{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":13868694,"defaultBranch":"master","name":"glibc","ownerLogin":"bminor","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2013-10-25T18:26:45.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/2345100?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1718068611.0","currentOid":""},"activityList":{"items":[{"before":"e00ff0233807361ff898817c6f355806fe569a81","after":"98f5e8eb9613399ceeaadeb46acb61fa99a34ab5","ref":"refs/heads/pranavk/grte_v5_plus","pushedAt":"2024-06-12T01:16:59.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Avoid cancellable I/O primitives in ld.so.\n\nNeither the <dlfcn.h> entry points, nor lazy symbol resolution, nor\ninitial shared library load-up, are cancellation points, so ld.so\nshould exclusively use I/O primitives that are not cancellable.  We\ncurrently achieve this by having the cancellation hooks compile as\nno-ops when IS_IN(rtld); this patch changes to using exclusively\n_nocancel primitives in the source code instead, which makes the\nintent clearer and significantly reduces the amount of code compiled\nunder IS_IN(rtld) as well as IS_IN(libc) -- in particular,\nelf/Makefile no longer thinks we require a copy of unwind.c in\nrtld-libc.a.  (The older mechanism is preserved as a backstop.)\n\nThe bulk of the change is splitting up the files that define the\n_nocancel I/O functions, so they don't also define the variants that\n*are* cancellation points; after which, the existing logic for picking\nout the bits of libc that need to be recompiled as part of ld.so Just\nWorks.  I did this for all of the _nocancel functions, not just the\nones used by ld.so, for consistency.\n\nfcntl was a little tricky because it's only a cancellation point for\ncertain opcodes (F_SETLKW(64), which can block), and the existing\n__fcntl_nocancel wasn't applying the FCNTL_ADJUST_CMD hook, which\nstrikes me as asking for trouble, especially as the only nontrivial\ndefinition of FCNTL_ADJUST_CMD (for powerpc64) changes F_*LK* opcodes.\nTo fix this, fcntl_common moves to fcntl_nocancel.c along with\n__fcntl_nocancel, and changes its name to the extern (but hidden)\nsymbol __fcntl_nocancel_adjusted, so that regular fcntl can continue\ncalling it.  __fcntl_nocancel now applies FCNTL_ADJUST_CMD; so that\nboth both fcntl.c and fcntl_nocancel.c can see it, the only nontrivial\ndefinition moves from sysdeps/u/s/l/powerpc/powerpc64/fcntl.c to\n.../powerpc64/sysdep.h and becomes entirely a macro, instead of a macro\nthat calls an inline function.\n\nThe nptl version of libpthread also changes a little, because its\n\"compat-routines\" formerly included files that defined all the\n_nocancel functions it uses; instead of continuing to duplicate them,\nI exported the relevant ones from libc.so as GLIBC_PRIVATE.  Since the\nLinux fcntl.c calls a function defined by fcntl_nocancel.c, it can no\nlonger be used from libpthread.so; instead, introduce a custom\nforwarder, pt-fcntl.c, and export __libc_fcntl from libc.so as\nGLIBC_PRIVATE.  The nios2-linux ABI doesn't include a copy of vfork()\nin libpthread, and it was handling that by manipulating\nlibpthread-routines in .../linux/nios2/Makefile; it is cleaner to do\nwhat other such ports do, and have a pt-vfork.S that defines no symbols.\n\nRight now, it appears that Hurd does not implement _nocancel I/O, so\nsysdeps/generic/not-cancel.h will forward everything back to the\nregular functions.  This changed the names of some of the functions\nthat sysdeps/mach/hurd/dl-sysdep.c needs to interpose.\n\n\t* elf/dl-load.c, elf/dl-misc.c, elf/dl-profile.c, elf/rtld.c\n\t* sysdeps/unix/sysv/linux/dl-sysdep.c\n\tInclude not-cancel.h.  Use __close_nocancel instead of __close,\n\t__open64_nocancel instead of __open, __read_nocancel instead of\n\t__libc_read, and __write_nocancel instead of __libc_write.\n\n\t* csu/check_fds.c (check_one_fd)\n\t* sysdeps/posix/fdopendir.c (__fdopendir)\n\t* sysdeps/posix/opendir.c (__alloc_dir): Use __fcntl_nocancel\n        instead of __fcntl and/or __libc_fcntl.\n\n\t* sysdeps/unix/sysv/linux/pthread_setname.c (pthread_setname_np)\n\t* sysdeps/unix/sysv/linux/pthread_getname.c (pthread_getname_np)\n        * sysdeps/unix/sysv/linux/i386/smp.h (is_smp_system):\n\tUse __open64_nocancel instead of __open_nocancel.\n\n\t* sysdeps/unix/sysv/linux/not-cancel.h: Move all of the\n\thidden_proto declarations to the end and issue them if either\n\tIS_IN(libc) or IS_IN(rtld).\n\t* sysdeps/unix/sysv/linux/Makefile [subdir=io] (sysdep_routines):\n\tAdd close_nocancel, fcntl_nocancel, nanosleep_nocancel,\n\topen_nocancel, open64_nocancel, openat_nocancel, pause_nocancel,\n\tread_nocancel, waitpid_nocancel, write_nocancel.\n\n        * io/Versions [GLIBC_PRIVATE]: Add __libc_fcntl,\n        __fcntl_nocancel, __open64_nocancel, __write_nocancel.\n        * posix/Versions: Add __nanosleep_nocancel, __pause_nocancel.\n\n        * nptl/pt-fcntl.c: New file.\n        * nptl/Makefile (pthread-compat-wrappers): Remove fcntl.\n        (libpthread-routines): Add pt-fcntl.\n        * include/fcntl.h (__fcntl_nocancel_adjusted): New function.\n        (__libc_fcntl): Remove attribute_hidden.\n\t* sysdeps/unix/sysv/linux/fcntl.c (__libc_fcntl): Call\n\t__fcntl_nocancel_adjusted, not fcntl_common.\n        (__fcntl_nocancel): Move to new file fcntl_nocancel.c.\n\t(fcntl_common): Rename to __fcntl_nocancel_adjusted; also move\n\tto fcntl_nocancel.c.\n\t* sysdeps/unix/sysv/linux/fcntl_nocancel.c: New file.\n\t* sysdeps/unix/sysv/linux/powerpc/powerpc64/fcntl.c: Remove file.\n\t* sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h:\n\tDefine FCNTL_ADJUST_CMD here, as a self-contained macro.\n\n\t* sysdeps/unix/sysv/linux/close.c: Move __close_nocancel to...\n\t* sysdeps/unix/sysv/linux/close_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/nanosleep.c: Move __nanosleep_nocancel to...\n\t* sysdeps/unix/sysv/linux/nanosleep_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/open.c: Move __open_nocancel to...\n\t* sysdeps/unix/sysv/linux/open_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/open64.c: Move __open64_nocancel to...\n\t* sysdeps/unix/sysv/linux/open64_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/openat.c: Move __openat_nocancel to...\n\t* sysdeps/unix/sysv/linux/openat_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/openat64.c: Move __openat64_nocancel to...\n\t* sysdeps/unix/sysv/linux/openat64_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/pause.c: Move __pause_nocancel to...\n\t* sysdeps/unix/sysv/linux/pause_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/read.c: Move __read_nocancel to...\n\t* sysdeps/unix/sysv/linux/read_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/waitpid.c: Move __waitpid_nocancel to...\n\t* sysdeps/unix/sysv/linux/waitpid_nocancel.c: ...this new file.\n\t* sysdeps/unix/sysv/linux/write.c: Move __write_nocancel to...\n\t* sysdeps/unix/sysv/linux/write_nocancel.c: ...this new file.\n\n        * sysdeps/unix/sysv/linux/nios2/Makefile: Don't override\n        libpthread-routines.\n        * sysdeps/unix/sysv/linux/nios2/pt-vfork.S: New file which\n        defines nothing.\n\n        * sysdeps/mach/hurd/dl-sysdep.c: Define __read instead of\n        __libc_read, and __write instead of __libc_write.  Define\n        __open64 in addition to __open.\n\n(cherry picked from commit 329ea513b451ae8322aa7a24ed84da13992af2dd)","shortMessageHtmlLink":"Avoid cancellable I/O primitives in ld.so."}},{"before":"bef2a827a55fc759693ccc5b0f614353b8ad712d","after":"e7ac92e6ca9784b397189df0b2e1fb34f425bab8","ref":"refs/heads/master","pushedAt":"2024-06-12T01:16:59.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"<stdio.h>: Acknowledge that getdelim/getline are in POSIX\n\nThese comments were written in 2003 (added in 2c008571c3a), predating\nthe addition of getdelim(3)/getline(3) in POSIX.1-2008.\n\nReviewed-by: Sam James <sam@gentoo.org>\nReviewed-by: Florian Weimer <fweimer@redhat.com>","shortMessageHtmlLink":"&lt;stdio.h&gt;: Acknowledge that getdelim/getline are in POSIX"}},{"before":"00899eba260ff3edb62bc6f45c1860bc64fd59e0","after":"9de9cd17e73db0ba9af9ef11dc12d490fb59720c","ref":"refs/heads/release/2.39/master","pushedAt":"2024-06-11T01:16:51.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"malloc: New test to check malloc alternate path using memory obstruction\n\nThe test aims to ensure that malloc uses the alternate path to\nallocate memory when sbrk() or brk() fails.To achieve this,\nthe test first creates an obstruction at current program break,\ntests that obstruction with a failing sbrk(), then checks if malloc\nis still returning a valid ptr thus inferring that malloc() used\nmmap() instead of brk() or sbrk() to allocate the memory.\nReviewed-by: Arjun Shankar <arjun@redhat.com>\nReviewed-by: Zack Weinberg <zack@owlfolio.org>\n\n(cherry picked from commit 127fc56152347d73cb7c1c283e60e1cb1f15e9f9)","shortMessageHtmlLink":"malloc: New test to check malloc alternate path using memory obstruction"}},{"before":"eb370158794d7c64740a257ab2246ab46b90306a","after":"bef2a827a55fc759693ccc5b0f614353b8ad712d","ref":"refs/heads/master","pushedAt":"2024-06-11T01:16:51.000Z","pushType":"push","commitsCount":5,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"x86: Enable non-temporal memset tunable for AMD\n\nIn commit 46b5e98ef6f1 (\"x86: Add seperate non-temporal tunable for\nmemset\") a tunable threshold for enabling non-temporal memset was added,\nbut only for Intel hardware.\n\nSince that commit, new benchmark results suggest that non-temporal\nmemset is beneficial on AMD, as well, so allow this tunable to be set\nfor AMD.\n\nSee:\nhttps://docs.google.com/spreadsheets/d/1opzukzvum4n6-RUVHTGddV6RjAEil4P2uMjjQGLbLcU/edit?usp=sharing\nwhich has been updated to include data using different stategies for\nlarge memset on AMD Zen2, Zen3, and Zen4.\n\nSigned-off-by: Joe Damato <jdamato@fastly.com>\nReviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>","shortMessageHtmlLink":"x86: Enable non-temporal memset tunable for AMD"}},{"before":null,"after":"e00ff0233807361ff898817c6f355806fe569a81","ref":"refs/heads/pranavk/grte_v5_plus","pushedAt":"2024-06-11T01:16:51.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Add narrowing divide functions.\n\nThis patch adds the narrowing divide functions from TS 18661-1 to\nglibc's libm: fdiv, fdivl, ddivl, f32divf64, f32divf32x, f32xdivf64\nfor all configurations; f32divf64x, f32divf128, f64divf64x,\nf64divf128, f32xdivf64x, f32xdivf128, f64xdivf128 for configurations\nwith _Float64x and _Float128; __nldbl_ddivl for ldbl-opt.\n\nThe changes are mostly essentially the same as for the other narrowing\nfunctions, so the description of those generally applies to this patch\nas well.\n\nTested for x86_64, x86, mips64 (all three ABIs, both hard and soft\nfloat) and powerpc, and with build-many-glibcs.py.\n\n\t* math/Makefile (libm-narrow-fns): Add div.\n\t(libm-test-funcs-narrow): Likewise.\n\t* math/Versions (GLIBC_2.28): Add narrowing divide functions.\n\t* math/bits/mathcalls-narrow.h (div): Use __MATHCALL_NARROW.\n\t* math/gen-auto-libm-tests.c (test_functions): Add div.\n\t* math/math-narrow.h (CHECK_NARROW_DIV): New macro.\n\t(NARROW_DIV_ROUND_TO_ODD): Likewise.\n\t(NARROW_DIV_TRIVIAL): Likewise.\n\t* sysdeps/ieee754/float128/float128_private.h (__fdivl): New\n\tmacro.\n\t(__ddivl): Likewise.\n\t* sysdeps/ieee754/ldbl-opt/Makefile (libnldbl-calls): Add fdiv and\n\tddiv.\n\t(CFLAGS-nldbl-ddiv.c): New variable.\n\t(CFLAGS-nldbl-fdiv.c): Likewise.\n\t* sysdeps/ieee754/ldbl-opt/Versions (GLIBC_2.28): Add\n\t__nldbl_ddivl.\n\t* sysdeps/ieee754/ldbl-opt/nldbl-compat.h (__nldbl_ddivl): New\n\tprototype.\n\t* manual/arith.texi (Misc FP Arithmetic): Document fdiv, fdivl,\n\tddivl, fMdivfN, fMdivfNx, fMxdivfN and fMxdivfNx.\n\t* math/auto-libm-test-in: Add tests of div.\n\t* math/auto-libm-test-out-narrow-div: New generated file.\n\t* math/libm-test-narrow-div.inc: New file.\n\t* sysdeps/i386/fpu/s_f32xdivf64.c: Likewise.\n\t* sysdeps/ieee754/dbl-64/s_f32xdivf64.c: Likewise.\n\t* sysdeps/ieee754/dbl-64/s_fdiv.c: Likewise.\n\t* sysdeps/ieee754/float128/s_f32divf128.c: Likewise.\n\t* sysdeps/ieee754/float128/s_f64divf128.c: Likewise.\n\t* sysdeps/ieee754/float128/s_f64xdivf128.c: Likewise.\n\t* sysdeps/ieee754/ldbl-128/s_ddivl.c: Likewise.\n\t* sysdeps/ieee754/ldbl-128/s_f64xdivf128.c: Likewise.\n\t* sysdeps/ieee754/ldbl-128/s_fdivl.c: Likewise.\n\t* sysdeps/ieee754/ldbl-128ibm/s_ddivl.c: Likewise.\n\t* sysdeps/ieee754/ldbl-128ibm/s_fdivl.c: Likewise.\n\t* sysdeps/ieee754/ldbl-96/s_ddivl.c: Likewise.\n\t* sysdeps/ieee754/ldbl-96/s_fdivl.c: Likewise.\n\t* sysdeps/ieee754/ldbl-opt/nldbl-ddiv.c: Likewise.\n\t* sysdeps/ieee754/ldbl-opt/nldbl-fdiv.c: Likewise.\n\t* sysdeps/ieee754/soft-fp/s_ddivl.c: Likewise.\n\t* sysdeps/ieee754/soft-fp/s_fdiv.c: Likewise.\n\t* sysdeps/ieee754/soft-fp/s_fdivl.c: Likewise.\n\t* sysdeps/powerpc/fpu/libm-test-ulps: Update.\n\t* sysdeps/mach/hurd/i386/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/aarch64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/alpha/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/arm/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/hppa/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/i386/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/ia64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/microblaze/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/mips/mips32/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/mips/mips64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/nios2/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/riscv/rv64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/sh/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/x86_64/64/libm.abilist: Likewise.\n\t* sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist: Likewise.\n\n(cherry picked from commit 632a6cbe44cdd41dba7242887992cdca7b42922a)","shortMessageHtmlLink":"Add narrowing divide functions."}},{"before":"6ade91c21140d8c803c289932dbfc74537f65a1f","after":"00899eba260ff3edb62bc6f45c1860bc64fd59e0","ref":"refs/heads/release/2.39/master","pushedAt":"2024-06-10T01:18:11.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"x86/cet: fix shadow stack test scripts\n\nSome shadow stack test scripts use the '==' operator with the 'test'\ncommand to validate exit codes resulting in the following error:\n\n  sysdeps/x86_64/tst-shstk-legacy-1e.sh: 31: test: 139: unexpected operator\n\nThe '==' operator is invalid for the 'test' command, use '-eq' like the\nprevious call to 'test'.\n\nSigned-off-by: Michael Jeanson <mjeanson@efficios.com>\nReviewed-by: H.J. Lu <hjl.tools@gmail.com>\n(cherry picked from commit 155bb9d036646138348fee0ac045de601811e0c5)","shortMessageHtmlLink":"x86/cet: fix shadow stack test scripts"}},{"before":"54c1efdac55b756a4e2ea95590bcc3ba396a6568","after":"eb370158794d7c64740a257ab2246ab46b90306a","ref":"refs/heads/master","pushedAt":"2024-06-10T01:18:11.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"localedata: add new locales scn_IT\n\nSigned-off-by: David Paleino <dapal@debian.org>","shortMessageHtmlLink":"localedata: add new locales scn_IT"}},{"before":null,"after":"a45dbfef5a00e673f22812f7a0b8c8c73287265f","ref":"refs/heads/fw/x86-diagnostics","pushedAt":"2024-06-08T01:15:18.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"WIP observations from an XSAVE size mismatch kernel issue","shortMessageHtmlLink":"WIP observations from an XSAVE size mismatch kernel issue"}},{"before":"92c270d32caf3f8d5a02b8e46c7ec5d9d0315158","after":"54c1efdac55b756a4e2ea95590bcc3ba396a6568","ref":"refs/heads/master","pushedAt":"2024-06-06T01:15:07.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"support: Fix typo in xgetsockname error message\n\nThe error message in xgetsockname was incorrectly referring to a\ndifferent function.  This commit fixes that.\n\nSuggested-by: Arjun Shankar <arjun@redhat.com>\nSigned-off-by: Avinal Kumar <avinal.xlvii@gmail.com>\nReviewed-by: Carlos O'Donell <carlos@redhat.com>","shortMessageHtmlLink":"support: Fix typo in xgetsockname error message"}},{"before":"63c15758364a3e5f6fd0446491ff5b7b3e335b3b","after":"0f31bbfcd4e9f2cb1933439a1b154c427a3206fb","ref":"refs/heads/azanella/mseal","pushedAt":"2024-06-05T01:15:13.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"elf: Add support to memory sealing for audit modules\n\nIt is done after the the loading process (instead at loading time\nwith RTLD_NODELETE), because the module might be unloaded due\nnexistent or wrong la_version.  It also means that the la_version\nand l_objopen/la_activety might be called without the object memory\nsealed.\n\nChecked on x86_64-linux-gnu.","shortMessageHtmlLink":"elf: Add support to memory sealing for audit modules"}},{"before":"992daa0b4b5fa8a9f403c9575638cec288e12bfa","after":"92c270d32caf3f8d5a02b8e46c7ec5d9d0315158","ref":"refs/heads/master","pushedAt":"2024-06-05T01:15:13.000Z","pushType":"push","commitsCount":9,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Linux: Add epoll ioctls\n\nAs of Linux kernel 6.9, some ioctls and a parameters structure have been\nintroduced which allow user programs to control whether a particular\nepoll context will busy poll.\n\nUpdate the headers to include these for the convenience of user apps.\n\nThe ioctls were added in Linux kernel 6.9 commit 18e2bf0edf4dd\n(\"eventpoll: Add epoll ioctl for epoll_params\") [1] to\ninclude/uapi/linux/eventpoll.h.\n\n[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/diff/?h=v6.9&id=18e2bf0edf4dd\n\nSigned-off-by: Joe Damato <jdamato@fastly.com>\nReviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>","shortMessageHtmlLink":"Linux: Add epoll ioctls"}},{"before":null,"after":"63c15758364a3e5f6fd0446491ff5b7b3e335b3b","ref":"refs/heads/azanella/mseal","pushedAt":"2024-06-04T01:15:45.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"elf: Add support to memory sealing\n\nIt uses the new Linux mseal syscall to seal memory mappings to avoid\nfurther changes (such changing the memory protection or remap).  The\nsealing is done in multiple places where the memory is supported to\nbe immuatable over program execution:\n\n  * All DT_NEEDED dependencies from the binary, including the RELRO\n    seguments after PT_GNU_RELRO setup.\n\n  * The binary itself, including dynamic and static linked.  In both\n    cases it is up either to binary or the loader to setup the\n    sealing.\n\n  * The ld.so.cache cache.\n\n  * The vDSO segment (if existent).\n\nFor binary dependencies, the RTLD_NODELETE is used to signal the\nlink_map should be sealed.  It also make dlopen objects with the\nflag sealed as well.\n\nThe sealing is controlled by a new tunable, glibc.rtld.seal, with\nthree different states:\n\n  0. Disabled  where no sealing is done.  This is the default.\n\n  1. Enabled, where loader will issue the mseal syscall on the\n     memory mappings but any failure will be ignored.\n\n  2. Enforce, similar to Enabled but any failure from the mseal\n     will terminate the process.\n\nChecked on x86_64-linux-gnu.","shortMessageHtmlLink":"elf: Add support to memory sealing"}},{"before":"70f560fc22212f733647c9121c26bbb2307f2e10","after":"6ade91c21140d8c803c289932dbfc74537f65a1f","ref":"refs/heads/release/2.39/master","pushedAt":"2024-06-04T01:15:45.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"elf: Avoid some free (NULL) calls in _dl_update_slotinfo\n\nThis has been confirmed to work around some interposed mallocs.  Here\nis a discussion of the impact test ust/libc-wrapper/test_libc-wrapper\nin lttng-tools:\n\n  New TLS usage in libgcc_s.so.1, compatibility impact\n  <https://inbox.sourceware.org/libc-alpha/8734v1ieke.fsf@oldenburg.str.redhat.com/>\n\nReportedly, this patch also papers over a similar issue when tcmalloc\n2.9.1 is not compiled with -ftls-model=initial-exec.  Of course the\ngoal really should be to compile mallocs with the initial-exec TLS\nmodel, but this commit appears to be a useful interim workaround.\n\nFixes commit d2123d68275acc0f061e73d5f86ca504e0d5a344 (\"elf: Fix slow\ntls access after dlopen [BZ #19924]\").\n\nReviewed-by: Carlos O'Donell <carlos@redhat.com>\n(cherry picked from commit afe42e935b3ee97bac9a7064157587777259c60e)","shortMessageHtmlLink":"elf: Avoid some free (NULL) calls in _dl_update_slotinfo"}},{"before":"46b5e98ef6f1b9f4b53851f152ecb8209064b26c","after":"992daa0b4b5fa8a9f403c9575638cec288e12bfa","ref":"refs/heads/master","pushedAt":"2024-06-04T01:15:45.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"stdlib: Describe __cxa_finalize usage in function comment\n\nReviewed-by: Jonathan Wakely <jwakely@redhat.com>","shortMessageHtmlLink":"stdlib: Describe __cxa_finalize usage in function comment"}},{"before":null,"after":"c5f7f4fc8bb857cbe07972ff1e29970b101e9995","ref":"refs/heads/fw/x86-shstk-backtrace","pushedAt":"2024-06-01T01:19:07.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"x86_64: Use shadow stack for backtrace implementation\n\nTest failures:\n\nFAIL: debug/tst-backtrace4\nFAIL: misc/tst-sigcontext-get_pc\n\nThe return address of signal handlers is not on the shadow stack.","shortMessageHtmlLink":"x86_64: Use shadow stack for backtrace implementation"}},{"before":"0c1d2c277a59f08fd3232b33d18644ea890190ea","after":"46b5e98ef6f1b9f4b53851f152ecb8209064b26c","ref":"refs/heads/master","pushedAt":"2024-05-31T01:15:19.000Z","pushType":"push","commitsCount":4,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"x86: Add seperate non-temporal tunable for memset\n\nThe tuning for non-temporal stores for memset vs memcpy is not always\nthe same. This includes both the exact value and whether non-temporal\nstores are profitable at all for a given arch.\n\nThis patch add `x86_memset_non_temporal_threshold`. Currently we\ndisable non-temporal stores for non Intel vendors as the only\nbenchmarks showing its benefit have been on Intel hardware.\nReviewed-by: H.J. Lu <hjl.tools@gmail.com>","shortMessageHtmlLink":"x86: Add seperate non-temporal tunable for memset"}},{"before":"c9d8534406ab69b9bc1fd3fdfb9e88c9580d3f24","after":"70f560fc22212f733647c9121c26bbb2307f2e10","ref":"refs/heads/release/2.39/master","pushedAt":"2024-05-30T01:14:34.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"misc: Add support for Linux uio.h RWF_NOAPPEND flag\n\nIn Linux 6.9 a new flag is added to allow for Per-io operations to\ndisable append mode even if a file was opened with the flag O_APPEND.\nThis is done with the new RWF_NOAPPEND flag.\n\nThis caused two test failures as these tests expected the flag 0x00000020\nto be unused.  Adding the flag definition now fixes these tests on Linux\n6.9 (v6.9-rc1).\n\n  FAIL: misc/tst-preadvwritev2\n  FAIL: misc/tst-preadvwritev64v2\n\nThis patch adds the flag, adjusts the test and adds details to\ndocumentation.\n\nLink: https://lore.kernel.org/all/20200831153207.GO3265@brightrain.aerifal.cx/\nReviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>\n(cherry picked from commit 3db9d208dd5f30b12900989c6d2214782b8e2011)","shortMessageHtmlLink":"misc: Add support for Linux uio.h RWF_NOAPPEND flag"}},{"before":"697ab62d1176180bea49094130ad0af2b108874d","after":"c8cb4d2b86ece572793e31a3422ea29e88d77df5","ref":"refs/heads/release/2.38/master","pushedAt":"2024-05-30T01:14:34.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"misc: Add support for Linux uio.h RWF_NOAPPEND flag\n\nIn Linux 6.9 a new flag is added to allow for Per-io operations to\ndisable append mode even if a file was opened with the flag O_APPEND.\nThis is done with the new RWF_NOAPPEND flag.\n\nThis caused two test failures as these tests expected the flag 0x00000020\nto be unused.  Adding the flag definition now fixes these tests on Linux\n6.9 (v6.9-rc1).\n\n  FAIL: misc/tst-preadvwritev2\n  FAIL: misc/tst-preadvwritev64v2\n\nThis patch adds the flag, adjusts the test and adds details to\ndocumentation.\n\nLink: https://lore.kernel.org/all/20200831153207.GO3265@brightrain.aerifal.cx/\nReviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>\n(cherry picked from commit 3db9d208dd5f30b12900989c6d2214782b8e2011)","shortMessageHtmlLink":"misc: Add support for Linux uio.h RWF_NOAPPEND flag"}},{"before":"e0f7f1808fb024d35c3d9973e68245efc78fff40","after":"0c1d2c277a59f08fd3232b33d18644ea890190ea","ref":"refs/heads/master","pushedAt":"2024-05-29T01:17:26.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"LoongArch: Use \"$fcsr0\" instead of \"$r0\" in _FPU_{GET,SET}CW\n\nClang inline-asm parser does not allow using \"$r0\" in\nmovfcsr2gr/movgr2fcsr, so everything using _FPU_{GET,SET}CW is now\nfailing to build with Clang on LoongArch.  As we now requires Binutils\n>= 2.41 which supports using \"$fcsr0\" here, use it instead of \"$r0\" to\nfix the issue.\n\nLink: https://github.com/loongson-community/discussions/issues/53#issuecomment-2081507390\nLink: https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=4142b2368353\nSigned-off-by: Xi Ruoyao <xry111@xry111.site>","shortMessageHtmlLink":"LoongArch: Use \"$fcsr0\" instead of \"$r0\" in _FPU_{GET,SET}CW"}},{"before":"249646f12b08f3dbd9a8e8b8021e5c0e3a0b7e1e","after":"697ab62d1176180bea49094130ad0af2b108874d","ref":"refs/heads/release/2.38/master","pushedAt":"2024-05-29T01:17:26.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782)\n\nThis patch disables Intel Xeon Phi tests for GCC 15 and above.\n\nGCC 15 removed Intel Xeon Phi ISA support.\ncommit e1a7e2c54d52d0ba374735e285b617af44841ace\nAuthor: Haochen Jiang <haochen.jiang@intel.com>\nDate:   Mon May 20 10:43:44 2024 +0800\n\n    i386: Remove Xeon Phi ISA support\n\nFixes BZ 31782.\n\nReviewed-by: H.J. Lu <hjl.tools@gmail.com>\n(cherry picked from commit 1b713c9a5349ef3cd1a8ccf9de017c7865713c67)","shortMessageHtmlLink":"i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782)"}},{"before":"26e7005728f0eea2972474e6be2905c467661237","after":"c9d8534406ab69b9bc1fd3fdfb9e88c9580d3f24","ref":"refs/heads/release/2.39/master","pushedAt":"2024-05-29T01:17:26.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782)\n\nThis patch disables Intel Xeon Phi tests for GCC 15 and above.\n\nGCC 15 removed Intel Xeon Phi ISA support.\ncommit e1a7e2c54d52d0ba374735e285b617af44841ace\nAuthor: Haochen Jiang <haochen.jiang@intel.com>\nDate:   Mon May 20 10:43:44 2024 +0800\n\n    i386: Remove Xeon Phi ISA support\n\nFixes BZ 31782.\n\nReviewed-by: H.J. Lu <hjl.tools@gmail.com>\n(cherry picked from commit 1b713c9a5349ef3cd1a8ccf9de017c7865713c67)","shortMessageHtmlLink":"i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782)"}},{"before":"c7c3f5bf80ae86b34501f473f1a9fc545c911b7f","after":"26e7005728f0eea2972474e6be2905c467661237","ref":"refs/heads/release/2.39/master","pushedAt":"2024-05-28T01:14:06.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"socket: Use may_alias on sockaddr structs (bug 19622)\n\nThis supports common coding patterns.  The GCC C front end before\nversion 7 rejects the may_alias attribute on a struct definition\nif it was not present in a previous forward declaration, so this\nattribute can only be conditionally applied.\n\nThis implements the spirit of the change in Austin Group issue 1641.\n\nSuggested-by: Marek Polacek <polacek@redhat.com>\nSuggested-by: Jakub Jelinek <jakub@redhat.com>\nReviewed-by: Sam James <sam@gentoo.org>\nReviewed-by: Carlos O'Donell <carlos@redhat.com>\n(cherry picked from commit 8d7b6b4cb27d4dec1dd5f7960298c1699275f962)","shortMessageHtmlLink":"socket: Use may_alias on sockaddr structs (bug 19622)"}},{"before":"ab46c6ba6cc43b18daea6f197030e19303c9f012","after":"e0f7f1808fb024d35c3d9973e68245efc78fff40","ref":"refs/heads/master","pushedAt":"2024-05-28T01:14:06.000Z","pushType":"push","commitsCount":4,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"x86_64: Reformat elf_machine_rela\n\nA space is added before the left bracket of the x86_64 elf_machine_rela\nfunction, in order to harmonize with the rest of the implementation of\nthe function and to make it easier to retrieve the function. The lines\nwhere the function definition is located has been re-indented, as well\nas its left curly bracket placed in the correct position.\n\nSigned-off-by: Xin Wang <yw987194828@gmail.com>\nReviewed-by: H.J. Lu <hjl.tools@gmail.com>","shortMessageHtmlLink":"x86_64: Reformat elf_machine_rela"}},{"before":"caed1f5c0b2e31b5f4e0f21fea4b2c9ecd3b5b30","after":"ab46c6ba6cc43b18daea6f197030e19303c9f012","ref":"refs/heads/master","pushedAt":"2024-05-26T01:19:17.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"elf/Makefile: Split and sort PIE tests\n\nPut each item on a separate line and sort PIE tests.\n\nSigned-off-by: H.J. Lu <hjl.tools@gmail.com>","shortMessageHtmlLink":"elf/Makefile: Split and sort PIE tests"}},{"before":"8f58e412b1e26d2c7e65c13a0ce758fbaf18d83f","after":"caed1f5c0b2e31b5f4e0f21fea4b2c9ecd3b5b30","ref":"refs/heads/master","pushedAt":"2024-05-25T01:12:35.000Z","pushType":"push","commitsCount":6,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Revert \"Test fscanf of long double without <stdio.h>\"\n\nThis reverts commit 30a745450e01e4c6f484dc691c16a54a6ea72288.\n\nOn ppc64le, without <stdio.h>, vfscanf is used and with <stdio.h>\n__isoc23_vfscanfieee128 is used.  I am reverting this since it doesn't\nwork on all targets.","shortMessageHtmlLink":"Revert \"Test fscanf of long double without &lt;stdio.h&gt;\""}},{"before":"6d3b523eb54198d15af6e042054912cf9f5210de","after":"8f58e412b1e26d2c7e65c13a0ce758fbaf18d83f","ref":"refs/heads/master","pushedAt":"2024-05-24T01:13:38.000Z","pushType":"push","commitsCount":11,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"powerpc: Remove duplicated versionsort from libm.a (BZ 31789)\n\nThe powerpc32 have an extra versionsort provided by LFS\nversionsort64.o.  It seems that 5226a81f5517bcbc892679cca792006a6bafc53f\nused the wrong check to create the alias for the LFS to non-LFS version.\nIt should not matter for _DIRENT_MATCHES_DIRENT64 since both symbols\nhave the same implementation.\nReviewed-by: H.J. Lu <hjl.tools@gmail.com>","shortMessageHtmlLink":"powerpc: Remove duplicated versionsort from libm.a (BZ 31789)"}},{"before":"85472c20a55ea2a49a7fbdf71652b4009118b0ae","after":"6d3b523eb54198d15af6e042054912cf9f5210de","ref":"refs/heads/master","pushedAt":"2024-05-23T01:13:12.000Z","pushType":"push","commitsCount":5,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Define write_profiling functions only in profile library [BZ #31756]\n\nlibc.so doesn't use nor export write_profiling functions.  There is no\npoint to define them in libc.so nor in libc.a.  Fix BZ #31756 by defining\nthem only in profile library.\n\nSigned-off-by: H.J. Lu <hjl.tools@gmail.com>\nReviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>","shortMessageHtmlLink":"Define write_profiling functions only in profile library [BZ #31756]"}},{"before":"f83e461f1014598a5cb4c89407ce303b9f0bd8ac","after":"85472c20a55ea2a49a7fbdf71652b4009118b0ae","ref":"refs/heads/master","pushedAt":"2024-05-22T01:14:15.000Z","pushType":"push","commitsCount":12,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Change _IO_stderr_/_IO_stdin_/_IO_stdout to compat symbols [BZ #31766]\n\nSince Glibc never provides symbol binary compatibility for relocatable\nfiles, fix BZ #31766 by changing _IO_stderr_/_IO_stdin_/_IO_stdout to\ncompat symbols.\n\nSigned-off-by: H.J. Lu <hjl.tools@gmail.com>\nReviewed-by: Sunil K Pandey <skpgkp2@gmail.com>","shortMessageHtmlLink":"Change _IO_stderr_/_IO_stdin_/_IO_stdout to compat symbols [BZ #31766]"}},{"before":"2be3352f0b1ebaa39596393fffe1062275186669","after":"f83e461f1014598a5cb4c89407ce303b9f0bd8ac","ref":"refs/heads/master","pushedAt":"2024-05-21T01:13:20.000Z","pushType":"push","commitsCount":8,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"powerpc: Update ulps\n\nFor the log2p1 implementation.","shortMessageHtmlLink":"powerpc: Update ulps"}},{"before":"8d7b6b4cb27d4dec1dd5f7960298c1699275f962","after":"2be3352f0b1ebaa39596393fffe1062275186669","ref":"refs/heads/master","pushedAt":"2024-05-20T01:13:52.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"berenm","name":"Beren Minor","path":"/berenm","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/235072?s=80&v=4"},"commit":{"message":"Pass -nostdlib -nostartfiles together with -r [BZ #31753]\n\nSince -r in GCC 6/7/8 doesn't imply -nostdlib -nostartfiles, update the\nlink-static-libc.out rule to also pass -nostdlib -nostartfiles.  This\nfixes BZ #31753.\n\nSigned-off-by: H.J. Lu <hjl.tools@gmail.com>\nReviewed-by: Florian Weimer <fweimer@redhat.com>","shortMessageHtmlLink":"Pass -nostdlib -nostartfiles together with -r [BZ #31753]"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEYso7RwA","startCursor":null,"endCursor":null}},"title":"Activity · bminor/glibc"}