8dee2397c0
Backport 3 x86 optimization patches. The first one 'Reversing calculation of __x86_shared_non_temporal_threshold' patch reverse calculation of __x86_shared_non_temporal_threshold from 3/4 of the entire shared cache size of a multi-threaded system to 3/4 of one thread's share of the cache size, and then improve the memcpy performance. The second patch 'Optimizing memcpy for AMD Zen architecture' recomputing the shareable cache as 'L3 per CCX(Core-Complex)' and improve performance for amd. The third patch 'Add Hygon Dhyana support' fix Hygon Dhyana processor CPU Vendor ID detection problem in glibc sysdep module.
39 lines
1.5 KiB
Diff
39 lines
1.5 KiB
Diff
From 2b46bc9b5a148f6da198321a8396a6c2c6a1b070 Mon Sep 17 00:00:00 2001
|
|
From: Feifei Wang1994 <wangfeifei@hygon.cn>
|
|
Date: Tue, 3 Sep 2024 08:30:43 +0000
|
|
Subject: [PATCH] backport-x86-Add-Hygon-support
|
|
|
|
This patch fix Hygon processor CPU Vendor ID detection problem
|
|
in glibc sysdep module, current glibc-2.28 doesn't recognize
|
|
Hygon CPU Vendor ID("HygonGenuine") and sets kind to arch_kind_other,
|
|
which result in incorrect zero value for __cache_sysconf() syscall.
|
|
|
|
This patch add Hygon CPU Vendor ID check, setup kind to arch_kind_amd
|
|
and reuse AMD code path, which lead to correct return value in __cache_sysconf() syscall.
|
|
Test case shows no failure with this patch in Hygon arch.
|
|
|
|
Signed-off-by: Feifei Wang <wangfeifei@hygon.cn>
|
|
---
|
|
sysdeps/x86/cpu-features.c | 5 +++--
|
|
1 file changed, 3 insertions(+), 2 deletions(-)
|
|
|
|
diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c
|
|
index ea0b64fd..4b1a0169 100644
|
|
--- a/sysdeps/x86/cpu-features.c
|
|
+++ b/sysdeps/x86/cpu-features.c
|
|
@@ -344,8 +344,9 @@ init_cpu_features (struct cpu_features *cpu_features)
|
|
cpu_features->feature[index_arch_Prefer_No_AVX512]
|
|
|= bit_arch_Prefer_No_AVX512;
|
|
}
|
|
- /* This spells out "AuthenticAMD". */
|
|
- else if (ebx == 0x68747541 && ecx == 0x444d4163 && edx == 0x69746e65)
|
|
+ /* This spells out "AuthenticAMD" or "HygonGenuine". */
|
|
+ else if ((ebx == 0x68747541 && ecx == 0x444d4163 && edx == 0x69746e65)
|
|
+ || (ebx == 0x6f677948 && ecx == 0x656e6975 && edx == 0x6e65476e))
|
|
{
|
|
unsigned int extended_model, stepping;
|
|
|
|
--
|
|
2.27.0
|
|
|