forked from rpms/glibc
57 lines
1.6 KiB
Diff
57 lines
1.6 KiB
Diff
|
commit e68b1151f7460d5fa88c3a567c13f66052da79a7
|
||
|
|
Author: Paul A. Clarke <pc@us.ibm.com>
|
||
|
|
Date: Thu Sep 19 11:39:44 2019 -0500
|
||
|
|
|
||
|
|
[powerpc] __fesetround_inline optimizations
|
||
|
|
|
||
|
|
On POWER9, use more efficient means to update the 2-bit rounding mode
|
||
|
|
via the 'mffscrn' instruction (instead of two 'mtfsb0/1' instructions
|
||
|
|
or one 'mtfsfi' instruction that modifies 4 bits).
|
||
|
|
|
||
|
|
Suggested-by: Paul E. Murphy <murphyp@linux.ibm.com>
|
||
|
|
Reviewed-By: Paul E Murphy <murphyp@linux.ibm.com>
|
||
|
|
|
||
|
|
diff --git a/sysdeps/powerpc/fpu/fenv_libc.h b/sysdeps/powerpc/fpu/fenv_libc.h
|
||
|
|
index 86ae7fda016abd8b..c3f541c08440b20e 100644
|
||
|
|
--- a/sysdeps/powerpc/fpu/fenv_libc.h
|
||
|
|
+++ b/sysdeps/powerpc/fpu/fenv_libc.h
|
||
|
|
@@ -149,7 +149,12 @@ typedef union
|
||
|
|
static inline int
|
||
|
|
__fesetround_inline (int round)
|
||
|
|
{
|
||
|
|
- if ((unsigned int) round < 2)
|
||
|
|
+#ifdef _ARCH_PWR9
|
||
|
|
+ __fe_mffscrn (round);
|
||
|
|
+#else
|
||
|
|
+ if (__glibc_likely (GLRO(dl_hwcap2) & PPC_FEATURE2_ARCH_3_00))
|
||
|
|
+ __fe_mffscrn (round);
|
||
|
|
+ else if ((unsigned int) round < 2)
|
||
|
|
{
|
||
|
|
asm volatile ("mtfsb0 30");
|
||
|
|
if ((unsigned int) round == 0)
|
||
|
|
@@ -165,7 +170,7 @@ __fesetround_inline (int round)
|
||
|
|
else
|
||
|
|
asm volatile ("mtfsb1 31");
|
||
|
|
}
|
||
|
|
-
|
||
|
|
+#endif
|
||
|
|
return 0;
|
||
|
|
}
|
||
|
|
|
||
|
|
@@ -174,7 +179,14 @@ __fesetround_inline (int round)
|
||
|
|
static inline void
|
||
|
|
__fesetround_inline_nocheck (const int round)
|
||
|
|
{
|
||
|
|
- asm volatile ("mtfsfi 7,%0" : : "i" (round));
|
||
|
|
+#ifdef _ARCH_PWR9
|
||
|
|
+ __fe_mffscrn (round);
|
||
|
|
+#else
|
||
|
|
+ if (__glibc_likely (GLRO(dl_hwcap2) & PPC_FEATURE2_ARCH_3_00))
|
||
|
|
+ __fe_mffscrn (round);
|
||
|
|
+ else
|
||
|
|
+ asm volatile ("mtfsfi 7,%0" : : "i" (round));
|
||
|
|
+#endif
|
||
|
|
}
|
||
|
|
|
||
|
|
#define FPSCR_MASK(bit) (1 << (31 - (bit)))
|