y2038: linux: Provide ___adjtimex64 implementation
This patch provides new ___adjtimex64 explicit 64 bit function for adjusting
Linux kernel clock.
Internally, the __clock_adjtime64 syscall is used. This patch is necessary
for having architectures with __WORDSIZE == 32 Y2038 safe.
Moreover, a 32 bit version - ___adjtimex has been refactored to internally
use ___adjtimex64.
The ___adjtimex is now supposed to be used on systems still supporting 32
bit time (__TIMESIZE != 64) - hence the necessary conversions between struct
timex and 64 bit struct __timex64.
Last but not least, in ___adjtimex64 function the __clock_adjtime syscall has
been replaced with __clock_adjtime64 to support 64 bit time on architectures
with __WORDSIZE == 32 and __TIMESIZE != 64.
Run-time tests:
- Run specific tests on ARM/x86 32bit systems (qemu):
https://github.com/lmajewski/meta-y2038 and run tests:
https://github.com/lmajewski/y2038-tests/commits/master
Above tests were performed with Y2038 redirection applied as well as without to
test the proper usage of both ___adjtimex64 and ___adjtimex.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
y2038: linux: Provide __clock_adjtime64 implementation
This patch replaces auto generated wrapper (as described in
sysdeps/unix/sysv/linux/syscalls.list) for clock_adjtime with one which adds
extra support for reading 64 bit time values on machines with __TIMESIZE != 64.
To achieve this goal new __clock_adjtime64 explicit 64 bit function for
adjusting Linux clock has been added.
Moreover, a 32 bit version - __clock_adjtime has been refactored to internally
use __clock_adjtime64.
The __clock_adjtime is now supposed to be used on systems still supporting 32
bit time (__TIMESIZE != 64) - hence the necessary conversions between 64 bit
struct __timespec64 and struct timespec.
The new __clock_adjtime64 syscall available from Linux 5.1+ has been used, when
applicable.
Up till v5.4 in the Linux kernel there was a bug preventing this call from
obtaining correct struct's timex time.tv_sec time after time_t overflow
(i.e. not being Y2038 safe).
Run-time tests:
- Run specific tests on ARM/x86 32bit systems (qemu):
https://github.com/lmajewski/meta-y2038 and run tests:
https://github.com/lmajewski/y2038-tests/commits/master
Linux kernel, headers and minimal kernel version for glibc build test matrix:
- Linux v5.1 (with clock_adjtime64) and glibc build with v5.1 as
minimal kernel version (--enable-kernel="5.1.0")
The __ASSUME_TIME64_SYSCALLS flag defined.
- Linux v5.1 and default minimal kernel version
The __ASSUME_TIME64_SYSCALLS not defined, but kernel supports clock_adjtime64
syscall.
- Linux v4.19 (no clock_adjtime64 support) with default minimal kernel version
for contemporary glibc (3.2.0)
This kernel doesn't support clock_adjtime64 syscall, so the fallback to
clock_adjtime is tested.
Above tests were performed with Y2038 redirection applied as well as without
(so the __TIMESIZE != 64 execution path is checked as well).
Florian Weimer [Tue, 19 May 2020 12:09:38 +0000 (14:09 +0200)]
nss_compat: internal_end*ent may clobber errno, hiding ERANGE [BZ #25976]
During cleanup, before returning from get*_r functions, the end*ent
calls must not change errno. Otherwise, an ERANGE error from the
underlying implementation can be hidden, causing unexpected lookup
failures. This commit introduces an internal_end*ent_noerror
function which saves and restore errno, and marks the original
internal_end*ent function as warn_unused_result, so that it is used
only in contexts were errors from it can be handled explicitly.
Anton Blanchard [Thu, 14 May 2020 11:49:16 +0000 (21:49 +1000)]
powerpc: Optimized rawmemchr for POWER9
This version uses vector instructions and is up to 60% faster on medium
matches and up to 90% faster on long matches, compared to the POWER7
version. A few examples:
H.J. Lu [Mon, 27 Apr 2020 22:44:07 +0000 (15:44 -0700)]
x86: Add --enable-cet=permissive
When CET is enabled, it is an error to dlopen a non CET enabled shared
library in CET enabled application. It may be desirable to make CET
permissive, that is disable CET when dlopening a non CET enabled shared
library. With the new --enable-cet=permissive configure option, CET is
disabled when dlopening a non CET enabled shared library.
Add DEFAULT_DL_X86_CET_CONTROL to config.h.in:
/* The default value of x86 CET control. */
#define DEFAULT_DL_X86_CET_CONTROL cet_elf_property
which enables CET features based on ELF property note.
--enable-cet=permissive it to
/* The default value of x86 CET control. */
#define DEFAULT_DL_X86_CET_CONTROL cet_permissive
which enables CET features permissively.
Update tst-cet-legacy-5a, tst-cet-legacy-5b, tst-cet-legacy-6a and
tst-cet-legacy-6b to check --enable-cet and --enable-cet=permissive.
Since 2.22 is older than the minimum required binutils version
for building glibc, we no longer need this. (The changes do
not impact the statically linked startup code.)
Florian Weimer [Mon, 18 May 2020 13:21:04 +0000 (15:21 +0200)]
elf: Assert that objects are relocated before their constructors run
If we try to run constructors before relocation, this is always
a dynamic linker bug. An assert is easier to notice than a call
via an invalid function pointer (which may not even produce a valid
call stack).
Add stpcpy support to the POWER9 strcpy. This is up to 40% faster on
small strings and up to 90% faster on long relatively unaligned strings,
compared to the POWER8 version. A few examples:
__stpcpy_power9 __stpcpy_power8
Length 20, alignments in bytes 4/ 4: 2.58246 4.8788
Length 1024, alignments in bytes 1/ 6: 24.8186 47.8528
This version uses VSX store vector with length instructions and is
significantly faster on small strings and relatively unaligned large
strings, compared to the POWER8 version. A few examples:
__strcpy_power9 __strcpy_power8
Length 16, alignments in bytes 0/ 0: 2.52454 4.62695
Length 412, alignments in bytes 4/ 0: 11.6 22.9185
H.J. Lu [Tue, 28 Apr 2020 17:05:25 +0000 (10:05 -0700)]
x86: Move CET control to _dl_x86_feature_control [BZ #25887]
1. Include <dl-procruntime.c> to get architecture specific initializer in
rtld_global.
2. Change _dl_x86_feature_1[2] to _dl_x86_feature_1.
3. Add _dl_x86_feature_control after _dl_x86_feature_1, which is a
struct of 2 bitfields for IBT and SHSTK control
Samuel Thibault [Sat, 16 May 2020 17:29:50 +0000 (19:29 +0200)]
sunrpc/tst-udp-*: Fix timeout value
Following on 00c3da43a ('sunrpc/tst-udp-timeout: Fix timeout value').
While e.g. 2.5 timeout can be represented exactly with doubles, time stamps
may not be represented that exactly, and thus with coarse-grain clocks (e.g.
10ms) we may end up with 2.499-ish values due to rounding errors.
Florian Weimer [Fri, 15 May 2020 09:32:30 +0000 (11:32 +0200)]
Linux: Remove remnants of the getcpu cache
The getcpu cache was removed from the kernel in Linux 2.6.24. glibc
support from the sched_getcpu implementation was removed in commit dd26c44403582fdf10d663170f947dfe4b3207a0 ("Consolidate sched_getcpu").
Paul Eggert [Fri, 15 May 2020 15:52:25 +0000 (08:52 -0700)]
Update timezone code from tzcode 2020a
This patch updates files coming from tzcode to tzcode 2020a.
This is mostly for better support for Internet RFC 8536, by adding
support to zic for the Expires line (new to tzcode 2020a), the -b
option (new to 2019b) and the -r option (new to 2019a).
One trivial change to other glibc was needed.
* time/tzfile.c (__tzfile_read): Adjust to tzcode private.h renaming.
* timezone/private.h, timezone/tzfile.h, timezone/version:
* timezone/zdump.c, timezone/zic.c: Update from tzcode 2020a.
Lexi Shao [Fri, 15 May 2020 10:48:59 +0000 (18:48 +0800)]
aarch64: fix strcpy and strnlen for big-endian [BZ #25824]
This patch fixes the optimized implementation of strcpy and strnlen
on a big-endian arm64 machine.
The optimized method uses neon, which can process 128bit with one
instruction. On a big-endian machine, the bit order should be reversed
for the whole 128-bits double word. But with instuction
rev64 datav.16b, datav.16b
it reverses 64bits in the two halves rather than reversing 128bits.
There is no such instruction as rev128 to reverse the 128bits, but we
can fix this by loading the data registers accordingly.
Fixes 0237b61526e7("aarch64: Optimized implementation of strcpy") and 2911cb68ed3d("aarch64: Optimized implementation of strnlen").
Signed-off-by: Lexi Shao <shaolexi@huawei.com> Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
Florian Weimer [Thu, 14 May 2020 10:12:29 +0000 (12:12 +0200)]
string: Fix string/tst-memmove-overflow to compile with GCC 7
GCC 8 relaxed what kind of expressions can be used in initializers,
and the previous use of static const variables relied on that. Switch
to wide (non-int) enum constants instead, which is another GCC
extension that is more widely implemented.
Reviewed-by: Carlos O'Donell <carlos@redhat.com> Tested-by: Carlos O'Donell <carlos@redhat.com>
guojinhui [Thu, 14 May 2020 16:09:46 +0000 (18:09 +0200)]
Add arch-syscall.h dependency for generating sysd-syscalls file
After using "make update-syscall-lists" to update arch-syscall.h for
new kernel versions, sysd-syscalls will not be not be regenerated.
This will cause a compile error because the new data is not being
picked up.
Florian Weimer [Tue, 12 May 2020 17:06:18 +0000 (19:06 +0200)]
elf: Remove redundant add_to_global_resize_failure call from dl_open_args
The second call does not do anything because the data structures have
already been resized by the call that comes before the demarcation
point. Fixes commit a509eb117fac1d764b15eba64993f4bdb63d7f3c
("Avoid late dlopen failure due to scope, TLS slotinfo updates
[BZ #25112]").
Florian Weimer [Tue, 12 May 2020 09:30:30 +0000 (11:30 +0200)]
aarch64: Accept PLT calls to __getauxval within libc.so
When using outline atomics (-moutline-atomics, the default for ARMv8-A
starting with GCC 10), libgcc contains an ELF constructor which calls
__getauxval. This code is built outside of glibc, so none of its
internal PLT avoidance schemes can be applied to it. This change
suppresses the elf/check-localplt failure.
Use unsigned constants for ICMP6 filters [BZ #22489]
The core problem here is that the filter array elements are unsigned
but the computed constants are signed. This both causes a
signededness conversion at the &= step and may cause undefined
behavior if the MSB is being modified. This patch uses unsigned
constants to avoid both cases. - DJ
Florian Weimer [Mon, 11 May 2020 09:20:02 +0000 (11:20 +0200)]
POWER: Add context-synchronizing instructions to pkey_write [BZ #25954]
Sandipan Das reported that,
"The Power ISA mandates that all writes to the Authority
Mask Register (AMR) must always be preceded as well as
succeeded by a context-synchronizing instruction. This
applies to both the privileged and unprivileged variants
of the Move To AMR instruction.
This [requirement] is from Table 6 of Chapter 11 in page 1134 of Power
ISA 3.0B. The document can be found here:
<https://ibm.ent.box.com/s/1hzcwkwf8rbju5h9iyf44wm94amnlcrv>
"
Only alpha and ia64 do not support __NR_umount2 (defined as
__NR_umount), but recent kernel fixes (74cd2184833f for ia64, and 12b57c5c70f39 for alpha) add the required alias.
Vineet Gupta [Tue, 5 May 2020 19:53:41 +0000 (12:53 -0700)]
semaphore: consolidate arch headers into a generic one
This consolidates the copy-pasted arch specific semaphore header into
single version (based on s390) which suffices 32-bit and and 64-bit
arch/ABI based on the canonical WORDSIZE.
For now I've left out arches which use alternate defines to choose for
32 vs 64-bit builds (aarch64, mips) which in theory can also use the same
header.
Joseph Myers [Wed, 6 May 2020 15:41:55 +0000 (15:41 +0000)]
Use GCC 10 branch in build-many-glibcs.py.
This updates the default GCC version used in build-many-glibcs.py when
no version is specified explicitly. I'm replacing my bot using GCC 8
with one using GCC 10 (leaving the GCC 9 and GCC mainline bots running
as at present).
y2038: Provide conversion helpers for struct __timex64
Those functions allow easy conversion between Y2038 safe, glibc internal
struct __timex64 and struct timex.
Those functions are put in Linux specific sys/timex.h file, as putting
them into glibc's local include/time.h would cause build break on HURD as
it doesn't support struct timex related syscalls.
y2038: Introduce struct __timex64 - new internal glibc type
The introduced glibc's 'internal' struct __timex64 is a copy of Linux kernel's
struct __kernel_timex (v5.6) introduced for properly handling data for
clock_adjtime64 syscall.
As the struct's __kernel_timex size is the same as for archs with
__WORDSIZE == 64, proper padding and data types conversion (i.e. long to long
long) had to be added for architectures with __WORDSIZE == 32 &&
__TIMESIZE != 64.
Moreover, it stores time in struct __timeval64 rather than struct
timeval, which makes it Y2038-proof.
y2038: include: Move struct __timeval64 definition to a separate file
The struct __timeval64's definition has been moved from ./include/time.h to
./include/struct___timeval64.h.
This change would prevent from polluting other glibc namespaces (when
headers are modified to support 64 bit time on architectures with
__WORDSIZE==32).
Now it is possible to just include definition of this particular structure
when needed.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Lukasz Majewski [Mon, 16 Mar 2020 12:50:23 +0000 (13:50 +0100)]
y2038: nscd: Modify nscd_helper to use __clock_gettime64
The nscd/nscd_helper.c uses __clock_gettime to get current time and on this
basis calculate the relative timeout for poll.
By using __clock_gettime64 on systems with __WORDSIZE == 32 && __TIMESIZE != 64
the timeout is correctly calculated after time_t overflow.
Lukasz Majewski [Mon, 16 Mar 2020 09:09:13 +0000 (10:09 +0100)]
y2038: inet: Convert inet deadline to support 64 bit time
This change brings 64 bit time support to inet deadline related code for
architectures with __WORDSIZE == 32 && __TIMESIZE != 64.
It is also safe to replace struct timespec with struct __timespec64 in
deadline related structures as:
- The __deadline_to_ms () returns the number of miliseconds to deadline to
be used with __poll (and hence it is a relative value).
- To calculate the deadline from timeval (which will be converted latter)
the uintmax_t type is used (unsinged long long int).
Lukasz Majewski [Thu, 16 Jan 2020 13:27:27 +0000 (14:27 +0100)]
y2038: hurd: Provide __clock_gettime64 function
For Linux glibc ports the __TIMESIZE == 64 ensures proper aliasing for
__clock_gettime64 (to __clock_gettime).
When __TIMESIZE != 64 (like ARM32, PPC) the glibc expects separate definition
of the __clock_gettime64.
The HURD port only provides __clock_gettime, so this patch adds
__clock_gettime64 as a tiny wrapper on it. Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Lukasz Majewski [Tue, 24 Mar 2020 07:52:15 +0000 (08:52 +0100)]
y2038: Export __clock_gettime64 to be usable in other libraries
In the glibc project calls to clock_gettime shall be replaced with
__clock_gettime64, which is supporting 64 bit time.
To allow that the __clock_gettime64 needs to be exported as a GLIBC_PRIVATE
symbol.
Paul E. Murphy [Fri, 1 May 2020 19:30:42 +0000 (14:30 -0500)]
powerpc64le/power9: guard power9 strcmp against rtld usage [BZ# 25905]
strcmp is used while resolving PLT references. Vector registers
should not be used during this. The P9 strcmp makes heavy use of
vector registers, so it should be avoided in rtld.
This prevents quiet vector register corruption when glibc is configured
with --disable-multi-arch and --with-cpu=power9. This can be seen with
test-float64x-compat_totalordermag during the first call into
totalordermagf64x@GLIBC_2.27.
Add a guard to fallback to the power8 implementation when building
power9 strcmp for libraries other than libc.
Paul E. Murphy [Tue, 7 Apr 2020 21:20:35 +0000 (16:20 -0500)]
float128: use builtin_signbitf128 always
The minimum GCC version has been raised to 6.2 for building
glibc. Therefore, follow the advice inside the implementation
and remove the GCC < 6 codepath.
Likewise, remove the hidden_proto as all internal usages should
inline now.
Martin Sebor [Mon, 4 May 2020 17:21:50 +0000 (11:21 -0600)]
improve out-of-bounds checking with GCC 10 attribute access [BZ #25219]
Adds the access attribute newly introduced in GCC 10 to the subset of
function declarations that are already covered by _FORTIFY_SOURCE and
that don't have corresponding GCC built-in equivalents.
Commit 06436acf819d9e6ada7be3ca977d5c0a23d3f138 created a kernel-features.h
file with '#undef __ASSUME_SYSVIPC_DEFAULT_IPC_64', which is wrong for Nios II.
Deleting the header.
Carlos O'Donell [Tue, 25 Feb 2020 15:15:30 +0000 (10:15 -0500)]
localedef: Add tests-container test for --no-hard-links.
The new tst-localedef-hardlinks verifies that when compiling
two locales (with default output directory) one with
--no-hard-links and one without the option, results in the
expected behaviour. When --no-hard-links is used the link
counts on LC_CTYPE is 1, indicating that even thoug the two
locale are identical (though different named source files and
output direcotry) the localedef did not carry out the hard
link optimization. Then when --no-hard-links is omitted the
localedef hard link optimization is correctly carried out and
for 2 compiled locales the link count for LC_CTYPE is 2.
Carlos O'Donell [Thu, 23 Jan 2020 14:45:00 +0000 (09:45 -0500)]
test-container: Support $(complocaledir) and mkdirp.
Expand the support infrastructure:
- Create $(complocaledir) in the testroot.pristine to support localedef.
- Add the variable $complocaledir to script support.
- Add the script command 'mkdirp'.
All localedef tests which run with default paths need to have the
$(complocaledir) created in testroot.pristine. The localedef binary
will not by itself create the default path, but it will write into
the path. By adding this we can simplify the localedef tests.
The variable $complocaledir is the value of the configured
$(complocaledir) which is the location of the compiled locales that
will be searched by the runtime by default.
The command mkdirp will be available in script setup and will
be equivalent to running `mkdir -p`.
The variable and command can be used to write more complex tests.
i386: Remove unused variable in sysdeps/x86/cacheinfo.c
Commit a98dc92dd1e278df4c501deb07985018bc2b06de ("x86: Add cache
information support for Zhaoxin processors") introduced an unused
variable warning in the default i686-linux-gnu build:
In file included from ../sysdeps/i386/cacheinfo.c:3:
../sysdeps/x86/cacheinfo.c: In function 'init_cacheinfo':
../sysdeps/x86/cacheinfo.c:762:16: error: unused variable 'eax' [-Werror=unused-variable]
762 | unsigned int eax;
| ^~~
On platforms where long double may have two different formats, i.e.: the
same format as double (64-bits) or something else (128-bits), building
with -mlong-double-128 is the default and function calls in the user
program match the name of the function in Glibc. When building with
-mlong-double-64, Glibc installed headers redirect such calls to the
appropriate function.
Likewise, the internals of glibc are now built against IEEE long double.
However, the only (minimally) notable usage of long double is difftime.
Reviewed-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
GCC 7.5.0 (PR94200) will refuse to compile if both -mabi=% and
-mlong-double-128 are passed on the command line. Surprisingly,
it will work happily if the latter is not. For the sake of
maintaining status quo, test for and blacklist such compilers.
Tested with a GCC 8.3.1 and GCC 7.5.0 compiler for ppc64le.
Reviewed-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
Paul E. Murphy [Fri, 7 Feb 2020 20:08:23 +0000 (14:08 -0600)]
powerpc64le: raise GCC requirement to 7.4 for long double transition
Add compiler feature tests to ensure we can build ieee128 long double.
These test for -mabi=ieeelongdouble, -mno-gnu-attribute, and -Wno-psabi.
Likewise, verify some compiler bugs have been addressed. These aren't
helpful for building glibc, but may cause test failures when testing
the new long double. See notes below from Raji.
On powerpc64le, some older compiler versions give error for the function
signbit() for 128-bit floating point types. This is fixed by PR83862
in gcc 8.0 and backported to gcc6 and gcc7. This patch adds a test
to check compiler version to avoid compiler errors during make check.
Likewise, test for -mno-gnu-attribute support which was
On powerpc64le, a few files are built on IEEE long double mode
(-mabi=ieeelongdouble), whereas most are built on IBM long double mode
(-mabi=ibmlongdouble, the default for -mlong-double-128). Since binutils
2.31, linking object files with different long double modes causes
errors similar to:
ld: libc_pic.a(s_isinfl.os) uses IBM long double,
libc_pic.a(ieee128-qefgcvt.os) uses IEEE long double.
collect2: error: ld returned 1 exit status
make[2]: *** [../Makerules:649: libc_pic.os] Error 1
The warnings are fair and correct, but in order for glibc to have
support for both long double modes on powerpc64le, they have to be
ignored. This can be accomplished with the use of -mno-gnu-attribute
option when building the few files that require IEEE long double mode.
However, -mno-gnu-attribute is not available in GCC 6, the minimum
version required to build glibc, so this patch adds a test for this
feature in powerpc64le builds, and fails early if it's not available.
Co-Authored-By: Rajalakshmi Srinivasaraghavan <raji@linux.vnet.ibm.com> Co-Authored-By: Gabriel F. T. Gomes <gabrielftg@linux.ibm.com> Reviewed-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
Paul E. Murphy [Fri, 14 Feb 2020 22:41:11 +0000 (16:41 -0600)]
Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI
Improve the commentary to aid future developers who will stumble
upon this novel, yet not always perfect, mechanism to support
alternative formats for long double.
Likewise, rename __LONG_DOUBLE_USES_FLOAT128 to
__LDOUBLE_REDIRECTS_TO_FLOAT128_ABI now that development work
has settled down. The command used was
Paul E. Murphy [Fri, 6 Mar 2020 15:41:03 +0000 (09:41 -0600)]
ldbl-128ibm-compat: workaround GCC 9 C++ PR90731
GCC 9 has a bug (PR90731) whereby __typeof does not correctly copy
exception specifiers[1]. Surprisingly, this can be quieted by declaring
"#pragma system_header", or if the headers are installed in a system
directory.
Work around this by using the pragma for any gcc version between
9.0 and 9.2 to ensure tests continue to compile.
[1] Example error from g++ 9.2.1:
In file included from ../include/sys/cdefs.h:3,
from ../include/features.h:465,
from ../bits/libc-header-start.h:33,
from ../math/math.h:27,
from ../include/math.h:7,
from test-math-isinff.cc:21:
../libio/bits/stdio-ldbl.h:25:20: error: declaration of ‘int sprintf(char*, const char*, ...)’ has a different exception specifier
25 | __LDBL_REDIR_DECL (sprintf)
| ^~~~~~~
../misc/sys/cdefs.h:461:26: note: in definition of macro ‘__LDBL_REDIR_DECL’
461 | extern __typeof (name) name __asm (__ASMNAME ("__" #name "ieee128"));
| ^~~~
In file included from ../include/stdio.h:5,
from test-math-isinff.cc:22:
../libio/stdio.h:334:12: note: from previous declaration ‘int sprintf(char*, const char*, ...) throw ()’
334 | extern int sprintf (char *__restrict __s,
| ^~~~~~~
Reviewed-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
H.J. Lu [Wed, 29 Apr 2020 20:20:27 +0000 (13:20 -0700)]
Add C wrappers for process_vm_readv/process_vm_writev [BZ #25810]
Since the the U marker can only be applied to 2 unsigned long arguments
in syscalls.list files, add a C wrapper for process_vm_readv and
process_vm_writev syscals which have more than 2 unsigned long arguments.
Alistair Francis [Fri, 14 Feb 2020 19:02:29 +0000 (11:02 -0800)]
generic/typesizes.h: Add support for 32-bit arches with 64-bit types
Update the default typesizes.h to match the new kernel sizes for 32-bit
architectures with a 64-bit time_t and friends. This follows the sizes
used for RV32 which is a y2038 safe architecture added after Linux 5.1.
H.J. Lu [Wed, 29 Apr 2020 15:08:40 +0000 (08:08 -0700)]
Mark unsigned long arguments with U in more syscalls [BZ #25810]
Mark unsigned long arguments in mmap, read, recv, recvfrom, send, sendto,
write, ioperm, sendfile64, setxattr, lsetxattr, fsetxattr, getxattr,
lgetxattr, fgetxattr, listxattr, llistxattr and flistxattr with U in
syscalls.list files.
elf: Add initial flag argument to __libc_early_init
The rseq initialization should happen only for the libc in the base
namespace (in the dynamic case) or the statically linked libc. The
__libc_multiple_libcs flag does not quite cover this case at present,
so this commit introduces a flag argument to __libc_early_init,
indicating whether the libc being libc is the primary one (of the main
program).
H.J. Lu [Wed, 29 Apr 2020 12:35:34 +0000 (05:35 -0700)]
Add SYSCALL_ULONG_ARG_[12] to pass long to syscall [BZ #25810]
X32 has 32-bit long and pointer with 64-bit off_t. Since x32 psABI
requires that pointers passed in registers must be zero-extended to
64bit, x32 can share many syscall interfaces with LP64. When a LP64
syscall with long and unsigned long int arguments is used for x32, these
arguments must be properly extended to 64-bit. Otherwise if the upper
32 bits of the register have undefined value, such a syscall will be
rejected by kernel.
For syscalls implemented in assembly codes, 'U' is added to syscall
signature key letters for unsigned long, which is zero-extended to
64-bit types. SYSCALL_ULONG_ARG_1 and SYSCALL_ULONG_ARG_2 are passed
to syscall-template.S for the first and the second unsigned long int
arguments if PSEUDOS_HAVE_ULONG_INDICES is defined. They are used by
x32 to zero-extend 32-bit arguments to 64 bits.
Tested on i386, x86-64 and x32 as well as with build-many-glibcs.py.
misc: Remove sstk from the autogenerated system call list
This change should not have an effect because the system call was
never defined. Also add the misssing attribute_compat_text_section
attribute to the sstk function (a minor optimization). Also update the
NEWS file to document the change.
Joseph Myers [Tue, 28 Apr 2020 15:05:09 +0000 (15:05 +0000)]
Remove unused floating-point configuration from gmp-impl.h.
This patch removes the IEEE_DOUBLE_BIG_ENDIAN and
IEEE_DOUBLE_MIXED_ENDIAN macros from gmp-impl.h and gmp-mparam.h, and
the ieee_double_extract union from gmp-impl.h. The macros were used
only in defining the union, which was used nowhere in glibc. As GMP's
gmp-impl.h is over 5000 lines, the file in glibc is so far from the
GMP version that it doesn't seem to make sense to keep things there
that are not relevant in glibc. (I expect there is plenty more in the
header after this patch that is also not relevant in glibc and can be
cleaned up later.)
Tested with build-many-glibcs.py that installed stripped shared
libraries are unchanged by this patch.
It is not implemented anywhere. There is an osf_sstk system call on
alpha, but it is not used to implement sstk, and the system call
is not implemented on Linux, either.
nptl: Start new threads with all signals blocked [BZ #25098]
New threads inherit the signal mask from the current thread. This
means that signal handlers can run on the newly created thread
immediately after the kernel has created the userspace thread, even
before glibc has initialized the TCB. Consequently, new threads can
observe uninitialized ctype data, among other things.
To address this, block all signals before starting the thread, and
pass the original signal mask to the start routine wrapper. On the
new thread, first perform all thread initialization, and then unblock
signals.
The cost of doing this is two rt_sigprocmask system calls on the old
thread, and one rt_sigprocmask system call on the new thread. (If
there was a way to clone a new thread with a signals disabled, this
could be brought down to one system call each.) The thread descriptor
increases in size, too, and sigset_t is fairly large. This increase
could be brought down by reusing space the in the descriptor which is
not needed before running user code, or by switching to an internal
sigset_t definition which only covers the signals supported by the
kernel definition. (Part of the thread descriptor size increase is
already offset by reduced stack usage in the thread start wrapper
routine after this commit.)
Carlos O'Donell [Wed, 8 Jan 2020 18:24:42 +0000 (13:24 -0500)]
localedef: Add verbose messages for failure paths.
During testing of localedef running in a minimal container
there were several error cases which were hard to diagnose
since they appeared as strerror (errno) values printed by the
higher level functions. This change adds three new verbose
messages for potential failure paths. The new messages give
the user the opportunity to use -v and display additional
information about why localedef might be failing. I found
these messages useful myself while writing a localedef
container test for --no-hard-links.
Since the changes cleanup the code that handle codeset
normalization we add tst-localedef-path-norm which contains
many sub-tests to verify the correct expected normalization of
codeset strings both when installing to default paths (the
only time normalization is enabled) and installing to absolute
paths. During the refactoring I created at least one
buffer-overflow which valgrind caught, but these tests did not
catch because the exec in the container had a very clean heap
with zero-initialized memory. However, between valgrind and
the tests the results are clean.
The new tst-localedef-path-norm passes without regression on
x86_64.
Joseph Myers [Fri, 24 Apr 2020 22:08:59 +0000 (22:08 +0000)]
Remove most gmp-mparam.h headers.
Most gmp-mparam.h headers in glibc define various macros to the same
values they would be defined to by the generic version of that header,
plus macros IEEE_DOUBLE_BIG_ENDIAN or IEEE_DOUBLE_MIXED_ENDIAN related
to the representation of double. The latter macros are in turn only
used in gmp-impl.h to define union ieee_double_extract, which is not
used in glibc. Thus all of these headers, except for the generic one
and those that define _LONG_LONG_LIMB for ILP32 configurations with
64-bit registers, are redundant, and this patch removes them.
Tested with build-many-glibcs.py that installed stripped shared
libraries are unchanged by this patch.
This function is defined in libc.so, and the dynamic loader calls
right after relocation has been finished, before any ELF constructors
or the preinit function is invoked. It is also used in the static
build for initializing parts of the static libc.
To locate __libc_early_init, a direct symbol lookup function is used,
_dl_lookup_direct. It does not search the entire symbol scope and
consults merely a single link map. This function could also be used
to implement lookups in the vDSO (as an optimization).
A per-namespace variable (libc_map) is added for locating libc.so,
to avoid repeated traversals of the search scope. It is similar to
GL(dl_initfirst). An alternative would have been to thread a context
argument from _dl_open down to _dl_map_object_from_fd (where libc.so
is identified). This could have avoided the global variable, but
the change would be larger as a result. It would not have been
possible to use this to replace GL(dl_initfirst) because that global
variable is used to pass the function pointer past the stack switch
from dl_main to the main program. Replacing that requires adding
a new argument to _dl_init, which in turn needs changes to the
architecture-specific libc.so startup code written in assembler.
__libc_early_init should not be used to replace _dl_var_init (as
it exists today on some architectures). Instead, _dl_lookup_direct
should be used to look up a new variable symbol in libc.so, and
that should then be initialized from the dynamic loader, immediately
after the object has been loaded in _dl_map_object_from_fd (before
relocation is run). This way, more IFUNC resolvers which depend on
these variables will work.
MIPS needs to ignore certain existing symbols during symbol lookup.
The old scheme uses the ELF_MACHINE_SYM_NO_MATCH macro, with an
inline function, within its own header, with a sysdeps override for
MIPS. This allows re-use of the function from another file (without
having to include <dl-machine.h> or providing the default definition
for ELF_MACHINE_SYM_NO_MATCH).
signal: Only handle on NSIG signals on signal functions (BZ #25657)
The upper bits of the sigset_t s not fully initialized in the signal
mask calls that return information from kernel (sigprocmask,
sigpending, and pthread_sigmask), since the exported sigset_t size
(1024 bits) is larger than Linux support one (64 or 128 bits).
It might make sigisemptyset/sigorset/sigandset fail if the mask
is filled prior the call.
This patch changes the internal signal function to handle up to
supported Linux signal number (_NSIG), the remaining bits are
untouched.
Mike FABIAN [Tue, 14 Apr 2020 16:43:13 +0000 (18:43 +0200)]
Bug 25819: Update to Unicode 13.0.0
Unicode 13.0.0 Support: Character encoding, character type info, and
transliteration tables are all updated to Unicode 13.0.0, using
the generator scripts contributed by Mike FABIAN (Red Hat).
Total added characters in newly generated CHARMAP: 5930
Total added characters in newly generated WIDTH: 5536
The __sfp_handle_exceptions is not fully correct regarding raising
exceptions, since there is no direct way to raise only FP_EX_OVERFLOW
nor FP_EX_UNDERFLOW for SSE mode. Both libgcc and feraiseexcept rely
on x87 mode to accomplish it.
The exported x86_64 fenv.h functions operate on both i387 and SSE (since
they should work on both float, double, and long double) while the
internal libc_fe* set either SSE (float, double, and float128) or
i387 (long double).
The libgcc __sfp_handle_exceptions (used on float128 implementation),
however, will set either SEE or i387 exception depending of the
exception to raise. This broke the internal assumption of float128
where only SSE operations will be used.
This patch reimplements the libgcc __sfp_handle_exceptions to use only
SSE operations and sets libgcc to use it instead of its own
implementation.
And I think we should fix libgcc in a similar manner, since checking on
config/i386/64/sfp-machine.h it already only supports SSE rounding mode
and x86_64 ABI also expectes float128 to use SSE registers [1]
(although it is not clear on how future implementation might implement
it).