This patch introduces a test to make sure no function modifies the
xmm/ymm registers. With the exception of the auditing functions.
The test is probably too pessimistic. All code linked into ld.so
is checked. Perhaps at some point the callgraph starting from
_dl_fixup and _dl_profile_fixup is checked and we can start using
faster SSE-using functions in parts of ld.so.
There will be more than one function which, in multiarch mode, wants
to use SSSE3. We should not test in each of them for Atoms with
slow SSSE3. Instead, disable the SSSE3 bit in the startup code for
such machines.
The original AVX patch used a function pointer to handle the difference
between machines with and without AVX support. This is insecure. A
well-placed memory exploit could lead to redirection of the execution.
Using a variable and several tests is a bit slower but cannot be
exploited in this way.
Some symbols have to be identified process-wide by their name. This is
particularly important for some C++ features (e.g., class local static data
and static variables in inline functions). This cannot completely be
implemented with ELF functionality so far. The STB_GNU_UNIQUE binding
helps by ensuring the dynamic linker will always use the same definition for
all symbols with the same name and this binding.
Some of the new multi-arch string functions for x86-64 were
not aligned to 16 byte boundarie,s possibly creating unnecessary
cache line misses and delays.
This patch adds SSSE3 strcpy/stpcpy. I got up to 4X speed up on Core 2
and Core i7. I disabled it on Atom since SSSE3 version is slower for
shorter (<64byte) data.
The terminal output etc is not visible in a core file. The new
libc-internal variable __abort_msg will point to a string with the
message which has been printed before the abort in case abort is
called from inside libc. BZ #10217
The dl-lookup.c changes are needed for prelink (support in prelink
checked into SVN, tested for both i?86 and x86-64), dl-irel.h just
something I discovered by code inspection.
The test to call the indirect function now includes a subtest to
checked whether the symbol is defined. When coming to that point
this is almost always the case. The test for STT_GNU_IFUNC on the
other hand rarely is true. Move it to the front means we don't have
to perform the second test unless really necessary.
SO far Intel and AMD use exactly the same bits meaning the same
things in CPUID index 1. Simplify the code. Should an architecture
come along which doesn't use the same semantics then it must use a
different index value than COMMON_CPUID_INDEX_1.
The latest stratcliff extension exposed a bug in the IA-64 memchr which
uses non-speculative loads to prefetch data. Change the code to use
speculative loads with appropriate fixup. Fixes BZ 10162.
Due to a pasto the fallocate64 interface, introduced in glibc 2.10,
isn't exported for 32-bit Linux platforms. It is too late for this
now so exported them for glibc 2.11.
If longjmp restores the stack frame to an address which is beyond
the stack frame at the time of the longjmp call it would install
an uninitialized stack frame. If compiled with _FORTIFY_SOURCE
defined, longjmp will now bail out in this situation.
2009-05-06 Ryan S. Arnold <rsa@us.ibm.com>
[BZ #10118]
* Makeconfig (+asflags): New variable based upon ASFLAG or
asflags-cpu.
(ASFLAGS): Add override to set ASFLAGS to +asflags.
* config.make.in (asflags-cpu): Add variable based upon
@libc_cv_cc_submachine@ to propagate -mcpu=CPU from --with-cpu=CPU to
the assembler.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power6/fpu/setcontext.S:
Remove unneeded file now that the assembler emits _ARCH_PWR6 and
recognizes power6 instruction set due to passing -mcpu=power6 from
--with-cpu=power6 when compiling .S files.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power6/fpu/swapcontext.S:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power6/fpu/setcontext.S:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power6/fpu/swapcontext.S:
Likewise.
from definition.
* sysdeps/x86_64/dl-machine.h (elf_machine_rela): Don't define
label if it is not used.
* elf/dl-profile.c (_dl_start_profile): Define real-type variant
of gmon_hist_hdr and gmon_hdr structures and use them.
* elf/dl-load.c (open_verify): Add temporary variable to avoid
warning.
* nscd/nscd_helper.c (get_mapping): Avoid casts to avoid warnings.
* sunrpc/clnt_raw.c (clntraw_private_s): Use union in definition
to avoid cast.
* inet/rexec.c (rexec_af): Make sa2 a union to avoid warnings.
* inet/rcmd.c (rcmd_af): Make from a union of the various needed types
to avoid warnings.
(iruserok_af): Use ss_family instead of casts.
* gmon/gmon.c (write_hist): Define real-type variant of
gmon_hist_hdr structure and use it.
(write_gmon): Likewise for gmon_hdr.
* sysdeps/unix/sysv/linux/readv.c: Avoid declaration of replacement
function if we are not going to define it.
* sysdeps/unix/sysv/linux/writev.c: Likewise.
* inet/inet6_option.c (optin_alloc): Add temporary variable to
avoid warning.
* libio/strfile.h (struct _IO_streambuf): Use correct type and
name of VTable element.
* libio/iovsprintf.c: Avoid casts to avoid warnings.
* libio/iovsscanf.c: Likewise.
* libio/vasprintf.c: Likewise.
* libio/vsnprintf.c: Likewise.
* stdio-common/isoc99_vsscanf.c: Likewise.
* stdlib/strfmon_l.c: Likewise.
* debug/vasprintf_chk.c: Likewise.
* debug/vsnprintf_chk.c: Likewise.
* debug/vsprintf_chk.c: Likewise.
* stdio-common/reg-modifier.c: Likewise.
2009-04-22 Andrew Stubbs <ams@codesourcery.com>
* sysdeps/sh/sh4/setjmp.S (__sigsetjmp): Adjust JB_SIZE when there
is no FPU.
are always at least 4 bytes in the returned line.
2009-04-15 Jakub Jelinek <jakub@redhat.com>
* sysdeps/unix/sysv/linux/getsysstats.c (__get_nprocs): Check
__libc_use_alloca (8192), if the stack is too small use 512 bytes
instead of 8K. Stop searching in /proc/stat after hitting first
line not starting with cpu.
(next_line): Truncate too long
lines at buffer size * 3/4 instead of pretending there were line
breaks inside of large lines.
* sysdeps/x86_64/add_n.S: New file.
* sysdeps/x86_64/addmul_1.S: New file.
* sysdeps/x86_64/lshift.S: New file.
* sysdeps/x86_64/mul_1.S: New file.
* sysdeps/x86_64/rshift.S: New file.
* sysdeps/x86_64/sub_n.S: New file.
* sysdeps/x86_64/submul_1.S: New file.
* misc/Versions: Export preadv, preadv64, pwritev, pwritev64 for
GLIBC_2.10.
* misc/sys/uio.h: Declare preadv, preadv64, pwritev, pwritev64.
* sysdeps/unix/sysv/linux/kernel-features.h: Add entries for preadv
and pwritev.
* misc/preadv.c: New file.
* misc/preadv64.c: New file.
* misc/pwritev.c: New file.
* misc/pwritev64.c: New file.
* sysdeps/posx/preadv.c: New file.
* sysdeps/posx/preadv64.c: New file.
* sysdeps/posx/pwritev.c: New file.
* sysdeps/posx/pwritev64.c: New file.
* sysdeps/unix/sysv/linux/preadv.c: New file.
* sysdeps/unix/sysv/linux/preadv64.c: New file.
* sysdeps/unix/sysv/linux/pwritev.c: New file.
* sysdeps/unix/sysv/linux/pwritev64.c: New file.
* sysdeps/unix/sysv/linux/wordsize-64/preadv64.c: New file.
* sysdeps/unix/sysv/linux/wordsize-64/pwritev64.c: New file.
__ASSUME_COMPLETE_READV_WRITEV.
* sysdeps/unix/sysv/linux/readv.c: No need for userlevel fallback
with modern kernels.
* sysdeps/unix/sysv/linux/writev.c: Likewise.
* sysdeps/posix/readv.c: Since read is a cancellation point we have
to free a possible malloced buffer in case of cancellation.
* sysdeps/posix/writev.c: Likewise for write.
c2009-04-01 Ulrich Drepper <drepper@redhat.com>
* elf/dl-load.c (_dl_map_object_from_fd): Only call audit hooks
if we are not loading a new audit library.
* elf/dl-reloc (_dl_relocate_object): Third parameter is now a bitmask.
Only use profiling trampoline for auditing if we are not relocating
an audit library.
* elf/dl-open.c (dl_open_worker): Adjust _dl_relocate_object call.
* elf/rtld.c: Likewise.
* sysdeps/generic/ldsodefs.h: Adjust _dl_relocate_object prototype.
* inet/inet6_rth.c (inet6_rth_add): Add some error checking.
Patch mostly by Yang Hongyang <yanghy@cn.fujitsu.com>.
* inet/Makefile (tests): Add tst-inet6_rth.
* inet/tst-inet6_rth.c: New file.
alignment of La_x86_64_regs. Store xmm parameters.
(reloc_index): Define.
(_dl_fixup): Rename reloc_offset parameter to reloc_arg.
(_dl_fixup_profile): Likewise. Use reloc_index instead of
computing index from reloc_offset.
(_dl_call_pltexit): Likewise.
* sysdeps/x86_64/dl-trampoline.S (_dl_runtime_resolve): Just pass
the relocation index to _dl_fixup.
(_dl_runtime_profile): Likewise for _dl_fixup_profile and
_dl_call_pltexit.
* sysdeps/x86_64/dl-runtime.c: New file.
* sysdeps/x86_64/dl-trampoline.S (_dl_runtime_profile): Fix
alignement of La_x86_64_regs. Store xmm parameters.
Patch mostly by Jiri Olsa <olsajiri@gmail.com>.
