kernel_compiler_patch/lite-more-x86-64-ISA-levels-for-kernel-6.16+.patch at master · graysky2/kernel_compiler_patch · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
From 9575536bb72a1d9cec5bb94f12f2cb9127560e00 Mon Sep 17 00:00:00 2001
From: graysky <therealgraysky AT proton DOT me>
Date: Sun Aug 17 04:04:33 2025 -0400

This patch adds additional tunings via new x86-64 ISA levels to the
Linux kernel.

These are selectable under:
	Processor type and features ---> x86-64 compiler ISA level

• x86-64     A value of (1) is the default
• x86-64-v2  A value of (2) brings support for vector
             instructions up to Streaming SIMD Extensions 4.2 (SSE4.2)
	     and Supplemental Streaming SIMD Extensions 3 (SSSE3), the
	     POPCNT instruction, and CMPXCHG16B.
• x86-64-v3  A value of (3) adds vector instructions up to AVX2, MOVBE,
             and additional bit-manipulation instructions.

BENEFITS
Small but real speed increases are measurable using a make endpoint comparing
a generic kernel to one built with one of the respective microarchs.

See the following experimental evidence supporting this statement:
https://github.com/graysky2/kernel_compiler_patch?tab=readme-ov-file#benchmarks

REQUIREMENTS
linux version 6.16+
gcc version >=9.0 or clang version >=9.0

ACKNOWLEDGMENTS
This patch builds on the seminal work by Jeroen.[2]

REFERENCES
1.  https://gitlab.com/x86-psABIs/x86-64-ABI/-/commit/77566eb03bc6a326811cb7e9
2.  http://www.linuxforge.net/docs/linux/linux-gcc.php
---
 arch/x86/Kconfig.cpu | 24 ++++++++++++++++++++++++
 arch/x86/Makefile    |  8 ++++++++
 2 files changed, 32 insertions(+)

diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu
index f928cf6e3252..7bdc62f4620c 100644
--- a/arch/x86/Kconfig.cpu
+++ b/arch/x86/Kconfig.cpu
@@ -281,6 +281,30 @@ config X86_GENERIC
 	  This is really intended for distributors who need more
 	  generic optimizations.

+config X86_64_VERSION
+	int "x86-64 compiler ISA level"
+	range 1 3
+	depends on (CC_IS_GCC && GCC_VERSION > 110000) || (CC_IS_CLANG && CLANG_VERSION >= 120000)
+	depends on X86_64
+	help
+	  Specify a specific x86-64 compiler ISA level.
+
+	  There are three x86-64 ISA levels that work on top of
+	  the x86-64 baseline, namely: x86-64-v2, x86-64-v3, and x86-64-v4.
+
+	  x86-64-v2 brings support for vector instructions up to Streaming SIMD
+	  Extensions 4.2 (SSE4.2) and Supplemental Streaming SIMD Extensions 3
+	  (SSSE3), the POPCNT instruction, and CMPXCHG16B.
+
+	  x86-64-v3 adds vector instructions up to AVX2, MOVBE, and additional
+	  bit-manipulation instructions.
+
+	  x86-64-v4 is not included since the kernel does not use AVX512 instructions
+
+	  You can find the best version for your CPU by running one of the following:
+	  /lib/ld-linux-x86-64.so.2 --help | grep supported
+	  /lib64/ld-linux-x86-64.so.2 --help | grep supported
+
 #
 # Define implied options from the CPU selection here
 config X86_INTERNODE_CACHE_SHIFT
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 1913d342969b..6da7cd5cb3b3 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -181,6 +181,14 @@ else
         KBUILD_RUSTFLAGS += -Ctarget-cpu=x86-64 -Ztune-cpu=generic
 endif

+ifeq ($(CONFIG_X86_64_VERSION),1)
+        KBUILD_CFLAGS += -march=x86-64 -mtune=generic
+        KBUILD_RUSTFLAGS += -Ctarget-cpu=x86-64 -Ztune-cpu=generic
+else
+        KBUILD_CFLAGS +=-march=x86-64-v$(CONFIG_X86_64_VERSION)
+        KBUILD_RUSTFLAGS += -Ctarget-cpu=x86-64-v$(CONFIG_X86_64_VERSION)
+endif
+
         KBUILD_CFLAGS += -mno-red-zone
         KBUILD_CFLAGS += -mcmodel=kernel
         KBUILD_RUSTFLAGS += -Cno-redzone=y
--
2.50.1