Current implementation of strcmp for power10 has
performance regression for multiple small sizes
and alignment combination.
Most of these performance issues are fixed by this
patch. The compare loop is unrolled and page crosses
of unrolled loop is handled.
Thanks to Paul E. Murphy for helping in fixing the
performance issues.
Signed-off-by: Amrita H S <amritahs@linux.vnet.ibm.com>
Co-Authored-By: Paul E. Murphy <murphyp@linux.ibm.com>
Reviewed-by: Rajalakshmi Srinivasaraghavan <rajis@linux.ibm.com>
Optimized memchr for POWER10 based on existing rawmemchr and strlen.
Reordering instructions and loop unrolling helped in getting better performance.
Reviewed-by: Rajalakshmi Srinivasaraghavan <rajis@linux.ibm.com>
The PT_GNU_PROPERTY segment is scanned before PT_NOTE. For binaries
with the PT_GNU_PROPERTY segment, we can check it to avoid scan of
the PT_NOTE segment.
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
This patch is based on __strcmp_power9 and __strlen_power10.
Improvements from __strcmp_power9:
1. Uses new POWER10 instructions
- This code uses lxvp to decrease contention on load
by loading 32 bytes per instruction.
2. Performance implication
- This version has around 30% better performance on average.
- Performance regression is seen for a specific combination
of sizes and alignments. Some of them is observed without
changes also, while rest may be induced by the patch.
Signed-off-by: Amrita H S <amritahs@linux.vnet.ibm.com>
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
To avoid any environment variable to change setuid binaries
semantics.
Checked on x86_64-linux-gnu.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Loader already ignores LD_DEBUG, LD_DEBUG_OUTPUT, and
LD_TRACE_LOADED_OBJECTS. Both LD_WARN and LD_VERBOSE are similar to
LD_DEBUG, in the sense they enable additional checks and debug
information, so it makes sense to disable them.
Also add both LD_VERBOSE and LD_WARN on filtered environment variables
for setuid binaries.
Checked on x86_64-linux-gnu.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
The .cfi_return_column directive changes the return column for the whole
FDE range. But the actual intent is to tell the unwinder that the value
in x30 (lr) now resides in x15 after the move, and that is expressed by
the .cfi_register directive.
New implementation is based on the existing exp/exp2, with different
reduction constants and polynomial. Worst-case error in round-to-
nearest is 0.513 ULP.
The exp/exp2 shared table is reused for exp10 - .rodata size of
e_exp_data increases by 64 bytes.
As for exp/exp2, targets with single-instruction rounding/conversion
intrinsics can use them by toggling TOINT_INTRINSICS=1 and adding the
necessary code to their math_private.h.
Improvements on Neoverse V1 compared to current GLIBC master:
exp10 thruput: 3.3x in [-0x1.439b746e36b52p+8 0x1.34413509f79ffp+8]
exp10 latency: 1.8x in [-0x1.439b746e36b52p+8 0x1.34413509f79ffp+8]
Tested on:
aarch64-linux-gnu (TOINT_INTRINSICS, fma contraction) and
x86_64-linux-gnu (!TOINT_INTRINSICS, no fma contraction)
Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
It is not strictly required by the POSIX, since O_PATH is a Linux
extension, but it is QoI to fail early instead of at readdir. Also
the check is free, since fdopendir already checks if the file
descriptor is opened for read.
Checked on x86_64-linux-gnu.
The linker just concatenates the .init and .fini sections which
results in the complete _init and _fini functions. If needed the
linker adds padding bytes due to an alignment. GNU ld is adding
NOPs, which is fine. But e.g. mold is adding traps which results
in broken _init and _fini functions.
Thus this patch removes the alignment in .init and .fini sections
in crtn.S files.
We keep the 4 byte function alignment in crti.S files. As the
assembler now also outputs the start of _init and _fini functions
as multiples of 4 byte, it perhaps has to fill it. Although GNU as
is using NOPs here, to be sure, we just keep the alignment with
0x07 (=NOPs) at the end of crti.S.
In order to avoid an obvious NOP slide in _fini, this patch also
uses an lg instead of lgr instruction. Then the emitted instructions
needs a multiple of 4 bytes.
Current code aligns to 2x VEC_SIZE. Aligning to 2x has no affect on
performance other than potentially resulting in an additional
iteration of the loop.
1x maintains aligned stores (the only reason to align in this case)
and doesn't incur any unnecessary loop iterations.
Reviewed-by: Sunil K Pandey <skpgkp2@gmail.com>
This patch adds the TCP_MD5SIG_FLAG_IFINDEX constant from Linux 5.6 to
sysdeps/gnu/netinet/tcp.h and updates struct tcp_md5sig accordingly to
contain the device index.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
A recent commit, apparently commit
6c6fce572f "elf: Remove /etc/suid-debug
support", resulted in localplt failures for i686-gnu and x86_64-gnu:
Missing required PLT reference: ld.so: __access_noerrno
After that commit, __access_noerrno is actually no longer used at all.
So rather than just removing the localplt expectation for that symbol
for Hurd, completely remove all definitions of and references to that
symbol.
Tested for x86_64, and with build-many-glibcs.py for i686-gnu and
x86_64-gnu.
This restore the 2.33 semantic for arena_get2. It was changed by
11a02b035b to avoid arena_get2 call malloc (back when __get_nproc
was refactored to use an scratch_buffer - 903bc7dcc2). The
__get_nproc was refactored over then and now it also avoid to call
malloc.
The 11a02b035b did not take in consideration any performance
implication, which should have been discussed properly. The
__get_nprocs_sched is still used as a fallback mechanism if procfs
and sysfs is not acessible.
Checked on x86_64-linux-gnu.
Reviewed-by: DJ Delorie <dj@redhat.com>
These were broken by the new atan2 functions, as they were only
set up for univariate functions. Arity is now detected from the
input file - this revealed a mistake that the double-precision
inputs were being used for both single- and double-precision
routines, which is now remedied.
The _dl_non_dynamic_init does not parse LD_PROFILE, which does not
enable profile for dlopen objects. Since dlopen is deprecated for
static objects, it is better to remove the support.
It also allows to trim down libc.a of profile support.
Checked on x86_64-linux-gnu.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Using the memcmp symbol directly allows the compile to inline the
memcmp calls (especially because _dl_tunable_set_hwcaps uses constants
values), generating better code.
Checked with tst-tunables on s390x-linux-gnu (qemu system).
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
The dl-symbol-redir-ifunc.h redirects compiler-generated libcalls to
arch-specific memory implementations to avoid ifunc calls where it is not
yet possible. The memcmp-isa-default-impl.h aims to fix the same issue
by calling the specific memset implementation directly.
Using the memcmp symbol directly allows the compiler to inline the memset
calls (especially because _dl_tunable_set_hwcaps uses constants values),
generating better code.
Checked on x86_64-linux-gnu.
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
The strlen might trigger and invalid GOT entry if it used before
the process is self-relocated (for instance on dl-tunables if any
error occurs).
For i386, _dl_writev with PIE requires to use the old 'int $0x80'
syscall mode because the calling the TLS register (gs) is not yet
initialized.
Checked on x86_64-linux-gnu.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Some environment variables allow alteration of allocator behavior
across setuid boundaries, where a setuid program may ignore the
tunable, but its non-setuid child can read it and adjust the memory
allocator behavior accordingly.
Most library behavior tunings is limited to the current process and does
not bleed in scope; so it is unclear how pratical this misfeature is.
If behavior change across privilege boundaries is desirable, it would be
better done with a wrapper program around the non-setuid child that sets
these envvars, instead of using the setuid process as the messenger.
The patch as fixes tst-env-setuid, where it fail if any unsecvars is
set. It also adds a dynamic test, although it requires
--enable-hardcoded-path-in-tests so kernel correctly sets the setuid
bit (using the loader command directly would require to set the
setuid bit on the loader itself, which is not a usual deployment).
Co-authored-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Checked on x86_64-linux-gnu.
Reviewed-by: DJ Delorie <dj@redhat.com>
The tunable privilege levels were a retrofit to try and keep the malloc
tunable environment variables' behavior unchanged across security
boundaries. However, CVE-2023-4911 shows how tricky can be
tunable parsing in a security-sensitive environment.
Not only parsing, but the malloc tunable essentially changes some
semantics on setuid/setgid processes. Although it is not a direct
security issue, allowing users to change setuid/setgid semantics is not
a good security practice, and requires extra code and analysis to check
if each tunable is safe to use on all security boundaries.
It also means that security opt-in features, like aarch64 MTE, would
need to be explicit enabled by an administrator with a wrapper script
or with a possible future system-wide tunable setting.
Co-authored-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
setuid/setgid process now ignores any glibc tunables, and filters out
all environment variables that might changes its behavior. This patch
also adds GLIBC_TUNABLES, so any spawned process by setuid/setgid
processes should set tunable explicitly.
Checked on x86_64-linux-gnu.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Linux 6.6 (09da082b07bbae1c) added support for fchmodat2, which has
similar semantics as fchmodat with an extra flag argument. This
allows fchmodat to implement AT_SYMLINK_NOFOLLOW and AT_EMPTY_PATH
without the need for procfs.
The syscall is registered on all architectures (with value of 452
except on alpha which is 562, commit 78252deb023cf087).
The tst-lchmod.c requires a small fix where fchmodat checks two
contradictory assertions ('(st.st_mode & 0777) == 2' and
'(st.st_mode & 0777) == 3').
Checked on x86_64-linux-gnu on a 6.6 kernel.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
strrchr-evex-base used `vpcompress{b|d}` in the page cross logic but
was missing the CPU_FEATURE checks for VBMI2 in the
ifunc/ifunc-impl-list.
The fix is either to add those checks or change the logic to not use
`vpcompress{b|d}`. Choosing the latter here so that the strrchr-evex
implementation is usable on SKX.
New implementation is a bit slower, but this is in a cold path so its
probably okay.
Fixes commit a61933fe27 ("sparc: Remove bzero optimization") that
after moving code jumped to the wrong label 4.
Verfied by successfully running string/test-memset on sparc32.
Signed-off-by: Andreas Larsson <andreas@gaisler.com>
Signed-off-by: Ludwig Rydberg <ludwig.rydberg@gaisler.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The latest implementations of memcpy are actually faster than the Falkor
implementations [1], so remove the falkor/phecda ifuncs for memcpy and
the now unused IS_FALKOR/IS_PHECDA defines.
[1] https://sourceware.org/pipermail/libc-alpha/2022-December/144227.html
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Add a specialized memset for the common ZVA size of 64 to avoid the
overhead of reading the ZVA size. Since the code is identical to
__memset_falkor, remove the latter.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cleanup emag memset - merge the memset_base64.S file, remove
the unused ZVA code (since it is disabled on emag).
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The PR_SET_VMA_ANON_NAME support is only enabled through a configurable
kernel switch, mainly because assigning a name to a
anonymous virtual memory area might prevent that area from being
merged with adjacent virtual memory areas.
For instance, with the following code:
void *p1 = mmap (NULL,
1024 * 4096,
PROT_READ | PROT_WRITE,
MAP_PRIVATE | MAP_ANONYMOUS,
-1,
0);
void *p2 = mmap (p1 + (1024 * 4096),
1024 * 4096,
PROT_READ | PROT_WRITE,
MAP_PRIVATE | MAP_ANONYMOUS,
-1,
0);
The kernel will potentially merge both mappings resulting in only one
segment of size 0x800000. If the segment is names with
PR_SET_VMA_ANON_NAME with different names, it results in two mappings.
Although this will unlikely be an issue for pthread stacks and malloc
arenas (since for pthread stacks the guard page will result in
a PROT_NONE segment, similar to the alignment requirement for the arena
block), it still might prevent the mmap memory allocated for detail
malloc.
There is also another potential scalability issue, where the prctl
requires
to take the mmap global lock which is still not fully fixed in Linux
[1] (for pthread stacks and arenas, it is mitigated by the stack
cached and the arena reuse).
So this patch disables anonymous mapping annotations as default and
add a new tunable, glibc.mem.decorate_maps, can be used to enable
it.
[1] https://lwn.net/Articles/906852/
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
Reviewed-by: DJ Delorie <dj@redhat.com>
Linux 5.17 added support to naming anonymous virtual memory areas
through the prctl syscall. The __set_vma_name is a wrapper to avoid
optimizing the prctl call if the kernel does not support it.
If the kernel does not support PR_SET_VMA_ANON_NAME, prctl returns
EINVAL. And it also returns the same error for an invalid argument.
Since it is an internal-only API, it assumes well-formatted input:
aligned START, with (START, START+LEN) being a valid memory range,
and NAME with a limit of 80 characters without an invalid one
("\\`$[]").
Reviewed-by: DJ Delorie <dj@redhat.com>
When invoking sem_open with O_CREAT as one of its flags, we'll end up
in the second part of sem_open's "if ((oflag & O_CREAT) == 0 || (oflag
& O_EXCL) == 0)", which means that we don't expect the semaphore file
to exist.
In that part, open_flags is initialized as "O_RDWR | O_CREAT | O_EXCL
| O_CLOEXEC" and there's an attempt to open(2) the file, which will
likely fail because it won't exist. After that first (expected)
failure, some cleanup is done and we go back to the label "try_again",
which lives in the first part of the aforementioned "if".
The problem is that, in that part of the code, we expect the semaphore
file to exist, and as such O_CREAT (this time the flag we pass to
open(2)) needs to be cleaned from open_flags, otherwise we'll see
another failure (this time unexpected) when trying to open the file,
which will lead the call to sem_open to fail as well.
This can cause very strange bugs, especially with OpenMPI, which makes
extensive use of semaphores.
Fix the bug by simplifying the logic when choosing open(2) flags and
making sure O_CREAT is not set when the semaphore file is expected to
exist.
A regression test for this issue would require a complex and cpu time
consuming logic, since to trigger the wrong code path is not
straightforward due the racy condition. There is a somewhat reliable
reproducer in the bug, but it requires using OpenMPI.
This resolves BZ #30789.
See also: https://bugs.launchpad.net/ubuntu/+source/h5py/+bug/2031912
Signed-off-by: Sergio Durigan Junior <sergiodj@sergiodj.net>
Co-Authored-By: Simon Chopin <simon.chopin@canonical.com>
Co-Authored-By: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
Fixes: 533deafbdf ("Use O_CLOEXEC in more places (BZ #15722)")
