string/endian.h: Restore the __USE_MISC conditionals
Commit 69fd157a3 "time: Add padding for the timespec if required"
caused a breakage in the glibc tests as the endian.h include file was
kept in the networking headers while the __USE_MISC #ifdefs had been
removed. This resulted in namespace violations in the networking
headers.
This patche restores the __USE_MISC conditionals in endian.h to fix the
test failures.
* string/endian.h: Restore the __USE_MISC conditionals.
Joseph Myers [Wed, 2 Oct 2019 17:26:14 +0000 (17:26 +0000)]
Disable warnings in string/tester.c at top level.
string/tester.c contains code that correctly triggers various GCC
warnings about dubious uses of string functions (uses that are being
deliberately tested there), and duly disables those warnings around
the relevant code.
A change in GCC mainline resulted in this code failing to compile with
a -Warray-bounds error, despite the location with the error having
-Warray-bounds already disabled. This has been reported as
<https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91890>. This patch
avoids that problem and possible future issues with these diagnostics
by moving all the warning disabling in this file to top level, as
suggested by Florian in
<https://sourceware.org/ml/libc-alpha/2019-10/msg00033.html>, rather
than only doing it locally around specific function calls.
Tested with build-many-glibcs.py for aarch64-linux-gnu with GCC
mainline (with only the conform/ failures noted in
<https://sourceware.org/ml/libc-alpha/2019-10/msg00043.html>).
* string/tester.c: Ignore -Warray-bounds and
-Wmemset-transposed-args at top level.
[__GNUC_PREREQ (7, 0)]: Ignore -Wrestrict and -Wstringop-overflow=
at top level.
[__GNUC_PREREQ (8, 0)]: Ignore -Wstringop-truncation at top level.
(test_stpncpy): Do not ignore warnings here.
(test_strncat): Likewise.
(test_strncpy): Likewise.
(test_memset): Likewise.
Lukasz Majewski [Wed, 2 Oct 2019 08:19:55 +0000 (10:19 +0200)]
Y2038: Include proper header to provide support for struct timeval on HURD
The HURD requires explicit inclusion of <bits/types/struct_timeval.h> to use
struct timeval in ./include/time.h.
For this particular glibc port, the proper header hasn't been included before
inclusion of time.h.
Tested with build-many-glibcs.py with i686-gnu and x86_64-linux-gnu:
build-many-glibcs.py /home/lukma/work/glibc/glibc-many-build --keep all compilers i686-gnu
build-many-glibcs.py /home/lukma/work/glibc/glibc-many-build --keep all glibcs i686-gnu
Also run of xcheck on x86_64:
./src/configure --prefix=/usr
make PARALLELMFLAGS="-j12" && make xcheck PARALLELMFLAGS="-j12"
Paul A. Clarke [Thu, 19 Sep 2019 16:11:04 +0000 (11:11 -0500)]
[powerpc] No need to enter "Ignore Exceptions Mode"
Since at least POWER8, there is no performance advantage to entering
"Ignore Exceptions Mode", and doing so conditionally requires
- the conditional logic, and
- a system call.
Arjun Shankar [Wed, 2 Oct 2019 11:59:43 +0000 (13:59 +0200)]
Enable passing arguments to the inferior in debugglibc.sh
This patch adds the ability to run debugglibc.sh's inferior program with
arguments specified on the command line. This enables convenient debugging
of non-testcase programs such as iconv/iconv_prog or other dynamically
linked programs. Program arguments may be passed using `--' as a separator.
Alistair Francis [Wed, 18 Sep 2019 23:51:23 +0000 (16:51 -0700)]
time: Add padding for the timespec if required
If we are running on a 32-bit system with a 64-bit time_t we need to
ensure there is padding around the tv_nsec variable. This is requried as
the timespec is #defined to the __timespec64 struct.
* time/bits/types/struct_timespec.h: Add padding for the timespec if
required.
Alistair Francis [Fri, 20 Sep 2019 21:23:51 +0000 (14:23 -0700)]
Split up endian.h to minimize exposure of BYTE_ORDER.
With only two exceptions (sys/types.h and sys/param.h, both of which
historically might have defined BYTE_ORDER) the public headers that
include <endian.h> only want to be able to test __BYTE_ORDER against
__*_ENDIAN.
This patch creates a new bits/endian.h that can be included by any
header that wants to be able to test __BYTE_ORDER and/or
__FLOAT_WORD_ORDER against the __*_ENDIAN constants, or needs
__LONG_LONG_PAIR. It only defines macros in the implementation
namespace.
The existing bits/endian.h (which could not be included independently
of endian.h, and only defines __BYTE_ORDER and maybe __FLOAT_WORD_ORDER)
is renamed to bits/endianness.h. I also took the opportunity to
canonicalize the form of this header, which we are stuck with having
one copy of per architecture. Since they are so short, this means git
doesn’t understand that they were renamed from existing headers, sigh.
endian.h itself is a nonstandard header and its only remaining use
from a standard header is guarded by __USE_MISC, so I dropped the
__USE_MISC conditionals from around all of the public-namespace things
it defines. (This means, an application that requests strict library
conformance but includes endian.h will still see the definition of
BYTE_ORDER.)
A few changes to specific bits/endian(ness).h variants deserve
mention:
- sysdeps/unix/sysv/linux/ia64/bits/endian.h is moved to
sysdeps/ia64/bits/endianness.h. If I remember correctly, ia64 did
have selectable endianness, but we have assembly code in
sysdeps/ia64 that assumes it’s little-endian, so there is no reason
to treat the ia64 endianness.h as linux-specific.
- The C-SKY port does not fully support big-endian mode, the compile
will error out if __CSKYBE__ is defined.
- The PowerPC port had extra logic in its bits/endian.h to detect a
broken compiler, which strikes me as unnecessary, so I removed it.
- The only files that defined __FLOAT_WORD_ORDER always defined it to
the same value as __BYTE_ORDER, so I removed those definitions.
The SH bits/endian(ness).h had comments inconsistent with the
actual setting of __FLOAT_WORD_ORDER, which I also removed.
- I *removed* copyright boilerplate from the few bits/endian(ness).h
headers that had it; these files record a single fact in a fashion
dictated by an external spec, so I do not think they are copyrightable.
As long as I was changing every copy of ieee754.h in the tree, I
noticed that only the MIPS variant includes float.h, because it uses
LDBL_MANT_DIG to decide among three different versions of
ieee854_long_double. This patch makes it not include float.h when
GCC’s intrinsic __LDBL_MANT_DIG__ is available.
* string/endian.h: Unconditionally define LITTLE_ENDIAN,
BIG_ENDIAN, PDP_ENDIAN, and BYTE_ORDER. Condition byteswapping
macros only on !__ASSEMBLER__. Move the definitions of
__BIG_ENDIAN, __LITTLE_ENDIAN, __PDP_ENDIAN, __FLOAT_WORD_ORDER,
and __LONG_LONG_PAIR to...
* string/bits/endian.h: ...this new file, which includes
the renamed header bits/endianness.h for the definition of
__BYTE_ORDER and possibly __FLOAT_WORD_ORDER.
* string/Makefile: Install bits/endianness.h.
* include/bits/endian.h: New wrapper.
* bits/endian.h: Rename to bits/endianness.h.
Add multiple-include guard. Rewrite the comment explaining what
the machine-specific variants of this file should do.
* sysdeps/unix/sysv/linux/ia64/bits/endian.h:
Move to sysdeps/ia64.
* sysdeps/unix/sysv/linux/hppa/pthread.h: Don’t include endian.h.
* sysdeps/mips/ieee754/ieee754.h: Use __LDBL_MANT_DIG__
in ifdefs, instead of LDBL_MANT_DIG. Only include float.h
when __LDBL_MANT_DIG__ is not predefined, in which case
define __LDBL_MANT_DIG__ to equal LDBL_MANT_DIG.
Mike FABIAN [Tue, 17 Sep 2019 15:20:32 +0000 (17:20 +0200)]
Sync "language", "lang_name", "territory", "country_name" with CLDR/langtable
Sync these values with CLDR and langtable as much as possible. Add
missing values.
If possible, take the values from CLDR, if CLDR does not have it,
take it from langtable. The values from langtable which are not from
CLDR are from Wikipedia or native speakers.
Lukasz Majewski [Mon, 25 Mar 2019 07:38:02 +0000 (08:38 +0100)]
y2038: Provide conversion helpers for struct __timespec64
Those functions allow easy conversion between Y2038 safe struct
__timespec64 and other time related data structures (like struct timeval
or struct timespec).
Joseph Myers [Mon, 30 Sep 2019 15:49:25 +0000 (15:49 +0000)]
Update bits/mman.h constants and tst-mman-consts.py for Linux 5.3.
The Linux 5.3 uapi headers have some rearrangement relating to MAP_*
constants, which includes the effect of adding definitions of MAP_SYNC
on powerpc and sparc. This patch updates the corresponding glibc
bits/mman.h headers accordingly, and updates the Linux kernel version
number in tst-mman-consts.py to reflect that these constants are now
current with that kernel version.
Tested with build-many-glibcs.py.
* sysdeps/unix/sysv/linux/powerpc/bits/mman.h [__USE_MISC]
(MAP_SYNC): New macro.
* sysdeps/unix/sysv/linux/sparc/bits/mman.h [__USE_MISC]
(MAP_SYNC): Likewise.
* sysdeps/unix/sysv/linux/tst-mman-consts.py (main): Update Linux
kernel version number to 5.3.
This patch adds a new make rule that generates a helper script for
debugging glibc test cases. The new script, debugglibc.sh, is similar
to testrun.sh, in the sense that it allows the execution of the
specified test case, however, it opens the test case in GDB, setting the
library path the same way that testrun.sh does. The commands are based
on the instructions on the wiki for glibc debugging [1,2].
By default, the script tells GDB to load the test case for symbol
information, so that, when a breakpoint is hit, the call stack is
displayed correctly (instead of printing lots of '??'s). For instance,
after running 'make' and 'make check', one could do the following:
$ ./debugglibc.sh nptl/tst-exec1 -b pthread_join
Reading symbols from /home/gabriel/build/powerpc64le/glibc//elf/ld.so...done.
Breakpoint 1 at 0x1444
add symbol table from file "nptl/tst-exec1"
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/home/gabriel/build/powerpc64le/glibc//nptl_db/libthread_db.so.1".
Breakpoint 1, 0x00007ffff7fb1444 in _dl_start_user () from /home/gabriel/build/powerpc64le/glibc/elf/ld.so
Breakpoint 2 at 0x7ffff7f49d48: file pthread_join.c, line 23.
Notice that the script will always start GDB with the program running
and halted at _dl_start_user. So, in order to reach the actual
breakpoint of interest, one should hit 'c', not 'r':
>>> c
Continuing.
[New Thread 0x7ffff7d1f180 (LWP 76443)]
[Switching to Thread 0x7ffff7d1f180 (LWP 76443)]
Thread 2 "ld.so" hit Breakpoint 2, __pthread_join (threadid=140737354087616, thread_return=0x0) at pthread_join.c:24
24 return __pthread_timedjoin_ex (threadid, thread_return, NULL, true);
Then inspect the call stack with 'bt', as usual, and see symbols from
both the test case and from the libraries themselves:
>>> bt
#0 __pthread_join (threadid=140737354087616, thread_return=0x0) at pthread_join.c:24
#1 0x0000000010001f4c in tf (arg=<optimized out>) at tst-exec1.c:37
#2 0x00007ffff7f487e8 in start_thread (arg=0x7ffff7510000) at pthread_create.c:479
#3 0x00007ffff7e523a8 in clone () at ../sysdeps/unix/sysv/linux/powerpc/powerpc64/clone.S:82
Reviewed-by: Carlos O'Donell <carlos@redhat.com> Reviewed-by: Dmitry V. Levin <ldv@altlinux.org> Reviewed-by: Joseph Myers <joseph@codesourcery.com> Reviewed-by: Andreas Schwab <schwab@suse.de>
Paul A. Clarke [Thu, 19 Sep 2019 19:04:45 +0000 (14:04 -0500)]
[powerpc] Rename fesetenv_mode to fesetenv_control
fesetenv_mode is used variously to write the FPSCR exception enable
bits and rounding mode bits. These are referred to as the control
bits in the POWER ISA. Change the name to be reflective of its
current and expected use, and match up well with fegetenv_control.
libc_feholdsetround_noex_ppc_ctx currently performs:
1. Read FPSCR, save to context.
2. Create new FPSCR value: clear enables and set new rounding mode.
3. Write new value to FPSCR.
Since other bits just pass through, there is no need to write them.
Instead, write just the changed values (enables and rounding mode),
which can be a bit more efficient.
Paul A. Clarke [Thu, 19 Sep 2019 16:58:46 +0000 (11:58 -0500)]
[powerpc] Rename fegetenv_status to fegetenv_control
fegetenv_status is used variously to retrieve the FPSCR exception enable
bits, rounding mode bits, or both. These are referred to as the control
bits in the POWER ISA. FPSCR status bits are also returned by the
'mffs' and 'mffsl' instructions, but they are uniformly ignored by all
uses of fegetenv_status. Change the name to be reflective of its
current and expected use.
Reviewed-By: Paul E Murphy <murphyp@linux.ibm.com>
Paul A. Clarke [Thu, 19 Sep 2019 16:39:44 +0000 (11:39 -0500)]
[powerpc] __fesetround_inline optimizations
On POWER9, use more efficient means to update the 2-bit rounding mode
via the 'mffscrn' instruction (instead of two 'mtfsb0/1' instructions
or one 'mtfsfi' instruction that modifies 4 bits).
Suggested-by: Paul E. Murphy <murphyp@linux.ibm.com> Reviewed-By: Paul E Murphy <murphyp@linux.ibm.com>
ROUND_TO_ODD and a couple of other places use libc_feupdateenv_test to
restore the rounding mode and exception enables, preserve exception flags,
and test whether given exception(s) were generated.
If the exception flags haven't changed, then it is sufficient and a bit
more efficient to just restore the rounding mode and enables, rather than
writing the full Floating-Point Status and Control Register (FPSCR).
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
Lukasz Majewski [Fri, 22 Mar 2019 10:53:45 +0000 (11:53 +0100)]
y2038: Introduce struct __timespec64 - new internal glibc type
This type is a glibc's "internal" type similar to struct timespec but
whose tv_sec field is a __time64_t rather than a time_t, which makes it
Y2038-proof and usable to pass syscalls between user code and Y2038-proof
kernel.
To support passing this structure to the kernel - the unnamed 32 bit
padding bit-field has been introduced. The placement of it depends on
the endianness of the SoC.
The total_deadline variable inside the clntudp_call() function inside
sunrpc/clnt_udp.c can cause uninitalised variable warnings when building
with GCC 8.3 or 9.2 on a platform with a 64-bit tv_nsec on a 32-bit
architecture. To fix the warning let's use the DIAG_* macros to hide the
warning.
A GCC bug case has also been submitted:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91691
2019-09-24 Alistair Francis <alistair.francis@wdc.com>
Script to generate ChangeLog-like output from git log
Co-authored-by: Gabriel F. T. Gomes <gabriel@inconstante.net.br> Reviewed-by: Gabriel F. T. Gomes <gabriel@inconstante.net.br> Reviewed-by: Joseph Myers <joseph@codesourcery.com>
The utility of a ChangeLog file has been discussed in various mailing
list threads and GNU Tools Cauldrons in the past years and the general
consensus is that while the file may have been very useful in the past
when revision control did not exist or was not as powerful as it is
today, it's current utility is fast diminishing. Further, the
ChangeLog format gets in the way of modernisation of processes since
it almost always results in rewriting of a commit, thus preventing use
of any code review tools to automatically manage patches in the glibc
project.
There is consensus in the glibc community that documentation of why a
change was done (i.e. a detailed description in a git commit) is more
useful than what changed (i.e. a ChangeLog entry) since the latter can
be deduced from the patch. The GNU community would however like to
keep the option of ascertaining what changed through a ChangeLog-like
output and as a compromise, it was proposed that a script be developed
that generates this output.
The script below is the result of these discussions. This script
takes two git revisions references as input and generates the git log
between those revisions in a form that resembles a ChangeLog. Its
capabilities and limitations are listed in a comment in the script.
On a high level it is capable of parsing C code and telling what
changed at the top level, but not within constructs such as functions.
Design
------
At a high level, the script analyses the raw output of a VCS, parses
the source files that have changed and attempts to determine what
changed. The script driver needs three distinct components to be
fully functional for a repository:
- A vcstocl_quirks.py file that helps it parse weird patterns in
sources that may result from preprocessor defines.
- A VCS plugin backend; the git backend is implemented for glibc
- A programming language parser plugin. C is currently implemented.
Additional programming language parsers can be added to give more
detailed output for changes in those types of files.
For input in languages other than those that have a parser, the script
only identifies if a file has been added, removed, modified,
permissions changed, etc. but cannot understand the change in content.
The C Parser
------------
The C parser is capable of parsing C programs with preprocessor macros
in place, as if they were part of the language. This presents some
challenges with parsing code that expands macros on the fly and to
help work around that, a vcstocl_quirks.py file has transformations to
ease things.
The C parser currently can identify macro definitions and scopes and
all global and static declarations and definitions. It cannot parse
(and compare) changes inside functions yet, it could be a future
enhancement if the need for it arises.
Testing
-------
The script has been tested with the glibc repository up to glibc-2.29
and also in the past with emacs. While it would be ideal to have
something like this in a repository like gnulib, that should not be a
bottleneck for glibc to start using this, so this patch proposes to
add these scripts into glibc.
And here is (hopefully!) one of the last ChangeLog entries we'd have
to write for glibc:
* scripts/gitlog_to_changelog.py: New script to auto-generate
ChangeLog.
* scripts/vcs_to_changelog/frontend_c.py: New file.
* scripts/vcs_to_changelog/misc_util.py: New file.
* scripts/vcs_to_changelog/vcs_git.py: New file.
* scripts/vcs_to_changelog/vcstocl_quirks.py: Likewise.
Paul A. Clarke [Thu, 19 Sep 2019 13:35:16 +0000 (08:35 -0500)]
[powerpc] SET_RESTORE_ROUND optimizations and bug fix
SET_RESTORE_ROUND brackets a block of code, temporarily setting and
restoring the rounding mode and letting everything else, including
exceptions generated within the block, pass through.
On powerpc, the current code clears the exception enables, which will hide
exceptions generated within the block. This issue was introduced by me
in commit e905212627350d54b58426214b5a54ddc852b0c9.
Fix this by not clearing exception enable bits in the prologue.
Also, since we are no longer changing the enable bits in either the
prologue or the epilogue, there is no need to test for entering/exiting
non-stop mode.
Also, optimize the prologue get/save/set rounding mode operations for
POWER9 and later by using 'mffscrn' when possible.
Suggested-by: Paul E. Murphy <murphyp@linux.ibm.com> Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com> Fixes: e905212627350d54b58426214b5a54ddc852b0c9
2019-09-19 Paul A. Clarke <pc@us.ibm.com>
* sysdeps/powerpc/fpu/fenv_libc.h (fegetenv_and_set_rn): New.
(__fe_mffscrn): New.
* sysdeps/powerpc/fpu/fenv_private.h (libc_feholdsetround_ppc_ctx):
Do not clear enable bits, remove obsolete code, use
fegetenv_and_set_rn.
(libc_feresetround_ppc): Remove obsolete code, use
fegetenv_and_set_rn.
Stefan Liebler [Thu, 19 Sep 2019 10:26:18 +0000 (12:26 +0200)]
Fix building support_ptrace.c on i686-gnu.
On i686-gnu the build is broken:
In file included from support_ptrace.c:22:
../include/sys/prctl.h:2:15: fatal error: sys/prctl.h: No such file or directory
#include_next <sys/prctl.h>
This patch just removes the unused prctl.h inclusion.
ChangeLog:
* support/support_ptrace.c: Remove inclusion of sys/prctl.h.
Joseph Myers [Wed, 18 Sep 2019 13:22:24 +0000 (13:22 +0000)]
Fix RISC-V vfork build with Linux 5.3 kernel headers.
Building glibc for RISC-V with Linux 5.3 kernel headers fails because
<linux/sched.h>, included in vfork.S for CLONE_* constants, contains a
structure definition not safe for inclusion in assembly code.
All other architectures already avoid use of that header in vfork.S,
either defining the CLONE_* constants locally or embedding the
required values directly in the relevant instruction, where they
implement vfork using the clone syscall (see the implementations for
aarch64, ia64, mips and nios2). This patch makes the RISC-V version
define the constants locally like the other architectures.
Tested build for all three RISC-V configurations in
build-many-glibcs.py with Linux 5.3 headers.
* sysdeps/unix/sysv/linux/riscv/vfork.S: Do not include
<linux/sched.h>.
(CLONE_VM): New macro.
(CLONE_VFORK): Likewise.
Stefan Liebler [Wed, 18 Sep 2019 10:40:00 +0000 (12:40 +0200)]
Add UNSUPPORTED check in elf/tst-pldd.
The testcase forks a child process and runs pldd with PID of
this child. On systems where /proc/sys/kernel/yama/ptrace_scope
differs from zero, pldd will fail with
/usr/bin/pldd: cannot attach to process 3: Operation not permitted
This patch checks if ptrace_scope exists, is zero "classic ptrace permissions"
or one "restricted ptrace". If ptrace_scope exists and has a higher
restriction, then the test is marked as UNSUPPORTED.
The case "restricted ptrace" is handled by rearranging the processes involved
during the test. Now we have the following process tree:
-parent: do_test (performs output checks)
--subprocess 1: pldd_process (becomes pldd via execve)
---subprocess 2: target_process (ptraced via pldd)
There is no need to sparc64 provide an arch-specific implementation to
route to POSIX one (which uses gettimeofday). Linux one already handles
the case for architecture that does not have __NR_time.
No semantic changes, checked against a build for sparc64-linux-gnu.
This patch simplifies the powerpc internal macros for vDSO calls
by:
- Removing INTERNAL_VSYSCALL_NO_SYSCALL_FALLBACK, used solely on
get_timebase_freq.
- Adjust INTERNAL_VSYSCALL_CALL_TYPE powerpc32 to follow powerpc64
argument ordering.
- Use HAVE_*_VSYSCALL instead of explicit strings.
- Make powerpc libc-vdso.h include generic implementation.
No semantic change expected, checked on powerpc-linux-gnu-power4,
powerpc64-linux-gnu, and powerpc64le-linux-gnu.
* sysdeps/unix/sysv/linux/libc-vdso.h (VDSO_IFUNC_RET): Define if not
defined.
* sysdeps/unix/sysv/linux/powerpc/get_timebase_freq.c
(__get_timebase_freq): Remove use of
INTERNAL_VSYSCALL_NO_SYSCALL_FALLBACK.
(get_timebase_freq_fallback): New symbol.
* sysdeps/unix/sysv/linux/powerpc/gettimeofday.c (time): Use
HAVE_GETTIMEOFDAY_VSYSCALL.
* sysdeps/unix/sysv/linux/powerpc/time.c (gettimeofday): Use
HAVE_TIME_VSYSCALL.
* sysdeps/unix/sysv/linux/powerpc/libc-vdso.h: Include generic
implementation.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/sysdep.h
(INTERNAL_VSYSCALL_CALL_TYPE): Make calling convention similar to
powerpc64.
(INTERNAL_VSYSCALL_NO_SYSCALL_FALLBACK): Remove macro.
* .../sysv/linux/powerpc/powerpc64/sysdep.h
(INTERNAL_VSYSCALL_NO_SYSCALL_FALLBACK): Likewise.
* sysdeps/unix/sysv/linux/powerpc/sysdep.h
(HAVE_GETTIMEOFDAY_VSYSCALL): Define.
Linux vDSO initialization code the internal function pointers require a
lot of duplicated boilerplate over different architectures. This patch
aims to simplify not only the code but the required definition to enable
a vDSO symbol.
The changes are:
1. Consolidate all init-first.c on only one implementation and enable
the symbol based on HAVE_*_VSYSCALL existence.
2. Set the HAVE_*_VSYSCALL to the architecture expected names string.
3. Add a new internal implementation, get_vdso_mangle_symbol, which
returns a mangled function pointer.
Currently the clock_gettime, clock_getres, gettimeofday, getcpu, and time
are handled in an arch-independent way, powerpc still uses some
arch-specific vDSO symbol handled in a specific init-first implementation.
Checked on aarch64-linux-gnu, arm-linux-gnueabihf, i386-linux-gnu,
mips64-linux-gnu, powerpc64le-linux-gnu, s390x-linux-gnu,
sparc64-linux-gnu, and x86_64-linux-gnu.
This patch removes the PREPARE_VERSION and PREPARE_VERSION_KNOW macro
and uses a static inline function instead, get_vdso_symbol. Each
architecture that supports vDSO must define the Linux version and its
hash for symbol resolution (VDSO_NAME and VDSO_HASH macro respectively).
It also organizes the HAVE_*_VSYSCALL for mips, powerpc, and s390 to
define them on a common header.
The idea is to require less code to configure and enable vDSO support
for newer ports. No semantic changes are expected.
Checked with a build against all affected architectures.
Fix small error in HP_TIMING_PRINT trailing null char setting
Fix a small error in the HP_TIMING_PRINT trailing zero setting; the '\0'
should be set at MIN(Len,string length), instead of always at the 'Len'
position.
* sysdeps/generic/hp-timing-common.h (HP_TIMING_PRINT): Correct
position of string null termination.
alpha: force old OSF1 syscalls for getegid, geteuid and getppid [BZ #24986]
On alpha, Linux kernel 5.1 added the standard getegid, geteuid and
getppid syscalls (commit ecf7e0a4ad15287). Up to now alpha was using
the corresponding OSF1 syscalls through:
- sysdeps/unix/alpha/getegid.S
- sysdeps/unix/alpha/geteuid.S
- sysdeps/unix/alpha/getppid.S
When building against kernel headers >= 5.1, the glibc now use the new
syscalls through sysdeps/unix/sysv/linux/syscalls.list. When it is then
used with an older kernel, the corresponding 3 functions fail.
A quick fix is to move the OSF1 wrappers under the
sysdeps/unix/sysv/linux/alpha directory so they override the standard
linux ones. A better fix would be to try the new syscalls and fallback
to the old OSF1 in case the new ones fail. This can be implemented in
a later commit.
Changelog:
[BZ #24986]
* sysdeps/unix/alpha/getegid.S: Move to ...
* sysdeps/unix/sysv/linux/alpha/getegid.S: ... here.
* sysdeps/unix/alpha/geteuid.S: Move to ...
* sysdeps/unix/sysv/linux/alpha/geteuid.S: ... here.
* sysdeps/unix/alpha/getppid.S: Move to ...
* sysdeps/unix/sysv/linux/alpha/getppid.S: ... here
Paul Eggert [Mon, 9 Sep 2019 21:08:46 +0000 (14:08 -0700)]
Regenerate charmap-kw.h, locfile-kw.h
This propagates the recent http->https URL changes.
Since I used gperf 3.1 to regenerate, this is also a minor
internal-to-localedef API change.
URL problem reported by Joseph Myers in:
https://www.sourceware.org/ml/libc-alpha/2019-09/msg00143.html
* locale/programs/charmap-kw.h, locale/programs/locfile-kw.h:
Regenerate with gperf 3.1.
* locale/programs/linereader.h (kw_hash_fct_t):
* locale/programs/repertoire.c (repertoiremap_hash):
2nd arg is now size_t not unsigned, for compatibility with gperf 3.1.
Paul Eggert [Sat, 7 Sep 2019 05:40:42 +0000 (22:40 -0700)]
Prefer https to http for gnu.org and fsf.org URLs
Also, change sources.redhat.com to sourceware.org.
This patch was automatically generated by running the following shell
script, which uses GNU sed, and which avoids modifying files imported
from upstream:
and then by running 'make dist-prepare' to regenerate files built
from the altered files, and then executing the following to cleanup:
chmod a+x sysdeps/unix/sysv/linux/riscv/configure
# Omit irrelevant whitespace and comment-only changes,
# perhaps from a slightly-different Autoconf version.
git checkout -f \
sysdeps/csky/configure \
sysdeps/hppa/configure \
sysdeps/riscv/configure \
sysdeps/unix/sysv/linux/csky/configure
# Omit changes that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/powerpc/powerpc64/ppc-mcount.S: trailing lines
git checkout -f \
sysdeps/powerpc/powerpc64/ppc-mcount.S \
sysdeps/unix/sysv/linux/s390/s390-64/syscall.S
# Omit change that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S: last line does not end in newline
git checkout -f sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S
When the const attribute of totalorder* functions was replaced with the
pure attribute, by commit ID ab41100bab12, it did not use the
__attribute_pure__ macro, but the __attribute__ ((__pure__)) statement.
All other uses of the pure attribute in glibc use the macro.
Make totalorder and totalordermag functions take pointer arguments.
the test case math/test-totalorderl-ldbl-128ibm fails on every input
pair, when compiled with -O2, which is the case for glibc test suite.
Debugging showed that the test case is passing arguments incorrectly to
totalorderl. This can also be inferred by the fact that compiling the
test case with -O0 hides the bug.
The documentation for the const attribute in GCC manual reads:
Note that a function that has pointer arguments and examines the data
pointed to must not be declared const if the pointed-to data might
change between successive invocations of the function. In general,
since a function cannot distinguish data that might change from data
that cannot, const functions should never take pointer or, in C++,
reference arguments. Likewise, a function that calls a non-const
function usually must not be const itself.
Since the pointed-to data is likely to be changed by user code between
invocations of totalorder*, this patch removes the const attribute from
the declarations of all totalorder functions, replacing it with the pure
attribute, as suggested in the manual:
The pure attribute imposes similar but looser restrictions on a
function’s definition than the const attribute: pure allows the
function to read any non-volatile memory, even if it changes in
between successive invocations of the function.
Lukasz Majewski [Wed, 28 Aug 2019 18:27:03 +0000 (14:27 -0400)]
y2038: Introduce the __ASSUME_TIME64_SYSCALLS define
Add a macro to linux/kernel-features.h, __ASSUME_TIME64_SYSCALLS, to
indicate whether the kernel can be assumed to provide a set of system
calls that process 64-bit time_t.
__ASSUME_TIME64_SYSCALLS does not indicate whether time_t is actually
64 bits (that's __TIMEBITS) and also does not indicate whether the
64-bit time_t system calls have "time64" suffixes on their names.
Code that uses __ASSUME_TIME64_SYSCALLS will be added in subsequent
patches.
* sysdeps/unix/sysv/linux/kernel-features.h
(__ASSUME_TIME64_SYSCALLS): New macro.
Reviewed-by: Alistair Francis <alistair23@gmail.com> Reviewed-by: Joseph Myers <joseph@codesourcery.com> Reviewed-by: Zack Weinberg <zackw@panix.com>
Finish move of clock_* functions to libc. [BZ #24959]
In glibc 2.17, the functions clock_getcpuclockid, clock_getres,
clock_gettime, clock_nanosleep, and clock_settime were moved from
librt.so to libc.so, leaving compatibility stubs behind. Now that the
dynamic linker no longer insists on finding versioned symbols in the
same library that originally defined them, we do not need the stubs
anymore, and this means we don't need GLIBC_PRIVATE __-prefix aliases
for most of the functions anymore either. (clock_gettime still needs
one.) For ports added before 2.17, libc.so needs to provide two
symbol versions for each, the default at GLIBC_2.17 plus a compat
version matching what librt had.
While I'm at it, move the clock_*.c files and their tests from rt/ to
time/.
localedef: Use initializer for flexible array member [BZ #24950]
struct charseq used a zero-length array instead of a flexible array
member. This required a strange construct to initialize struct
charseq objects, and GCC 10 warns about that:
cc1: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
In file included from programs/repertoire.h:24,
from programs/localedef.h:32,
from programs/ld-ctype.c:35:
programs/charmap.h:63:17: note: destination object declared here
63 | unsigned char bytes[0];
| ^~~~~
cc1: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
programs/charmap.h:63:17: note: destination object declared here
cc1: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
programs/charmap.h:63:17: note: destination object declared here
cc1: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
programs/charmap.h:63:17: note: destination object declared here
The change makes the object physically const, but it is not expected
to be modified.
Ian Kent [Mon, 2 Sep 2019 11:26:14 +0000 (13:26 +0200)]
Use autofs "ignore" mount hint in getmntent_r/getmntent
Historically autofs mounts were not included in mount table
listings. This is the case in other SysV autofs implementations
and was also the case with Linux autofs.
But now that /etc/mtab is a symlink to the proc filesystem
mount table the autofs mount entries appear in the mount table
on Linux.
Prior to the symlinking of /etc/mtab mount table it was
sufficient to call mount(2) and simply not update /etc/mtab
to exclude autofs mounts from mount listings.
Also, with the symlinking of /etc/mtab we have seen a shift in
usage toward using the proc mount tables directly.
But the autofs mount entries need to be retained when coming
from the proc file system for applications that need them
(largely autofs file system users themselves) so filtering out
these entries within the kernel itself can't be done. So it
needs be done in user space.
There are three reasons to omit the autofs mount entries.
One is that certain types of auto-mounts have an autofs mount
for every entry in their autofs mount map and these maps can
be quite large. This leads to mount table listings containing
a lot of unnecessary entries.
Also, this change in behaviour between autofs implementations
can cause problems for applications that use getmntent(3) in
other OS implementations as well as Linux.
Lastly, there's very little that user space can do with autofs
mount entries since this must be left to the autofs mount owner,
typically the automount daemon. But it can also lead to attempts
to access automount managed paths resulting mounts being triggered
when they aren't needed or mounts staying mounted for much longer
thay they need be. While the point of this change ins't to help
with these problems (and it can be quite a problem) it may be
a welcome side effect.
So the Linux autofs file system has been modified to accept a
pseudo mount option of "ignore" (as is used in other OS
implementations) so that user space can use this as a hint to
skip autofs entries on reading the mount table.
The Linux autofs automount daemon used getmntent(3) itself and
has been modified to use the proc file system directly so that
it can "ignore" mount option.
The use of this mount option is opt-in and a configuration
option has been added which defaults to not use this option
so if there are applications that need these entries, other
than autofs itself, they can be retained. Also, since this
filtering is based on an added mount option earlier versions
of Linux autofs iand other autofs file system users will not
use the option and so won't be affected by the change.
Samuel Thibault [Thu, 29 Aug 2019 23:48:38 +0000 (01:48 +0200)]
hurd: Fix SS_ONSTACK support
* sysdeps/mach/hurd/i386/sigreturn.c (__sigreturn2): New function,
unlocks SS and returns to the saved PC.
(__sigreturn): Do not unlock SS, and "return" into __sigreturn2 on the
thread stack instead of the saved PC.
Samuel Thibault [Thu, 29 Aug 2019 23:20:23 +0000 (01:20 +0200)]
hurd: Fix poll and select POSIX compliancy details about errors
This fixes the following:
- On error, poll must not return without polling, including EBADF, and instead
report POLLHUP/POLLERR/POLLNVAL
- Select must report EBADF if some set contains an invalid FD.
The idea is to move error management to after all select calls, in the
poll/select final treatment. The error is instead recorded in a new `error'
field, and a new SELECT_ERROR bit set.
Thanks Svante Signell for the initial version of the patch.
* hurd/hurdselect.c (SELECT_ERROR): New macro.
(_hurd_select):
- Add `error' field to `d' structures array.
- If a poll descriptor is bogus, set EBADF, but continue with a zero
timeout.
- Go through the whole fd_set, not only until _hurd_dtablesize. Return
EBADF there is any bit set above _hurd_dtablesize.
- Do not request io_select on bogus descriptors (SELECT_ERROR).
- On io_select request error, record the error.
- On io_select bogus reply, use EIO error code.
- On io_select bogus or error reply, record the error.
- Do not destroy reply port for bogus FDs.
- On error, make poll set POLLHUP in the EPIPE case, POLLNVAL in the
EBADF case, or else POLLERR.
- On error, make select simulated readiness.
Richard Braun [Thu, 29 Aug 2019 23:16:37 +0000 (01:16 +0200)]
hurd: Fix timeout handling in _hurd_select
Rely on servers to implement timeouts, so that very short values (including
0) don't make mach_msg return before valid replies can be received. The
purpose of this scheme is to guarantee a full client-server round-trip,
whatever the timeout value.
This change depends on the new io_select_timeout RPC being implemented by
servers.
* hurd/Makefile (user-interfaces): Add io_reply and io_request.
* hurd/hurdselect.c: Include <sys/time.h>, <hurd/io_request.h> and
<limits.h>.
(_hurd_select): Replace the call to __io_select with either
__io_select_request or __io_select_timeout_request, depending on the
timeout. Count the number of ready descriptors (replies for which at
least one type bit is set). Implement the timeout locally when there is
no file descriptor.
Samuel Thibault [Thu, 29 Aug 2019 23:08:09 +0000 (01:08 +0200)]
hurd getcwd: Allow unknown root directory
To be efficient, the remap translator simply returns ports from the underlying
filesystem, and thus the root directory found through browsing '..' is the
underlying root, not the remap root. This should not be a reason for getcwd to
fail.
* sysdeps/mach/hurd/getcwd.c (_hurd_canonicalize_directory_name_internal): Do
not remove the heading slash if we got an unknown root directory.
(__getcwd): Do not fail with EGRATUITOUS if we got an unknown root directory.
Samuel Thibault [Thu, 29 Aug 2019 23:04:17 +0000 (01:04 +0200)]
hurd: Fix implementation of setitimer.
The preemptor sigcode doesn't match since the POSIX sigcode SI_TIMER is
used when SIGALRM is sent. In addition, The inline version of
hurd_preempt_signals doesn't update _hurdsig_preempted_set. For these
reasons, the preemptor would be skipped by post_signal.
Richard Braun [Thu, 29 Aug 2019 22:58:14 +0000 (00:58 +0200)]
hurd: Fix _hurd_select for single fd sets
The function attempts to optimize this case by performing one IPC system
call with the timeout included among the parameters, but in the absence
of a reply, it will call mach_msg again with the same timeout later,
effectively doubling the total timeout of the select/poll call.
Remove this optimization for the time being.
* hurd/hurdselect.c (_hurd_select): Always call __io_select with no
timeout.
This patch is a reimplementation of [1], which was submitted back in
2015. Copyright issue has been sorted [2] last year. It proposed a new
section (.gnu.xhash) and related dynamic tag (GT_GNU_XHASH). The new
section would be virtually identical to the existing .gnu.hash except
for the translation table (xlat) which would contain correct MIPS
.dynsym indexes corresponding to the hashvals in chains. This is because
MIPS ABI imposes a different ordering of the dynsyms than the one
expected by the .gnu.hash section. Another addition would be a leading
word at the beggining of the section, which would contain the number of
entries in the translation table.
In this patch, the new section name and dynamic tag are changed to
reflect the fact that the section should be treated as MIPS specific
(.MIPS.xhash and DT_MIPS_XHASH).
This patch addresses the alignment issue reported in [3] which is caused
by the leading word of the .MIPS.xhash section. Leading word is now
removed in the corresponding binutils patch, and the number of entries
in the translation table is computed using DT_MIPS_SYMTABNO dynamic tag.
Since the MIPS specific dl-lookup.c file was removed following the
initial patch submission, I opted for the definition of three new macros
in the generic ldsodefs.h. ELF_MACHINE_GNU_HASH_ADDRIDX defines the
index of the dynamic tag in the l_info array. ELF_MACHINE_HASH_SYMIDX is
used to calculate the index of a symbol in GNU hash. On MIPS, it is
defined to look up the symbol index in the translation table.
ELF_MACHINE_XHASH_SETUP is defined for MIPS only. It initializes the
.MIPS.xhash pointer in the link_map_machine struct.
The other major change is bumping the highest EI_ABIVERSION value for
MIPS to suggest that the dynamic linker now supports GNU hash.
The patch was tested by running the glibc testsuite for the three MIPS
ABIs (o32, n32 and n64) and for x86_64-linux-gnu.
* elf/dl-addr.c (determine_info): Calculate the symbol index
using the newly defined ELF_MACHINE_HASH_SYMIDX macro.
* elf/dl-lookup.c (do_lookup_x): Ditto.
(_dl_setup_hash): Initialize MIPS xhash translation table.
* elf/elf.h (SHT_MIPS_XHASH): New define.
(DT_MIPS_XHASH): New define.
* sysdeps/generic/ldsodefs.h (ELF_MACHINE_GNU_HASH_ADDRIDX): New
define.
(ELF_MACHINE_HASH_SYMIDX): Ditto.
(ELF_MACHINE_XHASH_SETUP): Ditto.
* sysdeps/mips/ldsodefs.h (ELF_MACHINE_GNU_HASH_ADDRIDX): New
define.
(ELF_MACHINE_HASH_SYMIDX): Ditto.
(ELF_MACHINE_XHASH_SETUP): Ditto.
* sysdeps/mips/linkmap.h (struct link_map_machine): New member.
* sysdeps/unix/sysv/linux/mips/ldsodefs.h: Increment valid ABI
version.
* sysdeps/unix/sysv/linux/mips/libc-abis: New ABI version.
The fix for BZ#18231 requires new symbols only for sh4eb. This patch
adds the required folder and files for both BE and LE abilist. No
semantic changes are expected.
Checked with check-abi for sh4eb-linux-gnu and sh4-linux-gnu.
* sysdeps/sh/preconfigure.ac: New file.
* sysdeps/sh/preconfigure: Regenerate.
* sysdeps/sh/be/sh3/Implies: New file.
* sysdeps/sh/be/sh4/Implies: Likewise.
* sysdeps/sh/le/sh3/Implies: Likewise.
* sysdeps/sh/le/sh4/Implies: Likewise.
* sysdeps/unix/sysv/linux/sh/le/sh3/Implies: Likewise.
* sysdeps/unix/sysv/linux/sh/le/sh4/Implies: Likewise.
* sysdeps/unix/sysv/linux/sh/*.abilist: Move to
sysdeps/unix/sysv/linux/sh/le/*.abilist.
* sysdeps/unix/sysv/linux/sh/be/*.abilist: New files.
The fix for BZ#18231 requires new symbols only for microblaze. This patch
adds the required folder and files for both BE and LE abilist. No semantic
changes are expected.
Checked with check-abi for microblaze-linux-gnueabihf and
microblazeel-linux-gnueabihf.
* sysdeps/microblaze/preconfigure.ac: New file.
* sysdeps/microblaze/preconfigure: Regenerate.
* sysdeps/microblaze/be/implies: New file.
* sysdeps/microblaze/le/implies: Likewise.
* sysdeps/unix/sysv/linux/microblaze/be/implies: Likewise.
* sysdeps/unix/sysv/linux/microblaze/le/implies: Likewise.
* sysdeps/unix/sysv/linux/microblaze/*.abilist. Move to
sysdeps/unix/sysv/linux/microblaze/be/*.abilist.
* sysdeps/unix/sysv/linux/microblaze/le/*.abilist: New files.
The fix for BZ#18231 requires new symbols only for armeb. This patch
adds the required folder and files for both BE and LE abilist. No
semantic changes are expected.
Checked with check-abi for arm-linux-gnueabihf and armeb-linux-gnueabihf.
* sysdeps/arm/preconfigure.ac: Set machine based on endianness.
* sysdeps/arm/preconfigure: Regenerate.
* sysdeps/arm/be/Implies: New file.
* sysdeps/arm/be/armv6/Implies: Likewise.
* sysdeps/arm/be/armv6t2/Implies: Likewise.
* sysdeps/arm/be/armv7/Implies: Likewise.
* sysdeps/arm/le/Implies: Likewise.
* sysdeps/unix/sysv/linux/arm/be/Implies: Likewise.
* sysdeps/unix/sysv/linux/arm/le/Implies: Likewise.
* sysdeps/unix/sysv/linux/arm/*.abilist: Move to
sysdeps/unix/sysv/linux/arm/le/*.abilist.
* sysdeps/unix/sysv/linux/arm/be/l*.abilist: New files.
Paul Eggert [Thu, 29 Aug 2019 09:34:13 +0000 (02:34 -0700)]
Fix posix/tst-regex by using UTF-8 and own test input
Problem reported by Stefan Liebler in:
https://sourceware.org/ml/libc-alpha/2019-08/msg00658.html
* posix/tst-regex.c: Convert this file from Latin-1 to UTF-8.
(do_test, test_expr): Adjust to the fact that this source file,
and the test data in ChangeLog.8, is now UTF-8 instead of Latin-1.
* posix/tst-regex.input: Copy from ChangeLog.old/ChangeLog.8,
so that it is now UTF-8.
fegetenv_status() wants to use the lighter weight instruction 'mffsl'
for reading the Floating-Point Status and Control Register (FPSCR).
It currently will use it directly if compiled '-mcpu=power9', and will
perform a runtime check (cpu_supports("arch_3_00")) otherwise.
Nicely, it turns out that the 'mffsl' instruction will decode to
'mffs' on architectures older than "arch_3_00" because the additional
bits set for 'mffsl' are "don't care" for 'mffs'. 'mffs' is a superset
of 'mffsl'.
Paul A. Clarke [Tue, 6 Aug 2019 04:13:45 +0000 (00:13 -0400)]
[powerpc] fesetenv: optimize FPSCR access
fesetenv() reads the current value of the Floating-Point Status and Control
Register (FPSCR) to determine the difference between the current state of
exception enables and the newly requested state. All of these bits are also
returned by the lighter weight 'mffsl' instruction used by fegetenv_status().
Use that instead.
Also, remove a local macro _FPU_MASK_ALL in favor of a common macro,
FPU_ENABLES_MASK from fenv_libc.h.
Finally, use a local variable ('new') in favor of a pointer dereference
('*envp').
Paul A. Clarke [Sat, 3 Aug 2019 02:47:57 +0000 (22:47 -0400)]
[powerpc] SET_RESTORE_ROUND improvements
SET_RESTORE_ROUND uses libc_feholdsetround_ppc_ctx and
libc_feresetround_ppc_ctx to bracket a block of code where the floating point
rounding mode must be set to a certain value.
For the *prologue*, libc_feholdsetround_ppc_ctx is used and performs:
1. Read/save FPSCR.
2. Create new value for FPSCR with new rounding mode and enables cleared.
3. If new value is different than current value,
a. If transitioning from a state where some exceptions enabled,
enter "ignore exceptions / non-stop" mode.
b. Write new value to FPSCR.
c. Put a mark on the wall indicating the FPSCR was changed.
(1) uses the 'mffs' instruction. On POWER9, the lighter weight 'mffsl'
instruction can be used, but it doesn't return all of the bits in the FPSCR.
fegetenv_status uses 'mffsl' on POWER9, 'mffs' otherwise, and can thus be
used instead of fegetenv_register.
(3b) uses 'mtfsf 0b11111111' to write the entire FPSCR, so it must
instead use 'mtfsf 0b00000011' to write just the enables and the mode,
because some of the rest of the bits are not valid if 'mffsl' was used.
fesetenv_mode uses 'mtfsf 0b00000011' on POWER9, 'mtfsf 0b11111111'
otherwise.
For the *epilogue*, libc_feresetround_ppc_ctx checks the mark on the wall, then
calls libc_feresetround_ppc, which just calls __libc_femergeenv_ppc with
parameters such that it performs:
1. Retreive saved value of FPSCR, saved in prologue above.
2. Read FPSCR.
3. Create new value of FPSCR where:
- Summary bits and exception indicators = current OR saved.
- Rounding mode and enables = saved.
- Status bits = current.
4. If transitioning from some exceptions enabled to none,
enter "ignore exceptions / non-stop" mode.
5. If transitioning from no exceptions enabled to some,
enter "catch exceptions" mode.
6. Write new value to FPSCR.
The summary bits are hardwired to the exception indicators, so there is no
need to restore any saved summary bits.
The exception indicator bits, which are sticky and remain set unless
explicitly cleared, would only need to be restored if the code block
might explicitly clear any of them. This is certainly not expected.
So, the only bits that need to be restored are the enables and the mode.
If it is the case that only those bits are to be restored, there is no need to
read the FPSCR. Steps (2) and (3) are unnecessary, and step (6) only needs to
write the bits being restored.
We know we are transitioning out of "ignore exceptions" mode, so step (4) is
unnecessary, and in step (6), we only need to check the state we are
entering.
Since fe{en,dis}ableexcept() and fesetmode() read-modify-write just the
"mode" (exception enable and rounding mode) bits of the Floating Point Status
Control Register (FPSCR), the lighter weight 'mffsl' instruction can be used
to read the FPSCR (enables and rounding mode), and 'mtfsf 0b00000011' can be
used to write just those bits back to the FPSCR. The net is better performance.
In addition, fe{en,dis}ableexcept() read the FPSCR again after writing it, or
they determine that it doesn't need to be written because it is not changing.
In either case, the local variable holds the current values of the enable
bits in the FPSCR. This local variable can be used instead of again reading
the FPSCR.
Also, that value of the FPSCR which is read the second time is validated
against the requested enables. Since the write can't fail, this validation
step is unnecessary, and can be removed. Instead, the exceptions to be
enabled (or disabled) are transformed into available bits in the FPSCR,
then validated after being transformed back, to ensure that all requested
bits are actually being set. For example, FE_INVALID_SQRT can be
requested, but cannot actually be set. This bit is not mapped during the
transformations, so a test for that bit being set before and after
transformations will show the bit would not be set, and the function will
return -1 for failure.
Finally, convert the local macros in fesetmode.c to more generally useful
macros in fenv_libc.h.
Paul A. Clarke [Sat, 13 Jul 2019 01:13:58 +0000 (20:13 -0500)]
[powerpc] fe{en,dis}ableexcept optimize bit translations
The exceptions passed to fe{en,dis}ableexcept() are defined in the ABI
as a bitmask, a combination of FE_INVALID, FE_OVERFLOW, etc.
Within the functions, these bits must be translated to/from the corresponding
enable bits in the Floating Point Status Control Register (FPSCR).
This translation is currently done bit-by-bit. The compiler generates
a series of conditional bit operations. Nicely, the "FE" exception
bits are all a uniform offset from the FPSCR enable bits, so the bit-by-bit
operation can instead be performed by a shift with appropriate masking.
Florian Weimer [Wed, 28 Aug 2019 09:59:45 +0000 (11:59 +0200)]
login: pututxline could fail to overwrite existing entries [BZ #24902]
The internal_getut_r function updates the file_offset variable and
therefore must always update last_entry as well.
Previously, if pututxline could not upgrade the read lock to a
write lock, internal_getut_r would update file_offset only,
without updating last_entry, and a subsequent call would not
overwrite the existing utmpx entry at file_offset, instead
creating a new entry. This has been observed to cause unbounded
file growth in high-load situations.
This commit removes the buffer argument to internal_getut_r and
updates the last_entry variable directly, along with file_offset.
Initially reported and fixed by Ondřej Lysoněk.
Reviewed-by: Gabriel F. T. Gomes <gabrielftg@linux.ibm.com>
Stefan Liebler [Wed, 28 Aug 2019 08:29:24 +0000 (10:29 +0200)]
Fix posix/tst-regex by using a dedicated input-file.
The recent commit e6855a3bdfe147c52b29b5e7d70a95a8aa22ece0
changed the encoding of ChangeLog.old/ChangeLog.8 from ISO-8859 to UTF-8.
Unfortunately the test posix/tst-regex assumes the former encoding.
Furthermore Francesco Potortì is now written with 'ì' instead of 'i`'
which would lead to two further matches in the first call to test_expr.
This patch just copies the former ChangeLog.8 file to tst-regex.input
and adjusts the test in order to use this new input file.
ChangeLog:
* posix/tst-regex.c (do_test): Use tst-regex.input as input file.
* posix/tst-regex.input: New file.
Paul Eggert [Fri, 23 Aug 2019 21:14:29 +0000 (14:14 -0700)]
Don't put non-ASCII into installed headers
Move non-ASCII contributor names from installed headers
into contrib.texi when possible, and when it's not (the
copyright notice in sysdeps/unix/sysv/linux/mips/sys/user.h)
go back to ASCIIfied names. Problem reported by Joseph Myers in:
https://www.sourceware.org/ml/libc-alpha/2019-08/msg00646.html
[MIPS] Raise highest supported EI_ABIVERSION value [BZ #24916]
This bumps the highest valid EI_ABIVERSION value to ABSOLUTE ABI.
New testcase loads the symbol from the GOT with the "lb" instruction
so that the EI_ABIVERSION header field of the shared object is set
to ABSOLUTE (it doesn't actually check the value of the symbol), and
makes sure that the main executable is executed without "ABI version
invalid" error.
Tested for all three ABIs (o32, n32, n64) using both static linker which
handles undefined weak symbols correctly [1] (and sets the EI_ABIVERSION
of the test module) and the one that doesn't (EI_ABIVERSION left as 0).
mips: Force RWX stack for hard-float builds that can run on pre-4.8 kernels
Linux/Mips kernels prior to 4.8 could potentially crash the user
process when doing FPU emulation while running on non-executable
user stack.
Currently, gcc doesn't emit .note.GNU-stack for mips, but that will
change in the future. To ensure that glibc can be used with such
future gcc, without silently resulting in binaries that might crash
in runtime, this patch forces RWX stack for all built objects if
configured to run against minimum kernel version less than 4.8.
* sysdeps/unix/sysv/linux/mips/Makefile
(test-xfail-check-execstack):
Move under mips-has-gnustack != yes.
(CFLAGS-.o*, ASFLAGS-.o*): New rules.
Apply -Wa,-execstack if mips-force-execstack == yes.
* sysdeps/unix/sysv/linux/mips/configure: Regenerated.
* sysdeps/unix/sysv/linux/mips/configure.ac
(mips-force-execstack): New var.
Set to yes for hard-float builds with minimum_kernel < 4.8.0
or minimum_kernel not set at all.
(mips-has-gnustack): New var.
Use value of libc_cv_as_noexecstack
if mips-force-execstack != yes, otherwise set to no.
linux: Make profil_counter a compat_symbol (BZ#17726)
As indicated by Joseph's comment on BZ#17726, this symbol is most
likely a historical ABI accident. This patch make it on both arm
and sparc ABIs a compat_symbol.
Checked against a build arm-linux-gnueabihf, sparcv9-linux-gnu, adn
sparc64-linux-gnu to see if the symbol is still present.
* gmon/Versions (libc) [GLIBC_2.31]: New entry.
* sysdeps/unix/sysv/linux/arm/profil-counter.h (profil_counter):
Make a compat_symbol.
* sysdeps/unix/sysv/linux/sparc/profil-counter.h
(__profil_counter_global): Likewise.
This patch refactor sigcontextinfo.h header to use SA_SIGINFO as default
for both gmon and debug implementations. This allows simplify
profil-counter.h on Linux to use a single implementation and remove the
requirements for newer ports to redefine __sigaction/sigaction to use
SA_SIGINFO.
The GET_PC macro is also replaced with a function sigcontext_get_pc that
returns an uintptr_t instead of a void pointer. It allows easier convertion
to integer on ILP32 architecture, such as x32, without the need to suppress
compiler warnings.
The patch also requires some refactor of register-dump.h file for some
architectures (to reflect it is now called from a sa_sigaction instead of
sa_handler signal context).
- Alpha, i386, and s390 are straighfoward to take in consideration the
new argument type.
- ia64 takes in consideration the kernel pass a struct sigcontextt
as third argument for sa_sigaction.
- sparc take in consideration the kernel pass a pt_regs struct
as third argument for sa_sigaction.
- m68k dummy function is removed and the FP state is dumped on
register_dump itself.
- For SH the register-dump.h file is consolidate on a common implementation
and the floating-point state is checked based on ownedfp field.
The register_dump does not change its output format in any affected
architecture.
I checked on x86_64-linux-gnu, i686-linux-gnu, aarch64-linux-gnu,
arm-linux-gnueabihf, sparcv9-linux-gnu, sparc64-linux-gnu, powerpc-linux-gnu,
powerpc64-linux-gnu, and powerpc64le-linux-gnu.
I also checked the libSegFault.so through catchsegv on alpha-linux-gnu,
m68k-linux-gnu and sh4-linux-gnu to confirm the output has not changed.
It was previously stated
<https://sourceware.org/ml/libc-alpha/2018-05/msg00771.html> that "It
should fail to indicate you have bugs in your system libidn.".
However, the glibc testsuite should be indicating whether there are
bugs in glibc, not whether there are bugs in other system pieces - so
unless you consider it a glibc bug that it fails to work around the
libidn issues, these FAILs are not helpful. And as a general
principle, it's best for the expected glibc test results to be clean,
with Bugzilla used to track known bugs in glibc itself, rather than
people needing to know about the expected FAILs to tell if there are
problems with their glibc build. So, while there is an argument that
install.texi (not just the old NEWS entries for 2.28) should explain
the use of libidn2 and that 2.0.5 or later is recommended, test FAILs
are not the right way to indicate the presence of an old libidn2
version.
This patch accordingly makes those tests return UNSUPPORTED for older
libidn2 versions, just as they do when libidn2 isn't present at all.
As implied by that past discussion, it's possible this could result in
UNSUPPORTED for systems with older versions but whatever required
fixes backported so the tests previously passed, if there are any such
systems.
Tested for x86_64 on Ubuntu 18.04, including verifying that putting an
earlier version in place of 2.0.5 results in the tests FAILing whereas
using 2.0.5 as in the patch results in UNSUPPORTED. Florian reports
that the tests still run on Fedora 30, with libidn 2.2.0.
* resolv/tst-resolv-ai_idn-latin1.c (do_test): Mark test
unsupported with libidn2 before 2.0.5.
* resolv/tst-resolv-ai_idn.c (do_test): Likewise.
Joseph Myers [Thu, 22 Aug 2019 12:38:40 +0000 (12:38 +0000)]
Document strftime %Ob and %OB as C2X features.
C2X standardizes strftime %Ob and %OB support. This patch updates the
glibc manual to say these are C2X features, while noting that the
details of what is the alternative form of a month name are not
specified in C2X.
Note: C2X seems unclear to me about whether %h being equivalent to %b
means %Oh is thereby allowed and equivalent to %Ob; I've asked WG14.
Tested with "make info" and "make pdf".
* manual/time.texi (strftime): Document %Ob and %OB as C2X
features.
Paul Eggert [Wed, 21 Aug 2019 16:25:22 +0000 (09:25 -0700)]
Fix bad pointer / leak in regex code
This was found by Coverity (CID 1484201). [BZ#24844]
* posix/regex_internal.c (create_cd_newstate): Fix use of bad
pointer and/or memory leak when storage is exhausted.
Zack Weinberg [Sat, 17 Aug 2019 01:10:11 +0000 (21:10 -0400)]
Don't use the argument to time.
It doesn't make sense to remove all the internal uses of time.
It's still a standard ISO C function, and its callers don't need
sub-second resolution and would be unnecessarily complicated if
they had to declare a struct timespec instead of just a time_t.
However, a handful of places were using the vestigial "result"
argument instead of the return value, which is slightly less
efficient and also looks strange. Correct this.
* misc/syslog.c (__vsyslog_internal)
* time/getdate.c (__getdate_r)
* time/tst_wcsftime.c (main):
Use return value of time, not its argument.
* string/strfry.c (strfry)
* sysdeps/mach/sleep.c (__sleep):
Remove unnecessary casts of NULL in calls to time.
Joseph Myers [Wed, 21 Aug 2019 12:06:44 +0000 (12:06 +0000)]
Add tgmath.h macros for narrowing functions.
When adding some of the TS 18661 narrowing functions for glibc 2.28, I
deferred adding corresponding <tgmath.h> support because of unresolved
questions about the specification for those type-generic macros,
especially in relation to _FloatN and _FloatNx types.
Those issues are now clarified in the response to Clarification
Request 13 to TS 18661-3, and this patch adds the deferred tgmath.h
support. As with other tgmath.h macros, there are fairly
straightforward implementations based on __builtin_tgmath for GCC 8
and later, which result in exactly the right function being called in
each case, and more complicated implementations for GCC 7 and earlier,
which generally result in a function being called whose arguments have
the right format (i.e. an alias for the right function), but which
might not be exactly the function name specified by TS 18661.
In one case with older compilers (f32x* macros, where the type
_Float64x exists and all the arguments have type _Float32 or
_Float32x), there is a further relaxation and the function called may
have arguments narrower than the one specified by the TS, but still
wide enough to represent the arguments exactly, so the result of the
call is unchanged (as this does not affect any case where rounding of
integer arguments might be involved). With GCC 6 or before this is
inherently unavoidable (but still harmless and not detectable by how
the compiled program behaves, unless it redefines the functions in
question like the testcases do) because _Float32x and _Float64 are
both typedefs for double in that case but the specified semantics
result in different functions, with different argument formats, being
called for those two argument types.
Tests for the new macros are handled through gen-tgmath-tests.py,
which deals with the special-case handling for older GCC.
Tested as follows: with the full glibc testsuite on x86_64 and x86
(with GCC 6, 7 and 8); with the math/ tests on aarch64 and arm (with
GCC 6, 7 and 8); with build-many-glibcs.py (with GCC 6, 7 and 9).
Carlos O'Donell [Thu, 21 Mar 2019 02:11:32 +0000 (22:11 -0400)]
nscd: Improve nscd.conf comments.
This change adds a warning to nscd.conf about running multiple caching
services together and that it may lead to unexpected behaviours. Also we
add a note that enabling the 'shared' option will cause cache hit rates
to be misreported (a side effect of the implementation).
v2
- Rewrite comment to avoid implementation details.
Carlos O'Donell [Wed, 20 Mar 2019 16:40:18 +0000 (12:40 -0400)]
nss: Make nsswitch.conf more distribution friendly.
The current default nsswitch.conf file provided by glibc is not very
distribution friendly. The file contains some minimal directives that no
real distribution uses. This update aims to provide a rich set of
comments which are useful for all distributions, and a broader set of
service defines which should work for all distributions.
Tested defaults on x86_64 and they work. The nsswitch.conf file more
closely matches what we have in Fedora now, and I'll adjust Fedora to
use this version with minor changes to enable Fedora-specific service
providers.
v2
- Add missing databases to manual.
- Add link to manual from default nsswitch.conf.
- Sort nsswitch.conf according to most used database first.
v3
- Only mention implemented services in 'NSS Basics.'
- Mention 'automount' in 'Services in the NSS configuration.'
- Sort services in alphabetical order.
v4
- Project name is 'Samba'.
v5
- Fix typo in manual/nss.texi.
v6
- Fix another typo in manual/nss.texi. Ran spell checker this time.
Florian Weimer [Mon, 19 Aug 2019 13:41:29 +0000 (15:41 +0200)]
Do not print backtraces on fatal glibc errors
If the process is in a bad state, we used to print backtraces in
many cases. This is problematic because doing so could involve
a lot of work, like loading libgcc_s using the dynamic linker,
and this could itself be targeted by exploit writers. For example,
if the crashing process was forked from a long-lived process, the
addresses in the error message could be used to bypass ASLR.
Commit ed421fca42fd9b4cab7c66e77894b8dd7ca57ed0 ("Avoid backtrace from
__stack_chk_fail [BZ #12189]"), backtraces where no longer printed
because backtrace_and_maps was always called with do_abort == 1.
Rather than fixing this logic error, this change removes the backtrace
functionality from the sources. With the prevalence of external crash
handlers, it does not appear to be particularly useful. The crash
handler may also destroy useful information for debugging.
Joseph Myers [Fri, 16 Aug 2019 19:23:30 +0000 (19:23 +0000)]
Correct TEST_COMPAT conditions in totalorder compat tests.
I noticed that the totalorder compat tests mixed TEST_COMPAT
conditions using COMPAT_VER (computed as the first symbol version for
the symbol being tested) and those with a hardcoded GLIBC_2_25.
COMPAT_VER is logically correct here, so this patch changes the tests
to use it. GLIBC_2_25 is harmless at present (even for _FloatN /
_FloatNx functions added in later glibc versions). However, if in
future we support _Float16 functions in glibc for any existing
configuration, the test using GLIBC_2_25 would get things wrong
(wrongly think that there are compat versions of totalorderf16 and
totalordermagf16 to test).
(The actual definitions of the compat / versioned symbols for the
totalorder functions make no attempt to deal with the possibility of
libm function support for a new floating-point format, already
supported for other architectures in glibc, being added in some glibc
configuration in future, for which the only vaguely plausible case
would be if some architecture gets _Float128 support it previously
lacked; this is much like functions added after glibc 2.4 not
generally attempting to deal with compat support for long double
changing away from having the same format as double.)