* Configurable build-details.json name
Linux distributions that co-install multiple Python versions in the same
path (e.g. multiarch on Debian, debug builds, and free-threading) need a
way to place multiple build-details.jsons side-by-side.
PEP-739 is being updated [0] to recommend renaming in this situation. To
ensure some standardization, this PR generates appropriate names for
distributions that need to use this feature.
[0]: https://github.com/python/peps/pull/4889
* NEWS entry
* Document configure option and add to whatsnew
The power ABI specification requires that compilers maintain a back
chain by default, so unwinding already works without a dedicated
frame pointer. Don't use -fno-omit-frame-pointer on ppc64le.
-fno-omit-frame-pointer is not enough to make every target walkable by the
simple manual frame pointer unwinder.
The helper used by test_frame_pointer_unwind used to assume the frame pointer
named a two-word record where fp[0] was the previous frame pointer and fp[1]
was the return address. That is only the generic layout used by some targets.
This patch keeps that default, but moves the slots behind named offsets so
architecture-specific layouts can describe where the backchain and return
address really live.
On s390x, GCC and Clang do not emit a usable backchain unless -mbackchain is
enabled. Without it, the unwinder stops at the current C frame and the test
reports no Python frames. Once backchains are present, the helper must also
stop at the current thread's known C stack bounds; otherwise it can follow the
final backchain far enough to dereference an invalid frame and segfault.
For Linux s390x backchain frames, the documented z/Architecture stack-frame
layout saves r14, the return-address register, at byte offset 112 from the
frame pointer, so read the return address from that named slot instead of fp[1].
The 112-byte offset comes from Linux's s390 debugging documentation: its Stack
Frame Layout table shows z/Architecture backchain frames with the backchain at
offset 0 and saved r14 of the caller function at offset 112:
https://www.kernel.org/doc/html/v5.3/s390/debugging390.html#stack-frame-layout
This helper remains scoped to Linux s390x backchain frames. GNU SFrame's s390x
notes state that the s390x ELF ABI does not generally mandate where RA and FP
are saved, or whether they are saved at all:
https://sourceware.org/binutils/docs/sframe-spec.html#s390x
As Jens Remus noted, -fno-omit-frame-pointer is not needed when -mbackchain is
present.
On 32-bit ARM, GCC defaults to Thumb mode on common armhf toolchains. The Thumb
prologue keeps the saved frame pointer and link register at offsets that depend
on the generated frame, which breaks the fp[0]/fp[1] walk used by the helper.
Use -marm when it is supported for frame-pointer builds, and teach the helper
the GCC ARM-mode slots where the previous frame pointer is at fp[-1] and the
saved LR return address is at fp[0].
Co-authored-by: Petr Viktorin <encukou@gmail.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
This option changes the behavior of --enable-shared to continue to build
the libpython3.x.so shared library, but not use it for linking the
python3 interpreter executable. Instead, the executable is linked
directly against the libpython .o files as it would be with
--disable-shared.
There are two benefits of this change. First, libpython uses
thread-local storage, which is noticeably slower when used in a loaded
module instead of in the main program, because the main program can take
advantage of constant offsets from the thread state pointer but loaded
modules have to dynamically call a function __tls_get_addr() to
potentially allocate their thread-local storage area. (There is another
thread-local storage model for dynamic libraries which mitigates most of
this performance hit, but it comes at the cost of preventing
dlopen("libpython3.x.so"), which is a use case we want to preserve.)
Second, this improves the user experience around relocatable Python a
little bit, in that we don't need to use an $ORIGIN-relative path to
locate libpython3.x.so, which has some mild benefits around musl (which
does not support $ORIGIN-relative DT_NEEDED, only $ORIGIN-relative
DT_RPATH/DT_RUNPATH), users who want to make the interpreter setuid or
setcap (which prevents processing $ORIGIN), etc.
* introduce executable specific linker flags
Add PY_CORE_EXE_LDFLAGS and EXE_LDFLAGS which stores executable specific
LDFLAGS, replacing PY_CORE_LDFLAGS for building
executable targets.
If PY_CORE_EXE_LDFLAGS / EXE_LDFLAGS is not provided, then it defaults
to the value of PY_CORE_LDFLAGS which is the existing behaviour.
If both flags are supplied, and there is a need
to distinguish between executable and shared specific LDFLAGS,
in particular, PY_CORE_LDFLAGS should contain the shared specific LDFLAGS.
* documentation for new linker flags
* update Misc folder documentation
* Update Makefile.pre.in
Co-authored-by: Victor Stinner <vstinner@python.org>
---------
Co-authored-by: Filipe Laíns <filipe.lains@gmail.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: Filipe Laíns <lains@riseup.net>
The pymalloc huge page support had two problems. First, on
architectures where the default huge page size exceeds the arena
size (e.g. 32 MiB on PPC, 512 MiB on ARM64 with 64 KB base
pages), mmap with MAP_HUGETLB silently allocates a full huge page
even when the requested size is smaller. The subsequent munmap
with the original arena size then fails with EINVAL, permanently
leaking the entire huge page. Second, huge pages were always
attempted when compiled in, with no way to disable them at
runtime. On Linux, if the huge page pool is exhausted, page
faults including copy-on-write faults after fork deliver SIGBUS
and kill the process.
The arena allocator now queries the system huge page size from
/proc/meminfo and skips MAP_HUGETLB when the arena size is not a
multiple of it. Huge pages also now require explicit opt-in at
runtime via the PYTHON_PYMALLOC_HUGEPAGES environment variable,
which is read through PyConfig and respects -E and -I flags.
The config field pymalloc_hugepages is propagated to the runtime
allocators struct so the low-level arena allocator can check it
without calling getenv directly.
Steve Dower