Noah Goldstein [Fri, 25 Mar 2022 22:13:33 +0000 (17:13 -0500)]
x86: Small improvements for wcslen
Just a few QOL changes.
1. Prefer `add` > `lea` as it has high execution units it can run
on.
2. Don't break macro-fusion between `test` and `jcc`
3. Reduce code size by removing gratuitous padding bytes (-90
bytes).
geometric_mean(N=20) of all benchmarks New / Original: 0.959
All string/memory tests pass. Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Noah Goldstein [Fri, 25 Mar 2022 22:13:32 +0000 (17:13 -0500)]
x86: Small improvements for wcscpy-ssse3
Just a few small QOL changes.
1. Prefer `add` > `lea` as it has high execution units it can run
on.
2. Don't break macro-fusion between `test` and `jcc`
geometric_mean(N=20) of all benchmarks New / Original: 0.973
All string/memory tests pass. Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Noah Goldstein [Wed, 23 Mar 2022 21:57:46 +0000 (16:57 -0500)]
x86: Remove AVX str{n}casecmp
The rational is:
1. SSE42 has nearly identical logic so any benefit is minimal (3.4%
regression on Tigerlake using SSE42 versus AVX across the
benchtest suite).
2. AVX2 version covers the majority of targets that previously
prefered it.
3. The targets where AVX would still be best (SnB and IVB) are
becoming outdated.
All in all the saving the code size is worth it.
All string/memory tests pass. Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
x86: Fix __wcsncmp_avx2 in strcmp-avx2.S [BZ# 28755]
Set the wrong fallback function for `__wcsncmp_avx2_rtm`. It was set
to fallback on to `__wcscmp_avx2` instead of `__wcscmp_avx2_rtm` which
can cause spurious aborts.
This change will need to be backported.
All string/memory tests pass. Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Noah Goldstein [Wed, 23 Mar 2022 21:57:14 +0000 (16:57 -0500)]
benchtests: Add random benchmark in bench-strchr.c
Add benchmark that randomizes whether return should be NULL or pointer
to CHAR. The rationale is on many architectures there is a choice
between a predicate execution option (i.e cmovcc on x86) or a branch.
On x86 the results for cmovcc vs branch are something along the lines
of the following:
Joseph Myers [Thu, 24 Mar 2022 15:35:27 +0000 (15:35 +0000)]
Update kernel version to 5.17 in tst-mman-consts.py
This patch updates the kernel version in the test tst-mman-consts.py
to 5.17. (There are no new MAP_* constants covered by this test in
5.17 that need any other header changes.)
Joseph Myers [Wed, 23 Mar 2022 17:11:56 +0000 (17:11 +0000)]
Update syscall lists for Linux 5.17
Linux 5.17 has one new syscall, set_mempolicy_home_node. Update
syscall-names.list and regenerate the arch-syscall.h headers with
build-many-glibcs.py update-syscalls.
It is not used on rtld and ldsodef interfaces are meant to be used
solely on loader. It also removes the only usage of gcc extension
__builtin_va_arg_pack.
Sam James [Tue, 22 Mar 2022 17:33:09 +0000 (17:33 +0000)]
configure.ac: fix bashisms in configure.ac
configure scripts need to be runnable with a POSIX-compliant /bin/sh.
On many (but not all!) systems, /bin/sh is provided by Bash, so errors
like this aren't spotted. Notably Debian defaults to /bin/sh provided
by dash which doesn't tolerate such bashisms as '=='.
This retains compatibility with bash.
Fixes configure warnings/errors like:
```
checking if compiler warns about alias for function with incompatible types... yes
/var/tmp/portage/sys-libs/glibc-2.34-r10/work/glibc-2.34/configure: 4209: test: xyes: unexpected operator
```
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org> Signed-off-by: Sam James <sam@gentoo.org>
resolv: Fix unaligned accesses to fields in HEADER struct
The structure HEADER is normally aligned to a word boundary but
sometimes it needs to be accessed when aligned on a byte boundary.
This change defines a new typedef, UHEADER, with alignment 1.
It is used to ensure the fields are accessed with byte loads and
stores when necessary.
V4: Change to res_mkquery.c deleted. Small whitespace fix.
V5: Move UHEADER typedef to resolv/resolv-internal.h. Replace all
HEADER usage with UHEADER in resolv/res_send.c.
Signed-off-by: John David Anglin <dave.anglin@bell.net> Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
All other cases of failures due to lack of memory return EAI_MEMORY, so
it seems wrong to return EAI_SYSTEM here. The only reason
convert_hostent_to_gaih_addrtuple could fail is on calloc failure.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org> Reviewed-by: DJ Delorie <dj@redhat.com>
Introduce the gaih_result structure and general paradigm for cleanups
that follow to process the lookup request and return a result. A lookup
function (like text_to_binary_address), should return an integer error
code and set members of gaih_result based on what it finds. If the
function does not have a result and no errors have occurred during the
lookup, it should return 0 and res.at should be set to NULL, allowing a
subsequent function to do the lookup until we run out of options.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org> Reviewed-by: DJ Delorie <dj@redhat.com>
Refactor the code to split out the service resolution code into a
separate function. Allocate the service tuples array just once to the
size of the typeproto array, thus avoiding the unnecessary pointer
chasing and stack allocations.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org> Reviewed-by: DJ Delorie <dj@redhat.com>
Use realloc in convert_hostent_to_gaih_addrtuple and fix up pointers in
the result list so that a single block is maintained for
hostbyname3_r/hostbyname2_r and freed in gaih_inet. This result is
never merged with any other results, since the hosts database does not
permit merging.
Resolves BZ #28852.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org> Reviewed-by: DJ Delorie <dj@redhat.com>
Simplify logic for allocation of canon to remove the canonbuf variable;
canon now always points to an allocated block. Also pull the canon name
set into a separate function.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org> Reviewed-by: DJ Delorie <dj@redhat.com>
Simplify allocations and fix merge and continue actions [BZ #28931]
Allocations for address tuples is currently a bit confusing because of
the pointer chasing through PAT, making it hard to observe the sequence
in which allocations have been made. Narrow scope of the pointer
chasing through PAT so that it is only used where necessary.
This also tightens actions behaviour with the hosts database in
getaddrinfo to comply with the manual text. The "continue" action
discards previous results and the "merge" action results in an immedate
lookup failure. Consequently, chaining of allocations across modules is
no longer necessary, thus opening up cleanup opportunities.
A test has been added that checks some combinations to ensure that they
work correctly.
Resolves: BZ #28931
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org> Reviewed-by: DJ Delorie <dj@redhat.com>
Max Gautier [Mon, 21 Mar 2022 12:25:05 +0000 (09:25 -0300)]
iconv: Add UTF-7-IMAP variant in utf-7.c
UTF-7-IMAP differs from UTF-7 in the followings ways (see RFC 3501[1]
for reference) :
- The shift character is '&' instead of '+'
- There is no "optional direct characters" and the "direct characters"
set is different
- There is no implicit shift back to US-ASCII from BASE64, all BASE64
sequences MUST be terminated with '-'
Max Gautier [Mon, 21 Mar 2022 12:12:05 +0000 (09:12 -0300)]
iconv: make utf-7.c able to use variants
Add infrastructure in utf-7.c to handle variants. The approach comes from
iso646.c
The variant is defined at gconv_init time and is passed as a
supplementary variable.
Signed-off-by: Max Gautier <mg@max.gautier.name> Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Max Gautier [Sun, 20 Mar 2022 16:41:05 +0000 (17:41 +0100)]
iconv: Better mapping to RFC for UTF-7
- Direct use of characters instead of arcane arrays
- isxbase64 is not the Modified BASE64 alphabet, but the characters who
needs to trigger an explicit shift back to US-ASCII. Make that clearer
Signed-off-by: Max Gautier <mg@max.gautier.name> Reviewed-by: Adhemerval Zanellla <adhemerval.zanella@linaro.org>
Florian Weimer [Fri, 18 Mar 2022 20:27:54 +0000 (21:27 +0100)]
libio: Flush-only _IO_str_overflow must not return EOF (bug 28949)
In general, _IO_str_overflow returns the character passed as an argument
on success. However, if flush-only operation is requested by passing
EOF, returning EOF looks like an error, and the caller cannot tell
whether the operation was successful or not.
_IO_wstr_overflow had the same bug regarding WEOF.
H.J. Lu [Wed, 16 Mar 2022 15:43:48 +0000 (08:43 -0700)]
benchtests: Use "=" instead of ":=" [BZ #28970]
Use "=" instead of ":=" to allow sysdeps Makefiles to add more benches
to bench and benchset. This fixes BZ #28970. Reviewed-by: Sunil K Pandey <skpgkp2@gmail.com>
Florian Weimer [Fri, 11 Mar 2022 07:23:56 +0000 (08:23 +0100)]
nss: Protect against errno changes in function lookup (bug 28953)
dlopen may clobber errno. The nss_test_errno module uses an ELF
constructor to achieve that, but there could be internal errors
during dlopen that cause this, too. Therefore, the NSS framework
has to guard against such errno clobbers.
__nss_module_get_function is currently the only function that calls
__nss_module_load, so it is sufficient to save and restore errno
around this call.
Florian Weimer [Fri, 11 Mar 2022 07:23:56 +0000 (08:23 +0100)]
nss: Do not mention NSS test modules in <gnu/lib-names.h>
They are not actually installed. Use the nss_files version instead
in nss/Makefile, similar to how __nss_shlib_revision is derived
from LIBNSS_FILES_SO.
José Bollo [Tue, 8 Mar 2022 08:58:16 +0000 (09:58 +0100)]
libio: Ensure output buffer for wchars (bug #28828)
The _IO_wfile_overflow does not check if the write pointer for wide
data is valid before access, different than _IO_file_overflow. This
leads to crash on some cases, as described by bug 28828.
The "fgetwc(stdin);" is necessary since it triggers the bug by setting
the flag _IO_CURRENTLY_PUTTING on stdout indirectly (file wfileops.c,
function _IO_wfile_underflow, line 213).
The 'not_first' is accessed on nrl_domainname() in a non atomically
way, although it is only updated after the lock is taken.
This patch fix the double-checked locking by using acquire-release
atomic operation instead of plain load and by moving the 'not_first'
store only after 'domain' is actually set.
Su Lifan [Fri, 25 Feb 2022 08:58:07 +0000 (16:58 +0800)]
benchtests: make compare_strings.py accept string as attribute value
Commit ac759b1fbf28a82d99afde9046f8b72c7cba5dae added attribute
"overlap" to bench-memmove-walk, whose value is a string. This change
makes compare_strings.py fail since benchout_strings.schema.json
requires the values of attributes to be number.
This patch relaxes such constraint. Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
1. Instructions proceeded by a tab.
2. Instruction less than 8 characters in length have a tab
between it and the first operand.
3. Instruction greater than 7 characters in length have a
space between it and the first operand.
4. Tabs after `#define`d names and their value.
5. 8 space at the beginning of line replaced by tab.
6. Indent comments with code.
7. Remove redundent .text section.
8. 1 space between line content and line comment.
9. Space after all commas.
1. Instructions proceeded by a tab.
2. Instruction less than 8 characters in length have a tab
between it and the first operand.
3. Instruction greater than 7 characters in length have a
space between it and the first operand.
4. Tabs after `#define`d names and their value.
5. 8 space at the beginning of line replaced by tab.
6. Indent comments with code.
7. Remove redundent .text section.
8. 1 space between line content and line comment.
9. Space after all commas.
1. Instructions proceeded by a tab.
2. Instruction less than 8 characters in length have a tab
between it and the first operand.
3. Instruction greater than 7 characters in length have a
space between it and the first operand.
4. Tabs after `#define`d names and their value.
5. 8 space at the beginning of line replaced by tab.
6. Indent comments with code.
7. Remove redundent .text section.
8. 1 space between line content and line comment.
9. Space after all commas.
1. Instructions proceeded by a tab.
2. Instruction less than 8 characters in length have a tab
between it and the first operand.
3. Instruction greater than 7 characters in length have a
space between it and the first operand.
4. Tabs after `#define`d names and their value.
5. 8 space at the beginning of line replaced by tab.
6. Indent comments with code.
7. Remove redundent .text section.
8. 1 space between line content and line comment.
9. Space after all commas.