As analyzed in the PR, the old code was with the assumption that assembler
won't set bit0 of thumb function address if it comes from PC-relative
instructions and the calculation can be finished during assembling. This
assumption however does not hold after PR gas/21458.
* sysdeps/arm/dl-machine.h (elf_machine_load_address): Also strip bit 0
of pcrel_address under Thumb mode.
powerpc64le: Iterate over all object suffixes when appending -mfloat128
On powerpc64le, the compilation of the files related to float128 support
requires the option -mfloat128 to be passed to gcc. However, not all
possible object suffixes were covered in the Makefile. This patch uses
$(all-object-suffixes) in all remaining rules.
Tested for powerpc64le.
* sysdeps/powerpc/powerpc64le/Makefile: Use $(all-object-suffixes)
to iterate over all possible object suffixes. Add a comment
explaining the use of sysdep-CFLAGS instead of CFLAGS.
H.J. Lu [Tue, 11 Jul 2017 14:44:01 +0000 (07:44 -0700)]
Avoid backtrace from __stack_chk_fail [BZ #12189]
__stack_chk_fail is called on corrupted stack. Stack backtrace is very
unreliable against corrupted stack. __libc_message is changed to accept
enum __libc_message_action and call BEFORE_ABORT only if action includes
do_backtrace. __fortify_fail_abort is added to avoid backtrace from
__stack_chk_fail.
[BZ #12189]
* debug/Makefile (CFLAGS-tst-ssp-1.c): New.
(tests): Add tst-ssp-1 if -fstack-protector works.
* debug/fortify_fail.c: Include <stdbool.h>.
(_fortify_fail_abort): New function.
(__fortify_fail): Call _fortify_fail_abort.
(__fortify_fail_abort): Add a hidden definition.
* debug/stack_chk_fail.c: Include <stdbool.h>.
(__stack_chk_fail): Call __fortify_fail_abort, instead of
__fortify_fail.
* debug/tst-ssp-1.c: New file.
* include/stdio.h (__libc_message_action): New enum.
(__libc_message): Replace int with enum __libc_message_action.
(__fortify_fail_abort): New hidden prototype.
* malloc/malloc.c (malloc_printerr): Update __libc_message calls.
* sysdeps/posix/libc_fatal.c (__libc_message): Replace int
with enum __libc_message_action. Call BEFORE_ABORT only if
action includes do_backtrace.
(__libc_fatal): Update __libc_message call.
Stefan Liebler [Tue, 11 Jul 2017 08:37:03 +0000 (10:37 +0200)]
S390: Fix tst-ptrace-singleblock if kernel does not support PTRACE_SINGLEBLOCK.
The request PTRACE_SINGLEBLOCK was introduced in Linux 3.15. Thus the ptrace call
will fail on older kernels.
Thus the test is now testing PTRACE_SINGLEBLOCK with data argument pointing to a
buffer on stack which is assumed to fail. If the request would be interpreted as
PTRACE_GETREGS, then the ptrace call will not fail and the regs are written to buf.
If we run with a kernel with support for PTRACE_SINGLEBLOCK a ptrace call with
data=NULL, returns zero with no error. If we run with a kernel without support for
PTRACE_SINGLEBLOCK a ptrace call with data=NULL reports an error.
In the latter case, the test is just continuing with PTRACE_CONT.
ChangeLog:
* sysdeps/unix/sysv/linux/s390/tst-ptrace-singleblock.c:
Support running on kernels without PTRACE_SINGLEBLOCK.
During Locale verification I observed that
Incorrect Full Weekday names for ks_IN@devanagari
Reference is taken from
http://www.mkraina.com/PDF/3-Self-authored%20Works%20(English)/15.pdf
And kashmiri devanagari travel book and other sources
[BZ #21721]
* locales/ks_IN@devanagari: Full weekday name Fix.
Joseph Myers [Thu, 6 Jul 2017 17:01:03 +0000 (17:01 +0000)]
Increase some test timeouts.
This patch increases the timeouts for some tests that I've seen timing
out on slow systems in my 2.26 release testing. (In the case of
tst-tsearch.c, increasing the timeout means removing a setting of 10
that was put there before the default timeout was increased to 20
seconds, so putting the default into effect.)
* iconvdata/tst-loading.c (TIMEOUT): Define to 30.
* misc/tst-tsearch.c (TIMEOUT): Remove.
* nptl/tst-create-detached.c (TIMEOUT): Define to 100.
* nptl/tst-robust-fork.c (TIMEOUT): Likewise.
* nptl/tst-rwlock19.c (TIMEOUT): Likewise.
* string/tst-cmp.c (TIMEOUT): Define to 600.
continue to work even if vector instructions are used in glibc which
require the ABI stack realignment.
__tls_get_addr_slow is added to handle the slow paths in the default
implementation of__tls_get_addr in elf/dl-tls.c. The new __tls_get_addr
calls __tls_get_addr_slow after realigning the stack. Internal calls
within ld.so go directly to the default implementation of __tls_get_addr
because they do not need stack realignment.
Szabolcs Nagy [Tue, 4 Jul 2017 15:05:12 +0000 (16:05 +0100)]
Single threaded stdio optimization
Locking overhead can be significant in some stdio operations
that are common in single threaded applications.
This patch adds the _IO_FLAGS2_NEED_LOCK flag to indicate if
an _IO_FILE object needs to be locked and some of the stdio
functions just jump to their _unlocked variant when not. The
flag is set on all _IO_FILE objects when the first thread is
created. A new GLIBC_PRIVATE libc symbol, _IO_enable_locks,
was added to do this from libpthread.
The optimization can be applied to more stdio functions,
currently it is only applied to single flag check or single
non-wide-char standard operations. The flag should probably
be never set for files with _IO_USER_LOCK, but that's just a
further optimization, not a correctness requirement.
The optimization is valid in a single thread because stdio
operations are non-as-safe (so lock state is not observable
from a signal handler) and stdio locks are recursive (so lock
state is not observable via deadlock). The optimization is not
valid if a thread may be created while an stdio lock is taken
and thus it should be disabled if any user code may run during
an stdio operation (interposed malloc, printf hooks, etc).
This makes the optimization more complicated for some stdio
operations (e.g. printf), but those are bigger and thus less
important to optimize so this patch does not try to do that.
A dot-less host name without an /etc/resolv.conf file caused an
assertion failure in update_from_conf because the function would not
deal correctly with the empty search list case.
Thanks to Andreas Schwab for debugging assistence.
resolv: Automatically reload a changed /etc/resolv.conf file [BZ #984]
This commit enhances the stub resolver to reload the configuration
in the per-thread _res object if the /etc/resolv.conf file has
changed. The resolver checks whether the application has modified
_res and will not overwrite the _res object in that case.
The struct resolv_context mechanism is used to check the
configuration file only once per name lookup.
Florian Weimer [Fri, 30 Jun 2017 18:19:10 +0000 (20:19 +0200)]
resolv: Mirror the entire resolver configuration in struct resolv_conf
This commit adds the remaining unchanging members (which are loaded
from /etc/resolv.conf) to struct resolv_conf.
The extended name server list is currently not used by the stub
resolver. The switch depends on a cleanup: The _u._ext.nssocks
array stores just a single socket, and needs to be replaced with
a single socket value.
(The compatibility gethostname implementation does not use the
extended addres sort list, either. Updating the compat code is
not worthwhile.)
Florian Weimer [Fri, 30 Jun 2017 22:53:05 +0000 (00:53 +0200)]
resolv: Lift domain search list limits [BZ #19569] [BZ #21475]
This change uses the extended resolver state in struct resolv_conf to
store the search list. If applications have not patched the _res
object directly, this extended search list will be used by the stub
resolver during name resolution.
resolv: Introduce struct resolv_conf with extended resolver state
This change provides additional resolver configuration state which
is not exposed through the _res ABI. It reuses the existing
initstamp field in the supposedly-private part of _res. Some effort
is undertaken to avoid memory safety issues introduced by applications
which directly patch the _res object.
With this commit, only the initstamp field is moved into struct
resolv_conf. Additional members will be added later, eventually
migrating the entire resolver configuration.
struct resolv_context objects provide a temporary resolver context
which does not change during a name lookup operation. Only when the
outmost context is created, the stub resolver configuration is
verified to be current (at present, only against previous res_init
calls). Subsequent attempts to obtain the context will reuse the
result of the initial verification operation.
struct resolv_context can also be extended in the future to store
data which needs to be deallocated during thread cancellation.
Rafal Luzynski [Thu, 29 Jun 2017 00:23:28 +0000 (02:23 +0200)]
More fixes after the recent import from CLDR-31
After the recent import of month names from CLDR (bug 21217) more
imports are also needed, mostly abbreviated month names.
* localedata/locales/br_FR (abmon): Reworded "Eve " to "Mezh".
* localedata/locales/fy_NL (abmon): Reworded "Maa" (March) to
"Mrt" and "Maa" (May) to "Mai".
* localedata/locales/lg_UG (abmon): Reworded "Jun" to "Juu".
* localedata/locales/ln_CD (abmon): "yan", "fbl", "msi",
and so on.
* localedata/locales/mn_MN (abmon): "1-р сар", "2-р сар",
"3-р сар", and so on.
* localedata/locales/vi_VN (abmon): Reworded "Th01" to "Thg 1",
"Th02" to "Thg 2" and so on.
* localedata/locales/yo_NG (abday): "Àìkú", "Ajé", "Ìsẹ́gun",
and so on, also comment updated to match the new content.
(day): "Ọjọ́ Àìkú", "Ọjọ́ Ajé", "Ọjọ́ Ìsẹ́gun", and so on.
(abmon): "Ṣẹ́rẹ́", "Èrèlè", "Ẹrẹ̀nà", and so on.
(mon): Comment updated to match the actual content.
(d_t_fmt): Changed "%A" to "%a" and "%B" to "%b".
* localedata/locales/zu_ZA (abmon): "Jan", "Feb", "Mas",
and so on, also comment updated to match the new content.
(mon): comment updated to match the actual content.
H.J. Lu [Sat, 1 Jul 2017 14:29:19 +0000 (07:29 -0700)]
Use __builtin_popcount in __sched_cpucount [BZ #21696]
posix/sched_cpucount.c assumes that size of __cpu_mask == size of long,
which is incorrect for x32. This patch uses __builtin_popcount, which
is availabe in GCC 4.9, in posix/sched_cpucount.c.
Tested on i686, x86-64 and x32 with multi-arch disabled.
[BZ #21696]
* posix/sched_cpucount.c: Don't include <limits.h>.
(__sched_cpucount): Use __builtin_popcount.
float128: Add signbit alternative for old compilers
In math/math.h, __MATH_TG will expand signbit to __builtin_signbit*,
e.g.: __builtin_signbitf128, before GCC 6. However, there has never
been a __builtin_signbitf128 in GCC and the type-generic builtin is
only available since GCC 6. For older GCC, this patch defines
__builtin_signbitf128 to __signbitf128, so that the internal function
is used instead of the non-existent builtin.
This patch also changes the implementation of __signbitf128, because
it was reusing the implementation of __signbitl from ldbl-128, which
calls __builtin_signbitl. Using the long double version of the
builtin is not correct on machines where _Float128 is ABI-distinct
from long double (i.e.: ia64, powerpc64le, x86, x86_84). The new
implementation does not rely on builtins when being built with GCC
versions older than 6.0.
The new code does not currently affect powerpc64le builds, because
only GCC 6.2 fulfills the requirements from configure. It might
affect powerpc64le builds if those requirements are backported to
older versions of the compiler. The new code affects x86_64 builds,
since glibc is supposed to build correctly with older versions of GCC.
Tested for powerpc64le and x86_64.
* include/math.h (__signbitf128): Define as hidden.
* sysdeps/ieee754/float128/s_signbitf128.c (__signbitf128):
Reimplement without builtins.
* sysdeps/ia64/bits/floatn.h [!__GNUC_PREREQ (6, 0)]
(__builtin_signbitf128): Define to __signbitf128.
* sysdeps/powerpc/bits/floatn.h: Likewise.
* sysdeps/x86/bits/floatn.h: Likewise.
This patch consolidates the fcntl Linux syscall generation on
sysdeps/unix/sysv/linux/fcntl.c. It basically removes all the
architecture specific implementations.
Checked on i686-linux-gnu, x86_64-linux-gnu, x86_64-linux-gnux32,
aarch64-linux-gnu, arm-linux-gnueabihf, and powerpc64le-linux-gnu.
Add a new tunable (glibc.tune.cpu) to override CPU identification on
aarch64. This is useful in two cases: one where it is desirable to
pretend to be another CPU for purposes of testing or because routines
written for that CPU are beneficial for specific workloads and second
where the underlying kernel does not support emulation of MRS to get
the MIDR of the CPU.
* elf/dl-tunables.h (tunable_is_name): Move from...
* elf/dl-tunables.c (is_name): ... here.
(parse_tunables, __tunables_init): Adjust.
* manual/tunables.texi: Document glibc.tune.cpu.
* sysdeps/aarch64/dl-tunables.list: New file.
* sysdeps/unix/sysv/linux/aarch64/cpu-features.c (struct
cpu_list): New type.
(cpu_list): New list of CPU names and their MIDR.
(get_midr_from_mcpu): New function.
(init_cpu_features): Override MIDR if necessary.
aarch64: Call all string function implementations in tests
The string function implementations implemented so far do not use any
instructions that may deviate from standard aarch64, so it is possible
for all routines to run on all armv8 hardware. Select all
implementations in the benchmarks and tests.
* sysdeps/aarch64/multiarch/ifunc-impl-list.c
(__libc_ifunc_impl_list): Unconditionally select thunderx
routine for testing.
As a result, glibc malloc no longer returns memory blocks which are as
aligned as max_align_t requires.
This causes malloc/tst-malloc-thread-fail to fail with an error like this
one:
error: allocation function 0, size 144 not aligned to 16
This patch moves the MALLOC_ALIGNMENT definition to <malloc-alignment.h>
and increases the malloc alignment to 16 for i386.
[BZ #21120]
* malloc/malloc-internal.h (MALLOC_ALIGNMENT): Moved to ...
* sysdeps/generic/malloc-alignment.h: Here. New file.
* sysdeps/i386/malloc-alignment.h: Likewise.
* sysdeps/generic/malloc-machine.h: Include <malloc-alignment.h>.
This updates a bunch of locales based on CLDR v29 data:
az_AZ: changing Azərbaycanca to azərbaycan dili
be_BY: changing беларуская мова to беларуская
bem_ZM: changing iciBemba to Ichibemba
bg_BG: changing български език to български
bo_CN: changing པོད་སྐད་ to བོད་སྐད་
bo_IN: changing པོད་སྐད་ to བོད་སྐད་
br_FR: changing Brezhoneg to brezhoneg
brx_IN: lang_name: setting to बड़ो
ce_RU: changing нохчийн мотт to нохчийн
cs_CZ: changing Čeština to čeština
dz_BT: changing (རྫོང་ཁ to རྫོང་ཁ
el_CY: changing ελληνικά to Ελληνικά
el_GR: changing ελληνικά to Ελληνικά
es_AR: changing Español to español
es_BO: changing Español to español
es_CL: changing Español to español
es_CO: changing Español to español
es_CR: changing Español to español
es_CU: changing Español to español
es_DO: changing Español to español
es_EC: changing Español to español
es_ES: changing Español to español
es_GT: changing Español to español
es_HN: changing Español to español
es_MX: changing Español to español
es_NI: changing Español to español
es_PA: changing Español to español
es_PE: changing Español to español
es_PR: changing Español to español
es_PY: changing Español to español
es_SV: changing Español to español
es_US: changing Español to español
es_UY: changing Español to español
es_VE: changing Español to español
et_EE: changing eesti keel to eesti
eu_ES: changing Euskara to euskara
fr_BE: changing Français to français
fr_CA: changing Français to français
fr_CH: changing Français to français
fr_FR: changing Français to français
fr_LU: changing Français to français
fur_IT: changing Furlan to furlan
fy_NL: changing Frysk to West-Frysk
gl_ES: changing Galego to galego
gv_GB: changing y Ghaelg to Gaelg
he_IL: lang_name: setting to עברית
hsb_DE: changing Hornjoserbšćina to hornjoserbšćina
hy_AM: changing Հայերեն to հայերեն
id_ID: changing Bahasa Indonesia to Bahasa Indonesia
it_CH: changing Italiano to italiano
it_IT: changing Italiano to italiano
kl_GL: changing Kalaallisut to kalaallisut
km_KH: changing ភាសាខ្មែរ to ខ្មែរ
ko_KR: changing 한국말 to 한국어
ks_IN: changing kạ̄šur to کٲشُر
kw_GB: changing Kernowek to kernewek
ky_KG: changing Кыргызча to кыргызча
lg_UG: changing Oluganda to Luganda
lt_LT: changing lietuvių kalba to lietuvių
lv_LV: changing latviešu valoda to latviešu
mk_MK: changing македонск/и јазик to македонски
mn_MN: changing Монгол хэл to монгол
nb_NO: changing Bokmål to norsk bokmål
nn_NO: changing Nynorsk to nynorsk
os_RU: lang_name: setting to ирон
ru_RU: lang_name: setting to русский
ru_UA: lang_name: setting to русский
se_NO: changing Davvisámegiella to davvisámegiella
sk_SK: lang_name: setting to slovenčina
ta_IN: lang_name: setting to தமிழ்
ta_LK: lang_name: setting to தமிழ்
tk_TM: changing Türkmençe to türkmençe
tr_CY: changing Turkish to Türkçe
tr_TR: changing Turkish to Türkçe
ur_IN: lang_name: setting to {اردو}
ur_PK: lang_name: setting to {اردو}
vi_VN: changing Việt ngữ to Tiếng Việt
yo_NG: changing Yorùbá to Èdè Yorùbá
zu_ZA: changing IsiZulu to isiZulu
Most of these are simple case changes, but they match the CLDR db.
A search for a few of the others suggests they're also correct.
Joseph Myers [Thu, 29 Jun 2017 19:44:43 +0000 (19:44 +0000)]
SPARC sys/ucontext.h namespace fixes (bug 21457).
This patch fixes various miscellaneous namespace issues in the SPARC
sys/ucontext.h header. These are similar to changes made previous to
other sys/ucontext.h headers, where the SPARC header was excluded from
those previous patches because of its complexity.
Tested for SPARC with build-many-glibcs.py.
[BZ #21457]
* sysdeps/unix/sysv/linux/sparc/sys/ucontext.h (__ctx): New macro.
[__WORDSIZE == 64] (MC_TSTATE): Define only for [__USE_MISC].
[__WORDSIZE == 64] (MC_PC): Likewise.
[__WORDSIZE == 64] (MC_NPC): Likewise.
[__WORDSIZE == 64] (MC_Y): Likewise.
[__WORDSIZE == 64] (MC_G1): Likewise.
[__WORDSIZE == 64] (MC_G2): Likewise.
[__WORDSIZE == 64] (MC_G3): Likewise.
[__WORDSIZE == 64] (MC_G4): Likewise.
[__WORDSIZE == 64] (MC_G5): Likewise.
[__WORDSIZE == 64] (MC_G6): Likewise.
[__WORDSIZE == 64] (MC_G7): Likewise.
[__WORDSIZE == 64] (MC_O0): Likewise.
[__WORDSIZE == 64] (MC_O1): Likewise.
[__WORDSIZE == 64] (MC_O2): Likewise.
[__WORDSIZE == 64] (MC_O3): Likewise.
[__WORDSIZE == 64] (MC_O4): Likewise.
[__WORDSIZE == 64] (MC_O5): Likewise.
[__WORDSIZE == 64] (MC_O6): Likewise.
[__WORDSIZE == 64] (MC_O7): Likewise.
[__WORDSIZE == 64] (MC_NGREG): Rename to __MC_NGREG and define to
__MC_NGREG if [__USE_MISC].
[__WORDSIZE == 64] (MC_MAXFPQ): Define only for [__USE_MISC].
[__WORDSIZE == 64] (mc_gregset_t): Define using __MC_NGREG.
[__WORDSIZE == 64] (struct mc_fq): Rename to struct __mc_fq.
Define fields using __ctx.
[__WORDSIZE == 64] (mc_fpu_t): Remove struct tag. Define fields
using __ctx.
[__WORDSIZE == 64] (mcontext_t): Define fields using __ctx.
(REG_PSR): Define only for [__USE_MISC].
(REG_PC): Likewise.
(REG_nPC): Likewise.
(REG_Y): Likewise.
(REG_G1): Likewise.
(REG_G2): Likewise.
(REG_G3): Likewise.
(REG_G4): Likewise.
(REG_G5): Likewise.
(REG_G6): Likewise.
(REG_G7): Likewise.
(REG_O0): Likewise.
(REG_O1): Likewise.
(REG_O2): Likewise.
(REG_O3): Likewise.
(REG_O4): Likewise.
(REG_O5): Likewise.
(REG_O6): Likewise.
(REG_O7): Likewise.
[__WORDSIZE == 64] (REG_ASI): Define only for [__USE_MISC].
[__WORDSIZE == 64] (REG_FPRS): Likewise.
(NGREG): Rename to __NGREG and define to __NGREG if [__USE_MISC].
(gregset_t): Define using __NGREG.
(SPARC_MAXREGWINDOW): Rename to __SPARC_MAXREGWINDOW and define to
__SPARC_MAXREGWINDOW if [__USE_MISC].
(struct rwindow): Rename to struct __rwindow. Define fields using
__ctx.
(rw_fp): Define only for [__USE_MISC].
(rw_rtn): Likewise.
(gwindows_t): Remove struct tag. Define fields using __ctx and
__SPARC_MAXREGWINDOW.
(MAXFPQ): Define only for [__USE_MISC].
(struct fpq): Rename to struct __fpq. Define fields using __ctx.
(struct fq): Rename to struct __fq. Define fields using __ctx.
(FPU_REGS_TYPE): Define only for [__USE_MISC].
(FPU_DREGS_TYPE): Likewise.
(V7_FPU_FSR_TYPE): Likewise.
(V9_FPU_FSR_TYPE): Likewise.
(V9_FPU_FPRS_TYPE): Likewise.
[__WORDSIZE == 64] (fpregset_t): Remove struct tag. Define fields
using __ctx.
[__WORDSIZE != 64] (fpregset_t): Likewise.
[__WORDSIZE != 64] (xrs_t): Define fields using __ctx.
[__WORDSIZE != 64] (XRS_ID): Define only for [__USE_MISC].
[__WORDSIZE != 64] (mcontext_t): Define fields using __ctx.
Rename field filler to __glibc_reserved1.
* sysdeps/unix/sysv/linux/sparc/sparc32/ucontext_i.sym
(MC_FILLER): Remove.
This patch improves the default posix implementation of posix_spawn{p}
and align with Linux one. The main idea is to fix some issues already
fixed in Linux code, and deprecated vfork internal usage (source of
various bug reports). In a short:
- It moves POSIX_SPAWN_USEVFORK usage and sets it a no-op. Since
the process that actually spawn the new process do not share
memory with parent (with vfork), it fixes BZ#14750 for this
implementation.
- It uses a pipe to correctly obtain the return upon failure
of execution (BZ#18433).
- It correctly enable/disable asynchronous cancellation (checked
on ptl/tst-exec5.c).
- It correctly disable/enable signal handling.
Using this version instead of Linux shows only one regression,
posix/tst-spawn3, because of pipe2 usage which increase total
number of file descriptor.
* sysdeps/posix/spawni.c (__spawni_child): New function.
(__spawni): Rename to __spawnix.
git://sourceware.org / glibc.git / log