This is the mail archive of the
glibc-cvs@sourceware.org
mailing list for the glibc project.
GNU C Library master sources branch, master, updated. glibc-2.10-268-g4e1e2f4
- From: drepper at sourceware dot org
- To: glibc-cvs at sourceware dot org
- Date: 8 Aug 2009 17:55:22 -0000
- Subject: GNU C Library master sources branch, master, updated. glibc-2.10-268-g4e1e2f4
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GNU C Library master sources".
The branch, master has been updated
via 4e1e2f42472744569f1540dd8410d23180e24bf9 (commit)
via fc1870e6a484ad3211648c9ae51bc076913518aa (commit)
via a3d4298d8354f4ea232f027ba2ebfd560afecc1b (commit)
via 49eea97b00d367aa3eac3dfd047c259e7bd04732 (commit)
from 8e436522e108229eeb96d8921c659c5a19eaf455 (commit)
Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.
- Log -----------------------------------------------------------------
http://sources.redhat.com/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=4e1e2f42472744569f1540dd8410d23180e24bf9
commit 4e1e2f42472744569f1540dd8410d23180e24bf9
Author: H.J. Lu <hongjiu.lu@intel.com>
Date: Sat Aug 8 10:54:42 2009 -0700
Support mixed SSE/AVX audit and check AVX only once.
This patch fixes mixed SSE/AVX audit and checks AVX only once in
_dl_runtime_profile. When an AVX or SSE register value in pltenter is
modified, we have to make sure that the SSE part value is the same in both
lr_xmm and lr_vector fields so that pltexit will get the correct value
from either lr_xmm or lr_vector fields. AVX-enabled pltenter should
update both lr_xmm and lr_vector fields to support stacked AVX/SSE
pltenter functions.
diff --git a/ChangeLog b/ChangeLog
index 072af87..12fa4cd 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,31 @@
+2009-08-01 H.J. Lu <hongjiu.lu@intel.com>
+
+ * elf/Makefile (distribute): Add tst-audit6.c tst-auditmod6a.c
+ tst-auditmod6b.c tst-auditmod6c.c tst-audit7.c tst-auditmod7a.c
+ tst-auditmod7b.c.
+ (tests): Add tst-audit6 tst-audit7.
+ (modules-names): Add st-auditmod6a tst-auditmod6b tst-auditmod6c
+ tst-auditmod7a tst-auditmod7b.
+ ($(objpfx)tst-audit6): New.
+ ($(objpfx)tst-audit6.out): Likewise.
+ ($(objpfx)tst-audit7): Likewise.
+ ($(objpfx)tst-audit7.out): Likewise.
+ (tst-audit6-ENV): Likewise.
+ (tst-audit7-ENV): Likewise.
+ (CFLAGS-tst-auditmod6b.c): Likewise.
+ (CFLAGS-tst-auditmod6c.c): Likewise.
+ (CFLAGS-tst-auditmod7b.c): Likewise.
+ * elf/tst-audit6.c: New file.
+ * elf/tst-audit7.c: New file.
+ * elf/tst-auditmod6a.c: New file.
+ * elf/tst-auditmod6b.c: New file.
+ * elf/tst-auditmod6c.c: New file.
+ * elf/tst-auditmod7a.c: New file.
+ * elf/tst-auditmod7b.c: New file.
+ * sysdeps/x86_64/dl-trampoline.S (_dl_runtime_profile): Move
+ saving and restoring SSE/AVX registers to ...
+ * sysdeps/x86_64/dl-trampoline.h: This. New file.
+
2009-08-07 H.J. Lu <hongjiu.lu@intel.com>
* sysdeps/i386/i686/multiarch/strcspn.S (STRCSPN): Use PIC
diff --git a/elf/Makefile b/elf/Makefile
index 3baad96..d57c7fe 100644
--- a/elf/Makefile
+++ b/elf/Makefile
@@ -93,6 +93,9 @@ distribute := rtld-Rules \
tst-auditmod1.c tst-auditmod3a.c tst-auditmod3b.c \
tst-auditmod4a.c tst-auditmod4b.c \
tst-audit5.c tst-auditmod5a.c tst-auditmod5b.c \
+ tst-audit6.c tst-auditmod6a.c tst-auditmod6b.c \
+ tst-auditmod6c.c \
+ tst-audit7.c tst-auditmod7a.c tst-auditmod7b.c \
order2mod1.c order2mod2.c order2mod3.c order2mod4.c \
tst-stackguard1.c tst-stackguard1-static.c \
tst-array5.c tst-array5-static.c tst-array5dep.c \
@@ -200,7 +203,7 @@ tests += loadtest restest1 preloadtest loadfail multiload origtest resolvfail \
test-srcs = tst-pathopt
tests-execstack-yes = tst-execstack tst-execstack-needed tst-execstack-prog
ifeq (x86_64,$(config-machine))
-tests += tst-audit3 tst-audit4 tst-audit5
+tests += tst-audit3 tst-audit4 tst-audit5 tst-audit6 tst-audit7
endif
endif
ifeq (yesyes,$(have-fpie)$(build-shared))
@@ -255,7 +258,9 @@ endif
ifeq (x86_64,$(config-machine))
modules-names += tst-auditmod3a tst-auditmod3b \
tst-auditmod4a tst-auditmod4b \
- tst-auditmod5a tst-auditmod5b
+ tst-auditmod5a tst-auditmod5b \
+ tst-auditmod6a tst-auditmod6b tst-auditmod6c \
+ tst-auditmod7a tst-auditmod7b
endif
modules-execstack-yes = tst-execstack-mod
extra-test-objs += $(addsuffix .os,$(strip $(modules-names)))
@@ -987,6 +992,15 @@ $(objpfx)tst-audit5: $(objpfx)tst-auditmod5a.so
$(objpfx)tst-audit5.out: $(objpfx)tst-auditmod5b.so
tst-audit5-ENV = LD_AUDIT=$(objpfx)tst-auditmod5b.so
+$(objpfx)tst-audit6: $(objpfx)tst-auditmod6a.so
+$(objpfx)tst-audit6.out: $(objpfx)tst-auditmod6b.so \
+ $(objpfx)tst-auditmod6c.so
+tst-audit6-ENV = LD_AUDIT=$(objpfx)tst-auditmod6b.so:$(objpfx)tst-auditmod6c.so
+
+$(objpfx)tst-audit7: $(objpfx)tst-auditmod7a.so
+$(objpfx)tst-audit7.out: $(objpfx)tst-auditmod7b.so
+tst-audit7-ENV = LD_AUDIT=$(objpfx)tst-auditmod7b.so
+
$(objpfx)tst-global1: $(libdl)
$(objpfx)tst-global1.out: $(objpfx)testobj6.so $(objpfx)testobj2.so
@@ -1134,4 +1148,7 @@ ifeq (yes,$(config-cflags-avx))
CFLAGS-tst-audit4.c += -mavx
CFLAGS-tst-auditmod4a.c += -mavx
CFLAGS-tst-auditmod4b.c += -mavx
+CFLAGS-tst-auditmod6b.c += -mavx
+CFLAGS-tst-auditmod6c.c += -mavx
+CFLAGS-tst-auditmod7b.c += -mavx
endif
diff --git a/elf/tst-audit6.c b/elf/tst-audit6.c
new file mode 100644
index 0000000..1f6dcb1
--- /dev/null
+++ b/elf/tst-audit6.c
@@ -0,0 +1,28 @@
+/* Test case for x86-64 preserved registers in dynamic linker. */
+
+#include <stdlib.h>
+#include <string.h>
+#include <cpuid.h>
+#include <emmintrin.h>
+
+extern __m128i audit_test (__m128i, __m128i, __m128i, __m128i,
+ __m128i, __m128i, __m128i, __m128i);
+
+int
+main (void)
+{
+ unsigned int eax, ebx, ecx, edx;
+
+ /* Run AVX test only if AVX is supported. */
+ if (__get_cpuid (1, &eax, &ebx, &ecx, &edx)
+ && (ecx & bit_AVX))
+ {
+ __m128i xmm = _mm_setzero_si128 ();
+ __m128i ret = audit_test (xmm, xmm, xmm, xmm, xmm, xmm, xmm, xmm);
+
+ xmm = _mm_set1_epi32 (0x98abcdef);
+ if (memcmp (&xmm, &ret, sizeof (ret)))
+ abort ();
+ }
+ return 0;
+}
diff --git a/elf/tst-audit7.c b/elf/tst-audit7.c
new file mode 100644
index 0000000..1d2a7de
--- /dev/null
+++ b/elf/tst-audit7.c
@@ -0,0 +1 @@
+#include "tst-audit6.c"
diff --git a/elf/tst-auditmod6a.c b/elf/tst-auditmod6a.c
new file mode 100644
index 0000000..c3a850c
--- /dev/null
+++ b/elf/tst-auditmod6a.c
@@ -0,0 +1,46 @@
+/* Test case for x86-64 preserved registers in dynamic linker. */
+
+#include <stdlib.h>
+#include <string.h>
+#include <emmintrin.h>
+
+__m128i
+audit_test (__m128i x0, __m128i x1, __m128i x2, __m128i x3,
+ __m128i x4, __m128i x5, __m128i x6, __m128i x7)
+{
+ __m128i xmm;
+
+ xmm = _mm_set1_epi32 (0x100);
+ if (memcmp (&xmm, &x0, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x101);
+ if (memcmp (&xmm, &x1, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x102);
+ if (memcmp (&xmm, &x2, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x103);
+ if (memcmp (&xmm, &x3, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x104);
+ if (memcmp (&xmm, &x4, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x105);
+ if (memcmp (&xmm, &x5, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x106);
+ if (memcmp (&xmm, &x6, sizeof (xmm)))
+ abort ();
+
+ xmm = _mm_set1_epi32 (0x107);
+ if (memcmp (&xmm, &x7, sizeof (xmm)))
+ abort ();
+
+ return _mm_setzero_si128 ();
+}
diff --git a/elf/tst-auditmod6b.c b/elf/tst-auditmod6b.c
new file mode 100644
index 0000000..f756b50
--- /dev/null
+++ b/elf/tst-auditmod6b.c
@@ -0,0 +1,220 @@
+/* Verify that changing AVX registers in audit library won't affect
+ function parameter passing/return. */
+
+#include <dlfcn.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <bits/wordsize.h>
+#include <gnu/lib-names.h>
+
+unsigned int
+la_version (unsigned int v)
+{
+ setlinebuf (stdout);
+
+ printf ("version: %u\n", v);
+
+ char buf[20];
+ sprintf (buf, "%u", v);
+
+ return v;
+}
+
+void
+la_activity (uintptr_t *cookie, unsigned int flag)
+{
+ if (flag == LA_ACT_CONSISTENT)
+ printf ("activity: consistent\n");
+ else if (flag == LA_ACT_ADD)
+ printf ("activity: add\n");
+ else if (flag == LA_ACT_DELETE)
+ printf ("activity: delete\n");
+ else
+ printf ("activity: unknown activity %u\n", flag);
+}
+
+char *
+la_objsearch (const char *name, uintptr_t *cookie, unsigned int flag)
+{
+ char buf[100];
+ const char *flagstr;
+ if (flag == LA_SER_ORIG)
+ flagstr = "LA_SET_ORIG";
+ else if (flag == LA_SER_LIBPATH)
+ flagstr = "LA_SER_LIBPATH";
+ else if (flag == LA_SER_RUNPATH)
+ flagstr = "LA_SER_RUNPATH";
+ else if (flag == LA_SER_CONFIG)
+ flagstr = "LA_SER_CONFIG";
+ else if (flag == LA_SER_DEFAULT)
+ flagstr = "LA_SER_DEFAULT";
+ else if (flag == LA_SER_SECURE)
+ flagstr = "LA_SER_SECURE";
+ else
+ {
+ sprintf (buf, "unknown flag %d", flag);
+ flagstr = buf;
+ }
+ printf ("objsearch: %s, %s\n", name, flagstr);
+
+ return (char *) name;
+}
+
+unsigned int
+la_objopen (struct link_map *l, Lmid_t lmid, uintptr_t *cookie)
+{
+ printf ("objopen: %ld, %s\n", lmid, l->l_name);
+
+ return 3;
+}
+
+void
+la_preinit (uintptr_t *cookie)
+{
+ printf ("preinit\n");
+}
+
+unsigned int
+la_objclose (uintptr_t *cookie)
+{
+ printf ("objclose\n");
+ return 0;
+}
+
+uintptr_t
+la_symbind64 (Elf64_Sym *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, unsigned int *flags, const char *symname)
+{
+ printf ("symbind64: symname=%s, st_value=%#lx, ndx=%u, flags=%u\n",
+ symname, (long int) sym->st_value, ndx, *flags);
+
+ return sym->st_value;
+}
+
+#define pltenter la_x86_64_gnu_pltenter
+#define pltexit la_x86_64_gnu_pltexit
+#define La_regs La_x86_64_regs
+#define La_retval La_x86_64_retval
+#define int_retval lrv_rax
+
+#include <tst-audit.h>
+
+#ifdef __AVX__
+#include <immintrin.h>
+#include <cpuid.h>
+
+static int avx = -1;
+
+static int
+__attribute ((always_inline))
+check_avx (void)
+{
+ if (avx == -1)
+ {
+ unsigned int eax, ebx, ecx, edx;
+
+ if (__get_cpuid (1, &eax, &ebx, &ecx, &edx)
+ && (ecx & bit_AVX))
+ avx = 1;
+ else
+ avx = 0;
+ }
+ return avx;
+}
+#else
+#include <emmintrin.h>
+#endif
+
+ElfW(Addr)
+pltenter (ElfW(Sym) *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, La_regs *regs, unsigned int *flags,
+ const char *symname, long int *framesizep)
+{
+ printf ("pltenter: symname=%s, st_value=%#lx, ndx=%u, flags=%u\n",
+ symname, (long int) sym->st_value, ndx, *flags);
+
+#ifdef __AVX__
+ if (check_avx () && strcmp (symname, "audit_test") == 0)
+ {
+ int i;
+
+ __m128i xmm = _mm_setzero_si128 ();
+ for (i = 0; i < 8; i++)
+ if (memcmp (®s->lr_xmm[i], &xmm, sizeof (xmm))
+ || memcmp (®s->lr_vector[i], &xmm, sizeof (xmm)))
+ abort ();
+
+ for (i = 0; i < 8; i += 2)
+ {
+ regs->lr_xmm[i] = (La_x86_64_xmm) _mm_set1_epi32 (i + 1);
+ regs->lr_vector[i].xmm[0] = regs->lr_xmm[i];
+ regs->lr_vector[i + 1].ymm[0]
+ = (La_x86_64_ymm) _mm256_set1_epi32 (i + 2);
+ regs->lr_xmm[i + 1] = regs->lr_vector[i + 1].xmm[0];
+ }
+
+ __m256i ymm = _mm256_set1_epi32 (-1);
+ asm volatile ("vmovdqa %0, %%ymm0" : : "x" (ymm) : "xmm0" );
+ asm volatile ("vmovdqa %0, %%ymm1" : : "x" (ymm) : "xmm1" );
+ asm volatile ("vmovdqa %0, %%ymm2" : : "x" (ymm) : "xmm2" );
+ asm volatile ("vmovdqa %0, %%ymm3" : : "x" (ymm) : "xmm3" );
+ asm volatile ("vmovdqa %0, %%ymm4" : : "x" (ymm) : "xmm4" );
+ asm volatile ("vmovdqa %0, %%ymm5" : : "x" (ymm) : "xmm5" );
+ asm volatile ("vmovdqa %0, %%ymm6" : : "x" (ymm) : "xmm6" );
+ asm volatile ("vmovdqa %0, %%ymm7" : : "x" (ymm) : "xmm7" );
+
+ *framesizep = 1024;
+ }
+#endif
+
+ return sym->st_value;
+}
+
+unsigned int
+pltexit (ElfW(Sym) *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, const La_regs *inregs, La_retval *outregs,
+ const char *symname)
+{
+ printf ("pltexit: symname=%s, st_value=%#lx, ndx=%u, retval=%tu\n",
+ symname, (long int) sym->st_value, ndx, outregs->int_retval);
+
+#ifdef __AVX__
+ if (check_avx () && strcmp (symname, "audit_test") == 0)
+ {
+ int i;
+
+ __m128i xmm = _mm_setzero_si128 ();
+ if (memcmp (&outregs->lrv_xmm0, &xmm, sizeof (xmm))
+ || memcmp (&outregs->lrv_vector0, &xmm, sizeof (xmm)))
+ abort ();
+
+ __m256i ymm;
+
+ for (i = 0; i < 8; i += 2)
+ {
+ xmm = _mm_set1_epi32 (i + 0x100);
+ if (memcmp (&inregs->lr_xmm[i], &xmm, sizeof (xmm))
+ || memcmp (&inregs->lr_vector[i], &xmm, sizeof (xmm)))
+ abort ();
+
+ ymm = _mm256_set1_epi32 (i + 0x101);
+ if (memcmp (&inregs->lr_xmm[i + 1],
+ &inregs->lr_vector[i + 1].xmm[0], sizeof (xmm))
+ || memcmp (&inregs->lr_vector[i + 1], &ymm, sizeof (ymm)))
+ abort ();
+ }
+
+ outregs->lrv_vector0.ymm[0]
+ = (La_x86_64_ymm) _mm256_set1_epi32 (0x12349876);
+
+ ymm = _mm256_set1_epi32 (-1);
+ asm volatile ("vmovdqa %0, %%ymm0" : : "x" (ymm) : "xmm0" );
+ asm volatile ("vmovdqa %0, %%ymm1" : : "x" (ymm) : "xmm1" );
+ }
+#endif
+
+ return 0;
+}
diff --git a/elf/tst-auditmod6c.c b/elf/tst-auditmod6c.c
new file mode 100644
index 0000000..49cbf05
--- /dev/null
+++ b/elf/tst-auditmod6c.c
@@ -0,0 +1,225 @@
+/* Verify that changing AVX registers in audit library won't affect
+ function parameter passing/return. */
+
+#include <dlfcn.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <bits/wordsize.h>
+#include <gnu/lib-names.h>
+
+unsigned int
+la_version (unsigned int v)
+{
+ setlinebuf (stdout);
+
+ printf ("version: %u\n", v);
+
+ char buf[20];
+ sprintf (buf, "%u", v);
+
+ return v;
+}
+
+void
+la_activity (uintptr_t *cookie, unsigned int flag)
+{
+ if (flag == LA_ACT_CONSISTENT)
+ printf ("activity: consistent\n");
+ else if (flag == LA_ACT_ADD)
+ printf ("activity: add\n");
+ else if (flag == LA_ACT_DELETE)
+ printf ("activity: delete\n");
+ else
+ printf ("activity: unknown activity %u\n", flag);
+}
+
+char *
+la_objsearch (const char *name, uintptr_t *cookie, unsigned int flag)
+{
+ char buf[100];
+ const char *flagstr;
+ if (flag == LA_SER_ORIG)
+ flagstr = "LA_SET_ORIG";
+ else if (flag == LA_SER_LIBPATH)
+ flagstr = "LA_SER_LIBPATH";
+ else if (flag == LA_SER_RUNPATH)
+ flagstr = "LA_SER_RUNPATH";
+ else if (flag == LA_SER_CONFIG)
+ flagstr = "LA_SER_CONFIG";
+ else if (flag == LA_SER_DEFAULT)
+ flagstr = "LA_SER_DEFAULT";
+ else if (flag == LA_SER_SECURE)
+ flagstr = "LA_SER_SECURE";
+ else
+ {
+ sprintf (buf, "unknown flag %d", flag);
+ flagstr = buf;
+ }
+ printf ("objsearch: %s, %s\n", name, flagstr);
+
+ return (char *) name;
+}
+
+unsigned int
+la_objopen (struct link_map *l, Lmid_t lmid, uintptr_t *cookie)
+{
+ printf ("objopen: %ld, %s\n", lmid, l->l_name);
+
+ return 3;
+}
+
+void
+la_preinit (uintptr_t *cookie)
+{
+ printf ("preinit\n");
+}
+
+unsigned int
+la_objclose (uintptr_t *cookie)
+{
+ printf ("objclose\n");
+ return 0;
+}
+
+uintptr_t
+la_symbind64 (Elf64_Sym *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, unsigned int *flags, const char *symname)
+{
+ printf ("symbind64: symname=%s, st_value=%#lx, ndx=%u, flags=%u\n",
+ symname, (long int) sym->st_value, ndx, *flags);
+
+ return sym->st_value;
+}
+
+#define pltenter la_x86_64_gnu_pltenter
+#define pltexit la_x86_64_gnu_pltexit
+#define La_regs La_x86_64_regs
+#define La_retval La_x86_64_retval
+#define int_retval lrv_rax
+
+#include <tst-audit.h>
+
+#ifdef __AVX__
+#include <immintrin.h>
+#include <cpuid.h>
+
+static int avx = -1;
+
+static int
+__attribute ((always_inline))
+check_avx (void)
+{
+ if (avx == -1)
+ {
+ unsigned int eax, ebx, ecx, edx;
+
+ if (__get_cpuid (1, &eax, &ebx, &ecx, &edx)
+ && (ecx & bit_AVX))
+ avx = 1;
+ else
+ avx = 0;
+ }
+ return avx;
+}
+#else
+#include <emmintrin.h>
+#endif
+
+ElfW(Addr)
+pltenter (ElfW(Sym) *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, La_regs *regs, unsigned int *flags,
+ const char *symname, long int *framesizep)
+{
+ printf ("pltenter: symname=%s, st_value=%#lx, ndx=%u, flags=%u\n",
+ symname, (long int) sym->st_value, ndx, *flags);
+
+#ifdef __AVX__
+ if (check_avx () && strcmp (symname, "audit_test") == 0)
+ {
+ int i;
+ __m128i xmm;
+ __m256i ymm;
+
+ for (i = 0; i < 8; i += 2)
+ {
+ xmm = _mm_set1_epi32 (i + 1);
+ if (memcmp (®s->lr_xmm[i], &xmm, sizeof (xmm))
+ || memcmp (®s->lr_vector[i], &xmm, sizeof (xmm)))
+ abort ();
+ regs->lr_xmm[i] = (La_x86_64_xmm) _mm_set1_epi32 (i + 0x100);
+ regs->lr_vector[i].xmm[0] = regs->lr_xmm[i];
+
+ ymm = _mm256_set1_epi32 (i + 2);
+ if (memcmp (®s->lr_xmm[i + 1],
+ ®s->lr_vector[i + 1].xmm[0], sizeof (xmm))
+ || memcmp (®s->lr_vector[i + 1], &ymm, sizeof (ymm)))
+ abort ();
+ regs->lr_vector[i + 1].ymm[0]
+ = (La_x86_64_ymm) _mm256_set1_epi32 (i + 0x101);
+ regs->lr_xmm[i + 1] = regs->lr_vector[i + 1].xmm[0];
+ }
+
+ ymm = _mm256_set1_epi32 (-1);
+ asm volatile ("vmovdqa %0, %%ymm0" : : "x" (ymm) : "xmm0" );
+ asm volatile ("vmovdqa %0, %%ymm1" : : "x" (ymm) : "xmm1" );
+ asm volatile ("vmovdqa %0, %%ymm2" : : "x" (ymm) : "xmm2" );
+ asm volatile ("vmovdqa %0, %%ymm3" : : "x" (ymm) : "xmm3" );
+ asm volatile ("vmovdqa %0, %%ymm4" : : "x" (ymm) : "xmm4" );
+ asm volatile ("vmovdqa %0, %%ymm5" : : "x" (ymm) : "xmm5" );
+ asm volatile ("vmovdqa %0, %%ymm6" : : "x" (ymm) : "xmm6" );
+ asm volatile ("vmovdqa %0, %%ymm7" : : "x" (ymm) : "xmm7" );
+
+ *framesizep = 1024;
+ }
+#endif
+
+ return sym->st_value;
+}
+
+unsigned int
+pltexit (ElfW(Sym) *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, const La_regs *inregs, La_retval *outregs,
+ const char *symname)
+{
+ printf ("pltexit: symname=%s, st_value=%#lx, ndx=%u, retval=%tu\n",
+ symname, (long int) sym->st_value, ndx, outregs->int_retval);
+
+#ifdef __AVX__
+ if (check_avx () && strcmp (symname, "audit_test") == 0)
+ {
+ int i;
+
+ __m256i ymm = _mm256_set1_epi32 (0x12349876);;
+ if (memcmp (&outregs->lrv_vector0, &ymm, sizeof (ymm)))
+ abort ();
+
+ __m128i xmm;
+
+ for (i = 0; i < 8; i += 2)
+ {
+ xmm = _mm_set1_epi32 (i + 0x100);
+ if (memcmp (&inregs->lr_xmm[i], &xmm, sizeof (xmm))
+ || memcmp (&inregs->lr_vector[i], &xmm, sizeof (xmm)))
+ abort ();
+
+ ymm = _mm256_set1_epi32 (i + 0x101);
+ if (memcmp (&inregs->lr_xmm[i + 1],
+ &inregs->lr_vector[i + 1].xmm[0], sizeof (xmm))
+ || memcmp (&inregs->lr_vector[i + 1], &ymm, sizeof (ymm)))
+ abort ();
+ }
+
+ outregs->lrv_vector0.ymm[0]
+ = (La_x86_64_ymm) _mm256_set1_epi32 (0x98abcdef);
+
+ ymm = _mm256_set1_epi32 (-1);
+ asm volatile ("vmovdqa %0, %%ymm0" : : "x" (ymm) : "xmm0" );
+ asm volatile ("vmovdqa %0, %%ymm1" : : "x" (ymm) : "xmm1" );
+ }
+#endif
+
+ return 0;
+}
diff --git a/elf/tst-auditmod7a.c b/elf/tst-auditmod7a.c
new file mode 100644
index 0000000..b379df7
--- /dev/null
+++ b/elf/tst-auditmod7a.c
@@ -0,0 +1 @@
+#include "tst-auditmod6a.c"
diff --git a/elf/tst-auditmod7b.c b/elf/tst-auditmod7b.c
new file mode 100644
index 0000000..eb23758
--- /dev/null
+++ b/elf/tst-auditmod7b.c
@@ -0,0 +1,218 @@
+/* Verify that changing AVX registers in audit library won't affect
+ function parameter passing/return. */
+
+#include <dlfcn.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <bits/wordsize.h>
+#include <gnu/lib-names.h>
+
+unsigned int
+la_version (unsigned int v)
+{
+ setlinebuf (stdout);
+
+ printf ("version: %u\n", v);
+
+ char buf[20];
+ sprintf (buf, "%u", v);
+
+ return v;
+}
+
+void
+la_activity (uintptr_t *cookie, unsigned int flag)
+{
+ if (flag == LA_ACT_CONSISTENT)
+ printf ("activity: consistent\n");
+ else if (flag == LA_ACT_ADD)
+ printf ("activity: add\n");
+ else if (flag == LA_ACT_DELETE)
+ printf ("activity: delete\n");
+ else
+ printf ("activity: unknown activity %u\n", flag);
+}
+
+char *
+la_objsearch (const char *name, uintptr_t *cookie, unsigned int flag)
+{
+ char buf[100];
+ const char *flagstr;
+ if (flag == LA_SER_ORIG)
+ flagstr = "LA_SET_ORIG";
+ else if (flag == LA_SER_LIBPATH)
+ flagstr = "LA_SER_LIBPATH";
+ else if (flag == LA_SER_RUNPATH)
+ flagstr = "LA_SER_RUNPATH";
+ else if (flag == LA_SER_CONFIG)
+ flagstr = "LA_SER_CONFIG";
+ else if (flag == LA_SER_DEFAULT)
+ flagstr = "LA_SER_DEFAULT";
+ else if (flag == LA_SER_SECURE)
+ flagstr = "LA_SER_SECURE";
+ else
+ {
+ sprintf (buf, "unknown flag %d", flag);
+ flagstr = buf;
+ }
+ printf ("objsearch: %s, %s\n", name, flagstr);
+
+ return (char *) name;
+}
+
+unsigned int
+la_objopen (struct link_map *l, Lmid_t lmid, uintptr_t *cookie)
+{
+ printf ("objopen: %ld, %s\n", lmid, l->l_name);
+
+ return 3;
+}
+
+void
+la_preinit (uintptr_t *cookie)
+{
+ printf ("preinit\n");
+}
+
+unsigned int
+la_objclose (uintptr_t *cookie)
+{
+ printf ("objclose\n");
+ return 0;
+}
+
+uintptr_t
+la_symbind64 (Elf64_Sym *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, unsigned int *flags, const char *symname)
+{
+ printf ("symbind64: symname=%s, st_value=%#lx, ndx=%u, flags=%u\n",
+ symname, (long int) sym->st_value, ndx, *flags);
+
+ return sym->st_value;
+}
+
+#define pltenter la_x86_64_gnu_pltenter
+#define pltexit la_x86_64_gnu_pltexit
+#define La_regs La_x86_64_regs
+#define La_retval La_x86_64_retval
+#define int_retval lrv_rax
+
+#include <tst-audit.h>
+
+#ifdef __AVX__
+#include <immintrin.h>
+#include <cpuid.h>
+
+static int avx = -1;
+
+static int
+__attribute ((always_inline))
+check_avx (void)
+{
+ if (avx == -1)
+ {
+ unsigned int eax, ebx, ecx, edx;
+
+ if (__get_cpuid (1, &eax, &ebx, &ecx, &edx)
+ && (ecx & bit_AVX))
+ avx = 1;
+ else
+ avx = 0;
+ }
+ return avx;
+}
+#else
+#include <emmintrin.h>
+#endif
+
+ElfW(Addr)
+pltenter (ElfW(Sym) *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, La_regs *regs, unsigned int *flags,
+ const char *symname, long int *framesizep)
+{
+ printf ("pltenter: symname=%s, st_value=%#lx, ndx=%u, flags=%u\n",
+ symname, (long int) sym->st_value, ndx, *flags);
+
+#ifdef __AVX__
+ if (check_avx () && strcmp (symname, "audit_test") == 0)
+ {
+ int i;
+
+ __m128i xmm = _mm_setzero_si128 ();
+ for (i = 0; i < 8; i++)
+ if (memcmp (®s->lr_xmm[i], &xmm, sizeof (xmm))
+ || memcmp (®s->lr_vector[i], &xmm, sizeof (xmm)))
+ abort ();
+
+ for (i = 0; i < 8; i += 2)
+ {
+ regs->lr_xmm[i] = (La_x86_64_xmm) _mm_set1_epi32 (i + 0x100);
+ regs->lr_vector[i + 1].ymm[0]
+ = (La_x86_64_ymm) _mm256_set1_epi32 (i + 0x101);
+ }
+
+ __m256i ymm = _mm256_set1_epi32 (-1);
+ asm volatile ("vmovdqa %0, %%ymm0" : : "x" (ymm) : "xmm0" );
+ asm volatile ("vmovdqa %0, %%ymm1" : : "x" (ymm) : "xmm1" );
+ asm volatile ("vmovdqa %0, %%ymm2" : : "x" (ymm) : "xmm2" );
+ asm volatile ("vmovdqa %0, %%ymm3" : : "x" (ymm) : "xmm3" );
+ asm volatile ("vmovdqa %0, %%ymm4" : : "x" (ymm) : "xmm4" );
+ asm volatile ("vmovdqa %0, %%ymm5" : : "x" (ymm) : "xmm5" );
+ asm volatile ("vmovdqa %0, %%ymm6" : : "x" (ymm) : "xmm6" );
+ asm volatile ("vmovdqa %0, %%ymm7" : : "x" (ymm) : "xmm7" );
+
+ *framesizep = 1024;
+ }
+#endif
+
+ return sym->st_value;
+}
+
+unsigned int
+pltexit (ElfW(Sym) *sym, unsigned int ndx, uintptr_t *refcook,
+ uintptr_t *defcook, const La_regs *inregs, La_retval *outregs,
+ const char *symname)
+{
+ printf ("pltexit: symname=%s, st_value=%#lx, ndx=%u, retval=%tu\n",
+ symname, (long int) sym->st_value, ndx, outregs->int_retval);
+
+#ifdef __AVX__
+ if (check_avx () && strcmp (symname, "audit_test") == 0)
+ {
+ int i;
+
+ __m128i xmm = _mm_setzero_si128 ();
+ if (memcmp (&outregs->lrv_xmm0, &xmm, sizeof (xmm))
+ || memcmp (&outregs->lrv_vector0, &xmm, sizeof (xmm)))
+ abort ();
+
+ __m256i ymm;
+
+ for (i = 0; i < 8; i += 2)
+ {
+ xmm = _mm_set1_epi32 (i + 0x100);
+ if (memcmp (&inregs->lr_xmm[i], &xmm, sizeof (xmm))
+ || memcmp (&inregs->lr_vector[i], &xmm, sizeof (xmm)))
+ abort ();
+
+ ymm = _mm256_set1_epi32 (i + 0x101);
+ if (memcmp (&inregs->lr_xmm[i + 1],
+ &inregs->lr_vector[i + 1].xmm[0], sizeof (xmm))
+ || memcmp (&inregs->lr_vector[i + 1], &ymm, sizeof (ymm)))
+ abort ();
+ }
+
+ outregs->lrv_vector0.ymm[0]
+ = (La_x86_64_ymm) _mm256_set1_epi32 (0x98abcdef);
+
+ ymm = _mm256_set1_epi32 (-1);
+ asm volatile ("vmovdqa %0, %%ymm0" : : "x" (ymm) : "xmm0" );
+ asm volatile ("vmovdqa %0, %%ymm1" : : "x" (ymm) : "xmm1" );
+ }
+#endif
+
+ return 0;
+}
diff --git a/sysdeps/x86_64/dl-trampoline.S b/sysdeps/x86_64/dl-trampoline.S
index 20da695..f9c60ad 100644
--- a/sysdeps/x86_64/dl-trampoline.S
+++ b/sysdeps/x86_64/dl-trampoline.S
@@ -146,247 +146,17 @@ L(have_avx):
2: movl %eax, L(have_avx)(%rip)
cmpl $0, %eax
-1: js L(no_avx1)
+1: js L(no_avx)
- /* This is to support AVX audit modules. */
- vmovdqu %ymm0, (LR_VECTOR_OFFSET)(%rsp)
- vmovdqu %ymm1, (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp)
- vmovdqu %ymm2, (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp)
- vmovdqu %ymm3, (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp)
- vmovdqu %ymm4, (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp)
- vmovdqu %ymm5, (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp)
- vmovdqu %ymm6, (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp)
- vmovdqu %ymm7, (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp)
+# define RESTORE_AVX
+# include "dl-trampoline.h"
- /* Save xmm0-xmm7 registers to detect if any of them are
- changed by audit module. */
- vmovdqa %xmm0, (LR_SIZE)(%rsp)
- vmovdqa %xmm1, (LR_SIZE + XMM_SIZE)(%rsp)
- vmovdqa %xmm2, (LR_SIZE + XMM_SIZE*2)(%rsp)
- vmovdqa %xmm3, (LR_SIZE + XMM_SIZE*3)(%rsp)
- vmovdqa %xmm4, (LR_SIZE + XMM_SIZE*4)(%rsp)
- vmovdqa %xmm5, (LR_SIZE + XMM_SIZE*5)(%rsp)
- vmovdqa %xmm6, (LR_SIZE + XMM_SIZE*6)(%rsp)
- vmovdqa %xmm7, (LR_SIZE + XMM_SIZE*7)(%rsp)
-
-L(no_avx1):
-# endif
-
- movq %rsp, %rcx # La_x86_64_regs pointer to %rcx.
- movq 48(%rbx), %rdx # Load return address if needed.
- movq 40(%rbx), %rsi # Copy args pushed by PLT in register.
- movq 32(%rbx), %rdi # %rdi: link_map, %rsi: reloc_index
- leaq 16(%rbx), %r8
- call _dl_profile_fixup # Call resolver.
-
- movq %rax, %r11 # Save return value.
-
- movq 8(%rbx), %rax # Get back register content.
- movq LR_RDX_OFFSET(%rsp), %rdx
- movq LR_R8_OFFSET(%rsp), %r8
- movq LR_R9_OFFSET(%rsp), %r9
-
- movaps (LR_XMM_OFFSET)(%rsp), %xmm0
- movaps (LR_XMM_OFFSET + XMM_SIZE)(%rsp), %xmm1
- movaps (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp), %xmm2
- movaps (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp), %xmm3
- movaps (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp), %xmm4
- movaps (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp), %xmm5
- movaps (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp), %xmm6
- movaps (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp), %xmm7
-
-# ifdef HAVE_AVX_SUPPORT
- cmpl $0, L(have_avx)(%rip)
- js L(no_avx2)
-
- /* Check if any xmm0-xmm7 registers are changed by audit
- module. */
- vpcmpeqq (LR_SIZE)(%rsp), %xmm0, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET)(%rsp), %ymm0
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp), %ymm1
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE*2)(%rsp), %xmm2, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp), %ymm2
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE*3)(%rsp), %xmm3, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp), %ymm3
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE*4)(%rsp), %xmm4, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp), %ymm4
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE*5)(%rsp), %xmm5, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp), %ymm5
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE*6)(%rsp), %xmm6, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp), %ymm6
-
-1: vpcmpeqq (LR_SIZE + XMM_SIZE*7)(%rsp), %xmm7, %xmm8
- vpmovmskb %xmm8, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp), %ymm7
-
-L(no_avx2):
-1:
-# endif
- movq 16(%rbx), %r10 # Anything in framesize?
- testq %r10, %r10
- jns 3f
-
- /* There's nothing in the frame size, so there
- will be no call to the _dl_call_pltexit. */
-
- /* Get back registers content. */
- movq LR_RCX_OFFSET(%rsp), %rcx
- movq LR_RSI_OFFSET(%rsp), %rsi
- movq LR_RDI_OFFSET(%rsp), %rdi
-
- movq %rbx, %rsp
- movq (%rsp), %rbx
- cfi_restore(rbx)
- cfi_def_cfa_register(%rsp)
-
- addq $48, %rsp # Adjust the stack to the return value
- # (eats the reloc index and link_map)
- cfi_adjust_cfa_offset(-48)
- jmp *%r11 # Jump to function address.
-
-3:
- cfi_adjust_cfa_offset(48)
- cfi_rel_offset(%rbx, 0)
- cfi_def_cfa_register(%rbx)
-
- /* At this point we need to prepare new stack for the function
- which has to be called. We copy the original stack to a
- temporary buffer of the size specified by the 'framesize'
- returned from _dl_profile_fixup */
-
- leaq LR_RSP_OFFSET(%rbx), %rsi # stack
- addq $8, %r10
- andq $0xfffffffffffffff0, %r10
- movq %r10, %rcx
- subq %r10, %rsp
- movq %rsp, %rdi
- shrq $3, %rcx
- rep
- movsq
-
- movq 24(%rdi), %rcx # Get back register content.
- movq 32(%rdi), %rsi
- movq 40(%rdi), %rdi
-
- call *%r11
-
- mov 24(%rbx), %rsp # Drop the copied stack content
-
- /* Now we have to prepare the La_x86_64_retval structure for the
- _dl_call_pltexit. The La_x86_64_regs is being pointed by rsp now,
- so we just need to allocate the sizeof(La_x86_64_retval) space on
- the stack, since the alignment has already been taken care of. */
-# ifdef HAVE_AVX_SUPPORT
- /* sizeof(La_x86_64_retval). Need extra space for 2 SSE
- registers to detect if xmm0/xmm1 registers are changed
- by audit module. */
- subq $(LRV_SIZE + XMM_SIZE*2), %rsp
-# else
- subq $LRV_SIZE, %rsp # sizeof(La_x86_64_retval)
-# endif
- movq %rsp, %rcx # La_x86_64_retval argument to %rcx.
-
- /* Fill in the La_x86_64_retval structure. */
- movq %rax, LRV_RAX_OFFSET(%rcx)
- movq %rdx, LRV_RDX_OFFSET(%rcx)
-
- movaps %xmm0, LRV_XMM0_OFFSET(%rcx)
- movaps %xmm1, LRV_XMM1_OFFSET(%rcx)
-
-# ifdef HAVE_AVX_SUPPORT
- cmpl $0, L(have_avx)(%rip)
- js L(no_avx3)
-
- /* This is to support AVX audit modules. */
- vmovdqu %ymm0, LRV_VECTOR0_OFFSET(%rcx)
- vmovdqu %ymm1, LRV_VECTOR1_OFFSET(%rcx)
-
- /* Save xmm0/xmm1 registers to detect if they are changed
- by audit module. */
- vmovdqa %xmm0, (LRV_SIZE)(%rcx)
- vmovdqa %xmm1, (LRV_SIZE + XMM_SIZE)(%rcx)
-
-L(no_avx3):
-# endif
-
- fstpt LRV_ST0_OFFSET(%rcx)
- fstpt LRV_ST1_OFFSET(%rcx)
-
- movq 24(%rbx), %rdx # La_x86_64_regs argument to %rdx.
- movq 40(%rbx), %rsi # Copy args pushed by PLT in register.
- movq 32(%rbx), %rdi # %rdi: link_map, %rsi: reloc_index
- call _dl_call_pltexit
-
- /* Restore return registers. */
- movq LRV_RAX_OFFSET(%rsp), %rax
- movq LRV_RDX_OFFSET(%rsp), %rdx
-
- movaps LRV_XMM0_OFFSET(%rsp), %xmm0
- movaps LRV_XMM1_OFFSET(%rsp), %xmm1
-
-# ifdef HAVE_AVX_SUPPORT
- cmpl $0, L(have_avx)(%rip)
- js L(no_avx4)
-
- /* Check if xmm0/xmm1 registers are changed by audit module. */
- vpcmpeqq (LRV_SIZE)(%rsp), %xmm0, %xmm2
- vpmovmskb %xmm2, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu LRV_VECTOR0_OFFSET(%rsp), %ymm0
-
-1: vpcmpeqq (LRV_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm2
- vpmovmskb %xmm2, %esi
- cmpl $0xffff, %esi
- jne 1f
- vmovdqu LRV_VECTOR1_OFFSET(%rsp), %ymm1
-
-L(no_avx4):
-1:
+ .align 16
+L(no_avx):
# endif
- fldt LRV_ST1_OFFSET(%rsp)
- fldt LRV_ST0_OFFSET(%rsp)
-
- movq %rbx, %rsp
- movq (%rsp), %rbx
- cfi_restore(rbx)
- cfi_def_cfa_register(%rsp)
-
- addq $48, %rsp # Adjust the stack to the return value
- # (eats the reloc index and link_map)
- cfi_adjust_cfa_offset(-48)
- retq
+# undef RESTORE_AVX
+# include "dl-trampoline.h"
cfi_endproc
.size _dl_runtime_profile, .-_dl_runtime_profile
diff --git a/sysdeps/x86_64/dl-trampoline.h b/sysdeps/x86_64/dl-trampoline.h
new file mode 100644
index 0000000..5d49ed4
--- /dev/null
+++ b/sysdeps/x86_64/dl-trampoline.h
@@ -0,0 +1,269 @@
+/* Partial PLT profile trampoline to save and restore x86-64 vector
+ registers.
+ Copyright (C) 2009 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#ifdef RESTORE_AVX
+ /* This is to support AVX audit modules. */
+ vmovdqu %ymm0, (LR_VECTOR_OFFSET)(%rsp)
+ vmovdqu %ymm1, (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp)
+ vmovdqu %ymm2, (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp)
+ vmovdqu %ymm3, (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp)
+ vmovdqu %ymm4, (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp)
+ vmovdqu %ymm5, (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp)
+ vmovdqu %ymm6, (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp)
+ vmovdqu %ymm7, (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp)
+
+ /* Save xmm0-xmm7 registers to detect if any of them are
+ changed by audit module. */
+ vmovdqa %xmm0, (LR_SIZE)(%rsp)
+ vmovdqa %xmm1, (LR_SIZE + XMM_SIZE)(%rsp)
+ vmovdqa %xmm2, (LR_SIZE + XMM_SIZE*2)(%rsp)
+ vmovdqa %xmm3, (LR_SIZE + XMM_SIZE*3)(%rsp)
+ vmovdqa %xmm4, (LR_SIZE + XMM_SIZE*4)(%rsp)
+ vmovdqa %xmm5, (LR_SIZE + XMM_SIZE*5)(%rsp)
+ vmovdqa %xmm6, (LR_SIZE + XMM_SIZE*6)(%rsp)
+ vmovdqa %xmm7, (LR_SIZE + XMM_SIZE*7)(%rsp)
+#endif
+
+ movq %rsp, %rcx # La_x86_64_regs pointer to %rcx.
+ movq 48(%rbx), %rdx # Load return address if needed.
+ movq 40(%rbx), %rsi # Copy args pushed by PLT in register.
+ movq 32(%rbx), %rdi # %rdi: link_map, %rsi: reloc_index
+ leaq 16(%rbx), %r8
+ call _dl_profile_fixup # Call resolver.
+
+ movq %rax, %r11 # Save return value.
+
+ movq 8(%rbx), %rax # Get back register content.
+ movq LR_RDX_OFFSET(%rsp), %rdx
+ movq LR_R8_OFFSET(%rsp), %r8
+ movq LR_R9_OFFSET(%rsp), %r9
+
+ movaps (LR_XMM_OFFSET)(%rsp), %xmm0
+ movaps (LR_XMM_OFFSET + XMM_SIZE)(%rsp), %xmm1
+ movaps (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp), %xmm2
+ movaps (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp), %xmm3
+ movaps (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp), %xmm4
+ movaps (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp), %xmm5
+ movaps (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp), %xmm6
+ movaps (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp), %xmm7
+
+#ifdef RESTORE_AVX
+ /* Check if any xmm0-xmm7 registers are changed by audit
+ module. */
+ vpcmpeqq (LR_SIZE)(%rsp), %xmm0, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm0, (LR_VECTOR_OFFSET)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET)(%rsp), %ymm0
+ vmovdqa %xmm0, (LR_XMM_OFFSET)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm1, (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp), %ymm1
+ vmovdqa %xmm1, (LR_XMM_OFFSET + XMM_SIZE)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE*2)(%rsp), %xmm2, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm2, (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp), %ymm2
+ vmovdqa %xmm2, (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE*3)(%rsp), %xmm3, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm3, (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp), %ymm3
+ vmovdqa %xmm3, (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE*4)(%rsp), %xmm4, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm4, (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp), %ymm4
+ vmovdqa %xmm4, (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE*5)(%rsp), %xmm5, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm5, (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp), %ymm5
+ vmovdqa %xmm5, (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE*6)(%rsp), %xmm6, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm6, (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp), %ymm6
+ vmovdqa %xmm6, (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp)
+
+1: vpcmpeqq (LR_SIZE + XMM_SIZE*7)(%rsp), %xmm7, %xmm8
+ vpmovmskb %xmm8, %esi
+ cmpl $0xffff, %esi
+ je 2f
+ vmovdqa %xmm7, (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp)
+ jmp 1f
+2: vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp), %ymm7
+ vmovdqa %xmm7, (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp)
+
+1:
+#endif
+ movq 16(%rbx), %r10 # Anything in framesize?
+ testq %r10, %r10
+ jns 3f
+
+ /* There's nothing in the frame size, so there
+ will be no call to the _dl_call_pltexit. */
+
+ /* Get back registers content. */
+ movq LR_RCX_OFFSET(%rsp), %rcx
+ movq LR_RSI_OFFSET(%rsp), %rsi
+ movq LR_RDI_OFFSET(%rsp), %rdi
+
+ movq %rbx, %rsp
+ movq (%rsp), %rbx
+ cfi_restore(rbx)
+ cfi_def_cfa_register(%rsp)
+
+ addq $48, %rsp # Adjust the stack to the return value
+ # (eats the reloc index and link_map)
+ cfi_adjust_cfa_offset(-48)
+ jmp *%r11 # Jump to function address.
+
+3:
+ cfi_adjust_cfa_offset(48)
+ cfi_rel_offset(%rbx, 0)
+ cfi_def_cfa_register(%rbx)
+
+ /* At this point we need to prepare new stack for the function
+ which has to be called. We copy the original stack to a
+ temporary buffer of the size specified by the 'framesize'
+ returned from _dl_profile_fixup */
+
+ leaq LR_RSP_OFFSET(%rbx), %rsi # stack
+ addq $8, %r10
+ andq $0xfffffffffffffff0, %r10
+ movq %r10, %rcx
+ subq %r10, %rsp
+ movq %rsp, %rdi
+ shrq $3, %rcx
+ rep
+ movsq
+
+ movq 24(%rdi), %rcx # Get back register content.
+ movq 32(%rdi), %rsi
+ movq 40(%rdi), %rdi
+
+ call *%r11
+
+ mov 24(%rbx), %rsp # Drop the copied stack content
+
+ /* Now we have to prepare the La_x86_64_retval structure for the
+ _dl_call_pltexit. The La_x86_64_regs is being pointed by rsp now,
+ so we just need to allocate the sizeof(La_x86_64_retval) space on
+ the stack, since the alignment has already been taken care of. */
+# ifdef RESTORE_AVX
+ /* sizeof(La_x86_64_retval). Need extra space for 2 SSE
+ registers to detect if xmm0/xmm1 registers are changed
+ by audit module. */
+ subq $(LRV_SIZE + XMM_SIZE*2), %rsp
+# else
+ subq $LRV_SIZE, %rsp # sizeof(La_x86_64_retval)
+# endif
+ movq %rsp, %rcx # La_x86_64_retval argument to %rcx.
+
+ /* Fill in the La_x86_64_retval structure. */
+ movq %rax, LRV_RAX_OFFSET(%rcx)
+ movq %rdx, LRV_RDX_OFFSET(%rcx)
+
+ movaps %xmm0, LRV_XMM0_OFFSET(%rcx)
+ movaps %xmm1, LRV_XMM1_OFFSET(%rcx)
+
+# ifdef RESTORE_AVX
+ /* This is to support AVX audit modules. */
+ vmovdqu %ymm0, LRV_VECTOR0_OFFSET(%rcx)
+ vmovdqu %ymm1, LRV_VECTOR1_OFFSET(%rcx)
+
+ /* Save xmm0/xmm1 registers to detect if they are changed
+ by audit module. */
+ vmovdqa %xmm0, (LRV_SIZE)(%rcx)
+ vmovdqa %xmm1, (LRV_SIZE + XMM_SIZE)(%rcx)
+# endif
+
+ fstpt LRV_ST0_OFFSET(%rcx)
+ fstpt LRV_ST1_OFFSET(%rcx)
+
+ movq 24(%rbx), %rdx # La_x86_64_regs argument to %rdx.
+ movq 40(%rbx), %rsi # Copy args pushed by PLT in register.
+ movq 32(%rbx), %rdi # %rdi: link_map, %rsi: reloc_index
+ call _dl_call_pltexit
+
+ /* Restore return registers. */
+ movq LRV_RAX_OFFSET(%rsp), %rax
+ movq LRV_RDX_OFFSET(%rsp), %rdx
+
+ movaps LRV_XMM0_OFFSET(%rsp), %xmm0
+ movaps LRV_XMM1_OFFSET(%rsp), %xmm1
+
+# ifdef RESTORE_AVX
+ /* Check if xmm0/xmm1 registers are changed by audit module. */
+ vpcmpeqq (LRV_SIZE)(%rsp), %xmm0, %xmm2
+ vpmovmskb %xmm2, %esi
+ cmpl $0xffff, %esi
+ jne 1f
+ vmovdqu LRV_VECTOR0_OFFSET(%rsp), %ymm0
+
+1: vpcmpeqq (LRV_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm2
+ vpmovmskb %xmm2, %esi
+ cmpl $0xffff, %esi
+ jne 1f
+ vmovdqu LRV_VECTOR1_OFFSET(%rsp), %ymm1
+
+1:
+# endif
+
+ fldt LRV_ST1_OFFSET(%rsp)
+ fldt LRV_ST0_OFFSET(%rsp)
+
+ movq %rbx, %rsp
+ movq (%rsp), %rbx
+ cfi_restore(rbx)
+ cfi_def_cfa_register(%rsp)
+
+ addq $48, %rsp # Adjust the stack to the return value
+ # (eats the reloc index and link_map)
+ cfi_adjust_cfa_offset(-48)
+ retq
http://sources.redhat.com/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=fc1870e6a484ad3211648c9ae51bc076913518aa
commit fc1870e6a484ad3211648c9ae51bc076913518aa
Author: H.J. Lu <hongjiu.lu@intel.com>
Date: Sat Aug 8 10:29:29 2009 -0700
Use PIC only if SHARED is defined.
For ia32, we can avoid PIC in IFUNC functions in static library.
diff --git a/ChangeLog b/ChangeLog
index 0a3df8a..072af87 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+2009-08-07 H.J. Lu <hongjiu.lu@intel.com>
+
+ * sysdeps/i386/i686/multiarch/strcspn.S (STRCSPN): Use PIC
+ only if SHARED is defined.
+ * sysdeps/i386/i686/multiarch/strspn.S (strspn): Likewise.
+
2009-08-03 Jim Meyering <meyering@redhat.com>
* sysdeps/i386/configure.in: Use AC_HEADER_CHECK.
diff --git a/sysdeps/i386/i686/multiarch/strcspn.S b/sysdeps/i386/i686/multiarch/strcspn.S
index ca9c3f5..73e7eb4 100644
--- a/sysdeps/i386/i686/multiarch/strcspn.S
+++ b/sysdeps/i386/i686/multiarch/strcspn.S
@@ -42,6 +42,7 @@
define multiple versions for strpbrk in static library since we
need strpbrk before the initialization happened. */
#if (defined SHARED || !defined USE_AS_STRPBRK) && !defined NOT_IN_libc
+# ifdef SHARED
.section .gnu.linkonce.t.__i686.get_pc_thunk.bx,"ax",@progbits
.globl __i686.get_pc_thunk.bx
.hidden __i686.get_pc_thunk.bx
@@ -71,6 +72,20 @@ ENTRY(STRCSPN)
cfi_restore (ebx)
ret
END(STRCSPN)
+# else
+ .text
+ENTRY(STRCSPN)
+ .type STRCSPN, @gnu_indirect_function
+ cmpl $0, KIND_OFFSET+__cpu_features
+ jne 1f
+ call __init_cpu_features
+1: leal STRCSPN_IA32, %eax
+ testl $(1<<20), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET+__cpu_features
+ jz 2f
+ leal STRCSPN_SSE42, %eax
+2: ret
+END(STRCSPN)
+# endif
# undef ENTRY
# define ENTRY(name) \
diff --git a/sysdeps/i386/i686/multiarch/strspn.S b/sysdeps/i386/i686/multiarch/strspn.S
index f11d3ea..f306d2d 100644
--- a/sysdeps/i386/i686/multiarch/strspn.S
+++ b/sysdeps/i386/i686/multiarch/strspn.S
@@ -27,6 +27,7 @@
/* Define multiple versions only for the definition in libc. */
#ifndef NOT_IN_libc
+# ifdef SHARED
.section .gnu.linkonce.t.__i686.get_pc_thunk.bx,"ax",@progbits
.globl __i686.get_pc_thunk.bx
.hidden __i686.get_pc_thunk.bx
@@ -56,6 +57,20 @@ ENTRY(strspn)
cfi_restore (ebx)
ret
END(strspn)
+# else
+ .text
+ENTRY(strspn)
+ .type strspn, @gnu_indirect_function
+ cmpl $0, KIND_OFFSET+__cpu_features
+ jne 1f
+ call __init_cpu_features
+1: leal __strspn_ia32, %eax
+ testl $(1<<20), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET+__cpu_features
+ jz 2f
+ leal __strspn_sse42, %eax
+2: ret
+END(strspn)
+# endif
# undef ENTRY
# define ENTRY(name) \
http://sources.redhat.com/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=a3d4298d8354f4ea232f027ba2ebfd560afecc1b
commit a3d4298d8354f4ea232f027ba2ebfd560afecc1b
Author: Jim Meyering <meyering@redhat.com>
Date: Sat Aug 8 10:25:58 2009 -0700
Simplify cpuid.h configure test.
diff --git a/ChangeLog b/ChangeLog
index 524a3aa..0a3df8a 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,7 @@
+2009-08-03 Jim Meyering <meyering@redhat.com>
+
+ * sysdeps/i386/configure.in: Use AC_HEADER_CHECK.
+
2009-08-08 Ulrich Drepper <drepper@redhat.com>
* sysdeps/x86_64/multiarch/strlen.S: Move SSE4.2 version into the same
diff --git a/sysdeps/i386/configure b/sysdeps/i386/configure
index ced0b31..d1c4f7f 100755
--- a/sysdeps/i386/configure
+++ b/sysdeps/i386/configure
@@ -1,13 +1,468 @@
+# Factoring default headers for most tests.
+ac_includes_default="\
+#include <stdio.h>
+#ifdef HAVE_SYS_TYPES_H
+# include <sys/types.h>
+#endif
+#ifdef HAVE_SYS_STAT_H
+# include <sys/stat.h>
+#endif
+#ifdef STDC_HEADERS
+# include <stdlib.h>
+# include <stddef.h>
+#else
+# ifdef HAVE_STDLIB_H
+# include <stdlib.h>
+# endif
+#endif
+#ifdef HAVE_STRING_H
+# if !defined STDC_HEADERS && defined HAVE_MEMORY_H
+# include <memory.h>
+# endif
+# include <string.h>
+#endif
+#ifdef HAVE_STRINGS_H
+# include <strings.h>
+#endif
+#ifdef HAVE_INTTYPES_H
+# include <inttypes.h>
+#endif
+#ifdef HAVE_STDINT_H
+# include <stdint.h>
+#endif
+#ifdef HAVE_UNISTD_H
+# include <unistd.h>
+#endif"
+
# This file is generated from configure.in by Autoconf. DO NOT EDIT!
# Local configure fragment for sysdeps/i386.
-{ echo "$as_me:$LINENO: checking if gcc provides <cpuid.h>" >&5
-echo $ECHO_N "checking if gcc provides <cpuid.h>... $ECHO_C" >&6; }
-if test "${libc_cv_gcc_cpuid+set}" = set; then
+
+{ echo "$as_me:$LINENO: checking for grep that handles long lines and -e" >&5
+echo $ECHO_N "checking for grep that handles long lines and -e... $ECHO_C" >&6; }
+if test "${ac_cv_path_GREP+set}" = set; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+ # Extract the first word of "grep ggrep" to use in msg output
+if test -z "$GREP"; then
+set dummy grep ggrep; ac_prog_name=$2
+if test "${ac_cv_path_GREP+set}" = set; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+ ac_path_GREP_found=false
+# Loop through the user's path and test for each of PROGNAME-LIST
+as_save_IFS=$IFS; IFS=$PATH_SEPARATOR
+for as_dir in $PATH$PATH_SEPARATOR/usr/xpg4/bin
+do
+ IFS=$as_save_IFS
+ test -z "$as_dir" && as_dir=.
+ for ac_prog in grep ggrep; do
+ for ac_exec_ext in '' $ac_executable_extensions; do
+ ac_path_GREP="$as_dir/$ac_prog$ac_exec_ext"
+ { test -f "$ac_path_GREP" && $as_test_x "$ac_path_GREP"; } || continue
+ # Check for GNU ac_path_GREP and select it if it is found.
+ # Check for GNU $ac_path_GREP
+case `"$ac_path_GREP" --version 2>&1` in
+*GNU*)
+ ac_cv_path_GREP="$ac_path_GREP" ac_path_GREP_found=:;;
+*)
+ ac_count=0
+ echo $ECHO_N "0123456789$ECHO_C" >"conftest.in"
+ while :
+ do
+ cat "conftest.in" "conftest.in" >"conftest.tmp"
+ mv "conftest.tmp" "conftest.in"
+ cp "conftest.in" "conftest.nl"
+ echo 'GREP' >> "conftest.nl"
+ "$ac_path_GREP" -e 'GREP$' -e '-(cannot match)-' < "conftest.nl" >"conftest.out" 2>/dev/null || break
+ diff "conftest.out" "conftest.nl" >/dev/null 2>&1 || break
+ ac_count=`expr $ac_count + 1`
+ if test $ac_count -gt ${ac_path_GREP_max-0}; then
+ # Best one so far, save it but keep looking for a better one
+ ac_cv_path_GREP="$ac_path_GREP"
+ ac_path_GREP_max=$ac_count
+ fi
+ # 10*(2^10) chars as input seems more than enough
+ test $ac_count -gt 10 && break
+ done
+ rm -f conftest.in conftest.tmp conftest.nl conftest.out;;
+esac
+
+
+ $ac_path_GREP_found && break 3
+ done
+done
+
+done
+IFS=$as_save_IFS
+
+
+fi
+
+GREP="$ac_cv_path_GREP"
+if test -z "$GREP"; then
+ { { echo "$as_me:$LINENO: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&5
+echo "$as_me: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&2;}
+ { (exit 1); exit 1; }; }
+fi
+
+else
+ ac_cv_path_GREP=$GREP
+fi
+
+
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_path_GREP" >&5
+echo "${ECHO_T}$ac_cv_path_GREP" >&6; }
+ GREP="$ac_cv_path_GREP"
+
+
+{ echo "$as_me:$LINENO: checking for egrep" >&5
+echo $ECHO_N "checking for egrep... $ECHO_C" >&6; }
+if test "${ac_cv_path_EGREP+set}" = set; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+ if echo a | $GREP -E '(a|b)' >/dev/null 2>&1
+ then ac_cv_path_EGREP="$GREP -E"
+ else
+ # Extract the first word of "egrep" to use in msg output
+if test -z "$EGREP"; then
+set dummy egrep; ac_prog_name=$2
+if test "${ac_cv_path_EGREP+set}" = set; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+ ac_path_EGREP_found=false
+# Loop through the user's path and test for each of PROGNAME-LIST
+as_save_IFS=$IFS; IFS=$PATH_SEPARATOR
+for as_dir in $PATH$PATH_SEPARATOR/usr/xpg4/bin
+do
+ IFS=$as_save_IFS
+ test -z "$as_dir" && as_dir=.
+ for ac_prog in egrep; do
+ for ac_exec_ext in '' $ac_executable_extensions; do
+ ac_path_EGREP="$as_dir/$ac_prog$ac_exec_ext"
+ { test -f "$ac_path_EGREP" && $as_test_x "$ac_path_EGREP"; } || continue
+ # Check for GNU ac_path_EGREP and select it if it is found.
+ # Check for GNU $ac_path_EGREP
+case `"$ac_path_EGREP" --version 2>&1` in
+*GNU*)
+ ac_cv_path_EGREP="$ac_path_EGREP" ac_path_EGREP_found=:;;
+*)
+ ac_count=0
+ echo $ECHO_N "0123456789$ECHO_C" >"conftest.in"
+ while :
+ do
+ cat "conftest.in" "conftest.in" >"conftest.tmp"
+ mv "conftest.tmp" "conftest.in"
+ cp "conftest.in" "conftest.nl"
+ echo 'EGREP' >> "conftest.nl"
+ "$ac_path_EGREP" 'EGREP$' < "conftest.nl" >"conftest.out" 2>/dev/null || break
+ diff "conftest.out" "conftest.nl" >/dev/null 2>&1 || break
+ ac_count=`expr $ac_count + 1`
+ if test $ac_count -gt ${ac_path_EGREP_max-0}; then
+ # Best one so far, save it but keep looking for a better one
+ ac_cv_path_EGREP="$ac_path_EGREP"
+ ac_path_EGREP_max=$ac_count
+ fi
+ # 10*(2^10) chars as input seems more than enough
+ test $ac_count -gt 10 && break
+ done
+ rm -f conftest.in conftest.tmp conftest.nl conftest.out;;
+esac
+
+
+ $ac_path_EGREP_found && break 3
+ done
+done
+
+done
+IFS=$as_save_IFS
+
+
+fi
+
+EGREP="$ac_cv_path_EGREP"
+if test -z "$EGREP"; then
+ { { echo "$as_me:$LINENO: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&5
+echo "$as_me: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&2;}
+ { (exit 1); exit 1; }; }
+fi
+
+else
+ ac_cv_path_EGREP=$EGREP
+fi
+
+
+ fi
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_path_EGREP" >&5
+echo "${ECHO_T}$ac_cv_path_EGREP" >&6; }
+ EGREP="$ac_cv_path_EGREP"
+
+
+{ echo "$as_me:$LINENO: checking for ANSI C header files" >&5
+echo $ECHO_N "checking for ANSI C header files... $ECHO_C" >&6; }
+if test "${ac_cv_header_stdc+set}" = set; then
echo $ECHO_N "(cached) $ECHO_C" >&6
else
cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+#include <stdlib.h>
+#include <stdarg.h>
+#include <string.h>
+#include <float.h>
+
+int
+main ()
+{
+
+ ;
+ return 0;
+}
+_ACEOF
+rm -f conftest.$ac_objext
+if { (ac_try="$ac_compile"
+case "(($ac_try" in
+ *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+ *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+ (eval "$ac_compile") 2>conftest.er1
+ ac_status=$?
+ grep -v '^ *+' conftest.er1 >conftest.err
+ rm -f conftest.er1
+ cat conftest.err >&5
+ echo "$as_me:$LINENO: \$? = $ac_status" >&5
+ (exit $ac_status); } && {
+ test -z "$ac_c_werror_flag" ||
+ test ! -s conftest.err
+ } && test -s conftest.$ac_objext; then
+ ac_cv_header_stdc=yes
+else
+ echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+ ac_cv_header_stdc=no
+fi
+
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+
+if test $ac_cv_header_stdc = yes; then
+ # SunOS 4.x string.h does not declare mem*, contrary to ANSI.
+ cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+#include <string.h>
+
+_ACEOF
+if (eval "$ac_cpp conftest.$ac_ext") 2>&5 |
+ $EGREP "memchr" >/dev/null 2>&1; then
+ :
+else
+ ac_cv_header_stdc=no
+fi
+rm -f conftest*
+
+fi
+
+if test $ac_cv_header_stdc = yes; then
+ # ISC 2.0.2 stdlib.h does not declare free, contrary to ANSI.
+ cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+#include <stdlib.h>
+
+_ACEOF
+if (eval "$ac_cpp conftest.$ac_ext") 2>&5 |
+ $EGREP "free" >/dev/null 2>&1; then
+ :
+else
+ ac_cv_header_stdc=no
+fi
+rm -f conftest*
+
+fi
+
+if test $ac_cv_header_stdc = yes; then
+ # /bin/cc in Irix-4.0.5 gets non-ANSI ctype macros unless using -ansi.
+ if test "$cross_compiling" = yes; then
+ :
+else
+ cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+#include <ctype.h>
+#include <stdlib.h>
+#if ((' ' & 0x0FF) == 0x020)
+# define ISLOWER(c) ('a' <= (c) && (c) <= 'z')
+# define TOUPPER(c) (ISLOWER(c) ? 'A' + ((c) - 'a') : (c))
+#else
+# define ISLOWER(c) \
+ (('a' <= (c) && (c) <= 'i') \
+ || ('j' <= (c) && (c) <= 'r') \
+ || ('s' <= (c) && (c) <= 'z'))
+# define TOUPPER(c) (ISLOWER(c) ? ((c) | 0x40) : (c))
+#endif
+
+#define XOR(e, f) (((e) && !(f)) || (!(e) && (f)))
+int
+main ()
+{
+ int i;
+ for (i = 0; i < 256; i++)
+ if (XOR (islower (i), ISLOWER (i))
+ || toupper (i) != TOUPPER (i))
+ return 2;
+ return 0;
+}
+_ACEOF
+rm -f conftest$ac_exeext
+if { (ac_try="$ac_link"
+case "(($ac_try" in
+ *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+ *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+ (eval "$ac_link") 2>&5
+ ac_status=$?
+ echo "$as_me:$LINENO: \$? = $ac_status" >&5
+ (exit $ac_status); } && { ac_try='./conftest$ac_exeext'
+ { (case "(($ac_try" in
+ *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+ *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+ (eval "$ac_try") 2>&5
+ ac_status=$?
+ echo "$as_me:$LINENO: \$? = $ac_status" >&5
+ (exit $ac_status); }; }; then
+ :
+else
+ echo "$as_me: program exited with status $ac_status" >&5
+echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+( exit $ac_status )
+ac_cv_header_stdc=no
+fi
+rm -f core *.core core.conftest.* gmon.out bb.out conftest$ac_exeext conftest.$ac_objext conftest.$ac_ext
+fi
+
+
+fi
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_header_stdc" >&5
+echo "${ECHO_T}$ac_cv_header_stdc" >&6; }
+if test $ac_cv_header_stdc = yes; then
+
+cat >>confdefs.h <<\_ACEOF
+#define STDC_HEADERS 1
+_ACEOF
+
+fi
+
+# On IRIX 5.3, sys/types and inttypes.h are conflicting.
+
+
+
+
+
+
+
+
+
+for ac_header in sys/types.h sys/stat.h stdlib.h string.h memory.h strings.h \
+ inttypes.h stdint.h unistd.h
+do
+as_ac_Header=`echo "ac_cv_header_$ac_header" | $as_tr_sh`
+{ echo "$as_me:$LINENO: checking for $ac_header" >&5
+echo $ECHO_N "checking for $ac_header... $ECHO_C" >&6; }
+if { as_var=$as_ac_Header; eval "test \"\${$as_var+set}\" = set"; }; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+ cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+$ac_includes_default
+
+#include <$ac_header>
+_ACEOF
+rm -f conftest.$ac_objext
+if { (ac_try="$ac_compile"
+case "(($ac_try" in
+ *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+ *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+ (eval "$ac_compile") 2>conftest.er1
+ ac_status=$?
+ grep -v '^ *+' conftest.er1 >conftest.err
+ rm -f conftest.er1
+ cat conftest.err >&5
+ echo "$as_me:$LINENO: \$? = $ac_status" >&5
+ (exit $ac_status); } && {
+ test -z "$ac_c_werror_flag" ||
+ test ! -s conftest.err
+ } && test -s conftest.$ac_objext; then
+ eval "$as_ac_Header=yes"
+else
+ echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+ eval "$as_ac_Header=no"
+fi
+
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+ac_res=`eval echo '${'$as_ac_Header'}'`
+ { echo "$as_me:$LINENO: result: $ac_res" >&5
+echo "${ECHO_T}$ac_res" >&6; }
+if test `eval echo '${'$as_ac_Header'}'` = yes; then
+ cat >>confdefs.h <<_ACEOF
+#define `echo "HAVE_$ac_header" | $as_tr_cpp` 1
+_ACEOF
+
+fi
+
+done
+
+
+if test "${ac_cv_header_cpuid_h+set}" = set; then
+ { echo "$as_me:$LINENO: checking for cpuid.h" >&5
+echo $ECHO_N "checking for cpuid.h... $ECHO_C" >&6; }
+if test "${ac_cv_header_cpuid_h+set}" = set; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_header_cpuid_h" >&5
+echo "${ECHO_T}$ac_cv_header_cpuid_h" >&6; }
+else
+ # Is the header compilable?
+{ echo "$as_me:$LINENO: checking cpuid.h usability" >&5
+echo $ECHO_N "checking cpuid.h usability... $ECHO_C" >&6; }
+cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+$ac_includes_default
#include <cpuid.h>
_ACEOF
rm -f conftest.$ac_objext
@@ -27,24 +482,103 @@ eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
- libc_cv_gcc_cpuid=yes
+ ac_header_compiler=yes
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
- libc_cv_gcc_cpuid=no
+ ac_header_compiler=no
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+{ echo "$as_me:$LINENO: result: $ac_header_compiler" >&5
+echo "${ECHO_T}$ac_header_compiler" >&6; }
+
+# Is the header present?
+{ echo "$as_me:$LINENO: checking cpuid.h presence" >&5
+echo $ECHO_N "checking cpuid.h presence... $ECHO_C" >&6; }
+cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h. */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h. */
+#include <cpuid.h>
+_ACEOF
+if { (ac_try="$ac_cpp conftest.$ac_ext"
+case "(($ac_try" in
+ *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+ *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+ (eval "$ac_cpp conftest.$ac_ext") 2>conftest.er1
+ ac_status=$?
+ grep -v '^ *+' conftest.er1 >conftest.err
+ rm -f conftest.er1
+ cat conftest.err >&5
+ echo "$as_me:$LINENO: \$? = $ac_status" >&5
+ (exit $ac_status); } >/dev/null && {
+ test -z "$ac_c_preproc_warn_flag$ac_c_werror_flag" ||
+ test ! -s conftest.err
+ }; then
+ ac_header_preproc=yes
+else
+ echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+ ac_header_preproc=no
fi
-{ echo "$as_me:$LINENO: result: $libc_cv_gcc_cpuid" >&5
-echo "${ECHO_T}$libc_cv_gcc_cpuid" >&6; }
-if test $libc_cv_gcc_cpuid != yes; then
+
+rm -f conftest.err conftest.$ac_ext
+{ echo "$as_me:$LINENO: result: $ac_header_preproc" >&5
+echo "${ECHO_T}$ac_header_preproc" >&6; }
+
+# So? What about this header?
+case $ac_header_compiler:$ac_header_preproc:$ac_c_preproc_warn_flag in
+ yes:no: )
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: accepted by the compiler, rejected by the preprocessor!" >&5
+echo "$as_me: WARNING: cpuid.h: accepted by the compiler, rejected by the preprocessor!" >&2;}
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: proceeding with the compiler's result" >&5
+echo "$as_me: WARNING: cpuid.h: proceeding with the compiler's result" >&2;}
+ ac_header_preproc=yes
+ ;;
+ no:yes:* )
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: present but cannot be compiled" >&5
+echo "$as_me: WARNING: cpuid.h: present but cannot be compiled" >&2;}
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: check for missing prerequisite headers?" >&5
+echo "$as_me: WARNING: cpuid.h: check for missing prerequisite headers?" >&2;}
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: see the Autoconf documentation" >&5
+echo "$as_me: WARNING: cpuid.h: see the Autoconf documentation" >&2;}
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: section \"Present But Cannot Be Compiled\"" >&5
+echo "$as_me: WARNING: cpuid.h: section \"Present But Cannot Be Compiled\"" >&2;}
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: proceeding with the preprocessor's result" >&5
+echo "$as_me: WARNING: cpuid.h: proceeding with the preprocessor's result" >&2;}
+ { echo "$as_me:$LINENO: WARNING: cpuid.h: in the future, the compiler will take precedence" >&5
+echo "$as_me: WARNING: cpuid.h: in the future, the compiler will take precedence" >&2;}
+
+ ;;
+esac
+{ echo "$as_me:$LINENO: checking for cpuid.h" >&5
+echo $ECHO_N "checking for cpuid.h... $ECHO_C" >&6; }
+if test "${ac_cv_header_cpuid_h+set}" = set; then
+ echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+ ac_cv_header_cpuid_h=$ac_header_preproc
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_header_cpuid_h" >&5
+echo "${ECHO_T}$ac_cv_header_cpuid_h" >&6; }
+
+fi
+if test $ac_cv_header_cpuid_h = yes; then
+ :
+else
{ { echo "$as_me:$LINENO: error: gcc must provide the <cpuid.h> header" >&5
echo "$as_me: error: gcc must provide the <cpuid.h> header" >&2;}
{ (exit 1); exit 1; }; }
fi
+
+
{ echo "$as_me:$LINENO: checking if -g produces usable source locations for assembler-with-cpp" >&5
echo $ECHO_N "checking if -g produces usable source locations for assembler-with-cpp... $ECHO_C" >&6; }
if test "${libc_cv_cpp_asm_debuginfo+set}" = set; then
diff --git a/sysdeps/i386/configure.in b/sysdeps/i386/configure.in
index 800f928..12dceaf 100644
--- a/sysdeps/i386/configure.in
+++ b/sysdeps/i386/configure.in
@@ -1,12 +1,8 @@
GLIBC_PROVIDES dnl See aclocal.m4 in the top level source directory.
# Local configure fragment for sysdeps/i386.
-AC_CACHE_CHECK([if gcc provides <cpuid.h>], libc_cv_gcc_cpuid, [dnl
-AC_COMPILE_IFELSE([#include <cpuid.h>], libc_cv_gcc_cpuid=yes,
- libc_cv_gcc_cpuid=no)])
-if test $libc_cv_gcc_cpuid != yes; then
- AC_MSG_ERROR([gcc must provide the <cpuid.h> header])
-fi
+AC_HEADER_CHECK([cpuid.h], ,
+ [AC_MSG_ERROR([gcc must provide the <cpuid.h> header])])
AC_CACHE_CHECK(if -g produces usable source locations for assembler-with-cpp,
libc_cv_cpp_asm_debuginfo, [dnl
http://sources.redhat.com/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=49eea97b00d367aa3eac3dfd047c259e7bd04732
commit 49eea97b00d367aa3eac3dfd047c259e7bd04732
Author: Ulrich Drepper <drepper@redhat.com>
Date: Sat Aug 8 10:21:46 2009 -0700
Another minor optimization of x86-64 pthread_cond_wait.
diff --git a/nptl/ChangeLog b/nptl/ChangeLog
index 0f5c231..48fcc0f 100644
--- a/nptl/ChangeLog
+++ b/nptl/ChangeLog
@@ -1,3 +1,9 @@
+2009-08-08 Ulrich Drepper <drepper@redhat.com>
+
+ * sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S
+ (__pthread_cond_wait): Optimize by avoiding use of callee-safe
+ register.
+
2009-08-07 Ulrich Drepper <drepper@redhat.com>
* sysdeps/unix/sysv/linux/x86_64/sem_wait.S: Little optimizations
diff --git a/nptl/sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S b/nptl/sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S
index 45116b1..f5b929e 100644
--- a/nptl/sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S
+++ b/nptl/sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S
@@ -45,9 +45,6 @@ __pthread_cond_wait:
cfi_lsda(DW_EH_PE_udata4, .LexceptSTART)
#endif
- pushq %r13
- cfi_adjust_cfa_offset(8)
- cfi_rel_offset(%r13, 0)
#define FRAME_SIZE 32
leaq -FRAME_SIZE(%rsp), %rsp
cfi_adjust_cfa_offset(FRAME_SIZE)
@@ -140,7 +137,7 @@ __pthread_cond_wait:
movl $SYS_futex, %eax
syscall
- movl $1, %r13d
+ movl $1, %r8d
#ifdef __ASSUME_REQUEUE_PI
jmp 62f
#else
@@ -158,7 +155,7 @@ __pthread_cond_wait:
#else
orl %fs:PRIVATE_FUTEX, %esi
#endif
-60: xorl %r13d, %r13d
+60: xorl %r8d, %r8d
movl $SYS_futex, %eax
syscall
@@ -233,20 +230,18 @@ __pthread_cond_wait:
/* If requeue_pi is used the kernel performs the locking of the
mutex. */
11: movq 16(%rsp), %rdi
- testl %r13d, %r13d
+ testl %r8d, %r8d
jnz 18f
callq __pthread_mutex_cond_lock
-14: movq FRAME_SIZE(%rsp), %r13
- leaq FRAME_SIZE+8(%rsp), %rsp
- cfi_adjust_cfa_offset(-(FRAME_SIZE + 8))
+14: leaq FRAME_SIZE(%rsp), %rsp
+ cfi_adjust_cfa_offset(-FRAME_SIZE)
/* We return the result of the mutex_lock operation. */
retq
- cfi_adjust_cfa_offset(8 + FRAME_SIZE)
- cfi_rel_offset(%r13, FRAME_SIZE)
+ cfi_adjust_cfa_offset(FRAME_SIZE)
18: callq __pthread_mutex_cond_lock_adjust
xorl %eax, %eax
@@ -341,9 +336,7 @@ versioned_symbol (libpthread, __pthread_cond_wait, pthread_cond_wait,
__condvar_cleanup1:
/* Stack frame:
- rsp + 40
- +--------------------------+
- rsp + 32 | %r13 |
+ rsp + 32
+--------------------------+
rsp + 24 | unused |
+--------------------------+
@@ -465,7 +458,6 @@ __condvar_cleanup1:
callq __pthread_mutex_cond_lock
movq 24(%rsp), %rdi
- movq 32(%rsp), %r13
.LcallUR:
call _Unwind_Resume@PLT
hlt
-----------------------------------------------------------------------
Summary of changes:
ChangeLog | 38 ++
elf/Makefile | 21 +-
elf/tst-audit6.c | 28 +
elf/tst-audit7.c | 1 +
elf/tst-auditmod6a.c | 46 ++
elf/tst-auditmod6b.c | 220 ++++++++
elf/tst-auditmod6c.c | 225 ++++++++
elf/tst-auditmod7a.c | 1 +
elf/tst-auditmod7b.c | 218 ++++++++
nptl/ChangeLog | 6 +
.../unix/sysv/linux/x86_64/pthread_cond_wait.S | 22 +-
sysdeps/i386/configure | 550 +++++++++++++++++++-
sysdeps/i386/configure.in | 8 +-
sysdeps/i386/i686/multiarch/strcspn.S | 15 +
sysdeps/i386/i686/multiarch/strspn.S | 15 +
sysdeps/x86_64/dl-trampoline.S | 244 +---------
sysdeps/x86_64/dl-trampoline.h | 269 ++++++++++
17 files changed, 1659 insertions(+), 268 deletions(-)
create mode 100644 elf/tst-audit6.c
create mode 100644 elf/tst-audit7.c
create mode 100644 elf/tst-auditmod6a.c
create mode 100644 elf/tst-auditmod6b.c
create mode 100644 elf/tst-auditmod6c.c
create mode 100644 elf/tst-auditmod7a.c
create mode 100644 elf/tst-auditmod7b.c
create mode 100644 sysdeps/x86_64/dl-trampoline.h
hooks/post-receive
--
GNU C Library master sources