Although the original motivation given for the workaround[1] is correct,
nlr.o and nlrthumb.o are linked with a small enough distance that the
problem does not occur, and the workaround isn't necessary. The distance
between the b instruction and its target (nlr_push_tail) is just 64
bytes[2], well within the ±2046 byte range addressable by an
unconditional branch instruction in Thumb mode.
The workaround induces a relocation in the text section (textrel), which
isn't supported everywhere, notably not on musl-libc[3], where it causes
a crash on start-up. With the workaround removed, micropython works on an
ARMv5T Linux system built with musl-libc.
This commit changes nlrthumb.c to use a direct jump by default, but
leaves the long jump workaround as an option for those cases where it's
actually needed.
[1]: commit dd376a239d
Author: Damien George <damien.p.george@gmail.com>
Date: Fri Sep 1 15:25:29 2017 +1000
py/nlrthumb: Get working again on standard Thumb arch (ie not Thumb2).
"b" on Thumb might not be long enough for the jump to nlr_push_tail so
it must be done indirectly.
[2]: Excerpt from objdump -d micropython:
000095c4 <nlr_push_tail>:
95c4: b510 push {r4, lr}
95c6: 0004 movs r4, r0
95c8: f02d fd42 bl 37050 <mp_thread_get_state>
95cc: 6943 ldr r3, [r0, #20]
95ce: 6023 str r3, [r4, #0]
95d0: 6144 str r4, [r0, #20]
95d2: 2000 movs r0, #0
95d4: bd10 pop {r4, pc}
000095d6 <nlr_pop>:
95d6: b510 push {r4, lr}
95d8: f02d fd3a bl 37050 <mp_thread_get_state>
95dc: 6943 ldr r3, [r0, #20]
95de: 681b ldr r3, [r3, #0]
95e0: 6143 str r3, [r0, #20]
95e2: bd10 pop {r4, pc}
000095e4 <nlr_push>:
95e4: 60c4 str r4, [r0, #12]
95e6: 6105 str r5, [r0, #16]
95e8: 6146 str r6, [r0, #20]
95ea: 6187 str r7, [r0, #24]
95ec: 4641 mov r1, r8
95ee: 61c1 str r1, [r0, #28]
95f0: 4649 mov r1, r9
95f2: 6201 str r1, [r0, #32]
95f4: 4651 mov r1, sl
95f6: 6241 str r1, [r0, #36] @ 0x24
95f8: 4659 mov r1, fp
95fa: 6281 str r1, [r0, #40] @ 0x28
95fc: 4669 mov r1, sp
95fe: 62c1 str r1, [r0, #44] @ 0x2c
9600: 4671 mov r1, lr
9602: 6081 str r1, [r0, #8]
9604: e7de b.n 95c4 <nlr_push_tail>
[3]: https://www.openwall.com/lists/musl/2020/09/25/4
Signed-off-by: J. Neuschäfer <j.ne@posteo.net>
Two cases, one assigning to a slice.
Closes https://github.com/micropython/micropython/issues/13283
Second is extending a slice from itself, similar logic.
In both cases the problem occurs when m_renew causes realloc to move the
buffer, leaving a dangling pointer behind.
There are more complex and hard to fix cases when either argument is a
memoryview into the buffer, currently resizing to a new address breaks
memoryviews into that object.
Reproducing this bug and confirming the fix was done by running the unix
port under valgrind with GC-aware extensions.
Note in default configurations with GIL this bug exists but has no impact
(the free buffer won't be reused while the function is still executing, and
is no longer referenced after it returns).
Signed-off-by: Angus Gratton <angus@redyak.com.au>
This commit swaps the order of the `flags` and `name` struct initialisers
for `mp_obj_type_t`, to fix an incompatibility with C++. The original
order of the initialiser didn't match the definition of the type, and
although that's still legal C, it's not legal C++.
Signed-off-by: Vonasmic <kasarkal123@gmail.com>
This is required because the .mpy native ABI was changed by the
introduction of `mp_proto_fun_t`, see commits:
- 416465d81e
- 5e3006f117
- e2ff00e811
And three `mp_binary` functions were added to `mp_fun_table` in
commit d2276f0d41.
Signed-off-by: Damien George <damien@micropython.org>
These are needed to read/write array.array objects, which is useful in
native code to provide fast extensions that work with big arrays of data.
Signed-off-by: Damien George <damien@micropython.org>
Newer versions of gcc (14 and up) have more sophisticated dead-code
detection, and the asm clobbers list needs to contain "memory" to inform
the compiler that the asm code actually does something.
Tested that adding this "memory" line does not change the generated code on
ARM Thumb2, x86-64 and Xtensa targets (using gcc 13.2).
Fixes issue #14115.
Signed-off-by: Damien George <damien@micropython.org>
Enabled by MICROPY_COMPILE_ALLOW_TOP_LEVEL_AWAIT. When enabled, this means
that scope such as module-level functions and REPL statements can yield.
The outer C code must then handle this yielded generator.
Signed-off-by: Damien George <damien@micropython.org>
This fixes a bug where a random Python object may become
un-garbage-collectable until an enclosing Python file (compiled on device)
finishes executing.
Details:
The mp_parse_tree_t structure is stored on the stack in top-level functions
such as parse_compile_execute() in pyexec.c (and others).
Although it quickly falls out of scope in these functions, it is usually
still in the current stack frame when the compiled code executes. (Compiler
dependent, but usually it's one stack push per function.)
This means if any Python object happens to allocate at the same address as
the (freed) root parse tree chunk, it's un-garbage-collectable as there's a
(dangling) pointer up the stack referencing this same address.
As reported by @GitHubsSilverBullet here:
https://github.com/orgs/micropython/discussions/14116#discussioncomment-8837214
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
This commit implements the 'e' half-float format: 10-bit mantissa, 5-bit
exponent. It uses native _Float16 if supported by the compiler, otherwise
uses custom bitshifting encoding/decoding routines.
Signed-off-by: Matthias Urlichs <matthias@urlichs.de>
Signed-off-by: Damien George <damien@micropython.org>
ASM_NOT_REG is optional, it can be synthesised by xor(reg, -1).
ASM_NEG_REG can also be synthesised with a subtraction, but most
architectures have a dedicated instruction for it.
Signed-off-by: Damien George <damien@micropython.org>
Add `pop()`, `appendleft()`, and `extend()` methods, support iteration
and indexing, and initializing from an existing sequence.
Iteration and indexing (subscription) have independent configuration flags
to enable them. They are enabled by default at the same level that
collections.deque is enabled (the extra features level).
Also add tests for checking new behavior.
Signed-off-by: Damien George <damien@micropython.org>
This new machine-module driver provides a "USBDevice" singleton object and
a shim TinyUSB "runtime" driver that delegates the descriptors and all of
the TinyUSB callbacks to Python functions. This allows writing arbitrary
USB devices in pure Python. It's also possible to have a base built-in
USB device implemented in C (eg CDC, or CDC+MSC) and a Python USB device
added on top of that.
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
Disabled by default, but enabled on all boards that previously had
`MICROPY_PY_MACHINE_BARE_METAL_FUNCS` enabled.
Signed-off-by: Damien George <damien@micropython.org>
The STATIC macro was introduced a very long time ago in commit
d5df6cd44a. The original reason for this was
to have the option to define it to nothing so that all static functions
become global functions and therefore visible to certain debug tools, so
one could do function size comparison and other things.
This STATIC feature is rarely (if ever) used. And with the use of LTO and
heavy inline optimisation, analysing the size of individual functions when
they are not static is not a good representation of the size of code when
fully optimised.
So the macro does not have much use and it's simpler to just remove it.
Then you know exactly what it's doing. For example, newcomers don't have
to learn what the STATIC macro is and why it exists. Reading the code is
also less "loud" with a lowercase static.
One other minor point in favour of removing it, is that it stops bugs with
`STATIC inline`, which should always be `static inline`.
Methodology for this commit was:
1) git ls-files | egrep '\.[ch]$' | \
xargs sed -Ei "s/(^| )STATIC($| )/\1static\2/"
2) Do some manual cleanup in the diff by searching for the word STATIC in
comments and changing those back.
3) "git-grep STATIC docs/", manually fixed those cases.
4) "rg -t python STATIC", manually fixed codegen lines that used STATIC.
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
This call used to be needed when there was an `emit_bc_pre()` function that
needed to be called at the start of each emitted bytecode. But in
8e7745eb31 that function was removed and now
the call to `mp_emit_bc_adjust_stack_size()` does nothing when adjusting by
0 entries, so it can be removed.
Signed-off-by: Damien George <damien@micropython.org>
For boards with MICROPY_CONFIG_ROM_LEVEL_AT_LEAST_CORE_FEATURES and up.
This gets samd21 boards working (which need the vfs module in _boot.py),
B_L072Z_LRWAN1, and nrf boards with smaller MCUs that use CORE or BASIC
feature levels.
Signed-off-by: robert-hh <robert@hammelrath.com>
This is now easy to support, since the first machine-word of a native
function tells how to find the prelude, from which the function name can be
extracted in the same way as for bytecode.
Signed-off-by: Damien George <damien@micropython.org>
Now native functions and native generators have similar behaviour: the
first machine-word of their code is an index to get to the prelude. This
simplifies the handling of these types of functions, and also reduces the
size of the emitted native machine code by no longer requiring special code
at the start of the function to load a pointer to the prelude.
Signed-off-by: Damien George <damien@micropython.org>
Viper functions are quite different to native functions and benefit from
being a separate type. For example, viper functions don't have a bytecode-
style prelude, and don't support generators or default arguments.
Signed-off-by: Damien George <damien@micropython.org>
They are no longer used. The new `mp_obj_malloc_with_finaliser()` macros
should be used instead, which force the setting of the `base.type` field.
And there's always `m_malloc_with_finaliser()` if needed.
Signed-off-by: Damien George <damien@micropython.org>
Following 709e8328d9.
Using this helps to reduce code size. And it ensure that the type is
always set as soon as the object is allocated, which is important for the
GC to function correctly.
Signed-off-by: Damien George <damien@micropython.org>
MICROPY_GIT_HASH was removed in 69e34b6b6b
but it is useful for, and used by, third-party code to tell which hash of
MicroPython is used.
Signed-off-by: Damien George <damien@micropython.org>
Allows bytecode itself to be used instead of an mp_raw_code_t in the simple
and common cases of a bytecode function without any children.
This can be used to further reduce frozen code size, and has the potential
to optimise other areas like importing.
Signed-off-by: Damien George <damien@micropython.org>
To simplify their access and reduce code size.
The `scope_flags` member is only ever used to determine if a function is a
generator or not, so make it reflect that fact as a bool type.
Signed-off-by: Damien George <damien@micropython.org>
The `asm_n_pos_args` and `asm_type_sig` members of `mp_raw_code_t` are only
used for raw codes of type MP_CODE_NATIVE_ASM, which are rare, for example
in frozen code. So using a truncated `mp_raw_code_t` in these cases helps
to reduce frozen code size on targets that have MICROPY_EMIT_INLINE_ASM
enabled.
With this, change in firmware size of RPI_PICO builds is -648.
Signed-off-by: Damien George <damien@micropython.org>
The mp_raw_code_t struct has been reordered and some members resized. The
`n_pos_args` member is renamed to `asm_n_pos_args`, and `type_sig` renamed
to `asm_type_sig` to indicate that these are used only for the inline-asm
emitters. These two members are also grouped together in the struct.
The justifications for resizing the members are:
- `fun_data_len` can be 32-bits without issue
- `n_children` is already limited to 16-bits by
`mp_emit_common_t::ct_cur_child`
- `scope_flags` is already limited to 16-bits by `scope_t::scope_flags`
- `prelude_offset` is already limited to 16-bits by the argument to
`mp_emit_glue_assign_native()`
- it's reasonable to limit the maximim number of inline-asm arguments to 12
(24 bits for `asm_type_sig` divided by 2)
This change helps to reduce frozen code size (and in some cases RAM usage)
in the following cases:
- 64-bit targets
- builds with MICROPY_PY_SYS_SETTRACE enabled
- builds with MICROPY_EMIT_MACHINE_CODE enabled but MICROPY_EMIT_INLINE_ASM
disabled
With this change, unix 64-bit builds are -4080 bytes in size. Bare-metal
ports like rp2 are unchanged (because mp_raw_code_t is still 32 bytes on
those 32-bit targets).
Signed-off-by: Damien George <damien@micropython.org>
If a return is executed within the try block of a try-finally then the
return value is stored on the top of the Python stack during the execution
of the finally block. In this case the Python stack is one larger than it
normally would be in the finally block.
Prior to this commit, the compiler was not taking this case into account
and could have a Python stack overflow if the Python stack used by the
finally block was more than that used elsewhere in the function. In such
a scenario the last argument of the function would be clobbered by the
top-most temporary value used in the deepest Python expression/statement.
This commit fixes that case by making sure enough Python stack is allocated
to the function.
Fixes issue #13562.
Signed-off-by: Damien George <damien@micropython.org>
stat_path is only called with stringified vstr_t objects.
Thus, pulling the stringification into the function replaces three
function calls with one, saving a few bytes.
Signed-off-by: Matthias Urlichs <matthias@urlichs.de>