i386: Replace assembly versions of e_logf with generic e_logf.c

This patch replaces i386 assembly versions of e_logf with generic
e_logf.c.  For workload-spec2017.wrf, on Nehalem, it improves
performance by:

                           Before            After     Improvement
reciprocal-throughput      73.3865          40.0454       83%
latency                    90.0985          54.4479       65%

On Skylake, it improves performance by:

                           Before            After     Improvement
reciprocal-throughput      75.1384          22.1452       239%
latency                    91.9441          50.7925       81%

On IvyBridge with --disable-multi-arch, it improves performance by:

                           Before            After     Improvement
reciprocal-throughput      84.5575          28.7879       193%
latency                    103.971          57.5231       80%

	* sysdeps/i386/fpu/e_logf.S: Removed.
	* sysdeps/i386/fpu/e_logf_data.c: Likewise.
	* sysdeps/i386/fpu/w_logf.c: Likewise.
	* sysdeps/i386/i686/fpu/e_logf.S: Likewise.
	* sysdeps/i386/fpu/libm-test-ulps: Updated for generic e_logf.c.
	* sysdeps/i386/i686/fpu/multiarch/libm-test-ulps: Likewise.
	* sysdeps/i386/i686/fpu/multiarch/Makefile (libm-sysdep_routines):
	Add e_logf-sse2.
	(CFLAGS-e_logf-sse2.c): New.
	* sysdeps/i386/i686/fpu/multiarch/e_logf-sse2.c: New file.
	* sysdeps/i386/i686/fpu/multiarch/e_logf.c: Likewise.

4 files changed, 10 insertions, 105 deletions

diff --git a/sysdeps/i386/fpu/e_logf.S b/sysdeps/i386/fpu/e_logf.S
deleted file mode 100644
index de967a31f5..0000000000
--- a/sysdeps/i386/fpu/e_logf.S
+++ /dev/null
@@ -1,93 +0,0 @@
-/*
- * Written by J.T. Conklin <jtc@netbsd.org>.
- * Public domain.
- * Adapted for float by Ulrich Drepper <drepper@cygnus.com>.
- *
- * Changed to use fyl2xp1 for values near 1, <drepper@cygnus.com>.
- */
-
-#include <machine/asm.h>
-
-	.section .rodata.cst8,"aM",@progbits,8
-
-	.p2align 3
-	.type one,@object
-one:	.double 1.0
-	ASM_SIZE_DIRECTIVE(one)
-	/* It is not important that this constant is precise.  It is only
-	   a value which is known to be on the safe side for using the
-	   fyl2xp1 instruction.  */
-	.type limit,@object
-limit:	.double 0.29
-	ASM_SIZE_DIRECTIVE(limit)
-
-
-#ifdef PIC
-# define MO(op) op##@GOTOFF(%edx)
-#else
-# define MO(op) op
-#endif
-
-	.text
-ENTRY(__ieee754_logf)
-	fldln2			// log(2)
-	flds	4(%esp)		// x : log(2)
-	fxam
-	fnstsw
-#ifdef PIC
-	LOAD_PIC_REG (dx)
-#endif
-	fld	%st		// x : x : log(2)
-	sahf
-	jc	3f		// in case x is NaN or +-Inf
-4:	fsubl	MO(one)		// x-1 : x : log(2)
-	fld	%st		// x-1 : x-1 : x : log(2)
-	fabs			// |x-1| : x-1 : x : log(2)
-	fcompl	MO(limit)	// x-1 : x : log(2)
-	fnstsw			// x-1 : x : log(2)
-	andb	$0x45, %ah
-	jz	2f
-	fxam
-	fnstsw
-	andb	$0x45, %ah
-	cmpb	$0x40, %ah
-	jne	5f
-	fabs			// log(1) is +0 in all rounding modes.
-5:	fstp	%st(1)		// x-1 : log(2)
-	fyl2xp1			// log(x)
-	ret
-
-2:	fstp	%st(0)		// x : log(2)
-	fyl2x			// log(x)
-	ret
-
-3:	jp	4b		// in case x is +-Inf
-	fstp	%st(1)
-	fstp	%st(1)
-	ret
-END (__ieee754_logf)
-
-ENTRY(__logf_finite)
-	fldln2			// log(2)
-	flds	4(%esp)		// x : log(2)
-#ifdef PIC
-	LOAD_PIC_REG (dx)
-#endif
-	fld	%st		// x : x : log(2)
-	fsubl	MO(one)		// x-1 : x : log(2)
-	fld	%st		// x-1 : x-1 : x : log(2)
-	fabs			// |x-1| : x-1 : x : log(2)
-	fcompl	MO(limit)	// x-1 : x : log(2)
-	fnstsw			// x-1 : x : log(2)
-	andb	$0x45, %ah
-	jz	2b
-	fxam
-	fnstsw
-	andb	$0x45, %ah
-	cmpb	$0x40, %ah
-	jne	6f
-	fabs			// log(1) is +0 in all rounding modes.
-6:	fstp	%st(1)		// x-1 : log(2)
-	fyl2xp1			// log(x)
-	ret
-END(__logf_finite)
diff --git a/sysdeps/i386/fpu/e_logf_data.c b/sysdeps/i386/fpu/e_logf_data.c
deleted file mode 100644
index 1cc8931700..0000000000
--- a/sysdeps/i386/fpu/e_logf_data.c
+++ /dev/null
@@ -1 +0,0 @@
-/* Not needed.  */
diff --git a/sysdeps/i386/fpu/libm-test-ulps b/sysdeps/i386/fpu/libm-test-ulps
index 0aa18ea9b7..ba9b9ec00f 100644
--- a/sysdeps/i386/fpu/libm-test-ulps
+++ b/sysdeps/i386/fpu/libm-test-ulps
@@ -2000,17 +2000,17 @@ ldouble: 4
 
 Function: "gamma_downward":
 double: 4
-float: 4
+float: 5
 idouble: 4
-ifloat: 4
+ifloat: 5
 ildouble: 7
 ldouble: 7
 
 Function: "gamma_towardzero":
 double: 4
-float: 2
+float: 3
 idouble: 4
-ifloat: 2
+ifloat: 3
 ildouble: 7
 ldouble: 7
 
@@ -2186,20 +2186,20 @@ ldouble: 4
 
 Function: "lgamma_downward":
 double: 4
-float: 4
+float: 5
 float128: 8
 idouble: 4
-ifloat: 4
+ifloat: 5
 ifloat128: 8
 ildouble: 7
 ldouble: 7
 
 Function: "lgamma_towardzero":
 double: 4
-float: 2
+float: 3
 float128: 5
 idouble: 4
-ifloat: 2
+ifloat: 3
 ifloat128: 5
 ildouble: 7
 ldouble: 7
@@ -2641,10 +2641,10 @@ ldouble: 5
 
 Function: "y0_towardzero":
 double: 2
-float: 2
+float: 3
 float128: 3
 idouble: 2
-ifloat: 2
+ifloat: 3
 ifloat128: 3
 ildouble: 5
 ldouble: 5
diff --git a/sysdeps/i386/fpu/w_logf.c b/sysdeps/i386/fpu/w_logf.c
deleted file mode 100644
index ea48d1356e..0000000000
--- a/sysdeps/i386/fpu/w_logf.c
+++ /dev/null
@@ -1 +0,0 @@
-#include <sysdeps/../math/w_logf.c>