Commit Graph

3207 Commits

Author SHA1 Message Date
Damien George d97906ca9a py/emit: Combine import from/name/star into one emit function.
Change in code size is:

   bare-arm:    +4
minimal x86:   -88
   unix x64:  -456
unix nanbox:   -88
      stm32:   -44
     cc3200:    +0
    esp8266:  -104
      esp32:    +8
2018-05-23 00:23:08 +10:00
Damien George 8a513da5a5 py/emit: Combine break_loop and continue_loop into one emit function.
Reduces code size by:

   bare-arm:    +0
minimal x86:    +0
   unix x64:   -80
unix nanbox:    +0
      stm32:   -12
     cc3200:    +0
    esp8266:   -28
      esp32:    +0
2018-05-23 00:23:04 +10:00
Damien George 6211d979ee py/emit: Combine load/store/delete attr into one emit function.
Reduces code size by:

   bare-arm:   -20
minimal x86:  -140
   unix x64:  -408
unix nanbox:  -140
      stm32:   -68
     cc3200:   -16
    esp8266:   -80
      esp32:   -32
2018-05-23 00:22:59 +10:00
Damien George a4941a8ba4 py/emit: Combine load/store/delete subscr into one emit function.
Reduces code size by:

   bare-arm:    -8
minimal x86:  -104
   unix x64:  -312
unix nanbox:  -120
      stm32:   -60
     cc3200:   -16
    esp8266:   -92
      esp32:   -24
2018-05-23 00:22:55 +10:00
Damien George d298013939 py/emit: Combine name and global into one func for load/store/delete.
Reduces code size by:

   bare-arm:   -56
minimal x86:  -300
   unix x64:  -576
unix nanbox:  -300
      stm32:  -164
     cc3200:   -56
    esp8266:  -236
      esp32:   -76
2018-05-23 00:22:47 +10:00
Damien George 26b5754092 py/emit: Combine build tuple/list/map emit funcs into one.
Reduces code size by:

   bare-arm:   -24
minimal x86:  -192
   unix x64:  -288
unix nanbox:  -184
      stm32:   -72
     cc3200:   -16
    esp8266:  -148
      esp32:   -32
2018-05-23 00:22:44 +10:00
Damien George e686c94052 py/emit: Combine yield value and yield-from emit funcs into one.
Reduces code size by:

   bare-arm:   -24
minimal x86:   -72
   unix x64:  -200
unix nanbox:   -72
      stm32:   -52
     cc3200:   -32
    esp8266:   -84
      esp32:   -24
2018-05-23 00:22:35 +10:00
Damien George 0a25fff956 py/emit: Combine fast and deref into one function for load/store/delete.
Reduces code size by:

   bare-arm:   -16
minimal x86:  -208
   unix x64:  -408
unix nanbox:  -248
      stm32:   -12
     cc3200:   -24
    esp8266:   -96
      esp32:   -44
2018-05-23 00:22:20 +10:00
Damien George 400273a799 py/objgenerator: Protect against reentering a generator.
Generators that are already executing cannot be reexecuted.  This patch
puts in a check for such a case.

Thanks to @jepler for finding the bug.
2018-05-22 16:54:03 +10:00
Damien George 771cb359af py/objgenerator: Save state in old_globals instead of local variable.
The code_state.old_globals variable is there to save the globals state so
should be used for this purpose, to avoid the need for additional local
variables on the C stack.
2018-05-22 16:39:19 +10:00
Jan Klusacek b318ebf101 py/modbuiltins: Add support for rounding integers.
As per CPython semantics.  This feature is controlled by
MICROPY_PY_BUILTINS_ROUND_INT which is disabled by default.
2018-05-22 14:18:16 +10:00
Damien George f2ec792554 py/parsenum: Adjust braces so they are balanced. 2018-05-22 13:20:00 +10:00
Damien George 6bd78741c1 py/gc: When GC threshold is hit don't unnecessarily collect twice.
Without this, if GC threshold is hit and there is not enough memory left to
satisfy the request, gc_collect() will run a second time and the search for
memory will happen again and will fail again.

Thanks to @adritium for pointing out this issue, see #3786.
2018-05-21 13:36:21 +10:00
Jeff Epler 95e43efc99 py/objfloat: Fix undefined integer behavior hashing negative zero.
Under ubsan, when evaluating hash(-0.) the following diagnostic occurs:

    ../../py/objfloat.c:102:15: runtime error: negation of
    -9223372036854775808 cannot be represented in type 'mp_int_t' (aka
    'long'); cast to an unsigned type to negate this value to itself

So do just that, to tell the compiler that we want to perform this
operation using modulo arithmetic rules.
2018-05-21 12:49:56 +10:00
Jeff Epler c4dafcef4f py/mpz: Avoid undefined behavior at integer overflow in mpz_hash.
Before this, ubsan would detect a problem when executing
hash(006699999999999999999999999999999999999999999999999999999999999999999999)

    ../../py/mpz.c:1539:20: runtime error: left shift of 1067371580458 by
    32 places cannot be represented in type 'mp_int_t' (aka 'long')

When the overflow does occur it now happens as defined by the rules of
unsigned arithmetic.
2018-05-21 12:48:26 +10:00
Jeff Epler 60eb5305f6 py/objfloat: Fix undefined shifting behavior in high-quality float hash.
When computing e.g. hash(0.4e3) with ubsan enabled, a diagnostic like the
following would occur:

    ../../py/objfloat.c:91:30: runtime error: shift exponent 44 is too
    large for 32-bit type 'int'

By casting constant "1" to the right type the intended value is preserved.
2018-05-21 12:42:22 +10:00
Jeff Epler 4f71a2a75a py/parsenum: Avoid undefined behavior parsing floats with large exponents.
Fuzz testing combined with the undefined behavior sanitizer found that
parsing unreasonable float literals like 1e+9999999999999 resulted in
undefined behavior due to overflow in signed integer arithmetic, and a
wrong result being returned.
2018-05-21 12:37:57 +10:00
Damien George 5efc575067 py/parsenum: Use int instead of mp_int_t for parsing float exponent.
There is no need to use the mp_int_t type which may be 64-bits wide, there
is enough bit-width in a normal int to parse reasonable exponents.  Using
int helps to reduce code size for 64-bit ports, especially nan-boxing
builds.  (Similarly for the "dig" variable which is now an unsigned int.)
2018-05-21 12:27:38 +10:00
Jeff Epler bc6c0b28bf py/emitbc: Avoid undefined behavior calling memset() with NULL 1st arg.
Calling memset(NULL, value, 0) is not standards compliant so we must add an
explicit check that emit->label_offsets is indeed not NULL before calling
memset (this pointer will be NULL on the first pass of the parse tree and
it's more logical / safer to check this pointer rather than check that the
pass is not the first one).

Code sanitizers will warn if NULL is passed as the first value to memset,
and compilers may optimise the code based on the knowledge that any pointer
passed to memset is guaranteed not to be NULL.
2018-05-21 12:04:20 +10:00
Damien George 828ce16dc8 py/compile: Change comment about ITER_BUF_NSLOTS to a static assertion. 2018-05-18 23:31:00 +10:00
Damien George 43d08d6dd6 py/misc.h: Add MP_STATIC_ASSERT macro to do static assertions. 2018-05-18 23:31:00 +10:00
Li Weiwei 3e6ab82179 py/repl: Fix handling of unmatched brackets and unfinished quotes.
Before this patch:

    >>> print(')
    ... ')
    Traceback (most recent call last):
      File "<stdin>", line 1
    SyntaxError: invalid syntax

After this patch:

    >>> print(')
    Traceback (most recent call last):
      File "<stdin>", line 1
    SyntaxError: invalid syntax

This matches CPython and prevents getting stuck in REPL continuation when a
1-quote is unmatched.
2018-05-18 15:23:02 +10:00
Damien George 869024dd6e py/vm: Improve performance of opcode dispatch when using switch stmt.
Before this patch, when using the switch statement for dispatch in the VM
(not computed goto) a pending exception check was done after each opcode.
This is not necessary and this patch makes the pending exception check only
happen when explicitly requested by certain opcodes, like jump.  This
improves performance of the VM by about 2.5% when using the switch.
2018-05-18 11:47:03 +10:00
Damien George 46ce395130 py/vm: Use enum names instead of magic numbers in multi-opcode dispatch. 2018-05-18 11:44:26 +10:00
Tom Collins a883fe12d9 py/objfun: Fix variable name in DECODE_CODESTATE_SIZE() macro.
This patch fixes the macro so you can pass any name in, and the macro will
make more sense if you're reading it on its own.  It worked previously
because n_state is always passed in as n_state_out_var.
2018-05-17 11:20:06 +10:00
Damien George 1b7487e519 py/vm: Adjust #if logic for gil_divisor so braces are balanced.
Having balanced braces { and } makes it easier to navigate the function.
2018-05-16 12:33:39 +10:00
Damien George c97607db5c py/nlrx86: Use naked attribute on nlr_push for gcc 8.0 and higher.
gcc 8.0 supports the naked attribute for x86 systems so it can now be used
here.  And in fact it is necessary to use this for nlr_push because gcc 8.0
no longer generates a prelude for this function (even without the naked
attribute).
2018-05-15 11:17:28 +10:00
Damien George 749b16174b py/mpstate.h: Adjust start of root pointer section to exclude non-ptrs.
This patch moves the start of the root pointer section in mp_state_ctx_t
so that it skips entries that are not pointers and don't need scanning.

Previously, the start of the root pointer section was at the very beginning
of the mp_state_ctx_t struct (which is the beginning of mp_state_thread_t).
This was the original assembler version of the NLR code was hard-coded to
have the nlr_top pointer at the start of this state structure.  But now
that the NLR code is partially written in C there is no longer this
restriction on the location of nlr_top (and a comment to this effect has
been removed in this patch).

So now the root pointer section starts part way through the
mp_state_thread_t structure, after the entries which are not root pointers.

This patch also moves the non-pointer entries for MICROPY_ENABLE_SCHEDULER
outside the root pointer section.

Moving non-pointer entries out of the root pointer section helps to make
the GC more precise and should help to prevent some cases of collectable
garbage being kept.

This patch also has a measurable improvement in performance of the
pystone.py benchmark: on unix x86-64 and stm32 there was an improvement of
roughly 0.6% (tested with both gcc 7.3 and gcc 8.1).
2018-05-13 22:53:28 +10:00
Damien George 9630376dbc py/mpconfig.h: Be stricter when autodetecting machine endianness.
This patch changes 2 things in the endianness detection:

1. Don't assume that __BYTE_ORDER__ not being __ORDER_LITTLE_ENDIAN__ means
   that the machine is big endian, so add an explicit check that this macro
   is indeed __ORDER_BIG_ENDIAN__ (same with __BYTE_ORDER, __LITTLE_ENDIAN
   and __BIG_ENDIAN).  A machine could have PDP endianness.

2. Remove the checks which base their autodetection decision on whether any
   little or big endian macros are defined (eg __LITTLE_ENDIAN__ or
   __BIG_ENDIAN__).  Just because a system defines these does not mean it
   has that endianness.

See issue #3760.
2018-05-11 21:51:34 +10:00
Damien George 6046e68fe1 py/repl: Initialise q_last variable to prevent compiler warnings.
Some older compilers cannot deduce that q_last is always written to before
being read.
2018-05-11 13:48:47 +10:00
Damien George 095d397017 py/objdeque: Fix sign extension bug when computing len of deque object.
For cases where size_t is smaller than mp_int_t (eg nan-boxing builds) the
difference between two size_t's is not sign extended into mp_int_t and so
the result is never negative.  This patch fixes this bug by using ssize_t
for the type of the result.
2018-05-11 13:44:50 +10:00
Damien George 3678a6bdc6 py/modbuiltins: Make built-in dir support the __dir__ special method.
If MICROPY_PY_ALL_SPECIAL_METHODS is enabled then dir() will now delegate
to the special method __dir__ if the object it is listing has this method.
2018-05-10 23:14:23 +10:00
Damien George 29d28c2574 py/modbuiltins: In built-in dir make use of mp_load_method_protected.
This gives dir() better behaviour when listing the attributes of a user
type that defines __getattr__: it will now not list those attributes for
which __getattr__ raises AttributeError (meaning the attribute is not
supported by the object).
2018-05-10 23:07:19 +10:00
Damien George 7241d90272 py/repl: Use mp_load_method_protected to prevent leaking of exceptions.
This patch fixes the possibility of a crash of the REPL when tab-completing
an object which raises an exception when its attributes are accessed.

See issue #3729.
2018-05-10 23:05:43 +10:00
Damien George 529860643b py/modbuiltins: Make built-in hasattr work properly for user types.
It now allows __getattr__ in a user type to raise AttributeError when the
attribute does not exist.
2018-05-10 23:03:30 +10:00
Damien George bc87b862fd py/runtime: Add mp_load_method_protected helper which catches exceptions
This new helper function acts like mp_load_method_maybe but is wrapped in
an NLR handler so it can catch exceptions.  It prevents AttributeError from
propagating out, and optionally all other exceptions.  This helper can be
used to fully implement hasattr (see follow-up commit), and also for cases
where mp_load_method_maybe is used but it must now raise an exception.
2018-05-10 23:00:04 +10:00
Damien George eb88803ac8 py/{modbuiltins,repl}: Start qstr probing from after empty qstr.
The list of qstrs starts with MP_QSTR_NULL followed by MP_QSTR_, and these
should never appear in dir() or REPL tab completion, so skip them.
2018-05-09 16:15:02 +10:00
Damien George 3cf02be4e0 py/emitnx86: Fix 32-bit x86 native emitter build by including header. 2018-05-04 20:39:16 +10:00
Damien George 60db80920a py/builtinhelp: Change occurrence of mp_uint_t to size_t. 2018-05-02 16:50:28 +10:00
Damien George 6b4b6d388b py/obj.h: Fix math.e constant for nan-boxing builds.
Due to a typo, math.e was too small by around 6e-11.
2018-05-01 23:25:18 +10:00
Ayke van Laethem d43c737756 py/stream: Use uPy errno instead of system's for non-blocking check.
This is a more consistent use of errno codes.  For example, it may be that
a stream returns MP_EAGAIN but the mp_is_nonblocking_error() macro doesn't
catch this value because it checks for EAGAIN instead (which may have a
different value than MP_EAGAIN when MICROPY_USE_INTERNAL_ERRNO is enabled).
2018-05-01 15:54:50 +10:00
Damien George 96740be357 py/mperrno: Define MP_EWOULDBLOCK as EWOULDBLOCK, not EAGAIN.
Most modern systems have EWOULDBLOCK aliased to EAGAIN, ie they have the
same value.  But some systems use different values for these errnos and if
a uPy port is using the system errno values (ie not the internal uPy
values) then it's important to be able to distinguish EWOULDBLOCK from
EAGAIN.  Eg if a system call returned EWOULDBLOCK it must be possible to
check for this return value, and this patch makes this now possible.
2018-05-01 15:53:25 +10:00
Ayke van Laethem deaa46aa66 py/nlrthumb: Fix Clang support wrt use of "return 0".
Clang defines __GNUC__ so we have to check for it specifically.
2018-04-27 15:10:42 +10:00
Damien George ef12a4bd05 py: Refactor how native emitter code is compiled with a file per arch.
Instead of emitnative.c having configuration code for each supported
architecture, and then compiling this file multiple times with different
macros defined, this patch adds a file per architecture with the necessary
code to configure the native emitter.  These files then #include the
emitnative.c file.

This simplifies emitnative.c (which is already very large), and simplifies
the build system because emitnative.c no longer needs special handling for
compilation and qstr extraction.
2018-04-10 15:06:47 +10:00
Jeff Epler cbf981f330 py/objgenerator: Check stack before resuming a generator.
This turns a hard crash in a recursive generator into a 'maximum recursion
depth exceeded' exception.
2018-04-10 14:06:26 +10:00
Damien George cf31d384f1 py/stream: Switch stream close operation from method to ioctl.
This patch moves the implementation of stream closure from a dedicated
method to the ioctl of the stream protocol, for each type that implements
closing.  The benefits of this are:

1. Rounds out the stream ioctl function, which already includes flush,
   seek and poll (among other things).

2. Makes calling mp_stream_close() on an object slightly more efficient
   because it now no longer needs to lookup the close method and call it,
   rather it just delegates straight to the ioctl function (if it exists).

3. Reduces code size and allows future types that implement the stream
   protocol to be smaller because they don't need a dedicated close method.

Code size reduction is around 200 bytes smaller for x86 archs and around
30 bytes smaller for the bare-metal archs.
2018-04-10 13:41:32 +10:00
Jeff Epler d6cf5c6749 py/objstr: In find/rfind, don't crash when end < start. 2018-04-05 16:14:17 +10:00
Damien George f1df86a017 py/objint: Simplify LHS arg type checking in int binary op functions.
The LHS passed to mp_obj_int_binary_op() will always be an integer, either
a small int or a big int, so the test for this type doesn't need to include
an "other, unsupported type" case.
2018-04-05 01:11:26 +10:00
Damien George 3f420c0c27 py: Don't include mp_optimise_value or opt_level() if compiler disabled.
Without the compiler enabled the mp_optimise_value is unused, and the
micropython.opt_level() function is not useful, so exclude these from the
build to save RAM and code size.
2018-04-04 14:24:03 +10:00
Damien George 323b5f7270 py/modsys: Don't compile getsizeof function if feature is disabled. 2018-04-04 14:23:25 +10:00