| 1 |
commit: 18f6a4706fd8339bf905e5a36d5fcff525915340 |
| 2 |
Author: Mike Pagano <mpagano <AT> gentoo <DOT> org> |
| 3 |
AuthorDate: Sat Mar 21 20:00:01 2015 +0000 |
| 4 |
Commit: Mike Pagano <mpagano <AT> gentoo <DOT> org> |
| 5 |
CommitDate: Sat Mar 21 20:00:01 2015 +0000 |
| 6 |
URL: https://gitweb.gentoo.org/proj/linux-patches.git/commit/?id=18f6a470 |
| 7 |
|
| 8 |
Update gcc >= 4.9 optimization patch. See bug #544028. |
| 9 |
|
| 10 |
...-additional-cpu-optimizations-for-gcc-4.9.patch | 67 +++++++++++++--------- |
| 11 |
1 file changed, 41 insertions(+), 26 deletions(-) |
| 12 |
|
| 13 |
diff --git a/5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch b/5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch |
| 14 |
index f931f75..c4efd06 100644 |
| 15 |
--- a/5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch |
| 16 |
+++ b/5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch |
| 17 |
@@ -18,13 +18,14 @@ should use the newer 'march=bonnell' flag for atom processors. |
| 18 |
I have made that change to this patch set as well. See the following kernel |
| 19 |
bug report to see if I'm right: https://bugzilla.kernel.org/show_bug.cgi?id=77461 |
| 20 |
|
| 21 |
-This patch will expand the number of microarchitectures to include new |
| 22 |
+This patch will expand the number of microarchitectures to include newer |
| 23 |
processors including: AMD K10-family, AMD Family 10h (Barcelona), AMD Family |
| 24 |
14h (Bobcat), AMD Family 15h (Bulldozer), AMD Family 15h (Piledriver), AMD |
| 25 |
Family 16h (Jaguar), Intel 1st Gen Core i3/i5/i7 (Nehalem), Intel 1.5 Gen Core |
| 26 |
i3/i5/i7 (Westmere), Intel 2nd Gen Core i3/i5/i7 (Sandybridge), Intel 3rd Gen |
| 27 |
-Core i3/i5/i7 (Ivybridge), Intel 4th Gen Core i3/i5/i7 (Haswell), and Intel 5th |
| 28 |
-Gen Core i3/i5/i7 (Broadwell). It also offers the compiler the 'native' flag. |
| 29 |
+Core i3/i5/i7 (Ivybridge), Intel 4th Gen Core i3/i5/i7 (Haswell), Intel 5th |
| 30 |
+Gen Core i3/i5/i7 (Broadwell), and the low power Silvermont series of Atom |
| 31 |
+processors (Silvermont). It also offers the compiler the 'native' flag. |
| 32 |
|
| 33 |
Small but real speed increases are measurable using a make endpoint comparing |
| 34 |
a generic kernel to one built with one of the respective microarchs. |
| 35 |
@@ -36,9 +37,9 @@ REQUIREMENTS |
| 36 |
linux version >=3.15 |
| 37 |
gcc version >=4.9 |
| 38 |
|
| 39 |
---- a/arch/x86/include/asm/module.h 2014-08-03 18:25:02.000000000 -0400 |
| 40 |
-+++ b/arch/x86/include/asm/module.h 2014-09-13 09:37:16.721385247 -0400 |
| 41 |
-@@ -15,6 +15,20 @@ |
| 42 |
+--- a/arch/x86/include/asm/module.h 2014-06-16 16:44:27.000000000 -0400 |
| 43 |
++++ b/arch/x86/include/asm/module.h 2015-03-07 03:27:32.556672424 -0500 |
| 44 |
+@@ -15,6 +15,22 @@ |
| 45 |
#define MODULE_PROC_FAMILY "586MMX " |
| 46 |
#elif defined CONFIG_MCORE2 |
| 47 |
#define MODULE_PROC_FAMILY "CORE2 " |
| 48 |
@@ -48,6 +49,8 @@ gcc version >=4.9 |
| 49 |
+#define MODULE_PROC_FAMILY "NEHALEM " |
| 50 |
+#elif defined CONFIG_MWESTMERE |
| 51 |
+#define MODULE_PROC_FAMILY "WESTMERE " |
| 52 |
++#elif defined CONFIG_MSILVERMONT |
| 53 |
++#define MODULE_PROC_FAMILY "SILVERMONT " |
| 54 |
+#elif defined CONFIG_MSANDYBRIDGE |
| 55 |
+#define MODULE_PROC_FAMILY "SANDYBRIDGE " |
| 56 |
+#elif defined CONFIG_MIVYBRIDGE |
| 57 |
@@ -59,7 +62,7 @@ gcc version >=4.9 |
| 58 |
#elif defined CONFIG_MATOM |
| 59 |
#define MODULE_PROC_FAMILY "ATOM " |
| 60 |
#elif defined CONFIG_M686 |
| 61 |
-@@ -33,6 +47,20 @@ |
| 62 |
+@@ -33,6 +49,20 @@ |
| 63 |
#define MODULE_PROC_FAMILY "K7 " |
| 64 |
#elif defined CONFIG_MK8 |
| 65 |
#define MODULE_PROC_FAMILY "K8 " |
| 66 |
@@ -80,8 +83,8 @@ gcc version >=4.9 |
| 67 |
#elif defined CONFIG_MELAN |
| 68 |
#define MODULE_PROC_FAMILY "ELAN " |
| 69 |
#elif defined CONFIG_MCRUSOE |
| 70 |
---- a/arch/x86/Kconfig.cpu 2014-08-03 18:25:02.000000000 -0400 |
| 71 |
-+++ b/arch/x86/Kconfig.cpu 2014-09-13 09:37:16.721385247 -0400 |
| 72 |
+--- a/arch/x86/Kconfig.cpu 2014-06-16 16:44:27.000000000 -0400 |
| 73 |
++++ b/arch/x86/Kconfig.cpu 2015-03-07 03:32:14.337713226 -0500 |
| 74 |
@@ -137,9 +137,8 @@ config MPENTIUM4 |
| 75 |
-Paxville |
| 76 |
-Dempsey |
| 77 |
@@ -185,7 +188,7 @@ gcc version >=4.9 |
| 78 |
---help--- |
| 79 |
|
| 80 |
Select this for Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and |
| 81 |
-@@ -260,14 +318,55 @@ config MCORE2 |
| 82 |
+@@ -260,14 +318,63 @@ config MCORE2 |
| 83 |
family in /proc/cpuinfo. Newer ones have 6 and older ones 15 |
| 84 |
(not a typo) |
| 85 |
|
| 86 |
@@ -213,6 +216,14 @@ gcc version >=4.9 |
| 87 |
+ |
| 88 |
+ Enables -march=westmere |
| 89 |
+ |
| 90 |
++config MSILVERMONT |
| 91 |
++ bool "Intel Silvermont" |
| 92 |
++ ---help--- |
| 93 |
++ |
| 94 |
++ Select this for the Intel Silvermont platform. |
| 95 |
++ |
| 96 |
++ Enables -march=silvermont |
| 97 |
++ |
| 98 |
+config MSANDYBRIDGE |
| 99 |
+ bool "Intel Sandy Bridge" |
| 100 |
+ ---help--- |
| 101 |
@@ -247,7 +258,7 @@ gcc version >=4.9 |
| 102 |
|
| 103 |
config GENERIC_CPU |
| 104 |
bool "Generic-x86-64" |
| 105 |
-@@ -276,6 +375,19 @@ config GENERIC_CPU |
| 106 |
+@@ -276,6 +383,19 @@ config GENERIC_CPU |
| 107 |
Generic x86-64 CPU. |
| 108 |
Run equally well on all x86-64 CPUs. |
| 109 |
|
| 110 |
@@ -267,53 +278,53 @@ gcc version >=4.9 |
| 111 |
endchoice |
| 112 |
|
| 113 |
config X86_GENERIC |
| 114 |
-@@ -300,7 +412,7 @@ config X86_INTERNODE_CACHE_SHIFT |
| 115 |
+@@ -300,7 +420,7 @@ config X86_INTERNODE_CACHE_SHIFT |
| 116 |
config X86_L1_CACHE_SHIFT |
| 117 |
int |
| 118 |
default "7" if MPENTIUM4 || MPSC |
| 119 |
- default "6" if MK7 || MK8 || MPENTIUMM || MCORE2 || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU |
| 120 |
-+ default "6" if MK7 || MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MJAGUAR || MPENTIUMM || MCORE2 || MNEHALEM || MWESTMERE || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || BROADWELL || MNATIVE || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU |
| 121 |
++ default "6" if MK7 || MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MJAGUAR || MPENTIUMM || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || BROADWELL || MNATIVE || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU |
| 122 |
default "4" if MELAN || M486 || MGEODEGX1 |
| 123 |
default "5" if MWINCHIP3D || MWINCHIPC6 || MCRUSOE || MEFFICEON || MCYRIXIII || MK6 || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || M586 || MVIAC3_2 || MGEODE_LX |
| 124 |
|
| 125 |
-@@ -331,11 +443,11 @@ config X86_ALIGNMENT_16 |
| 126 |
+@@ -331,11 +451,11 @@ config X86_ALIGNMENT_16 |
| 127 |
|
| 128 |
config X86_INTEL_USERCOPY |
| 129 |
def_bool y |
| 130 |
- depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK7 || MEFFICEON || MCORE2 |
| 131 |
-+ depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK8SSE3 || MK7 || MEFFICEON || MCORE2 || MK10 || MBARCELONA || MNEHALEM || MWESTMERE || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MNATIVE |
| 132 |
++ depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK8SSE3 || MK7 || MEFFICEON || MCORE2 || MK10 || MBARCELONA || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MNATIVE |
| 133 |
|
| 134 |
config X86_USE_PPRO_CHECKSUM |
| 135 |
def_bool y |
| 136 |
- depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MATOM |
| 137 |
-+ depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MK10 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MATOM || MNATIVE |
| 138 |
++ depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MK10 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MATOM || MNATIVE |
| 139 |
|
| 140 |
config X86_USE_3DNOW |
| 141 |
def_bool y |
| 142 |
-@@ -359,17 +471,17 @@ config X86_P6_NOP |
| 143 |
+@@ -359,17 +479,17 @@ config X86_P6_NOP |
| 144 |
|
| 145 |
config X86_TSC |
| 146 |
def_bool y |
| 147 |
- depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MATOM) || X86_64 |
| 148 |
-+ depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MNATIVE || MATOM) || X86_64 |
| 149 |
++ depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MNATIVE || MATOM) || X86_64 |
| 150 |
|
| 151 |
config X86_CMPXCHG64 |
| 152 |
def_bool y |
| 153 |
- depends on X86_PAE || X86_64 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM |
| 154 |
-+ depends on X86_PAE || X86_64 || MCORE2 || MNEHALEM || MWESTMERE || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM || MNATIVE |
| 155 |
++ depends on X86_PAE || X86_64 || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM || MNATIVE |
| 156 |
|
| 157 |
# this should be set for all -march=.. options where the compiler |
| 158 |
# generates cmov. |
| 159 |
config X86_CMOV |
| 160 |
def_bool y |
| 161 |
- depends on (MK8 || MK7 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MATOM || MGEODE_LX) |
| 162 |
-+ depends on (MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MJAGUAR || MK7 || MCORE2 || MNEHALEM || MWESTMERE || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MNATIVE || MATOM || MGEODE_LX) |
| 163 |
++ depends on (MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MJAGUAR || MK7 || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MNATIVE || MATOM || MGEODE_LX) |
| 164 |
|
| 165 |
config X86_MINIMUM_CPU_FAMILY |
| 166 |
int |
| 167 |
---- a/arch/x86/Makefile 2014-08-03 18:25:02.000000000 -0400 |
| 168 |
-+++ b/arch/x86/Makefile 2014-09-13 09:37:16.721385247 -0400 |
| 169 |
-@@ -92,13 +92,33 @@ else |
| 170 |
+--- a/arch/x86/Makefile 2014-06-16 16:44:27.000000000 -0400 |
| 171 |
++++ b/arch/x86/Makefile 2015-03-07 03:33:27.650843211 -0500 |
| 172 |
+@@ -92,13 +92,35 @@ else |
| 173 |
KBUILD_CFLAGS += $(call cc-option,-mpreferred-stack-boundary=3) |
| 174 |
|
| 175 |
# FIXME - should be integrated in Makefile.cpu (Makefile_32.cpu) |
| 176 |
@@ -337,6 +348,8 @@ gcc version >=4.9 |
| 177 |
+ $(call cc-option,-march=nehalem,$(call cc-option,-mtune=nehalem)) |
| 178 |
+ cflags-$(CONFIG_MWESTMERE) += \ |
| 179 |
+ $(call cc-option,-march=westmere,$(call cc-option,-mtune=westmere)) |
| 180 |
++ cflags-$(CONFIG_MSILVERMONT) += \ |
| 181 |
++ $(call cc-option,-march=silvermont,$(call cc-option,-mtune=silvermont)) |
| 182 |
+ cflags-$(CONFIG_MSANDYBRIDGE) += \ |
| 183 |
+ $(call cc-option,-march=sandybridge,$(call cc-option,-mtune=sandybridge)) |
| 184 |
+ cflags-$(CONFIG_MIVYBRIDGE) += \ |
| 185 |
@@ -350,8 +363,8 @@ gcc version >=4.9 |
| 186 |
cflags-$(CONFIG_GENERIC_CPU) += $(call cc-option,-mtune=generic) |
| 187 |
KBUILD_CFLAGS += $(cflags-y) |
| 188 |
|
| 189 |
---- a/arch/x86/Makefile_32.cpu 2014-08-03 18:25:02.000000000 -0400 |
| 190 |
-+++ b/arch/x86/Makefile_32.cpu 2014-09-13 09:37:16.721385247 -0400 |
| 191 |
+--- a/arch/x86/Makefile_32.cpu 2014-06-16 16:44:27.000000000 -0400 |
| 192 |
++++ b/arch/x86/Makefile_32.cpu 2015-03-07 03:34:15.203586024 -0500 |
| 193 |
@@ -23,7 +23,15 @@ cflags-$(CONFIG_MK6) += -march=k6 |
| 194 |
# Please note, that patches that add -march=athlon-xp and friends are pointless. |
| 195 |
# They make zero difference whatsosever to performance at this time. |
| 196 |
@@ -368,7 +381,7 @@ gcc version >=4.9 |
| 197 |
cflags-$(CONFIG_MCRUSOE) += -march=i686 $(align)-functions=0 $(align)-jumps=0 $(align)-loops=0 |
| 198 |
cflags-$(CONFIG_MEFFICEON) += -march=i686 $(call tune,pentium3) $(align)-functions=0 $(align)-jumps=0 $(align)-loops=0 |
| 199 |
cflags-$(CONFIG_MWINCHIPC6) += $(call cc-option,-march=winchip-c6,-march=i586) |
| 200 |
-@@ -32,8 +40,14 @@ cflags-$(CONFIG_MCYRIXIII) += $(call cc- |
| 201 |
+@@ -32,8 +40,15 @@ cflags-$(CONFIG_MCYRIXIII) += $(call cc- |
| 202 |
cflags-$(CONFIG_MVIAC3_2) += $(call cc-option,-march=c3-2,-march=i686) |
| 203 |
cflags-$(CONFIG_MVIAC7) += -march=i686 |
| 204 |
cflags-$(CONFIG_MCORE2) += -march=i686 $(call tune,core2) |
| 205 |
@@ -376,6 +389,7 @@ gcc version >=4.9 |
| 206 |
- $(call cc-option,-mtune=atom,$(call cc-option,-mtune=generic)) |
| 207 |
+cflags-$(CONFIG_MNEHALEM) += -march=i686 $(call tune,nehalem) |
| 208 |
+cflags-$(CONFIG_MWESTMERE) += -march=i686 $(call tune,westmere) |
| 209 |
++cflags-$(CONFIG_MSILVERMONT) += -march=i686 $(call tune,silvermont) |
| 210 |
+cflags-$(CONFIG_MSANDYBRIDGE) += -march=i686 $(call tune,sandybridge) |
| 211 |
+cflags-$(CONFIG_MIVYBRIDGE) += -march=i686 $(call tune,ivybridge) |
| 212 |
+cflags-$(CONFIG_MHASWELL) += -march=i686 $(call tune,haswell) |
| 213 |
@@ -385,3 +399,4 @@ gcc version >=4.9 |
| 214 |
|
| 215 |
# AMD Elan support |
| 216 |
cflags-$(CONFIG_MELAN) += -march=i486 |
| 217 |
+ |
Archives