2012-01-28 Chris Metcalf <cmetcalf@tilera.com>
+ * scripts/firstversions.awk: Fix bug in version range handling.
+
+ * sysdeps/unix/sysv/linux/grantpt.c: Use <> brackets for not-cancel.h.
+
+ * sysdeps/unix/sysv/linux/faccessat.c (faccessat): Call __fxstatat64.
+
+ * include/sys/epoll.h: New file.
+ * sysdeps/unix/sysv/linux/epoll_pwait.c (epoll_pwait): Mark as
+ libc_hidden_def.
+
2012-01-28 Ulrich Drepper <drepper@gmail.com>
+ * sysdeps/x86_64/fpu/bits/mathinline.h (__signbitl): Optimize a bit.
+ Avoid unnecessary __WORDSIZE == 64 test.
+ (fmaxf): Use VEX format if possible.
+ (fmax): Likewise.
+ (fminf): Likewise.
+ (fmin): Likewise.
+
+ * config.h.in: Define HAVE_SSE2AVX_SUPPORT.
+ * math/math_private.h: Remove libc_fegetround* and
+ libc_fesetround*.
+ * sysdeps/i386/configure.in: Check for -msse2avx.
+ * sysdeps/x86_64/fpu/math_private.h: Use VEX-encoded instructions
+ also if SSE2AVX is defined.
+ Remove libc_fegetround* and libc_fesetround*.
+ * sysdeps/x86_64/fpu/multiarch/Makefile: Compile *-avx functions
+ if config-cflags-sse2avx is yes. Also add -DSSE2AVX to defines.
+ * sysdeps/x86_64/fpu/multiarch/e_atan2.c: Use HAS_AVX again instead
+ of HAS_YMM_USABLE.
+ * sysdeps/x86_64/fpu/multiarch/e_exp.c: Likewise.
+ * sysdeps/x86_64/fpu/multiarch/e_log.c: Likewise.
+ * sysdeps/x86_64/fpu/multiarch/s_atan.c: Likewise.
+ * sysdeps/x86_64/fpu/multiarch/s_sin.c: Likewise.
+ * sysdeps/x86_64/fpu/multiarch/s_tan.c: Likewise.
+
* sysdeps/x86_64/fpu/math_private.h: Simplify use of AVX instructions.
2012-01-19 Adhemerval Zanella <azanella@linux.vnet.ibm.com>