG_2);
|
|
|
|
} else {
|
|
|
|
emit_pre_pop_reg(emit, &vtype_index, REG_ARG_1);
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_NATIVE_TO_OBJ, vtype_index, REG_ARG_2); // arg2 = type
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_2, REG_RET);
|
|
|
|
}
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1);
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_OBJ_SUBSCR, (mp_uint_t)MP_OBJ_SENTINEL, REG_ARG_3);
|
2014-04-17 22:10:53 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
} else {
|
2014-10-12 16:59:29 +01:00
|
|
|
// viper load
|
|
|
|
// TODO The different machine architectures have very different
|
|
|
|
// capabilities and requirements for loads, so probably best to
|
|
|
|
// write a completely separate load-optimiser for each one.
|
|
|
|
stack_info_t *top = peek_stack(emit, 0);
|
|
|
|
if (top->vtype == VTYPE_INT && top->kind == STACK_IMM) {
|
|
|
|
// index is an immediate
|
2015-01-24 23:14:12 +00:00
|
|
|
mp_int_t index_value = top->data.u_imm;
|
2014-10-12 16:59:29 +01:00
|
|
|
emit_pre_pop_discard(emit); // discard index
|
|
|
|
int reg_base = REG_ARG_1;
|
|
|
|
int reg_index = REG_ARG_2;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_base, ®_base, reg_index, reg_index);
|
2021-09-13 13:30:24 +01:00
|
|
|
need_reg_single(emit, REG_RET, 0);
|
2014-10-12 16:59:29 +01:00
|
|
|
switch (vtype_base) {
|
|
|
|
case VTYPE_PTR8: {
|
|
|
|
// pointer to 8-bit memory
|
|
|
|
// TODO optimise to use thumb ldrb r1, [r2, r3]
|
|
|
|
if (index_value != 0) {
|
|
|
|
// index is non-zero
|
|
|
|
#if N_THUMB
|
|
|
|
if (index_value > 0 && index_value < 32) {
|
|
|
|
asm_thumb_ldrb_rlo_rlo_i5(emit->as, REG_RET, reg_base, index_value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2022-11-11 01:25:32 +00:00
|
|
|
need_reg_single(emit, reg_index, 0);
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value);
|
2014-10-12 16:59:29 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, reg_index, reg_base); // add index to base
|
|
|
|
reg_base = reg_index;
|
|
|
|
}
|
|
|
|
ASM_LOAD8_REG_REG(emit->as, REG_RET, reg_base); // load from (base+index)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case VTYPE_PTR16: {
|
|
|
|
// pointer to 16-bit memory
|
|
|
|
if (index_value != 0) {
|
|
|
|
// index is a non-zero immediate
|
|
|
|
#if N_THUMB
|
|
|
|
if (index_value > 0 && index_value < 32) {
|
|
|
|
asm_thumb_ldrh_rlo_rlo_i5(emit->as, REG_RET, reg_base, index_value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2022-11-11 01:25:32 +00:00
|
|
|
need_reg_single(emit, reg_index, 0);
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value << 1);
|
2014-10-12 16:59:29 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, reg_index, reg_base); // add 2*index to base
|
|
|
|
reg_base = reg_index;
|
|
|
|
}
|
|
|
|
ASM_LOAD16_REG_REG(emit->as, REG_RET, reg_base); // load from (base+2*index)
|
|
|
|
break;
|
|
|
|
}
|
2015-10-13 00:50:17 +01:00
|
|
|
case VTYPE_PTR32: {
|
|
|
|
// pointer to 32-bit memory
|
|
|
|
if (index_value != 0) {
|
|
|
|
// index is a non-zero immediate
|
|
|
|
#if N_THUMB
|
|
|
|
if (index_value > 0 && index_value < 32) {
|
|
|
|
asm_thumb_ldr_rlo_rlo_i5(emit->as, REG_RET, reg_base, index_value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2022-11-11 01:25:32 +00:00
|
|
|
need_reg_single(emit, reg_index, 0);
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value << 2);
|
2015-10-13 00:50:17 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, reg_index, reg_base); // add 4*index to base
|
|
|
|
reg_base = reg_index;
|
|
|
|
}
|
|
|
|
ASM_LOAD32_REG_REG(emit->as, REG_RET, reg_base); // load from (base+4*index)
|
|
|
|
break;
|
|
|
|
}
|
2014-10-12 16:59:29 +01:00
|
|
|
default:
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't load from '%q'"), vtype_to_qstr(vtype_base));
|
2014-10-12 16:59:29 +01:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
// index is not an immediate
|
|
|
|
vtype_kind_t vtype_index;
|
|
|
|
int reg_index = REG_ARG_2;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_index, ®_index, REG_ARG_1, REG_ARG_1);
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1);
|
2021-09-13 13:30:24 +01:00
|
|
|
need_reg_single(emit, REG_RET, 0);
|
2016-02-09 13:46:49 +00:00
|
|
|
if (vtype_index != VTYPE_INT && vtype_index != VTYPE_UINT) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't load with '%q' index"), vtype_to_qstr(vtype_index));
|
2016-02-09 13:46:49 +00:00
|
|
|
}
|
2014-10-12 16:59:29 +01:00
|
|
|
switch (vtype_base) {
|
|
|
|
case VTYPE_PTR8: {
|
|
|
|
// pointer to 8-bit memory
|
|
|
|
// TODO optimise to use thumb ldrb r1, [r2, r3]
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_LOAD8_REG_REG(emit->as, REG_RET, REG_ARG_1); // store value to (base+index)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case VTYPE_PTR16: {
|
|
|
|
// pointer to 16-bit memory
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_LOAD16_REG_REG(emit->as, REG_RET, REG_ARG_1); // load from (base+2*index)
|
|
|
|
break;
|
|
|
|
}
|
2015-10-13 00:50:17 +01:00
|
|
|
case VTYPE_PTR32: {
|
|
|
|
// pointer to word-size memory
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_LOAD32_REG_REG(emit->as, REG_RET, REG_ARG_1); // load from (base+4*index)
|
|
|
|
break;
|
|
|
|
}
|
2014-10-12 16:59:29 +01:00
|
|
|
default:
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't load from '%q'"), vtype_to_qstr(vtype_base));
|
2014-10-12 16:59:29 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
emit_post_push_reg(emit, VTYPE_INT, REG_RET);
|
2014-04-17 22:10:53 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_store_fast(emit_t *emit, qstr qst, mp_uint_t local_num) {
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype;
|
2022-05-20 05:31:56 +01:00
|
|
|
if (local_num < MAX_REGS_FOR_LOCAL_VARS && CAN_USE_REGS_FOR_LOCALS(emit)) {
|
2018-08-15 01:52:38 +01:00
|
|
|
emit_pre_pop_reg(emit, &vtype, reg_local_table[local_num]);
|
2013-10-08 09:05:10 +01:00
|
|
|
} else {
|
2014-09-29 16:25:04 +01:00
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_TEMP0);
|
2018-10-01 04:07:04 +01:00
|
|
|
emit_native_mov_state_reg(emit, LOCAL_IDX_LOCAL_VAR(emit, local_num), REG_TEMP0);
|
2014-08-16 21:55:53 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post(emit);
|
|
|
|
|
|
|
|
// check types
|
|
|
|
if (emit->local_vtype[local_num] == VTYPE_UNBOUND) {
|
|
|
|
// first time this local is assigned, so give it a type of the object stored in it
|
|
|
|
emit->local_vtype[local_num] = vtype;
|
|
|
|
} else if (emit->local_vtype[local_num] != vtype) {
|
|
|
|
// type of local is not the same as object stored in it
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("local '%q' has type '%q' but source is '%q'"),
|
2015-04-20 14:29:31 +01:00
|
|
|
qst, vtype_to_qstr(emit->local_vtype[local_num]), vtype_to_qstr(vtype));
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_store_deref(emit_t *emit, qstr qst, mp_uint_t local_num) {
|
2015-01-15 14:41:41 +00:00
|
|
|
DEBUG_printf("store_deref(%s, " UINT_FMT ")\n", qstr_str(qst), local_num);
|
|
|
|
need_reg_single(emit, REG_TEMP0, 0);
|
|
|
|
need_reg_single(emit, REG_TEMP1, 0);
|
|
|
|
emit_native_load_fast(emit, qst, local_num);
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
int reg_base = REG_TEMP0;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype, ®_base, -1, -1);
|
|
|
|
int reg_src = REG_TEMP1;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype, ®_src, reg_base, reg_base);
|
|
|
|
ASM_STORE_REG_REG_OFFSET(emit->as, reg_src, reg_base, 1);
|
|
|
|
emit_post(emit);
|
2013-12-11 00:41:43 +00:00
|
|
|
}
|
|
|
|
|
2018-05-18 15:11:04 +01:00
|
|
|
STATIC void emit_native_store_local(emit_t *emit, qstr qst, mp_uint_t local_num, int kind) {
|
|
|
|
if (kind == MP_EMIT_IDOP_LOCAL_FAST) {
|
|
|
|
emit_native_store_fast(emit, qst, local_num);
|
|
|
|
} else {
|
|
|
|
emit_native_store_deref(emit, qst, local_num);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-05-22 12:16:30 +01:00
|
|
|
STATIC void emit_native_store_global(emit_t *emit, qstr qst, int kind) {
|
|
|
|
MP_STATIC_ASSERT(MP_F_STORE_NAME + MP_EMIT_IDOP_GLOBAL_NAME == MP_F_STORE_NAME);
|
|
|
|
MP_STATIC_ASSERT(MP_F_STORE_NAME + MP_EMIT_IDOP_GLOBAL_GLOBAL == MP_F_STORE_GLOBAL);
|
|
|
|
if (kind == MP_EMIT_IDOP_GLOBAL_NAME) {
|
|
|
|
// mp_store_name, but needs conversion of object (maybe have mp_viper_store_name(obj, type))
|
|
|
|
vtype_kind_t vtype;
|
2014-08-15 23:47:59 +01:00
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_2);
|
2018-05-22 12:16:30 +01:00
|
|
|
assert(vtype == VTYPE_PYOBJ);
|
2014-08-15 23:47:59 +01:00
|
|
|
} else {
|
2018-05-22 12:16:30 +01:00
|
|
|
vtype_kind_t vtype = peek_vtype(emit, 0);
|
|
|
|
if (vtype == VTYPE_PYOBJ) {
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_2);
|
|
|
|
} else {
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_1);
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_NATIVE_TO_OBJ, vtype, REG_ARG_2); // arg2 = type
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_2, REG_RET);
|
|
|
|
}
|
2014-08-15 23:47:59 +01:00
|
|
|
}
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_STORE_NAME + kind, qst, REG_ARG_1); // arg1 = name
|
2014-08-15 23:47:59 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_store_attr(emit_t *emit, qstr qst) {
|
2022-07-08 15:05:59 +01:00
|
|
|
vtype_kind_t vtype_base;
|
|
|
|
vtype_kind_t vtype_val = peek_vtype(emit, 1);
|
|
|
|
if (vtype_val == VTYPE_PYOBJ) {
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_base, REG_ARG_1, &vtype_val, REG_ARG_3); // arg1 = base, arg3 = value
|
|
|
|
} else {
|
|
|
|
emit_access_stack(emit, 2, &vtype_val, REG_ARG_1); // arg1 = value
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_NATIVE_TO_OBJ, vtype_val, REG_ARG_2); // arg2 = type
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_3, REG_RET); // arg3 = value (converted)
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1); // arg1 = base
|
|
|
|
adjust_stack(emit, -1); // pop value
|
|
|
|
}
|
2013-10-10 11:24:39 +01:00
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_STORE_ATTR, qst, REG_ARG_2); // arg2 = attribute name
|
2013-10-10 11:24:39 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_store_subscr(emit_t *emit) {
|
2014-09-29 22:10:41 +01:00
|
|
|
DEBUG_printf("store_subscr\n");
|
|
|
|
// need to compile: base[index] = value
|
|
|
|
|
|
|
|
// pop: index, base, value
|
|
|
|
// optimise case where index is an immediate
|
|
|
|
vtype_kind_t vtype_base = peek_vtype(emit, 1);
|
|
|
|
|
|
|
|
if (vtype_base == VTYPE_PYOBJ) {
|
2015-06-04 11:52:16 +01:00
|
|
|
// standard Python subscr
|
|
|
|
vtype_kind_t vtype_index = peek_vtype(emit, 0);
|
|
|
|
vtype_kind_t vtype_value = peek_vtype(emit, 2);
|
|
|
|
if (vtype_index != VTYPE_PYOBJ || vtype_value != VTYPE_PYOBJ) {
|
|
|
|
// need to implicitly convert non-objects to objects
|
|
|
|
// TODO do this properly
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_1, 3);
|
|
|
|
adjust_stack(emit, 3);
|
|
|
|
}
|
2014-09-29 22:10:41 +01:00
|
|
|
emit_pre_pop_reg_reg_reg(emit, &vtype_index, REG_ARG_2, &vtype_base, REG_ARG_1, &vtype_value, REG_ARG_3);
|
|
|
|
emit_call(emit, MP_F_OBJ_SUBSCR);
|
|
|
|
} else {
|
2014-09-29 22:41:41 +01:00
|
|
|
// viper store
|
|
|
|
// TODO The different machine architectures have very different
|
|
|
|
// capabilities and requirements for stores, so probably best to
|
|
|
|
// write a completely separate store-optimiser for each one.
|
2014-09-29 22:10:41 +01:00
|
|
|
stack_info_t *top = peek_stack(emit, 0);
|
|
|
|
if (top->vtype == VTYPE_INT && top->kind == STACK_IMM) {
|
|
|
|
// index is an immediate
|
2015-01-24 23:14:12 +00:00
|
|
|
mp_int_t index_value = top->data.u_imm;
|
2014-09-29 22:10:41 +01:00
|
|
|
emit_pre_pop_discard(emit); // discard index
|
|
|
|
vtype_kind_t vtype_value;
|
|
|
|
int reg_base = REG_ARG_1;
|
|
|
|
int reg_index = REG_ARG_2;
|
|
|
|
int reg_value = REG_ARG_3;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_base, ®_base, reg_index, reg_value);
|
2020-11-24 12:37:58 +00:00
|
|
|
#if N_X64 || N_X86
|
2014-09-29 22:41:41 +01:00
|
|
|
// special case: x86 needs byte stores to be from lower 4 regs (REG_ARG_3 is EDX)
|
|
|
|
emit_pre_pop_reg(emit, &vtype_value, reg_value);
|
|
|
|
#else
|
2014-09-29 22:10:41 +01:00
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_value, ®_value, reg_base, reg_index);
|
2014-09-29 22:41:41 +01:00
|
|
|
#endif
|
2016-02-15 09:38:02 +00:00
|
|
|
if (vtype_value != VTYPE_BOOL && vtype_value != VTYPE_INT && vtype_value != VTYPE_UINT) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't store '%q'"), vtype_to_qstr(vtype_value));
|
2016-02-15 09:38:02 +00:00
|
|
|
}
|
2014-09-29 22:10:41 +01:00
|
|
|
switch (vtype_base) {
|
|
|
|
case VTYPE_PTR8: {
|
|
|
|
// pointer to 8-bit memory
|
|
|
|
// TODO optimise to use thumb strb r1, [r2, r3]
|
|
|
|
if (index_value != 0) {
|
|
|
|
// index is non-zero
|
|
|
|
#if N_THUMB
|
|
|
|
if (index_value > 0 && index_value < 32) {
|
|
|
|
asm_thumb_strb_rlo_rlo_i5(emit->as, reg_value, reg_base, index_value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value);
|
2014-10-03 23:53:46 +01:00
|
|
|
#if N_ARM
|
|
|
|
asm_arm_strb_reg_reg_reg(emit->as, reg_value, reg_base, reg_index);
|
|
|
|
return;
|
|
|
|
#endif
|
2014-09-29 22:10:41 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, reg_index, reg_base); // add index to base
|
|
|
|
reg_base = reg_index;
|
|
|
|
}
|
|
|
|
ASM_STORE8_REG_REG(emit->as, reg_value, reg_base); // store value to (base+index)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case VTYPE_PTR16: {
|
|
|
|
// pointer to 16-bit memory
|
|
|
|
if (index_value != 0) {
|
|
|
|
// index is a non-zero immediate
|
|
|
|
#if N_THUMB
|
|
|
|
if (index_value > 0 && index_value < 32) {
|
|
|
|
asm_thumb_strh_rlo_rlo_i5(emit->as, reg_value, reg_base, index_value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value << 1);
|
2014-09-29 22:10:41 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, reg_index, reg_base); // add 2*index to base
|
|
|
|
reg_base = reg_index;
|
|
|
|
}
|
|
|
|
ASM_STORE16_REG_REG(emit->as, reg_value, reg_base); // store value to (base+2*index)
|
|
|
|
break;
|
|
|
|
}
|
2015-10-13 00:50:17 +01:00
|
|
|
case VTYPE_PTR32: {
|
|
|
|
// pointer to 32-bit memory
|
|
|
|
if (index_value != 0) {
|
|
|
|
// index is a non-zero immediate
|
|
|
|
#if N_THUMB
|
|
|
|
if (index_value > 0 && index_value < 32) {
|
|
|
|
asm_thumb_str_rlo_rlo_i5(emit->as, reg_value, reg_base, index_value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
#if N_ARM
|
2018-08-16 16:12:05 +01:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value);
|
2015-10-13 00:50:17 +01:00
|
|
|
asm_arm_str_reg_reg_reg(emit->as, reg_value, reg_base, reg_index);
|
|
|
|
return;
|
|
|
|
#endif
|
2018-08-16 16:12:05 +01:00
|
|
|
ASM_MOV_REG_IMM(emit->as, reg_index, index_value << 2);
|
2015-10-13 00:50:17 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, reg_index, reg_base); // add 4*index to base
|
|
|
|
reg_base = reg_index;
|
|
|
|
}
|
|
|
|
ASM_STORE32_REG_REG(emit->as, reg_value, reg_base); // store value to (base+4*index)
|
|
|
|
break;
|
|
|
|
}
|
2014-09-29 22:10:41 +01:00
|
|
|
default:
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't store to '%q'"), vtype_to_qstr(vtype_base));
|
2014-09-29 22:10:41 +01:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
// index is not an immediate
|
|
|
|
vtype_kind_t vtype_index, vtype_value;
|
|
|
|
int reg_index = REG_ARG_2;
|
|
|
|
int reg_value = REG_ARG_3;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_index, ®_index, REG_ARG_1, reg_value);
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1);
|
2016-02-09 13:46:49 +00:00
|
|
|
if (vtype_index != VTYPE_INT && vtype_index != VTYPE_UINT) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't store with '%q' index"), vtype_to_qstr(vtype_index));
|
2016-02-09 13:46:49 +00:00
|
|
|
}
|
2020-11-24 12:37:58 +00:00
|
|
|
#if N_X64 || N_X86
|
2014-09-29 22:41:41 +01:00
|
|
|
// special case: x86 needs byte stores to be from lower 4 regs (REG_ARG_3 is EDX)
|
|
|
|
emit_pre_pop_reg(emit, &vtype_value, reg_value);
|
|
|
|
#else
|
2014-09-29 22:10:41 +01:00
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_value, ®_value, REG_ARG_1, reg_index);
|
2014-09-29 22:41:41 +01:00
|
|
|
#endif
|
2016-02-15 09:38:02 +00:00
|
|
|
if (vtype_value != VTYPE_BOOL && vtype_value != VTYPE_INT && vtype_value != VTYPE_UINT) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't store '%q'"), vtype_to_qstr(vtype_value));
|
2016-02-15 09:38:02 +00:00
|
|
|
}
|
2014-09-29 22:10:41 +01:00
|
|
|
switch (vtype_base) {
|
|
|
|
case VTYPE_PTR8: {
|
|
|
|
// pointer to 8-bit memory
|
|
|
|
// TODO optimise to use thumb strb r1, [r2, r3]
|
2014-10-03 23:53:46 +01:00
|
|
|
#if N_ARM
|
|
|
|
asm_arm_strb_reg_reg_reg(emit->as, reg_value, REG_ARG_1, reg_index);
|
|
|
|
break;
|
|
|
|
#endif
|
2014-09-29 22:10:41 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_STORE8_REG_REG(emit->as, reg_value, REG_ARG_1); // store value to (base+index)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case VTYPE_PTR16: {
|
|
|
|
// pointer to 16-bit memory
|
2014-10-03 23:53:46 +01:00
|
|
|
#if N_ARM
|
|
|
|
asm_arm_strh_reg_reg_reg(emit->as, reg_value, REG_ARG_1, reg_index);
|
|
|
|
break;
|
|
|
|
#endif
|
2014-09-29 22:10:41 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_STORE16_REG_REG(emit->as, reg_value, REG_ARG_1); // store value to (base+2*index)
|
|
|
|
break;
|
|
|
|
}
|
2015-10-13 00:50:17 +01:00
|
|
|
case VTYPE_PTR32: {
|
|
|
|
// pointer to 32-bit memory
|
|
|
|
#if N_ARM
|
|
|
|
asm_arm_str_reg_reg_reg(emit->as, reg_value, REG_ARG_1, reg_index);
|
|
|
|
break;
|
|
|
|
#endif
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_1, reg_index); // add index to base
|
|
|
|
ASM_STORE32_REG_REG(emit->as, reg_value, REG_ARG_1); // store value to (base+4*index)
|
|
|
|
break;
|
|
|
|
}
|
2014-09-29 22:10:41 +01:00
|
|
|
default:
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't store to '%q'"), vtype_to_qstr(vtype_base));
|
2014-09-29 22:10:41 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2018-05-18 15:11:04 +01:00
|
|
|
STATIC void emit_native_delete_local(emit_t *emit, qstr qst, mp_uint_t local_num, int kind) {
|
|
|
|
if (kind == MP_EMIT_IDOP_LOCAL_FAST) {
|
|
|
|
// TODO: This is not compliant implementation. We could use MP_OBJ_SENTINEL
|
|
|
|
// to mark deleted vars but then every var would need to be checked on
|
|
|
|
// each access. Very inefficient, so just set value to None to enable GC.
|
|
|
|
emit_native_load_const_tok(emit, MP_TOKEN_KW_NONE);
|
|
|
|
emit_native_store_fast(emit, qst, local_num);
|
|
|
|
} else {
|
|
|
|
// TODO implement me!
|
|
|
|
}
|
2013-12-11 00:41:43 +00:00
|
|
|
}
|
|
|
|
|
2018-05-22 12:16:30 +01:00
|
|
|
STATIC void emit_native_delete_global(emit_t *emit, qstr qst, int kind) {
|
|
|
|
MP_STATIC_ASSERT(MP_F_DELETE_NAME + MP_EMIT_IDOP_GLOBAL_NAME == MP_F_DELETE_NAME);
|
|
|
|
MP_STATIC_ASSERT(MP_F_DELETE_NAME + MP_EMIT_IDOP_GLOBAL_GLOBAL == MP_F_DELETE_GLOBAL);
|
2014-09-06 18:38:20 +01:00
|
|
|
emit_native_pre(emit);
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_DELETE_NAME + kind, qst, REG_ARG_1);
|
2014-09-06 18:38:20 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-06 18:38:20 +01:00
|
|
|
STATIC void emit_native_delete_attr(emit_t *emit, qstr qst) {
|
2014-06-22 18:35:04 +01:00
|
|
|
vtype_kind_t vtype_base;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1); // arg1 = base
|
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
2019-02-03 01:25:50 +00:00
|
|
|
ASM_XOR_REG_REG(emit->as, REG_ARG_3, REG_ARG_3); // arg3 = value (null for delete)
|
|
|
|
emit_call_with_qstr_arg(emit, MP_F_STORE_ATTR, qst, REG_ARG_2); // arg2 = attribute name
|
2014-06-22 18:35:04 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_delete_subscr(emit_t *emit) {
|
2014-04-17 22:10:53 +01:00
|
|
|
vtype_kind_t vtype_index, vtype_base;
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_index, REG_ARG_2, &vtype_base, REG_ARG_1); // index, base
|
|
|
|
assert(vtype_index == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_OBJ_SUBSCR, (mp_uint_t)MP_OBJ_NULL, REG_ARG_3);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2018-05-22 12:31:56 +01:00
|
|
|
STATIC void emit_native_subscr(emit_t *emit, int kind) {
|
|
|
|
if (kind == MP_EMIT_SUBSCR_LOAD) {
|
|
|
|
emit_native_load_subscr(emit);
|
|
|
|
} else if (kind == MP_EMIT_SUBSCR_STORE) {
|
|
|
|
emit_native_store_subscr(emit);
|
|
|
|
} else {
|
|
|
|
emit_native_delete_subscr(emit);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-05-22 12:43:41 +01:00
|
|
|
STATIC void emit_native_attr(emit_t *emit, qstr qst, int kind) {
|
|
|
|
if (kind == MP_EMIT_ATTR_LOAD) {
|
|
|
|
emit_native_load_attr(emit, qst);
|
|
|
|
} else if (kind == MP_EMIT_ATTR_STORE) {
|
|
|
|
emit_native_store_attr(emit, qst);
|
|
|
|
} else {
|
|
|
|
emit_native_delete_attr(emit, qst);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_dup_top(emit_t *emit) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("dup_top\n");
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype;
|
2014-10-19 19:00:51 +01:00
|
|
|
int reg = REG_TEMP0;
|
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype, ®, -1, -1);
|
|
|
|
emit_post_push_reg_reg(emit, vtype, reg, vtype, reg);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_dup_top_two(emit_t *emit) {
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype0, vtype1;
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype0, REG_TEMP0, &vtype1, REG_TEMP1);
|
|
|
|
emit_post_push_reg_reg_reg_reg(emit, vtype1, REG_TEMP1, vtype0, REG_TEMP0, vtype1, REG_TEMP1, vtype0, REG_TEMP0);
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_pop_top(emit_t *emit) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("pop_top\n");
|
2014-09-06 18:38:20 +01:00
|
|
|
emit_pre_pop_discard(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post(emit);
|
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_rot_two(emit_t *emit) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("rot_two\n");
|
2013-10-08 22:18:32 +01:00
|
|
|
vtype_kind_t vtype0, vtype1;
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype0, REG_TEMP0, &vtype1, REG_TEMP1);
|
|
|
|
emit_post_push_reg_reg(emit, vtype0, REG_TEMP0, vtype1, REG_TEMP1);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_rot_three(emit_t *emit) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("rot_three\n");
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype0, vtype1, vtype2;
|
|
|
|
emit_pre_pop_reg_reg_reg(emit, &vtype0, REG_TEMP0, &vtype1, REG_TEMP1, &vtype2, REG_TEMP2);
|
|
|
|
emit_post_push_reg_reg_reg(emit, vtype0, REG_TEMP0, vtype2, REG_TEMP2, vtype1, REG_TEMP1);
|
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_jump(emit_t *emit, mp_uint_t label) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("jump(label=" UINT_FMT ")\n", label);
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2014-05-07 18:30:52 +01:00
|
|
|
// need to commit stack because we are jumping elsewhere
|
|
|
|
need_stack_settled(emit);
|
2014-09-06 23:06:36 +01:00
|
|
|
ASM_JUMP(emit->as, label);
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post(emit);
|
2022-06-17 14:14:32 +01:00
|
|
|
mp_asm_base_suppress_code(&emit->as->base);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2018-08-04 12:45:24 +01:00
|
|
|
STATIC void emit_native_jump_helper(emit_t *emit, bool cond, mp_uint_t label, bool pop) {
|
2014-09-29 18:45:42 +01:00
|
|
|
vtype_kind_t vtype = peek_vtype(emit, 0);
|
2015-04-20 14:29:31 +01:00
|
|
|
if (vtype == VTYPE_PYOBJ) {
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_1);
|
|
|
|
if (!pop) {
|
|
|
|
adjust_stack(emit, 1);
|
|
|
|
}
|
|
|
|
emit_call(emit, MP_F_OBJ_IS_TRUE);
|
|
|
|
} else {
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_RET);
|
|
|
|
if (!pop) {
|
|
|
|
adjust_stack(emit, 1);
|
|
|
|
}
|
|
|
|
if (!(vtype == VTYPE_BOOL || vtype == VTYPE_INT || vtype == VTYPE_UINT)) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't implicitly convert '%q' to 'bool'"), vtype_to_qstr(vtype));
|
2015-04-20 14:29:31 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-10-19 19:00:51 +01:00
|
|
|
// For non-pop need to save the vtype so that emit_native_adjust_stack_size
|
|
|
|
// can use it. This is a bit of a hack.
|
|
|
|
if (!pop) {
|
|
|
|
emit->saved_stack_vtype = vtype;
|
|
|
|
}
|
2014-05-07 18:30:52 +01:00
|
|
|
// need to commit stack because we may jump elsewhere
|
|
|
|
need_stack_settled(emit);
|
2018-08-04 12:45:24 +01:00
|
|
|
// Emit the jump
|
2015-02-28 15:04:06 +00:00
|
|
|
if (cond) {
|
2018-08-04 13:03:49 +01:00
|
|
|
ASM_JUMP_IF_REG_NONZERO(emit->as, REG_RET, label, vtype == VTYPE_PYOBJ);
|
2015-02-28 15:04:06 +00:00
|
|
|
} else {
|
2018-08-04 13:03:49 +01:00
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_RET, label, vtype == VTYPE_PYOBJ);
|
2015-02-28 15:04:06 +00:00
|
|
|
}
|
2018-08-04 12:45:24 +01:00
|
|
|
if (!pop) {
|
|
|
|
adjust_stack(emit, -1);
|
|
|
|
}
|
2013-11-03 13:58:19 +00:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2013-11-03 13:58:19 +00:00
|
|
|
|
2018-08-04 12:45:24 +01:00
|
|
|
STATIC void emit_native_pop_jump_if(emit_t *emit, bool cond, mp_uint_t label) {
|
|
|
|
DEBUG_printf("pop_jump_if(cond=%u, label=" UINT_FMT ")\n", cond, label);
|
|
|
|
emit_native_jump_helper(emit, cond, label, true);
|
|
|
|
}
|
|
|
|
|
2015-02-28 15:04:06 +00:00
|
|
|
STATIC void emit_native_jump_if_or_pop(emit_t *emit, bool cond, mp_uint_t label) {
|
|
|
|
DEBUG_printf("jump_if_or_pop(cond=%u, label=" UINT_FMT ")\n", cond, label);
|
2018-08-04 12:45:24 +01:00
|
|
|
emit_native_jump_helper(emit, cond, label, false);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2018-05-22 12:50:22 +01:00
|
|
|
STATIC void emit_native_unwind_jump(emit_t *emit, mp_uint_t label, mp_uint_t except_depth) {
|
2018-09-04 05:31:28 +01:00
|
|
|
if (except_depth > 0) {
|
|
|
|
exc_stack_entry_t *first_finally = NULL;
|
|
|
|
exc_stack_entry_t *prev_finally = NULL;
|
|
|
|
exc_stack_entry_t *e = &emit->exc_stack[emit->exc_stack_size - 1];
|
|
|
|
for (; except_depth > 0; --except_depth, --e) {
|
|
|
|
if (e->is_finally && e->is_active) {
|
|
|
|
// Found an active finally handler
|
|
|
|
if (first_finally == NULL) {
|
|
|
|
first_finally = e;
|
|
|
|
}
|
|
|
|
if (prev_finally != NULL) {
|
|
|
|
// Mark prev finally as needed to unwind a jump
|
|
|
|
prev_finally->unwind_label = e->label;
|
|
|
|
}
|
|
|
|
prev_finally = e;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (prev_finally == NULL) {
|
|
|
|
// No finally, handle the jump ourselves
|
|
|
|
// First, restore the exception handler address for the jump
|
|
|
|
if (e < emit->exc_stack) {
|
|
|
|
ASM_XOR_REG_REG(emit->as, REG_RET, REG_RET);
|
|
|
|
} else {
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_RET, e->label);
|
|
|
|
}
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_PC(emit), REG_RET);
|
|
|
|
} else {
|
|
|
|
// Last finally should do our jump for us
|
|
|
|
// Mark finally as needing to decide the type of jump
|
|
|
|
prev_finally->unwind_label = UNWIND_LABEL_DO_FINAL_UNWIND;
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_RET, label & ~MP_EMIT_BREAK_FROM_FOR);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_UNWIND(emit), REG_RET);
|
2019-02-15 01:18:59 +00:00
|
|
|
// Cancel any active exception (see also emit_native_pop_except_jump)
|
2020-01-24 07:01:20 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_RET, (mp_uint_t)MP_OBJ_NULL);
|
2018-09-04 05:31:28 +01:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_VAL(emit), REG_RET);
|
|
|
|
// Jump to the innermost active finally
|
|
|
|
label = first_finally->label;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
emit_native_jump(emit, label & ~MP_EMIT_BREAK_FROM_FOR);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-05-07 18:30:52 +01:00
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_setup_with(emit_t *emit, mp_uint_t label) {
|
2016-04-07 08:53:24 +01:00
|
|
|
// the context manager is on the top of the stack
|
|
|
|
// stack: (..., ctx_mgr)
|
|
|
|
|
|
|
|
// get __exit__ method
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
emit_access_stack(emit, 1, &vtype, REG_ARG_1); // arg1 = ctx_mgr
|
|
|
|
assert(vtype == VTYPE_PYOBJ);
|
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_3, 2); // arg3 = dest ptr
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_LOAD_METHOD, MP_QSTR___exit__, REG_ARG_2);
|
2016-04-07 08:53:24 +01:00
|
|
|
// stack: (..., ctx_mgr, __exit__, self)
|
|
|
|
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_3); // self
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_2); // __exit__
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_1); // ctx_mgr
|
|
|
|
emit_post_push_reg(emit, vtype, REG_ARG_2); // __exit__
|
|
|
|
emit_post_push_reg(emit, vtype, REG_ARG_3); // self
|
|
|
|
// stack: (..., __exit__, self)
|
|
|
|
// REG_ARG_1=ctx_mgr
|
|
|
|
|
|
|
|
// get __enter__ method
|
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_3, 2); // arg3 = dest ptr
|
2019-02-03 01:25:50 +00:00
|
|
|
emit_call_with_qstr_arg(emit, MP_F_LOAD_METHOD, MP_QSTR___enter__, REG_ARG_2); // arg2 = method name
|
2016-04-07 08:53:24 +01:00
|
|
|
// stack: (..., __exit__, self, __enter__, self)
|
|
|
|
|
|
|
|
// call __enter__ method
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 2); // pointer to items, including meth and self
|
|
|
|
emit_call_with_2_imm_args(emit, MP_F_CALL_METHOD_N_KW, 0, REG_ARG_1, 0, REG_ARG_2);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET); // push return value of __enter__
|
|
|
|
// stack: (..., __exit__, self, as_value)
|
|
|
|
|
|
|
|
// need to commit stack because we may jump elsewhere
|
|
|
|
need_stack_settled(emit);
|
2018-09-04 05:31:28 +01:00
|
|
|
emit_native_push_exc_stack(emit, label, true);
|
2016-04-07 08:53:24 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_dup_top(emit);
|
|
|
|
// stack: (..., __exit__, self, as_value, as_value)
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-06-30 05:17:25 +01:00
|
|
|
|
2018-05-22 13:33:26 +01:00
|
|
|
STATIC void emit_native_setup_block(emit_t *emit, mp_uint_t label, int kind) {
|
|
|
|
if (kind == MP_EMIT_SETUP_BLOCK_WITH) {
|
|
|
|
emit_native_setup_with(emit, label);
|
|
|
|
} else {
|
|
|
|
// Set up except and finally
|
|
|
|
emit_native_pre(emit);
|
|
|
|
need_stack_settled(emit);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_push_exc_stack(emit, label, kind == MP_EMIT_SETUP_BLOCK_FINALLY);
|
2018-05-22 13:33:26 +01:00
|
|
|
emit_post(emit);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-04-07 08:50:38 +01:00
|
|
|
STATIC void emit_native_with_cleanup(emit_t *emit, mp_uint_t label) {
|
2018-09-04 05:31:28 +01:00
|
|
|
// Note: 3 labels are reserved for this function, starting at *emit->label_slot
|
2016-04-07 08:53:24 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// stack: (..., __exit__, self, as_value)
|
2016-04-07 08:53:24 +01:00
|
|
|
emit_native_pre(emit);
|
2018-09-04 05:31:28 +01:00
|
|
|
emit_native_leave_exc_stack(emit, false);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
adjust_stack(emit, -1);
|
2016-04-07 08:53:24 +01:00
|
|
|
// stack: (..., __exit__, self)
|
|
|
|
|
2018-09-04 05:31:28 +01:00
|
|
|
// Label for case where __exit__ is called from an unwind jump
|
|
|
|
emit_native_label_assign(emit, *emit->label_slot + 2);
|
|
|
|
|
2016-04-07 08:53:24 +01:00
|
|
|
// call __exit__
|
2018-10-14 13:23:25 +01:00
|
|
|
emit_post_push_imm(emit, VTYPE_PTR_NONE, 0);
|
|
|
|
emit_post_push_imm(emit, VTYPE_PTR_NONE, 0);
|
|
|
|
emit_post_push_imm(emit, VTYPE_PTR_NONE, 0);
|
2016-04-07 08:53:24 +01:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 5);
|
|
|
|
emit_call_with_2_imm_args(emit, MP_F_CALL_METHOD_N_KW, 3, REG_ARG_1, 0, REG_ARG_2);
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Replace exc with None and finish
|
|
|
|
emit_native_jump(emit, *emit->label_slot);
|
2016-04-07 08:53:24 +01:00
|
|
|
|
|
|
|
// nlr_catch
|
2018-09-04 05:31:28 +01:00
|
|
|
// Don't use emit_native_label_assign because this isn't a real finally label
|
|
|
|
mp_asm_base_label_assign(&emit->as->base, label);
|
2016-04-07 08:53:24 +01:00
|
|
|
|
2018-09-04 05:31:28 +01:00
|
|
|
// Leave with's exception handler
|
|
|
|
emit_native_leave_exc_stack(emit, true);
|
2016-04-07 08:53:24 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Adjust stack counter for: __exit__, self (implicitly discard as_value which is above self)
|
|
|
|
emit_native_adjust_stack_size(emit, 2);
|
2016-04-07 08:53:24 +01:00
|
|
|
// stack: (..., __exit__, self)
|
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_ARG_1, LOCAL_IDX_EXC_VAL(emit)); // get exc
|
2018-09-04 05:31:28 +01:00
|
|
|
|
2020-01-24 07:01:20 +00:00
|
|
|
// Check if exc is MP_OBJ_NULL (i.e. zero) and jump to non-exc handler if it is
|
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_ARG_1, *emit->label_slot + 2, false);
|
2018-09-04 05:31:28 +01:00
|
|
|
|
2016-04-07 08:53:24 +01:00
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_ARG_2, REG_ARG_1, 0); // get type(exc)
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_ARG_2); // push type(exc)
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_ARG_1); // push exc value
|
2018-10-14 13:23:25 +01:00
|
|
|
emit_post_push_imm(emit, VTYPE_PTR_NONE, 0); // traceback info
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Stack: (..., __exit__, self, type(exc), exc, traceback)
|
2016-04-07 08:53:24 +01:00
|
|
|
|
|
|
|
// call __exit__ method
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 5);
|
|
|
|
emit_call_with_2_imm_args(emit, MP_F_CALL_METHOD_N_KW, 3, REG_ARG_1, 0, REG_ARG_2);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// Stack: (...)
|
2016-04-07 08:53:24 +01:00
|
|
|
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
// If REG_RET is true then we need to replace exception with None (swallow exception)
|
2016-04-07 08:53:24 +01:00
|
|
|
if (REG_ARG_1 != REG_RET) {
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_1, REG_RET);
|
|
|
|
}
|
|
|
|
emit_call(emit, MP_F_OBJ_IS_TRUE);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_RET, *emit->label_slot + 1, true);
|
2016-04-07 08:53:24 +01:00
|
|
|
|
2020-01-24 07:01:20 +00:00
|
|
|
// Replace exception with MP_OBJ_NULL.
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_label_assign(emit, *emit->label_slot);
|
2020-01-24 07:01:20 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_TEMP0, (mp_uint_t)MP_OBJ_NULL);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_VAL(emit), REG_TEMP0);
|
2016-04-07 08:53:24 +01:00
|
|
|
|
|
|
|
// end of with cleanup nlr_catch block
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_label_assign(emit, *emit->label_slot + 1);
|
|
|
|
|
|
|
|
// Exception is in nlr_buf.ret_val slot
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-06-30 05:17:25 +01:00
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_end_finally(emit_t *emit) {
|
2015-01-21 17:00:01 +00:00
|
|
|
// logic:
|
|
|
|
// exc = pop_stack
|
|
|
|
// if exc == None: pass
|
|
|
|
// else: raise exc
|
|
|
|
// the check if exc is None is done in the MP_F_NATIVE_RAISE stub
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
emit_native_pre(emit);
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_ARG_1, LOCAL_IDX_EXC_VAL(emit));
|
2015-01-21 17:00:01 +00:00
|
|
|
emit_call(emit, MP_F_NATIVE_RAISE);
|
2018-09-04 05:31:28 +01:00
|
|
|
|
|
|
|
// Get state for this finally and see if we need to unwind
|
|
|
|
exc_stack_entry_t *e = emit_native_pop_exc_stack(emit);
|
|
|
|
if (e->unwind_label != UNWIND_LABEL_UNUSED) {
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_RET, LOCAL_IDX_EXC_HANDLER_UNWIND(emit));
|
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_RET, *emit->label_slot, false);
|
|
|
|
if (e->unwind_label == UNWIND_LABEL_DO_FINAL_UNWIND) {
|
|
|
|
ASM_JUMP_REG(emit->as, REG_RET);
|
|
|
|
} else {
|
|
|
|
emit_native_jump(emit, e->unwind_label);
|
|
|
|
}
|
|
|
|
emit_native_label_assign(emit, *emit->label_slot);
|
|
|
|
}
|
|
|
|
|
2014-09-06 18:38:20 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2013-10-16 23:58:48 +01:00
|
|
|
|
2016-01-09 23:59:52 +00:00
|
|
|
STATIC void emit_native_get_iter(emit_t *emit, bool use_stack) {
|
2013-10-08 09:05:10 +01:00
|
|
|
// perhaps the difficult one, as we want to rewrite for loops using native code
|
|
|
|
// in cases where we iterate over a Python object, can we use normal runtime calls?
|
2013-10-16 23:58:48 +01:00
|
|
|
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_1);
|
|
|
|
assert(vtype == VTYPE_PYOBJ);
|
2016-01-09 23:59:52 +00:00
|
|
|
if (use_stack) {
|
2017-03-23 05:36:08 +00:00
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_2, MP_OBJ_ITER_BUF_NSLOTS);
|
2017-01-17 04:27:37 +00:00
|
|
|
emit_call(emit, MP_F_NATIVE_GETITER);
|
2016-01-09 23:59:52 +00:00
|
|
|
} else {
|
|
|
|
// mp_getiter will allocate the iter_buf on the heap
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_2, 0);
|
2017-01-17 04:27:37 +00:00
|
|
|
emit_call(emit, MP_F_NATIVE_GETITER);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2016-01-09 23:59:52 +00:00
|
|
|
}
|
2013-10-16 23:58:48 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_for_iter(emit_t *emit, mp_uint_t label) {
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2017-03-23 05:36:08 +00:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_1, MP_OBJ_ITER_BUF_NSLOTS);
|
|
|
|
adjust_stack(emit, MP_OBJ_ITER_BUF_NSLOTS);
|
2017-01-17 04:27:37 +00:00
|
|
|
emit_call(emit, MP_F_NATIVE_ITERNEXT);
|
2019-03-08 04:48:20 +00:00
|
|
|
#if MICROPY_DEBUG_MP_OBJ_SENTINELS
|
2017-11-15 00:46:49 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_TEMP1, (mp_uint_t)MP_OBJ_STOP_ITERATION);
|
2014-09-06 23:06:36 +01:00
|
|
|
ASM_JUMP_IF_REG_EQ(emit->as, REG_RET, REG_TEMP1, label);
|
2019-03-08 04:48:20 +00:00
|
|
|
#else
|
|
|
|
MP_STATIC_ASSERT(MP_OBJ_STOP_ITERATION == 0);
|
|
|
|
ASM_JUMP_IF_REG_ZERO(emit->as, REG_RET, label, false);
|
2018-07-12 09:08:01 +01:00
|
|
|
#endif
|
2013-10-16 23:58:48 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2013-10-16 23:58:48 +01:00
|
|
|
|
2017-01-17 04:30:18 +00:00
|
|
|
STATIC void emit_native_for_iter_end(emit_t *emit) {
|
2013-10-16 23:58:48 +01:00
|
|
|
// adjust stack counter (we get here from for_iter ending, which popped the value for us)
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2017-03-23 05:36:08 +00:00
|
|
|
adjust_stack(emit, -MP_OBJ_ITER_BUF_NSLOTS);
|
2013-10-16 23:58:48 +01:00
|
|
|
emit_post(emit);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2019-02-15 01:18:59 +00:00
|
|
|
STATIC void emit_native_pop_except_jump(emit_t *emit, mp_uint_t label, bool within_exc_handler) {
|
|
|
|
if (within_exc_handler) {
|
|
|
|
// Cancel any active exception so subsequent handlers don't see it
|
2020-01-24 07:01:20 +00:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_TEMP0, (mp_uint_t)MP_OBJ_NULL);
|
2019-02-15 01:18:59 +00:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_VAL(emit), REG_TEMP0);
|
|
|
|
} else {
|
2018-09-04 05:31:28 +01:00
|
|
|
emit_native_leave_exc_stack(emit, false);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
}
|
2019-02-15 01:18:59 +00:00
|
|
|
emit_native_jump(emit, label);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-03-30 13:35:08 +01:00
|
|
|
STATIC void emit_native_unary_op(emit_t *emit, mp_unary_op_t op) {
|
2014-09-23 15:10:03 +01:00
|
|
|
vtype_kind_t vtype;
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_2);
|
2015-10-08 13:08:59 +01:00
|
|
|
if (vtype == VTYPE_PYOBJ) {
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_UNARY_OP, op, REG_ARG_1);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
} else {
|
|
|
|
adjust_stack(emit, 1);
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("unary op %q not implemented"), mp_unary_op_method_name[op]);
|
2014-06-30 05:17:25 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-03-30 13:35:08 +01:00
|
|
|
STATIC void emit_native_binary_op(emit_t *emit, mp_binary_op_t op) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("binary_op(" UINT_FMT ")\n", op);
|
2014-09-29 18:45:42 +01:00
|
|
|
vtype_kind_t vtype_lhs = peek_vtype(emit, 1);
|
|
|
|
vtype_kind_t vtype_rhs = peek_vtype(emit, 0);
|
2020-06-26 09:26:39 +01:00
|
|
|
if ((vtype_lhs == VTYPE_INT || vtype_lhs == VTYPE_UINT)
|
|
|
|
&& (vtype_rhs == VTYPE_INT || vtype_rhs == VTYPE_UINT)) {
|
2017-10-12 02:26:49 +01:00
|
|
|
// for integers, inplace and normal ops are equivalent, so use just normal ops
|
|
|
|
if (MP_BINARY_OP_INPLACE_OR <= op && op <= MP_BINARY_OP_INPLACE_POWER) {
|
|
|
|
op += MP_BINARY_OP_OR - MP_BINARY_OP_INPLACE_OR;
|
|
|
|
}
|
|
|
|
|
2014-09-29 18:45:42 +01:00
|
|
|
#if N_X64 || N_X86
|
|
|
|
// special cases for x86 and shifting
|
2017-10-12 02:26:49 +01:00
|
|
|
if (op == MP_BINARY_OP_LSHIFT || op == MP_BINARY_OP_RSHIFT) {
|
2014-09-29 18:45:42 +01:00
|
|
|
#if N_X64
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_rhs, ASM_X64_REG_RCX, &vtype_lhs, REG_RET);
|
|
|
|
#else
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_rhs, ASM_X86_REG_ECX, &vtype_lhs, REG_RET);
|
|
|
|
#endif
|
2017-10-12 02:26:49 +01:00
|
|
|
if (op == MP_BINARY_OP_LSHIFT) {
|
2014-09-29 18:45:42 +01:00
|
|
|
ASM_LSL_REG(emit->as, REG_RET);
|
|
|
|
} else {
|
2020-06-26 09:26:39 +01:00
|
|
|
if (vtype_lhs == VTYPE_UINT) {
|
|
|
|
ASM_LSR_REG(emit->as, REG_RET);
|
|
|
|
} else {
|
|
|
|
ASM_ASR_REG(emit->as, REG_RET);
|
|
|
|
}
|
2014-09-29 18:45:42 +01:00
|
|
|
}
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_RET);
|
2014-09-29 18:45:42 +01:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
#endif
|
2017-10-11 08:54:34 +01:00
|
|
|
|
|
|
|
// special cases for floor-divide and module because we dispatch to helper functions
|
2017-10-12 02:26:49 +01:00
|
|
|
if (op == MP_BINARY_OP_FLOOR_DIVIDE || op == MP_BINARY_OP_MODULO) {
|
2017-10-11 08:54:34 +01:00
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_rhs, REG_ARG_2, &vtype_lhs, REG_ARG_1);
|
2020-06-26 09:26:39 +01:00
|
|
|
if (vtype_lhs != VTYPE_INT) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
|
|
|
MP_ERROR_TEXT("div/mod not implemented for uint"), mp_binary_op_method_name[op]);
|
|
|
|
}
|
2017-10-12 02:26:49 +01:00
|
|
|
if (op == MP_BINARY_OP_FLOOR_DIVIDE) {
|
2017-10-11 08:54:34 +01:00
|
|
|
emit_call(emit, MP_F_SMALL_INT_FLOOR_DIVIDE);
|
|
|
|
} else {
|
|
|
|
emit_call(emit, MP_F_SMALL_INT_MODULO);
|
|
|
|
}
|
|
|
|
emit_post_push_reg(emit, VTYPE_INT, REG_RET);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2014-09-29 18:45:42 +01:00
|
|
|
int reg_rhs = REG_ARG_3;
|
2014-09-29 22:10:41 +01:00
|
|
|
emit_pre_pop_reg_flexible(emit, &vtype_rhs, ®_rhs, REG_RET, REG_ARG_2);
|
2014-09-29 18:45:42 +01:00
|
|
|
emit_pre_pop_reg(emit, &vtype_lhs, REG_ARG_2);
|
2020-02-19 13:17:13 +00:00
|
|
|
|
2014-09-29 18:45:42 +01:00
|
|
|
#if !(N_X64 || N_X86)
|
2020-02-19 13:17:13 +00:00
|
|
|
if (op == MP_BINARY_OP_LSHIFT || op == MP_BINARY_OP_RSHIFT) {
|
|
|
|
if (op == MP_BINARY_OP_LSHIFT) {
|
|
|
|
ASM_LSL_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
|
|
|
} else {
|
2020-06-26 09:26:39 +01:00
|
|
|
if (vtype_lhs == VTYPE_UINT) {
|
|
|
|
ASM_LSR_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
|
|
|
} else {
|
|
|
|
ASM_ASR_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
|
|
|
}
|
2020-02-19 13:17:13 +00:00
|
|
|
}
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2020-02-19 13:17:13 +00:00
|
|
|
return;
|
|
|
|
}
|
2014-09-29 18:45:42 +01:00
|
|
|
#endif
|
2020-02-19 13:17:13 +00:00
|
|
|
|
2019-05-05 04:29:43 +01:00
|
|
|
if (op == MP_BINARY_OP_OR) {
|
2014-10-12 14:21:06 +01:00
|
|
|
ASM_OR_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2017-10-12 02:26:49 +01:00
|
|
|
} else if (op == MP_BINARY_OP_XOR) {
|
2014-10-12 14:21:06 +01:00
|
|
|
ASM_XOR_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2017-10-12 02:26:49 +01:00
|
|
|
} else if (op == MP_BINARY_OP_AND) {
|
2014-10-12 14:21:06 +01:00
|
|
|
ASM_AND_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2017-10-12 02:26:49 +01:00
|
|
|
} else if (op == MP_BINARY_OP_ADD) {
|
2014-09-29 18:45:42 +01:00
|
|
|
ASM_ADD_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2017-10-12 02:26:49 +01:00
|
|
|
} else if (op == MP_BINARY_OP_SUBTRACT) {
|
2014-09-29 18:45:42 +01:00
|
|
|
ASM_SUB_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2017-10-12 02:26:49 +01:00
|
|
|
} else if (op == MP_BINARY_OP_MULTIPLY) {
|
2015-06-04 15:00:29 +01:00
|
|
|
ASM_MUL_REG_REG(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
emit_post_push_reg(emit, vtype_lhs, REG_ARG_2);
|
2023-02-25 16:01:28 +00:00
|
|
|
} else if (op == MP_BINARY_OP_LESS
|
|
|
|
|| op == MP_BINARY_OP_MORE
|
|
|
|
|| op == MP_BINARY_OP_EQUAL
|
|
|
|
|| op == MP_BINARY_OP_LESS_EQUAL
|
|
|
|
|| op == MP_BINARY_OP_MORE_EQUAL
|
|
|
|
|| op == MP_BINARY_OP_NOT_EQUAL) {
|
|
|
|
// comparison ops
|
2020-06-26 09:26:39 +01:00
|
|
|
|
|
|
|
if (vtype_lhs != vtype_rhs) {
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit, MP_ERROR_TEXT("comparison of int and uint"));
|
|
|
|
}
|
|
|
|
|
|
|
|
size_t op_idx = op - MP_BINARY_OP_LESS + (vtype_lhs == VTYPE_UINT ? 0 : 6);
|
|
|
|
|
2014-10-19 19:00:51 +01:00
|
|
|
need_reg_single(emit, REG_RET, 0);
|
2014-09-29 18:45:42 +01:00
|
|
|
#if N_X64
|
|
|
|
asm_x64_xor_r64_r64(emit->as, REG_RET, REG_RET);
|
|
|
|
asm_x64_cmp_r64_with_r64(emit->as, reg_rhs, REG_ARG_2);
|
2020-06-26 09:26:39 +01:00
|
|
|
static byte ops[6 + 6] = {
|
|
|
|
// unsigned
|
|
|
|
ASM_X64_CC_JB,
|
|
|
|
ASM_X64_CC_JA,
|
|
|
|
ASM_X64_CC_JE,
|
|
|
|
ASM_X64_CC_JBE,
|
|
|
|
ASM_X64_CC_JAE,
|
|
|
|
ASM_X64_CC_JNE,
|
|
|
|
// signed
|
2014-09-29 18:45:42 +01:00
|
|
|
ASM_X64_CC_JL,
|
|
|
|
ASM_X64_CC_JG,
|
|
|
|
ASM_X64_CC_JE,
|
|
|
|
ASM_X64_CC_JLE,
|
|
|
|
ASM_X64_CC_JGE,
|
|
|
|
ASM_X64_CC_JNE,
|
|
|
|
};
|
2020-06-26 09:26:39 +01:00
|
|
|
asm_x64_setcc_r8(emit->as, ops[op_idx], REG_RET);
|
2014-09-29 18:45:42 +01:00
|
|
|
#elif N_X86
|
|
|
|
asm_x86_xor_r32_r32(emit->as, REG_RET, REG_RET);
|
|
|
|
asm_x86_cmp_r32_with_r32(emit->as, reg_rhs, REG_ARG_2);
|
2020-06-26 09:26:39 +01:00
|
|
|
static byte ops[6 + 6] = {
|
|
|
|
// unsigned
|
|
|
|
ASM_X86_CC_JB,
|
|
|
|
ASM_X86_CC_JA,
|
|
|
|
ASM_X86_CC_JE,
|
|
|
|
ASM_X86_CC_JBE,
|
|
|
|
ASM_X86_CC_JAE,
|
|
|
|
ASM_X86_CC_JNE,
|
|
|
|
// signed
|
2014-09-29 18:45:42 +01:00
|
|
|
ASM_X86_CC_JL,
|
|
|
|
ASM_X86_CC_JG,
|
|
|
|
ASM_X86_CC_JE,
|
|
|
|
ASM_X86_CC_JLE,
|
|
|
|
ASM_X86_CC_JGE,
|
|
|
|
ASM_X86_CC_JNE,
|
|
|
|
};
|
2020-06-26 09:26:39 +01:00
|
|
|
asm_x86_setcc_r8(emit->as, ops[op_idx], REG_RET);
|
2014-09-29 18:45:42 +01:00
|
|
|
#elif N_THUMB
|
|
|
|
asm_thumb_cmp_rlo_rlo(emit->as, REG_ARG_2, reg_rhs);
|
2022-05-23 08:57:33 +01:00
|
|
|
if (asm_thumb_allow_armv7m(emit->as)) {
|
|
|
|
static uint16_t ops[6 + 6] = {
|
|
|
|
// unsigned
|
|
|
|
ASM_THUMB_OP_ITE_CC,
|
|
|
|
ASM_THUMB_OP_ITE_HI,
|
|
|
|
ASM_THUMB_OP_ITE_EQ,
|
|
|
|
ASM_THUMB_OP_ITE_LS,
|
|
|
|
ASM_THUMB_OP_ITE_CS,
|
|
|
|
ASM_THUMB_OP_ITE_NE,
|
|
|
|
// signed
|
|
|
|
ASM_THUMB_OP_ITE_LT,
|
|
|
|
ASM_THUMB_OP_ITE_GT,
|
|
|
|
ASM_THUMB_OP_ITE_EQ,
|
|
|
|
ASM_THUMB_OP_ITE_LE,
|
|
|
|
ASM_THUMB_OP_ITE_GE,
|
|
|
|
ASM_THUMB_OP_ITE_NE,
|
|
|
|
};
|
|
|
|
asm_thumb_op16(emit->as, ops[op_idx]);
|
|
|
|
asm_thumb_mov_rlo_i8(emit->as, REG_RET, 1);
|
|
|
|
asm_thumb_mov_rlo_i8(emit->as, REG_RET, 0);
|
|
|
|
} else {
|
|
|
|
static uint16_t ops[6 + 6] = {
|
|
|
|
// unsigned
|
|
|
|
ASM_THUMB_CC_CC,
|
|
|
|
ASM_THUMB_CC_HI,
|
|
|
|
ASM_THUMB_CC_EQ,
|
|
|
|
ASM_THUMB_CC_LS,
|
|
|
|
ASM_THUMB_CC_CS,
|
|
|
|
ASM_THUMB_CC_NE,
|
|
|
|
// signed
|
|
|
|
ASM_THUMB_CC_LT,
|
|
|
|
ASM_THUMB_CC_GT,
|
|
|
|
ASM_THUMB_CC_EQ,
|
|
|
|
ASM_THUMB_CC_LE,
|
|
|
|
ASM_THUMB_CC_GE,
|
|
|
|
ASM_THUMB_CC_NE,
|
|
|
|
};
|
|
|
|
asm_thumb_bcc_rel9(emit->as, ops[op_idx], 6);
|
|
|
|
asm_thumb_mov_rlo_i8(emit->as, REG_RET, 0);
|
|
|
|
asm_thumb_b_rel12(emit->as, 4);
|
|
|
|
asm_thumb_mov_rlo_i8(emit->as, REG_RET, 1);
|
|
|
|
}
|
2014-09-29 18:45:42 +01:00
|
|
|
#elif N_ARM
|
2014-10-03 23:53:46 +01:00
|
|
|
asm_arm_cmp_reg_reg(emit->as, REG_ARG_2, reg_rhs);
|
2020-06-26 09:26:39 +01:00
|
|
|
static uint ccs[6 + 6] = {
|
|
|
|
// unsigned
|
|
|
|
ASM_ARM_CC_CC,
|
|
|
|
ASM_ARM_CC_HI,
|
|
|
|
ASM_ARM_CC_EQ,
|
|
|
|
ASM_ARM_CC_LS,
|
|
|
|
ASM_ARM_CC_CS,
|
|
|
|
ASM_ARM_CC_NE,
|
|
|
|
// signed
|
2014-10-03 23:53:46 +01:00
|
|
|
ASM_ARM_CC_LT,
|
|
|
|
ASM_ARM_CC_GT,
|
|
|
|
ASM_ARM_CC_EQ,
|
|
|
|
ASM_ARM_CC_LE,
|
|
|
|
ASM_ARM_CC_GE,
|
|
|
|
ASM_ARM_CC_NE,
|
|
|
|
};
|
2020-06-26 09:26:39 +01:00
|
|
|
asm_arm_setcc_reg(emit->as, REG_RET, ccs[op_idx]);
|
2019-09-13 04:15:12 +01:00
|
|
|
#elif N_XTENSA || N_XTENSAWIN
|
2020-06-26 09:26:39 +01:00
|
|
|
static uint8_t ccs[6 + 6] = {
|
|
|
|
// unsigned
|
|
|
|
ASM_XTENSA_CC_LTU,
|
|
|
|
0x80 | ASM_XTENSA_CC_LTU, // for GTU we'll swap args
|
|
|
|
ASM_XTENSA_CC_EQ,
|
|
|
|
0x80 | ASM_XTENSA_CC_GEU, // for LEU we'll swap args
|
|
|
|
ASM_XTENSA_CC_GEU,
|
|
|
|
ASM_XTENSA_CC_NE,
|
|
|
|
// signed
|
2016-12-09 05:39:39 +00:00
|
|
|
ASM_XTENSA_CC_LT,
|
|
|
|
0x80 | ASM_XTENSA_CC_LT, // for GT we'll swap args
|
|
|
|
ASM_XTENSA_CC_EQ,
|
|
|
|
0x80 | ASM_XTENSA_CC_GE, // for LE we'll swap args
|
|
|
|
ASM_XTENSA_CC_GE,
|
|
|
|
ASM_XTENSA_CC_NE,
|
|
|
|
};
|
2020-06-26 09:26:39 +01:00
|
|
|
uint8_t cc = ccs[op_idx];
|
2016-12-09 05:39:39 +00:00
|
|
|
if ((cc & 0x80) == 0) {
|
|
|
|
asm_xtensa_setcc_reg_reg_reg(emit->as, cc, REG_RET, REG_ARG_2, reg_rhs);
|
|
|
|
} else {
|
|
|
|
asm_xtensa_setcc_reg_reg_reg(emit->as, cc & ~0x80, REG_RET, reg_rhs, REG_ARG_2);
|
|
|
|
}
|
2014-09-29 18:45:42 +01:00
|
|
|
#else
|
|
|
|
#error not implemented
|
|
|
|
#endif
|
2014-01-11 09:47:06 +00:00
|
|
|
emit_post_push_reg(emit, VTYPE_BOOL, REG_RET);
|
|
|
|
} else {
|
|
|
|
// TODO other ops not yet implemented
|
2015-06-04 11:52:16 +01:00
|
|
|
adjust_stack(emit, 1);
|
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("binary op %q not implemented"), mp_binary_op_method_name[op]);
|
2014-01-11 09:47:06 +00:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
} else if (vtype_lhs == VTYPE_PYOBJ && vtype_rhs == VTYPE_PYOBJ) {
|
2014-09-29 18:45:42 +01:00
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_rhs, REG_ARG_3, &vtype_lhs, REG_ARG_2);
|
2014-09-23 15:10:03 +01:00
|
|
|
bool invert = false;
|
|
|
|
if (op == MP_BINARY_OP_NOT_IN) {
|
|
|
|
invert = true;
|
|
|
|
op = MP_BINARY_OP_IN;
|
|
|
|
} else if (op == MP_BINARY_OP_IS_NOT) {
|
|
|
|
invert = true;
|
|
|
|
op = MP_BINARY_OP_IS;
|
|
|
|
}
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_BINARY_OP, op, REG_ARG_1);
|
2014-09-23 15:10:03 +01:00
|
|
|
if (invert) {
|
2014-09-29 18:45:42 +01:00
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_2, REG_RET);
|
2014-09-23 15:10:03 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_UNARY_OP, MP_UNARY_OP_NOT, REG_ARG_1);
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
} else {
|
2015-04-22 23:16:03 +01:00
|
|
|
adjust_stack(emit, -1);
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("can't do binary op between '%q' and '%q'"),
|
2015-04-20 14:29:31 +01:00
|
|
|
vtype_to_qstr(vtype_lhs), vtype_to_qstr(vtype_rhs));
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-05-22 13:18:42 +01:00
|
|
|
#if MICROPY_PY_BUILTINS_SLICE
|
|
|
|
STATIC void emit_native_build_slice(emit_t *emit, mp_uint_t n_args);
|
|
|
|
#endif
|
|
|
|
|
2018-05-18 15:41:40 +01:00
|
|
|
STATIC void emit_native_build(emit_t *emit, mp_uint_t n_args, int kind) {
|
2013-10-16 23:58:48 +01:00
|
|
|
// for viper: call runtime, with types of args
|
|
|
|
// if wrapped in byte_array, or something, allocates memory and fills it
|
2018-05-18 15:41:40 +01:00
|
|
|
MP_STATIC_ASSERT(MP_F_BUILD_TUPLE + MP_EMIT_BUILD_TUPLE == MP_F_BUILD_TUPLE);
|
|
|
|
MP_STATIC_ASSERT(MP_F_BUILD_TUPLE + MP_EMIT_BUILD_LIST == MP_F_BUILD_LIST);
|
|
|
|
MP_STATIC_ASSERT(MP_F_BUILD_TUPLE + MP_EMIT_BUILD_MAP == MP_F_BUILD_MAP);
|
2018-05-22 13:18:42 +01:00
|
|
|
MP_STATIC_ASSERT(MP_F_BUILD_TUPLE + MP_EMIT_BUILD_SET == MP_F_BUILD_SET);
|
|
|
|
#if MICROPY_PY_BUILTINS_SLICE
|
|
|
|
if (kind == MP_EMIT_BUILD_SLICE) {
|
|
|
|
emit_native_build_slice(emit, n_args);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
#endif
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
2018-05-22 13:18:42 +01:00
|
|
|
if (kind == MP_EMIT_BUILD_TUPLE || kind == MP_EMIT_BUILD_LIST || kind == MP_EMIT_BUILD_SET) {
|
2018-05-18 15:41:40 +01:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_2, n_args); // pointer to items
|
|
|
|
}
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_BUILD_TUPLE + kind, n_args, REG_ARG_1);
|
2018-05-22 13:18:42 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET); // new tuple/list/map/set
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_store_map(emit_t *emit) {
|
2013-10-08 09:05:10 +01:00
|
|
|
vtype_kind_t vtype_key, vtype_value, vtype_map;
|
|
|
|
emit_pre_pop_reg_reg_reg(emit, &vtype_key, REG_ARG_2, &vtype_value, REG_ARG_3, &vtype_map, REG_ARG_1); // key, value, map
|
|
|
|
assert(vtype_key == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_value == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_map == VTYPE_PYOBJ);
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call(emit, MP_F_STORE_MAP);
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET); // map
|
|
|
|
}
|
|
|
|
|
2014-12-27 17:20:41 +00:00
|
|
|
#if MICROPY_PY_BUILTINS_SLICE
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_build_slice(emit_t *emit, mp_uint_t n_args) {
|
2014-04-06 12:58:40 +01:00
|
|
|
DEBUG_printf("build_slice %d\n", n_args);
|
2014-06-30 05:17:25 +01:00
|
|
|
if (n_args == 2) {
|
|
|
|
vtype_kind_t vtype_start, vtype_stop;
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_stop, REG_ARG_2, &vtype_start, REG_ARG_1); // arg1 = start, arg2 = stop
|
|
|
|
assert(vtype_start == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_stop == VTYPE_PYOBJ);
|
2018-10-14 13:56:46 +01:00
|
|
|
emit_native_mov_reg_const(emit, REG_ARG_3, MP_F_CONST_NONE_OBJ); // arg3 = step
|
2014-06-30 05:17:25 +01:00
|
|
|
} else {
|
|
|
|
assert(n_args == 3);
|
|
|
|
vtype_kind_t vtype_start, vtype_stop, vtype_step;
|
|
|
|
emit_pre_pop_reg_reg_reg(emit, &vtype_step, REG_ARG_3, &vtype_stop, REG_ARG_2, &vtype_start, REG_ARG_1); // arg1 = start, arg2 = stop, arg3 = step
|
|
|
|
assert(vtype_start == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_stop == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_step == VTYPE_PYOBJ);
|
|
|
|
}
|
2018-10-14 13:56:46 +01:00
|
|
|
emit_call(emit, MP_F_NEW_SLICE);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-12-27 17:20:41 +00:00
|
|
|
#endif
|
2014-04-06 12:58:40 +01:00
|
|
|
|
2016-09-18 14:59:47 +01:00
|
|
|
STATIC void emit_native_store_comp(emit_t *emit, scope_kind_t kind, mp_uint_t collection_index) {
|
|
|
|
mp_fun_kind_t f;
|
|
|
|
if (kind == SCOPE_LIST_COMP) {
|
|
|
|
vtype_kind_t vtype_item;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_item, REG_ARG_2);
|
|
|
|
assert(vtype_item == VTYPE_PYOBJ);
|
|
|
|
f = MP_F_LIST_APPEND;
|
|
|
|
#if MICROPY_PY_BUILTINS_SET
|
|
|
|
} else if (kind == SCOPE_SET_COMP) {
|
|
|
|
vtype_kind_t vtype_item;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_item, REG_ARG_2);
|
|
|
|
assert(vtype_item == VTYPE_PYOBJ);
|
|
|
|
f = MP_F_STORE_SET;
|
|
|
|
#endif
|
|
|
|
} else {
|
|
|
|
// SCOPE_DICT_COMP
|
|
|
|
vtype_kind_t vtype_key, vtype_value;
|
|
|
|
emit_pre_pop_reg_reg(emit, &vtype_key, REG_ARG_2, &vtype_value, REG_ARG_3);
|
|
|
|
assert(vtype_key == VTYPE_PYOBJ);
|
|
|
|
assert(vtype_value == VTYPE_PYOBJ);
|
|
|
|
f = MP_F_STORE_MAP;
|
|
|
|
}
|
|
|
|
vtype_kind_t vtype_collection;
|
|
|
|
emit_access_stack(emit, collection_index, &vtype_collection, REG_ARG_1);
|
|
|
|
assert(vtype_collection == VTYPE_PYOBJ);
|
|
|
|
emit_call(emit, f);
|
|
|
|
emit_post(emit);
|
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_unpack_sequence(emit_t *emit, mp_uint_t n_args) {
|
2014-04-06 12:58:40 +01:00
|
|
|
DEBUG_printf("unpack_sequence %d\n", n_args);
|
|
|
|
vtype_kind_t vtype_base;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1); // arg1 = seq
|
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_3, n_args); // arg3 = dest ptr
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_UNPACK_SEQUENCE, n_args, REG_ARG_2); // arg2 = n_args
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-04-06 12:58:40 +01:00
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_unpack_ex(emit_t *emit, mp_uint_t n_left, mp_uint_t n_right) {
|
2014-05-07 18:30:52 +01:00
|
|
|
DEBUG_printf("unpack_ex %d %d\n", n_left, n_right);
|
|
|
|
vtype_kind_t vtype_base;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_base, REG_ARG_1); // arg1 = seq
|
|
|
|
assert(vtype_base == VTYPE_PYOBJ);
|
2014-06-30 05:17:25 +01:00
|
|
|
emit_get_stack_pointer_to_reg_for_push(emit, REG_ARG_3, n_left + n_right + 1); // arg3 = dest ptr
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_UNPACK_EX, n_left | (n_right << 8), REG_ARG_2); // arg2 = n_left + n_right
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_make_function(emit_t *emit, scope_t *scope, mp_uint_t n_pos_defaults, mp_uint_t n_kw_defaults) {
|
2013-10-08 09:05:10 +01:00
|
|
|
// call runtime, with type info for args, or don't support dict/default params, or only support Python objects for them
|
2014-03-27 23:30:26 +00:00
|
|
|
emit_native_pre(emit);
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_native_mov_reg_state(emit, REG_ARG_2, LOCAL_IDX_FUN_OBJ(emit));
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_ARG_2, REG_ARG_2, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
2014-05-07 18:30:52 +01:00
|
|
|
if (n_pos_defaults == 0 && n_kw_defaults == 0) {
|
2018-09-27 14:27:53 +01:00
|
|
|
need_reg_all(emit);
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_3, 0);
|
2014-05-07 18:30:52 +01:00
|
|
|
} else {
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 2);
|
2018-09-27 14:27:53 +01:00
|
|
|
need_reg_all(emit);
|
2014-05-07 18:30:52 +01:00
|
|
|
}
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_load_reg_with_child(emit, REG_ARG_1, scope->raw_code);
|
2018-10-13 05:13:55 +01:00
|
|
|
ASM_CALL_IND(emit->as, MP_F_MAKE_FUNCTION_FROM_RAW_CODE);
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_make_closure(emit_t *emit, scope_t *scope, mp_uint_t n_closed_over, mp_uint_t n_pos_defaults, mp_uint_t n_kw_defaults) {
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
// make function
|
2015-01-15 14:41:41 +00:00
|
|
|
emit_native_pre(emit);
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_native_mov_reg_state(emit, REG_ARG_2, LOCAL_IDX_FUN_OBJ(emit));
|
|
|
|
ASM_LOAD_REG_REG_OFFSET(emit->as, REG_ARG_2, REG_ARG_2, OFFSETOF_OBJ_FUN_BC_CONTEXT);
|
2015-01-15 14:41:41 +00:00
|
|
|
if (n_pos_defaults == 0 && n_kw_defaults == 0) {
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
need_reg_all(emit);
|
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_3, 0);
|
2015-01-15 14:41:41 +00:00
|
|
|
} else {
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 2 + n_closed_over);
|
|
|
|
adjust_stack(emit, 2 + n_closed_over);
|
|
|
|
need_reg_all(emit);
|
|
|
|
}
|
|
|
|
emit_load_reg_with_child(emit, REG_ARG_1, scope->raw_code);
|
|
|
|
ASM_CALL_IND(emit->as, MP_F_MAKE_FUNCTION_FROM_RAW_CODE);
|
|
|
|
|
|
|
|
// make closure
|
|
|
|
#if REG_ARG_1 != REG_RET
|
|
|
|
ASM_MOV_REG_REG(emit->as, REG_ARG_1, REG_RET);
|
|
|
|
#endif
|
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_2, n_closed_over);
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, n_closed_over);
|
|
|
|
if (n_pos_defaults != 0 || n_kw_defaults != 0) {
|
|
|
|
adjust_stack(emit, -2);
|
2015-01-15 14:41:41 +00:00
|
|
|
}
|
py: Rework bytecode and .mpy file format to be mostly static data.
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
2021-10-22 12:22:47 +01:00
|
|
|
ASM_CALL_IND(emit->as, MP_F_NEW_CLOSURE);
|
2015-01-15 14:41:41 +00:00
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_call_function(emit_t *emit, mp_uint_t n_positional, mp_uint_t n_keyword, mp_uint_t star_flags) {
|
2014-09-23 15:10:03 +01:00
|
|
|
DEBUG_printf("call_function(n_pos=" UINT_FMT ", n_kw=" UINT_FMT ", star_flags=" UINT_FMT ")\n", n_positional, n_keyword, star_flags);
|
|
|
|
|
2014-09-29 22:10:41 +01:00
|
|
|
// TODO: in viper mode, call special runtime routine with type info for args,
|
|
|
|
// and wanted type info for return, to remove need for boxing/unboxing
|
|
|
|
|
|
|
|
emit_native_pre(emit);
|
|
|
|
vtype_kind_t vtype_fun = peek_vtype(emit, n_positional + 2 * n_keyword);
|
|
|
|
if (vtype_fun == VTYPE_BUILTIN_CAST) {
|
|
|
|
// casting operator
|
|
|
|
assert(n_positional == 1 && n_keyword == 0);
|
2015-04-06 22:48:21 +01:00
|
|
|
assert(!star_flags);
|
2014-09-29 22:10:41 +01:00
|
|
|
DEBUG_printf(" cast to %d\n", vtype_fun);
|
2015-01-24 23:14:12 +00:00
|
|
|
vtype_kind_t vtype_cast = peek_stack(emit, 1)->data.u_imm;
|
2014-09-29 22:10:41 +01:00
|
|
|
switch (peek_vtype(emit, 0)) {
|
|
|
|
case VTYPE_PYOBJ: {
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_1);
|
|
|
|
emit_pre_pop_discard(emit);
|
2016-02-02 23:16:05 +00:00
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_OBJ_TO_NATIVE, vtype_cast, REG_ARG_2); // arg2 = type
|
2014-09-29 22:10:41 +01:00
|
|
|
emit_post_push_reg(emit, vtype_cast, REG_RET);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case VTYPE_BOOL:
|
|
|
|
case VTYPE_INT:
|
|
|
|
case VTYPE_UINT:
|
|
|
|
case VTYPE_PTR:
|
|
|
|
case VTYPE_PTR8:
|
|
|
|
case VTYPE_PTR16:
|
2015-10-13 00:50:17 +01:00
|
|
|
case VTYPE_PTR32:
|
2014-09-29 22:10:41 +01:00
|
|
|
case VTYPE_PTR_NONE:
|
|
|
|
emit_fold_stack_top(emit, REG_ARG_1);
|
|
|
|
emit_post_top_set_vtype(emit, vtype_cast);
|
|
|
|
break;
|
|
|
|
default:
|
2017-03-14 12:00:53 +00:00
|
|
|
// this can happen when casting a cast: int(int)
|
2020-03-02 11:35:22 +00:00
|
|
|
mp_raise_NotImplementedError(MP_ERROR_TEXT("casting"));
|
2014-09-29 22:10:41 +01:00
|
|
|
}
|
|
|
|
} else {
|
2013-10-08 09:05:10 +01:00
|
|
|
assert(vtype_fun == VTYPE_PYOBJ);
|
2015-04-06 22:48:21 +01:00
|
|
|
if (star_flags) {
|
2020-03-25 02:39:46 +00:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, n_positional + 2 * n_keyword + 2); // pointer to args
|
2015-04-06 22:48:21 +01:00
|
|
|
emit_call_with_2_imm_args(emit, MP_F_CALL_METHOD_N_KW_VAR, 0, REG_ARG_1, n_positional | (n_keyword << 8), REG_ARG_2);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
} else {
|
|
|
|
if (n_positional != 0 || n_keyword != 0) {
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, n_positional + 2 * n_keyword); // pointer to args
|
|
|
|
}
|
|
|
|
emit_pre_pop_reg(emit, &vtype_fun, REG_ARG_1); // the function
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_NATIVE_CALL_FUNCTION_N_KW, n_positional | (n_keyword << 8), REG_ARG_2);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
}
|
2014-02-02 13:11:48 +00:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_call_method(emit_t *emit, mp_uint_t n_positional, mp_uint_t n_keyword, mp_uint_t star_flags) {
|
2015-04-06 22:48:21 +01:00
|
|
|
if (star_flags) {
|
2020-03-25 02:39:46 +00:00
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, n_positional + 2 * n_keyword + 3); // pointer to args
|
2015-04-06 22:48:21 +01:00
|
|
|
emit_call_with_2_imm_args(emit, MP_F_CALL_METHOD_N_KW_VAR, 1, REG_ARG_1, n_positional | (n_keyword << 8), REG_ARG_2);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
} else {
|
|
|
|
emit_native_pre(emit);
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 2 + n_positional + 2 * n_keyword); // pointer to items, including meth and self
|
|
|
|
emit_call_with_2_imm_args(emit, MP_F_CALL_METHOD_N_KW, n_positional, REG_ARG_1, n_keyword, REG_ARG_2);
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_RET);
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-02-12 16:31:30 +00:00
|
|
|
STATIC void emit_native_return_value(emit_t *emit) {
|
2014-04-06 12:58:40 +01:00
|
|
|
DEBUG_printf("return_value\n");
|
2018-10-01 04:07:04 +01:00
|
|
|
|
|
|
|
if (emit->scope->scope_flags & MP_SCOPE_FLAG_GENERATOR) {
|
|
|
|
// Save pointer to current stack position for caller to access return value
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_TEMP0, 1);
|
2019-09-18 04:40:42 +01:00
|
|
|
emit_native_mov_state_reg(emit, OFFSETOF_CODE_STATE_SP, REG_TEMP0);
|
2018-10-01 04:07:04 +01:00
|
|
|
|
|
|
|
// Put return type in return value slot
|
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_TEMP0, MP_VM_RETURN_NORMAL);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_RET_VAL(emit), REG_TEMP0);
|
|
|
|
|
|
|
|
// Do the unwinding jump to get to the return handler
|
|
|
|
emit_native_unwind_jump(emit, emit->exit_label, emit->exc_stack_size);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
if (emit->do_viper_types) {
|
2018-09-15 03:41:25 +01:00
|
|
|
vtype_kind_t return_vtype = emit->scope->scope_flags >> MP_SCOPE_FLAG_VIPERRET_POS;
|
2014-09-29 22:10:41 +01:00
|
|
|
if (peek_vtype(emit, 0) == VTYPE_PTR_NONE) {
|
|
|
|
emit_pre_pop_discard(emit);
|
2018-09-15 03:41:25 +01:00
|
|
|
if (return_vtype == VTYPE_PYOBJ) {
|
2019-09-10 04:35:19 +01:00
|
|
|
emit_native_mov_reg_const(emit, REG_PARENT_RET, MP_F_CONST_NONE_OBJ);
|
2014-09-29 22:10:41 +01:00
|
|
|
} else {
|
2018-09-14 08:40:59 +01:00
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_ARG_1, 0);
|
2014-09-29 22:10:41 +01:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
vtype_kind_t vtype;
|
2019-09-10 04:35:19 +01:00
|
|
|
emit_pre_pop_reg(emit, &vtype, return_vtype == VTYPE_PYOBJ ? REG_PARENT_RET : REG_ARG_1);
|
2018-09-15 03:41:25 +01:00
|
|
|
if (vtype != return_vtype) {
|
2015-04-20 14:29:31 +01:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit,
|
2020-03-02 11:35:22 +00:00
|
|
|
MP_ERROR_TEXT("return expected '%q' but got '%q'"),
|
2018-09-15 03:41:25 +01:00
|
|
|
vtype_to_qstr(return_vtype), vtype_to_qstr(vtype));
|
2014-08-15 23:47:59 +01:00
|
|
|
}
|
2014-08-15 16:45:41 +01:00
|
|
|
}
|
2018-09-14 08:40:59 +01:00
|
|
|
if (return_vtype != VTYPE_PYOBJ) {
|
|
|
|
emit_call_with_imm_arg(emit, MP_F_CONVERT_NATIVE_TO_OBJ, return_vtype, REG_ARG_2);
|
2019-11-06 01:11:10 +00:00
|
|
|
#if REG_RET != REG_PARENT_RET
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_REG_REG(emit->as, REG_PARENT_RET, REG_RET);
|
|
|
|
#endif
|
2018-09-14 08:40:59 +01:00
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
} else {
|
2014-09-29 22:10:41 +01:00
|
|
|
vtype_kind_t vtype;
|
2019-09-10 04:35:19 +01:00
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_PARENT_RET);
|
2013-10-08 09:05:10 +01:00
|
|
|
assert(vtype == VTYPE_PYOBJ);
|
|
|
|
}
|
2018-09-04 05:31:28 +01:00
|
|
|
if (NEED_GLOBAL_EXC_HANDLER(emit)) {
|
|
|
|
// Save return value for the global exception handler to use
|
2019-09-10 04:35:19 +01:00
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_RET_VAL(emit), REG_PARENT_RET);
|
2018-09-04 05:31:28 +01:00
|
|
|
}
|
|
|
|
emit_native_unwind_jump(emit, emit->exit_label, emit->exc_stack_size);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-09-08 23:05:16 +01:00
|
|
|
STATIC void emit_native_raise_varargs(emit_t *emit, mp_uint_t n_args) {
|
2019-02-25 03:52:36 +00:00
|
|
|
(void)n_args;
|
2014-06-30 05:17:25 +01:00
|
|
|
assert(n_args == 1);
|
2014-08-16 22:06:11 +01:00
|
|
|
vtype_kind_t vtype_exc;
|
|
|
|
emit_pre_pop_reg(emit, &vtype_exc, REG_ARG_1); // arg1 = object to raise
|
|
|
|
if (vtype_exc != VTYPE_PYOBJ) {
|
2020-03-02 11:35:22 +00:00
|
|
|
EMIT_NATIVE_VIPER_TYPE_ERROR(emit, MP_ERROR_TEXT("must raise an object"));
|
2014-08-16 22:06:11 +01:00
|
|
|
}
|
|
|
|
// TODO probably make this 1 call to the runtime (which could even call convert, native_raise(obj, type))
|
2014-08-16 22:31:57 +01:00
|
|
|
emit_call(emit, MP_F_NATIVE_RAISE);
|
2022-06-17 14:14:32 +01:00
|
|
|
mp_asm_base_suppress_code(&emit->as->base);
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
2014-06-30 05:17:25 +01:00
|
|
|
|
2018-05-18 15:30:42 +01:00
|
|
|
STATIC void emit_native_yield(emit_t *emit, int kind) {
|
2018-10-01 04:07:04 +01:00
|
|
|
// Note: 1 (yield) or 3 (yield from) labels are reserved for this function, starting at *emit->label_slot
|
|
|
|
|
|
|
|
if (emit->do_viper_types) {
|
2020-03-02 11:35:22 +00:00
|
|
|
mp_raise_NotImplementedError(MP_ERROR_TEXT("native yield"));
|
2018-10-01 04:07:04 +01:00
|
|
|
}
|
|
|
|
emit->scope->scope_flags |= MP_SCOPE_FLAG_GENERATOR;
|
|
|
|
|
|
|
|
need_stack_settled(emit);
|
|
|
|
|
|
|
|
if (kind == MP_EMIT_YIELD_FROM) {
|
|
|
|
|
|
|
|
// Top of yield-from loop, conceptually implementing:
|
|
|
|
// for item in generator:
|
|
|
|
// yield item
|
|
|
|
|
|
|
|
// Jump to start of loop
|
|
|
|
emit_native_jump(emit, *emit->label_slot + 2);
|
|
|
|
|
|
|
|
// Label for top of loop
|
|
|
|
emit_native_label_assign(emit, *emit->label_slot + 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
// Save pointer to current stack position for caller to access yielded value
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_TEMP0, 1);
|
2019-09-18 04:40:42 +01:00
|
|
|
emit_native_mov_state_reg(emit, OFFSETOF_CODE_STATE_SP, REG_TEMP0);
|
2018-10-01 04:07:04 +01:00
|
|
|
|
|
|
|
// Put return type in return value slot
|
|
|
|
ASM_MOV_REG_IMM(emit->as, REG_TEMP0, MP_VM_RETURN_YIELD);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_RET_VAL(emit), REG_TEMP0);
|
|
|
|
|
|
|
|
// Save re-entry PC
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_TEMP0, *emit->label_slot);
|
|
|
|
emit_native_mov_state_reg(emit, LOCAL_IDX_GEN_PC(emit), REG_TEMP0);
|
|
|
|
|
|
|
|
// Jump to exit handler
|
|
|
|
ASM_JUMP(emit->as, emit->exit_label);
|
|
|
|
|
|
|
|
// Label re-entry point
|
|
|
|
mp_asm_base_label_assign(&emit->as->base, *emit->label_slot);
|
|
|
|
|
|
|
|
// Re-open any active exception handler
|
|
|
|
if (emit->exc_stack_size > 0) {
|
|
|
|
// Find innermost active exception handler, to restore as current handler
|
|
|
|
exc_stack_entry_t *e = &emit->exc_stack[emit->exc_stack_size - 1];
|
|
|
|
for (; e >= emit->exc_stack; --e) {
|
|
|
|
if (e->is_active) {
|
|
|
|
// Found active handler, get its PC
|
|
|
|
ASM_MOV_REG_PCREL(emit->as, REG_RET, e->label);
|
|
|
|
ASM_MOV_LOCAL_REG(emit->as, LOCAL_IDX_EXC_HANDLER_PC(emit), REG_RET);
|
2020-01-24 07:01:55 +00:00
|
|
|
break;
|
2018-10-01 04:07:04 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_native_adjust_stack_size(emit, 1); // send_value
|
|
|
|
|
|
|
|
if (kind == MP_EMIT_YIELD_VALUE) {
|
|
|
|
// Check LOCAL_IDX_EXC_VAL for any injected value
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_ARG_1, LOCAL_IDX_EXC_VAL(emit));
|
|
|
|
emit_call(emit, MP_F_NATIVE_RAISE);
|
|
|
|
} else {
|
|
|
|
// Label loop entry
|
|
|
|
emit_native_label_assign(emit, *emit->label_slot + 2);
|
|
|
|
|
|
|
|
// Get the next item from the delegate generator
|
|
|
|
vtype_kind_t vtype;
|
|
|
|
emit_pre_pop_reg(emit, &vtype, REG_ARG_2); // send_value
|
|
|
|
emit_access_stack(emit, 1, &vtype, REG_ARG_1); // generator
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_ARG_3, LOCAL_IDX_EXC_VAL(emit)); // throw_value
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_ARG_3);
|
|
|
|
emit_get_stack_pointer_to_reg_for_pop(emit, REG_ARG_3, 1); // ret_value
|
|
|
|
emit_call(emit, MP_F_NATIVE_YIELD_FROM);
|
|
|
|
|
|
|
|
// If returned non-zero then generator continues
|
|
|
|
ASM_JUMP_IF_REG_NONZERO(emit->as, REG_RET, *emit->label_slot + 1, true);
|
|
|
|
|
|
|
|
// Pop exhausted gen, replace with ret_value
|
|
|
|
emit_native_adjust_stack_size(emit, 1); // ret_value
|
|
|
|
emit_fold_stack_top(emit, REG_ARG_1);
|
|
|
|
}
|
2013-10-08 09:05:10 +01:00
|
|
|
}
|
|
|
|
|
2014-06-30 05:17:25 +01:00
|
|
|
STATIC void emit_native_start_except_handler(emit_t *emit) {
|
2018-09-04 05:31:28 +01:00
|
|
|
// Protected block has finished so leave the current exception handler
|
|
|
|
emit_native_leave_exc_stack(emit, true);
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
|
|
|
|
// Get and push nlr_buf.ret_val
|
|
|
|
ASM_MOV_REG_LOCAL(emit->as, REG_TEMP0, LOCAL_IDX_EXC_VAL(emit));
|
|
|
|
emit_post_push_reg(emit, VTYPE_PYOBJ, REG_TEMP0);
|
2014-06-30 05:17:25 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void emit_native_end_except_handler(emit_t *emit) {
|
py/emitnative: Optimise and improve exception handling in native code.
Prior to this patch, native code would use a full nlr_buf_t for each
exception handler (try-except, try-finally, with). For nested exception
handlers this would use a lot of C stack and be rather inefficient.
This patch changes how exceptions are handled in native code by setting up
only a single nlr_buf_t context for the entire function, and then manages a
state machine (using the PC) to work out which exception handler to run
when an exception is raised by an nlr_jump. This keeps the C stack usage
at a constant level regardless of the depth of Python exception blocks.
The patch also fixes an existing bug when local variables are written to
within an exception handler, then their value was incorrectly restored if
an exception was raised (since the nlr_jump would restore register values,
back to the point of the nlr_push).
And it also gets nested try-finally+with working with the viper emitter.
Broadly speaking, efficiency of executing native code that doesn't use
any exception blocks is unchanged, and emitted code size is only slightly
increased for such function. C stack usage of all native functions is
either equal or less than before. Emitted code size for native functions
that use exception blocks is increased by roughly 10% (due in part to
fixing of above-mentioned bugs).
But, most importantly, this patch allows to implement more Python features
in native code, like unwind jumps and yielding from within nested exception
blocks.
2018-08-16 04:56:36 +01:00
|
|
|
adjust_stack(emit, -1); // pop the exception (end_finally didn't use it)
|
2014-06-30 05:17:25 +01:00
|
|
|
}
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
const emit_method_table_t EXPORT_FUN(method_table) = {
|
2019-03-08 23:59:25 +00:00
|
|
|
#if MICROPY_DYNAMIC_COMPILER
|
|
|
|
EXPORT_FUN(new),
|
|
|
|
EXPORT_FUN(free),
|
|
|
|
#endif
|
|
|
|
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_start_pass,
|
|
|
|
emit_native_end_pass,
|
2014-04-10 18:28:54 +01:00
|
|
|
emit_native_adjust_stack_size,
|
2014-01-18 23:24:36 +00:00
|
|
|
emit_native_set_source_line,
|
2013-10-08 09:05:10 +01:00
|
|
|
|
2015-03-26 14:42:40 +00:00
|
|
|
{
|
2018-05-18 15:11:04 +01:00
|
|
|
emit_native_load_local,
|
2015-03-26 14:42:40 +00:00
|
|
|
emit_native_load_global,
|
|
|
|
},
|
|
|
|
{
|
2018-05-18 15:11:04 +01:00
|
|
|
emit_native_store_local,
|
2015-03-26 14:42:40 +00:00
|
|
|
emit_native_store_global,
|
|
|
|
},
|
|
|
|
{
|
2018-05-18 15:11:04 +01:00
|
|
|
emit_native_delete_local,
|
2015-03-26 14:42:40 +00:00
|
|
|
emit_native_delete_global,
|
|
|
|
},
|
2013-10-08 09:05:10 +01:00
|
|
|
|
|
|
|
emit_native_label_assign,
|
2018-05-22 12:58:25 +01:00
|
|
|
emit_native_import,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_load_const_tok,
|
|
|
|
emit_native_load_const_small_int,
|
|
|
|
emit_native_load_const_str,
|
2015-01-13 15:55:54 +00:00
|
|
|
emit_native_load_const_obj,
|
2014-04-20 17:50:40 +01:00
|
|
|
emit_native_load_null,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_load_method,
|
|
|
|
emit_native_load_build_class,
|
2018-05-22 12:31:56 +01:00
|
|
|
emit_native_subscr,
|
2018-05-22 12:43:41 +01:00
|
|
|
emit_native_attr,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_dup_top,
|
|
|
|
emit_native_dup_top_two,
|
|
|
|
emit_native_pop_top,
|
|
|
|
emit_native_rot_two,
|
|
|
|
emit_native_rot_three,
|
|
|
|
emit_native_jump,
|
2015-02-28 15:04:06 +00:00
|
|
|
emit_native_pop_jump_if,
|
|
|
|
emit_native_jump_if_or_pop,
|
2018-05-22 12:50:22 +01:00
|
|
|
emit_native_unwind_jump,
|
2018-05-22 13:33:26 +01:00
|
|
|
emit_native_setup_block,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_with_cleanup,
|
|
|
|
emit_native_end_finally,
|
|
|
|
emit_native_get_iter,
|
|
|
|
emit_native_for_iter,
|
|
|
|
emit_native_for_iter_end,
|
2019-02-15 01:18:59 +00:00
|
|
|
emit_native_pop_except_jump,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_unary_op,
|
|
|
|
emit_native_binary_op,
|
2018-05-18 15:41:40 +01:00
|
|
|
emit_native_build,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_store_map,
|
2016-09-18 14:59:47 +01:00
|
|
|
emit_native_store_comp,
|
2013-10-08 09:05:10 +01:00
|
|
|
emit_native_unpack_sequence,
|
|
|
|
emit_native_unpack_ex,
|
|
|
|
emit_native_make_function,
|
|
|
|
emit_native_make_closure,
|
|
|
|
emit_native_call_function,
|
|
|
|
emit_native_call_method,
|
|
|
|
emit_native_return_value,
|
|
|
|
emit_native_raise_varargs,
|
2018-05-18 15:30:42 +01:00
|
|
|
emit_native_yield,
|
2014-06-30 05:17:25 +01:00
|
|
|
|
|
|
|
emit_native_start_except_handler,
|
|
|
|
emit_native_end_except_handler,
|
2013-10-08 09:05:10 +01:00
|
|
|
};
|
|
|
|
|
2014-09-06 23:06:36 +01:00
|
|
|
#endif
|