On powerpc, atomic_exchange_and_add is implemented without any
barriers. This patchs adds the missing instruction and memory barrier
for acquire and release semanthics.
Linux commit dd58a092c4202f2bd490adab7285b3ff77f8e467 added the
PPC_FEATURE2_VEC_CRYPTO auvx capability to indicate whether to
hardware supports vector crypto hardware instructions. This patch
adds its definition to powerpc hwcap bits.
elf/tst-auxv.c includes misc/sys/auxv.h, which ends up not actually
being included due to the guard overlap, and getauxval becomes an
implicit declaration and implicit pointer conversion which means, at
best, the test isn't actually testing what it thinks it is and, at
worst, it'll crash and burn on platforms where implict pointer
conversion is a Very Bad Thing.
* sysdeps/powerpc/bits/hwcap.h: Allow _SYSDEPS_SYSDEP_H guard as a
synonym for _SYS_AUXV_H to allow direct inclusion.
* sysdeps/sparc/bits/hwcap.h: Likewise.
* sysdeps/powerpc/sysdep.h: Define _SYSDEPS_SYSDEP_H instead of
_SYS_AUXV_H so we can include sysdep.h and sys/auxv.h together.
* sysdeps/sparc/sysdep.h: Likewise.
The ELFv2 ABI changes the calling convention by passing and returning
structures in registers in more cases than the old ABI:
http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01145.htmlhttp://gcc.gnu.org/ml/gcc-patches/2013-11/msg01147.html
For the most part, this does not affect glibc, since glibc assembler
files do not use structure parameters / return values. However, one
place is affected: the LD_AUDIT interface provides a structure to
the audit routine that contains all registers holding function
argument and return values for the intercepted PLT call.
Since the new ABI now sometimes uses registers to return values
that were never used for this purpose in the old ABI, this structure
has to be extended. To force audit routines to be modified for the
new ABI if necessary, the patch defines v2 variants of the la_ppc64
types and routines.
In addition, the patch contains two unrelated changes to the
PLT trampoline routines: it fixes a bug where FPR return values
were stored in the wrong place, and it removes the unnecessary
save/restore of CR.
__fe_nomask_env.
* sysdeps/powerpc/fpu/fe_nomask.c: Add libm_hidden_def.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/fe_nomask.c: Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/fpu/fe_nomask.c: Likewise.
* sysdeps/powerpc/bits/fenv.h: Make safe for C++.
* sysdeps/unix/sysv/linux/powerpc/bits/mathinline.h: New file.
* sysdeps/powerpc/fpu/fegetexcept.c (__fegetexcept): Rename
function from fegetexcept and make old name weak alias.
* include/fenv.h: Declare __fegetexcept.
* sysdeps/powerpc/fpu/fedisblxcpt.c: Use __fegetexcept instead of
fegetexcept.
* sysdeps/powerpc/fpu/feenablxcpt.c: Likewise.
* sysdeps/powerpc/fpu/fraiseexcpt.c (__feraiseexcept): Avoid call
to fetestexcept.
* sysdeps/ieee754/ldbl-128ibm/s_log1pl.c (__log1pl): Use __frexpl
instead of frexpl to avoid local PLT.
* math/s_significandl.c (__significandl): Use __ilogbl instead of
ilogbl to avoid local PLT.
* sysdeps/ieee754/ldbl-128ibm/s_expm1l.c (__expm1l): Use __ldexpl
instead of ldexpl to avoid local PLT.
* sysdeps/ieee754/ldbl-128ibm/e_expl.c (__ieee754_expl): Use
__roundl not roundl to avoid local PLT.
* sysdeps/ieee754/ldbl-128/e_j0l.c: Use function names which avoid
local PLTs. Use __sincosl instead of separate sinl and cosl
calls.
* sysdeps/ieee754/ldbl-128/e_j1l.c: Likewise.
Peter Bergner <bergner@us.ibm.com>
* sysdeps/powerpc/bits/fenv.h: Declare __fe_mask_env extern.
Define FE_NOMASK_ENV as FE_EANBLED_ENV. Define FE_MASK_ENV.
* sysdeps/powerpc/fpu/Makefile: Add fe_mask to libm-support.
* sysdeps/powerpc/fpu/fe_mask.c: New file.
* sysdeps/powerpc/fpu/fe_nomask.c: Correct comment.
* sysdeps/powerpc/fpu/fedisblxcpt.c (fedisableexcept):
Call __fe_mask_env() if all FP exceptions disabled.
* sysdeps/powerpc/fpu/feholdexcpt.c (feholdexcept): Copy high 32-bits
from old FPSCR to new fenv to propagate DFP rounding modes.
Call __fe_mask_env() if FP exceptions previously enabled.
* sysdeps/powerpc/fpu/fesetenv.c (fesetenv): Change mask to merge
exceptions from env. Use __fe_nomask_env() or __fe_mask_env() when
transitioning from all exceptions disabled to any exception enabled
or visa versa.
* sysdeps/powerpc/fpu/feupdateenv.c (__feupdateenv): Change mask to
merge exceptions from env. Call __fe_nomask_env or __fe_mask_env
when transitioning from all exceptions disabled to any exception
enabled or visa versa.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/fe_nomask.c: Moved to...
* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/fe_nomask.c: ...here.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/fe_nomask.c: Moved to...
* sysdeps/unix/sysv/linux/powerpc/powerpc64/fpu/fe_nomask.c: ...here.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/fe_mask.c: New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/fpu/fe_mask.c: New file.
* math/multc3.c: New file.
* sysdeps/unix/sysv/linux/sparc/sparc32/Makefile
[$(subdir) = math] (libm-routines): Add multc3, divtc3.
* sysdeps/unix/sysv/linux/sparc/sparc32/fpu/Implies: New file.
* sysdeps/sparc/sparc32/fpu/s_fabsl.c: Include math.h and
math_ldbl_opt.h.
(fabsl): Use long_double_symbol instead of weak_alias.
* sysdeps/sparc/sparc32/fpu/s_fabs.c: Include math.h and
math_ldbl_opt.h.
[LONG_DOUBLE_COMPAT] (fabsl): Add compat_symbol.
2006-01-31 Martin Schwidefsky <schwidefsky@de.ibm.com>
Jakub Jelinek <jakub@redhat.com>
* sysdeps/unix/sysv/linux/s390/Implies: New file.
* sysdeps/unix/sysv/linux/s390/fpu/Implies: New file.
* sysdeps/unix/sysv/linux/s390/nldbl-abi.h: New file.
* sysdeps/s390/fpu/bits/mathinline.h (signbitl, sqrtl): New inlines.
* sysdeps/s390/bits/mathdef.h: New file.
* sysdeps/s390/fpu/e_sqrtl.c: New file.
* sysdeps/s390/s390-32/bits/wordsize.h (__LONG_DOUBLE_MATH_OPTIONAL,
__NO_LONG_DOUBLE_MATH): Define.
* sysdeps/s390/s390-64/bits/wordsize.h: Likewise.
* sysdeps/s390/Implies: Add ieee754/ldbl-128.
* sysdeps/s390/ldbl2mpn.c: File removed.
* sysdeps/s390/fpu/libm-test-ulps: Updated.
2006-01-31 Roland McGrath <roland@redhat.com>
* sysdeps/alpha/fpu/bits/mathdef.h: Moved to ...
* sysdeps/alpha/bits/mathdef.h: ... here.
* sysdeps/i386/fpu/bits/mathdef.h: Moved to ...
* sysdeps/i386/bits/mathdef.h: ... here.
* sysdeps/mips/fpu/bits/mathdef.h: Moved to ...
* sysdeps/mips/bits/mathdef.h: ... here.
* sysdeps/m68k/fpu/bits/mathdef.h: Moved to ...
* sysdeps/m68k/bits/mathdef.h: ... here.
* sysdeps/powerpc/fpu/bits/mathdef.h: Moved to ...
* sysdeps/powerpc/bits/mathdef.h: ... here.
* sysdeps/sparc/fpu/bits/mathdef.h: Moved to ...
* sysdeps/sparc/bits/mathdef.h: ... here.
* sysdeps/ia64/fpu/bits/mathdef.h: Moved to ...
* sysdeps/ia64/bits/mathdef.h: ... here.
* sysdeps/sh/sh4/fpu/bits/mathdef.h: Moved to ...
* sysdeps/sh/sh4/bits/mathdef.h: ... here.
* sysdeps/x86_64/fpu/bits/mathdef.h: Moved to ...
* sysdeps/x86_64/bits/mathdef.h: ... here.
2006-01-31 Jakub Jelinek <jakub@redhat.com>
(La_ppc64_retval): Correct size of lrc_fp.
* sysdeps/powerpc/powerpc64/dl-trampoline.S (_dl_profile_resolve):
Fix up ABI problems and complete function.
2004-02-13 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc64/__longjmp-common.S: New file.
* sysdeps/powerpc/powerpc64/__longjmp.S [NOT_IN_libc](__longjmp):
Non-versioned __longjmp for rtld-__longjmp.
[!NOT_IN_libc](__vmx__longjmp): Add VMX reg support and define as
default version of __longjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_3, GLIBC_2_3_4)]
(__novmx__longjmp): Original version of __longjmp.
* sysdeps/powerpc/powerpc64/bsd-_setjmp.S [NOT_IN_libc](_setjmp):
Non-versioned _setjmp for rtld-_setjmp.
[!NOT_IN_libc](__vmx_setjmp): Branch to __vmx__sigsetjmp and define
as default version of _setjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_3, GLIBC_2_3_4)]
(__novmx_setjmp): Original version of _setjmp.
* sysdeps/powerpc/powerpc64/bsd-setjmp.S (__vmxsetjmp): Branch to
__vmx__sigsetjmp and define as default version of setjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_3, GLIBC_2_3_4)]
(__novmxsetjmp): Original version of setjmp.
* sysdeps/powerpc/powerpc64/setjmp-common.S: New file.
* sysdeps/powerpc/powerpc64/setjmp.S [NOT_IN_libc](__setjmp):
Non-versioned __sigsetjmp for rtld-setjmp.
[!NOT_IN_libc](__vmx__sigsetjmp): Add VMX reg support and define
as default version of __sigsetjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_3, GLIBC_2_3_4)]
(__novmx__sigsetjmp): Original version of __sigsetjmp.
* sysdeps/powerpc/powerpc32/__longjmp-common.S: New File
* sysdeps/powerpc/powerpc32/__longjmp.S [NOT_IN_libc](__longjmp):
Non-versioned __longjmp for rtld-__longjmp.
[!NOT_IN_libc](__vmx__longjmp): Add VMX reg support and define as
default version of __longjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4)]
(__novmx__longjmp): Original version of __longjmp.
* sysdeps/powerpc/powerpc32/bsd-_setjmp.S [NOT_IN_libc](_setjmp):
Non-versioned _setjmp for rtld-_setjmp.
[!NOT_IN_libc](__vmx_setjmp): Branch to __vmx__sigsetjmp and define
as default version of _setjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4)]
(__novmx_setjmp): Original version of _setjmp.
* sysdeps/powerpc/powerpc32/bsd-setjmp.S (__vmxsetjmp): Branch to
__vmx__sigsetjmp and define as default version of setjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4)]
(__novmxsetjmp): Original version of setjmp.
* sysdeps/powerpc/powerpc32/fpu/__longjmp-common.S: New file.
* sysdeps/powerpc/powerpc32/fpu/__longjmp.S [NOT_IN_libc](__longjmp):
Non-versioned __longjmp for rtld-__longjmp.
[!NOT_IN_libc](__vmx__longjmp): Add VMX reg support and define as
default version of __longjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4)]
(__novmx__longjmp): Original version of __longjmp.
* sysdeps/powerpc/powerpc32/fpu/setjmp-common.S: New file.
* sysdeps/powerpc/powerpc32/fpu/setjmp.S [NOT_IN_libc](__setjmp):
Non-versioned __sigsetjmp for rtld-setjmp.
[!NOT_IN_libc](__vmx__sigsetjmp): Add VMX reg support and define
as default version of __sigsetjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4)]
(__novmx__sigsetjmp): Original version of __sigsetjmp.
* sysdeps/powerpc/powerpc32/setjmp-common.S: New file.
* sysdeps/powerpc/powerpc32/setjmp.S [NOT_IN_libc](__setjmp):
Non-versioned __sigsetjmp for rtld-setjmp.
[!NOT_IN_libc](__vmx__sigsetjmp): Add VMX reg support and define
as default version of __sigsetjmp.
[SHARED && SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4)]
(__novmx__sigsetjmp): Original version of __sigsetjmp.
* sysdeps/powerpc/Dist: New File.
* sysdeps/powerpc/Makefile (sysdep_routines): Add novmx-longjmp.c
and novmx-sigjmp.c
* sysdeps/powerpc/Versions (libc): To GLIBC_2.3_4 add _longjmp,
__sigsetjmp, _setjmp, longjmp, and setjmp. To GLIBC_PRIVATE add
__novmx__libc_longjmp, __novmx__libc_siglongjmp, __vmx__libc_longjmp,
and __vmx__libc_siglongjmp.
* sysdeps/powerpc/bits/setjmp.h: Define JB_VRSAVE, JB_VRS, and adjust
JB_SIZE to add VMX regs to __jmp_buf.
* sysdeps/powerpc/longjmp.c: New file.
* sysdeps/powerpc/novmxsetjmp.h: New file.
* sysdeps/powerpc/novmx-longjmp.c: New file.
* sysdeps/powerpc/novmx-sigjmp.c: New file.
* sysdeps/powerpc/sigjmp.c: New file.
2003-04-16 Jakub Jelinek <jakub@redhat.com>
* elf/Makefile (distribute): Add tst-tlsmod{[7-9],1[0-2]}.c and
tst-tls10.h.
(tests): Add tst-tls1[0-2].
(modules-names): Add tst-tlsmod{[7-8],1[0-2]}.
($(objpfx)tst-tlsmod8.so): Depend on tst-tlsmod7.so.
($(objpfx)tst-tlsmod10.so): Depend on tst-tlsmod9.so.
($(objpfx)tst-tlsmod12.so): Depend on tst-tlsmod11.so.
($(objpfx)tst-tls10): Depend on tst-tlsmod8.so.
($(objpfx)tst-tls11): Depend on tst-tlsmod10.so.
($(objpfx)tst-tls12): Depend on tst-tlsmod12.so.
* elf/tst-tls10.c: New test.
* elf/tst-tls11.c: New test.
* elf/tst-tls12.c: New test.
* elf/tst-tls10.h: New file.
* elf/tst-tlsmod7.c: New file.
* elf/tst-tlsmod8.c: New file.
* elf/tst-tlsmod9.c: New file.
* elf/tst-tlsmod10.c: New file.
* elf/tst-tlsmod11.c: New file.
* elf/tst-tlsmod12.c: New file.
2003-04-15 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/bits/atomic.h: Moved ppc32/64 specific code to ...
* sysdeps/powerpc/powerpc32/bits/atomic.h: New file.
* sysdeps/powerpc/powerpc64/bits/atomic.h: New file.
* posix/regex.h: Include <sys/types.h>.
2003-04-11 Ulrich Drepper <drepper@redhat.com>
* sysdeps/generic/libc-start.c: Cleanup MAIN_AUXVEC_ARG handling.
Remove HAVE_CANCELBUF code. Replace with code using the new
initializers for unwind-based cleanup handling.
* sysdeps/generic/unwind.h: Update from latest gcc version.
* sysdeps/unix/sysv/linux/i386/sysdep.h: Define labels in a few
places to allow unwind data generation.
* sysdeps/i386/bits/setjmp.h: Allow file to be included multiple times.
* sysdeps/x86_64/bits/setjmp.h: Likewise.
* sysdeps/sh/bits/setjmp.h: Likewise.
* sysdeps/powerpc/bits/setjmp.h: Likewise.
* sysdeps/unix/sysv/linux/ia64/bits/setjmp.h: Likewise.
* sysdeps/alpha/bits/setjmp.h: Likewise.
atomic_compare_and_exchange_val_acq, atomic_add_zero,
atomic_compare_and_exchange_bool_acq and atomic_add_negative tests.
* include/atomic.h (atomic_add_negative, atomic_add_zero):
Prefix local variable so that it doesn't clash with the one
in atomic_exchange_and_add.
* sysdeps/ia64/bits/atomic.h (atomic_exchange): Fix for long/void *
pointers.
(atomic_exchange_and_add): Implement using __sync_fetch_and_add_?i.
* sysdeps/powerpc/bits/atomic.h (atomic_exchange_and_add): Force
value into register.
* sysdeps/s390/bits/atomic.h (__arch_compare_and_exchange_val_64_acq):
Cast newval to long.
* sysdeps/x86_64/bits/atomic.h
(__arch_compare_and_exchange_val_64_acq): Cast newval and oldval to
long.
(atomic_exchange): Cast newvalue to long if sizeof == 8.
(atomic_exchange_and_add): Cast value to long if sizeof == 8.
(atomic_add, atomic_add_negative, atomic_add_zero): Likewise.
(atomic_bit_set): Shift 1L up in all cases to shut up warnings.
* sysdeps/powerpc/bits/atomic.h (__arch_atomic_exchange_32): New macro.
(__arch_atomic_exchange_64): New macro.
(atomic_exchange): Use them.
(__arch_atomic_exchange_and_add_32): New macro.
(__arch_atomic_exchange_and_add_64): New macro.
(atomic_exchange_and_add): Use them.
Original patch from Steven Munroe <sjmunroe@us.ibm.com>.
variable. Remove superfluous memory clobber.
* include/atomic.h: Syntax braino fix.
* posix/tst-nice.c (do_test): Use %m formats instead of printing errno
in decimal. Don't bail if niced at start. Just check that nice call
bumps the total at all.