Commit Graph

3535 Commits

Author SHA1 Message Date
Damien George 9adedce42e py: Add new Xtensa-Windowed arch for native emitter.
Enabled via the configuration MICROPY_EMIT_XTENSAWIN.
2019-10-05 13:44:53 +10:00
Damien George f7ddc94166 py/asmxtensa: Add support for Xtensa with windowed registers.
Window-specific asm emit functions are added, along with a new macro option
GENERIC_ASM_API_WIN.
2019-10-05 13:44:08 +10:00
Damien George 306ec5369a py/emitnative: Add support for archs that cannot read executable data.
In which case place the native function prelude in a bytes object, linked
from the const_table of that function.  An architecture should define
N_PRELUDE_AS_BYTES_OBJ to 1 before including py/emitnative.c to emit
correct machine code, then enable MICROPY_EMIT_NATIVE_PRELUDE_AS_BYTES_OBJ
so the runtime can correctly handle the prelude being in a bytes object.
2019-10-05 13:42:39 +10:00
Damien George 3504edc804 py/emitnative: Add support for using setjmp with native emitter.
To enable this feature the N_NLR_SETJMP macro should be set to 1 before
including py/emitnative.c.
2019-10-05 13:41:58 +10:00
Damien George 4107597b84 py/emitnative: Add support for archs with windowed registers.
Such that args/return regs for the parent are different to args/return regs
for child calls.  For an architecture to use this feature it should define
the REG_PARENT_xxx macros before including py/emitnative.c.
2019-10-05 13:41:14 +10:00
Damien George 809d89c794 py/runtime: Fix PEP479 behaviour throwing StopIteration into yield from.
Commit 3f6ffe059f implemented PEP479 but did
not catch the case fixed in this commit.  Found by coverage analysis, that
the VM had uncovered code.
2019-10-04 23:27:00 +10:00
Damien George 82c494a97e py/vm: Fix handling of unwind jump out of active finally.
Prior to this commit, when unwinding through an active finally the stack
was not being correctly popped/folded, which resulting in the VM crashing
for complicated unwinding of nested finallys.

This should be fixed with this commit, and more tests for return/break/
continue within a finally have been added to exercise this.
2019-10-04 23:01:29 +10:00
Jim Mussared a09fd04758 py/makeqstrdefs.py: Remove unused blacklist.
As of 7d58a197cf, `NULL` should no longer be
here because it's allowed (MP_QSTRnull took its place).  This entry was
preventing the use of MP_QSTR_NULL to mean "NULL" (although this is not
currently used).

A blacklist should not be needed because it should be possible to intern
all strings.

Fixes issue #5140.
2019-10-04 17:18:56 +10:00
Jim Mussared 4ddd46e6cf docs/develop/qstr.rst: Add documentation for string interning. 2019-10-04 17:13:13 +10:00
Petr Viktorin 25a9bccdee py/compile: Disallow 'import *' outside module level.
This check follows CPython's behaviour, because 'import *' always populates
the globals with the imported names, not locals.

Since it's safe to do this (doesn't lead to a crash or undefined behaviour)
the check is only enabled for MICROPY_CPYTHON_COMPAT.

Fixes issue #5121.
2019-10-04 16:46:47 +10:00
Damien George 1d0423419b py/bc: Don't include mp_decode_uint funcs when not needed.
These are now only needed when persistent code is disabled.
2019-10-01 12:26:22 +10:00
Damien George c8c0fd4ca3 py: Rework and compress second part of bytecode prelude.
This patch compresses the second part of the bytecode prelude which
contains the source file name, function name, source-line-number mapping
and cell closure information.  This part of the prelude now begins with a
single varible length unsigned integer which encodes 2 numbers, being the
byte-size of the following 2 sections in the header: the "source info
section" and the "closure section".  After decoding this variable unsigned
integer it's possible to skip over one or both of these sections very
easily.

This scheme saves about 2 bytes for most functions compared to the original
format: one in the case that there are no closure cells, and one because
padding was eliminated.
2019-10-01 12:26:22 +10:00
Damien George b5ebfadbd6 py: Compress first part of bytecode prelude.
The start of the bytecode prelude contains 6 numbers telling the amount of
stack needed for the Python values and exceptions, and the signature of the
function.  Prior to this patch these numbers were all encoded one after the
other (2x variable unsigned integers, then 4x bytes), but using so many
bytes is unnecessary.

An entropy analysis of around 150,000 bytecode functions from the CPython
standard library showed that the optimal Shannon coding would need about
7.1 bits on average to encode these 6 numbers, compared to the existing 48
bits.

This patch attempts to get close to this optimal value by packing the 6
numbers into a single, varible-length unsigned integer via bit-wise
interleaving.  The interleaving scheme is chosen to minimise the average
number of bytes needed, and at the same time keep the scheme simple enough
so it can be implemented without too much overhead in code size or speed.
The scheme requires about 10.5 bits on average to store the 6 numbers.

As a result most functions which originally took 6 bytes to encode these 6
numbers now need only 1 byte (in 80% of cases).
2019-10-01 12:26:22 +10:00
Damien George 81d04a0200 py: Add n_state to mp_code_state_t struct.
This value is used often enough that it is better to cache it instead of
decode it each time.
2019-10-01 12:26:22 +10:00
Damien George 4c5e1a0368 py/bc: Change mp_code_state_t.exc_sp to exc_sp_idx.
Change from a pointer to an index, to make space in mp_code_state_t.
2019-10-01 12:26:22 +10:00
Damien George 1d7afcce49 py/bc: Remove comments referring to obsolete currently_in_except_block.
It was made obsolete in 6f9e3ff719
2019-10-01 12:26:22 +10:00
Jim Mussared 16f8ceeaaa extmod/modbluetooth: Add low-level Python BLE API. 2019-10-01 09:51:02 +10:00
Jim Mussared 42e9bdf19b py/ringbuf: Add helpers for put16/get16. 2019-10-01 09:51:02 +10:00
Damien George 5716c5cf65 py/persistentcode: Bump .mpy version to 5.
The bytecode opcodes have changed (there are more, and they have been
reordered).
2019-09-26 16:39:37 +10:00
Damien George 96f2a38075 py/nativeglue: Make mp_fun_table fixed size regardless of config.
So that mpy files with native code will always work correctly, and raise an
exception if a feature is used that is not supported by the runtime.
2019-09-26 16:24:06 +10:00
Josh Lloyd 7d58a197cf py: Rename MP_QSTR_NULL to MP_QSTRnull to avoid intern collisions.
Fixes #5140.
2019-09-26 16:04:56 +10:00
Damien George b596638b9b mpy-cross: Set number of registers in nlr_buf_t based on native arch.
Fixes #5059.  Done in collaboration with Jim Mussared.
2019-09-26 15:53:05 +10:00
Damien George 74503107a7 py/emitnative: Factor sizeof/offsetof calculations to macros. 2019-09-26 15:52:19 +10:00
Damien George a48cdb57b2 py/nlr.h: Factor out constants to specific macros. 2019-09-26 15:52:19 +10:00
Damien George 02db91a7a3 py: Split RAISE_VARARGS opcode into 3 separate ones.
From the beginning of this project the RAISE_VARARGS opcode was named and
implemented following CPython, where it has an argument (to the opcode)
counting how many args the raise takes:

    raise # 0 args (re-raise previous exception)
    raise exc # 1 arg
    raise exc from exc2 # 2 args (chained raise)

In the bytecode this operation therefore takes 2 bytes, one for
RAISE_VARARGS and one for the number of args.

This patch splits this opcode into 3, where each is now a single byte.
This reduces bytecode size by 1 byte for each use of raise.  Every byte
counts!  It also has the benefit of reducing code size (on all ports except
nanbox).
2019-09-26 15:39:50 +10:00
Damien George 870e900d02 py: Introduce and use constants for multi-opcode sizes. 2019-09-26 15:27:11 +10:00
Damien George 1f7202d122 py/bc: Replace big opcode format table with simple macro. 2019-09-26 15:27:11 +10:00
Damien George 5889cf58db py/bc0: Order opcodes into groups based on their size and format. 2019-09-26 15:27:10 +10:00
Damien George 2069c563f9 py: Add support for matmul operator @ as per PEP 465.
To make progress towards MicroPython supporting Python 3.5, adding the
matmul operator is important because it's a really "low level" part of the
language, being a new token and modifications to the grammar.

It doesn't make sense to make it configurable because 1) it would make the
grammar and lexer complicated/messy; 2) no other operators are
configurable; 3) it's not a feature that can be "dynamically plugged in"
via an import.

And matmul can be useful as a general purpose user-defined operator, it
doesn't have to be just for numpy use.

Based on work done by Jim Mussared.
2019-09-26 15:12:39 +10:00
Damien George 14e203282a py/compile: Use calculation instead of switch to convert token to op. 2019-09-26 14:37:26 +10:00
Damien George 9bf2feba63 py/parse: Use calculation instead of table to convert token to operator. 2019-09-26 14:37:26 +10:00
Damien George 6ce7c051e8 py/lexer: Reorder operator tokens to match corresponding binary ops. 2019-09-26 14:37:26 +10:00
stijn 22131a6738 py/persistentcode: Enable persistent code saving for Windows ports. 2019-09-18 22:15:36 +10:00
Damien George 6e07fde895 py/mkrules.mk: Add QSTR_GLOBAL_REQUIREMENTS variable for qstr auto-gen. 2019-09-12 18:16:16 +10:00
Damien George ea060a42e9 py/vm: Factor cached map lookup code to inline function.
To reduce code duplication and allow to more easily modify this function.
2019-09-10 11:23:52 +10:00
Damien George fb73bdae62 py/mkenv.mk: Add GDB variable. 2019-09-05 22:59:06 +10:00
Damien George b29fae0c56 py/bc: Fix size calculation of UNWIND_JUMP opcode in mp_opcode_format.
Prior to this patch mp_opcode_format would calculate the incorrect size of
the MP_BC_UNWIND_JUMP opcode, missing the additional byte.  But, because
opcodes below 0x10 are unused and treated as bytes in the .mpy load/save
and freezing code, this bug did not show any symptoms, since nested unwind
jumps would rarely (if ever) reach a depth of 16 (so the extra byte of this
opcode would be between 0x01 and 0x0f and be correctly loaded/saved/frozen
simply as an undefined opcode).

This patch fixes this bug by correctly accounting for the additional byte.
        .
2019-09-02 13:30:16 +10:00
Damien George c348e79187 py/binary: Change mp_uint_t to size_t for index, size, align args.
Reduces code size for nan-box builds, otherwise changes nothing.
2019-09-02 13:14:27 +10:00
Damien George 24c3e9b283 py/modstruct: Fix struct.pack_into with unaligned offset of native type.
Following the same fix for unpack.
2019-09-02 13:14:16 +10:00
Tom McDermott 1022f9cc35 py/modstruct: Fix struct.unpack with unaligned offset of native type.
With this patch alignment is done relative to the start of the buffer that
is being unpacked, not the raw pointer value, as per CPython.

Fixes issue #3314.
2019-09-02 13:10:55 +10:00
Jeff Epler 12f13ee634 py/objtuple: Allow compatible subclasses of tuple in mp_obj_tuple_get.
As part of this patch a private macro mp_obj_is_tuple_compatible is
introduced to encapsulate the check, which is used in two locations.

Fixes #5005.
2019-08-31 00:14:11 +10:00
Milan Rossa b295df4b08 py/profile: Add debugging for sys.settrace feature. 2019-08-30 16:48:47 +10:00
Milan Rossa 310b3d1b81 py: Integrate sys.settrace feature into the VM and runtime.
This commit adds support for sys.settrace, allowing to install Python
handlers to trace execution of Python code.  The interface follows CPython
as closely as possible.  The feature is disabled by default and can be
enabled via MICROPY_PY_SYS_SETTRACE.
2019-08-30 16:44:12 +10:00
Milan Rossa c96aedad46 py/profile: Add initial implementation of sys.settrace feature. 2019-08-30 16:43:56 +10:00
Damien George dbf35d3da3 py/bc: Factor out code to get bytecode line number info into new func. 2019-08-30 16:43:46 +10:00
Damien George c7c6703950 py/compile: Improve the line numbering precision for lambdas.
Prior to this patch the line number for a lambda would be "line 1" if the
body of the lambda contained only a simple expression (with no line number
stored in the parse node).  Now the line number is always reported
correctly.
2019-08-30 16:43:46 +10:00
Damien George af20c2ead3 py: Add global default_emit_opt variable to make emit kind persistent.
mp_compile no longer takes an emit_opt argument, rather this setting is now
provided by the global default_emit_opt variable.

Now, when -X emit=native is passed as a command-line option, the emitter
will be set for all compiled modules (included imports), not just the
top-level script.

In the future there could be a way to also set this variable from a script.

Fixes issue #4267.
2019-08-28 12:47:58 +10:00
Damien George 08c1fe5569 py/vm: Don't add traceback info for exceptions that are re-raised.
With this patch exceptions that are re-raised have improved tracebacks
(less confusing, match CPython), and it makes re-raise slightly more
efficient (in time and RAM) because they no longer need to add a traceback.
Also general VM performance is not measurably affected.

Partially fixes issue #2928.
2019-08-28 12:31:53 +10:00
Damien George 16f6169c88 py/vm: Don't add traceback info for exc's propagated through a finally.
With this patch exception tracebacks that go through a finally are improved
(less confusing, match CPython), and it makes finally's slightly more
efficient (in time and RAM) because they no longer need to add a traceback.

Partially fixes issue #2928.
2019-08-28 12:31:49 +10:00
Damien George 3d7455a0bb py/py.mk: Remove trailing spaces at end of line. 2019-08-27 16:37:43 +10:00