Include SSE state in i386 fenv_t (bug 16064).

This patch fixes bug 16064, i386 fenv_t not including SSE state, using
the technique suggested there of storing the state in the existing
__eip field of fenv_t to avoid needing to increase the size of fenv_t
and add new symbol versions.  The included testcase, which previously
failed for i386 (but passed for x86_64), illustrates how the previous
state was buggy.

This patch causes the SSE state to be included *to the extent it is on
x86_64*.  Where some state should logically be included but isn't for
x86_64 (see bug 16068), this patch does not cause it to be included
for i386 either.  The idea is that any patch fixing that bug should
fix it for both x86_64 and i386 at once.

Tested i386 and x86_64.  (I haven't tested the case of a CPU without
SSE2 disabling the test.)

	[BZ #16064]
	* sysdeps/i386/fpu/fegetenv.c: Include <unistd.h>, <ldsodefs.h>
	and <dl-procinfo.h>.
	(__fegetenv): Save SSE state in envp->__eip if supported.
	* sysdeps/i386/fpu/feholdexcpt.c (feholdexcept): Save SSE state in
	envp->__eip if supported.
	* sysdeps/i386/fpu/fesetenv.c: Include <unistd.h>, <ldsodefs.h>
	and <dl-procinfo.h>.
	(__fesetenv): Always set __eip, __cs_selector, __opcode,
	__data_offset and __data_selector in environment to 0.  Set SSE
	state if supported.
	* sysdeps/x86/fpu/Makefile [$(subdir) = math] (tests): Add
	test-fenv-sse.
	[$(subdir) = math] (CFLAGS-test-fenv-sse.c): Add -msse2
	-mfpmath=sse.
	* sysdeps/x86/fpu/test-fenv-sse.c: New file.

3 files changed, 43 insertions, 17 deletions

diff --git a/sysdeps/i386/fpu/fegetenv.c b/sysdeps/i386/fpu/fegetenv.c
index 8dbdb5787a..8c45b6b8cb 100644
--- a/sysdeps/i386/fpu/fegetenv.c
+++ b/sysdeps/i386/fpu/fegetenv.c
@@ -18,6 +18,9 @@
    <http://www.gnu.org/licenses/>.  */
 
 #include <fenv.h>
+#include <unistd.h>
+#include <ldsodefs.h>
+#include <dl-procinfo.h>
 
 int
 __fegetenv (fenv_t *envp)
@@ -28,6 +31,9 @@ __fegetenv (fenv_t *envp)
      would block all exceptions.  */
   __asm__ ("fldenv %0" : : "m" (*envp));
 
+  if ((GLRO(dl_hwcap) & HWCAP_I386_XMM) != 0)
+    __asm__ ("stmxcsr %0" : "=m" (envp->__eip));
+
   /* Success.  */
   return 0;
 }
diff --git a/sysdeps/i386/fpu/feholdexcpt.c b/sysdeps/i386/fpu/feholdexcpt.c
index d475ca850c..dc9d7031ca 100644
--- a/sysdeps/i386/fpu/feholdexcpt.c
+++ b/sysdeps/i386/fpu/feholdexcpt.c
@@ -35,10 +35,10 @@ feholdexcept (fenv_t *envp)
       unsigned int xwork;
 
       /* Get the current control word.  */
-      __asm__ ("stmxcsr %0" : "=m" (*&xwork));
+      __asm__ ("stmxcsr %0" : "=m" (envp->__eip));
 
       /* Set all exceptions to non-stop and clear them.  */
-      xwork = (xwork | 0x1f80) & ~0x3f;
+      xwork = (envp->__eip | 0x1f80) & ~0x3f;
 
       __asm__ ("ldmxcsr %0" : : "m" (*&xwork));
     }
diff --git a/sysdeps/i386/fpu/fesetenv.c b/sysdeps/i386/fpu/fesetenv.c
index 95b2f0a1ab..2a84657030 100644
--- a/sysdeps/i386/fpu/fesetenv.c
+++ b/sysdeps/i386/fpu/fesetenv.c
@@ -19,6 +19,9 @@
 
 #include <fenv.h>
 #include <assert.h>
+#include <unistd.h>
+#include <ldsodefs.h>
+#include <dl-procinfo.h>
 
 
 int
@@ -40,21 +43,11 @@ __fesetenv (const fenv_t *envp)
       temp.__control_word |= FE_ALL_EXCEPT;
       temp.__control_word &= ~FE_TOWARDZERO;
       temp.__status_word &= ~FE_ALL_EXCEPT;
-      temp.__eip = 0;
-      temp.__cs_selector = 0;
-      temp.__opcode = 0;
-      temp.__data_offset = 0;
-      temp.__data_selector = 0;
     }
   else if (envp == FE_NOMASK_ENV)
     {
       temp.__control_word &= ~(FE_ALL_EXCEPT | FE_TOWARDZERO);
       temp.__status_word &= ~FE_ALL_EXCEPT;
-      temp.__eip = 0;
-      temp.__cs_selector = 0;
-      temp.__opcode = 0;
-      temp.__data_offset = 0;
-      temp.__data_selector = 0;
     }
   else
     {
@@ -63,15 +56,42 @@ __fesetenv (const fenv_t *envp)
 			      & (FE_ALL_EXCEPT | FE_TOWARDZERO));
       temp.__status_word &= ~FE_ALL_EXCEPT;
       temp.__status_word |= envp->__status_word & FE_ALL_EXCEPT;
-      temp.__eip = envp->__eip;
-      temp.__cs_selector = envp->__cs_selector;
-      temp.__opcode = envp->__opcode;
-      temp.__data_offset = envp->__data_offset;
-      temp.__data_selector = envp->__data_selector;
     }
+  temp.__eip = 0;
+  temp.__cs_selector = 0;
+  temp.__opcode = 0;
+  temp.__data_offset = 0;
+  temp.__data_selector = 0;
 
   __asm__ ("fldenv %0" : : "m" (temp));
 
+  if ((GLRO(dl_hwcap) & HWCAP_I386_XMM) != 0)
+    {
+      unsigned int mxcsr;
+      __asm__ ("stmxcsr %0" : "=m" (mxcsr));
+
+      if (envp == FE_DFL_ENV)
+	{
+	  /* Set mask for SSE MXCSR.  */
+	  mxcsr |= (FE_ALL_EXCEPT << 7);
+	  /* Set rounding to FE_TONEAREST.  */
+	  mxcsr &= ~0x6000;
+	  mxcsr |= (FE_TONEAREST << 3);
+	}
+      else if (envp == FE_NOMASK_ENV)
+	{
+	  /* Do not mask exceptions.  */
+	  mxcsr &= ~(FE_ALL_EXCEPT << 7);
+	  /* Set rounding to FE_TONEAREST.  */
+	  mxcsr &= ~0x6000;
+	  mxcsr |= (FE_TONEAREST << 3);
+	}
+      else
+	mxcsr = envp->__eip;
+
+      __asm__ ("ldmxcsr %0" : : "m" (mxcsr));
+    }
+
   /* Success.  */
   return 0;
 }