[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <477173b7784bc28afb3e53d76ae5ef143917e8dd.1612924255.git.luto@kernel.org>
Date: Tue, 9 Feb 2021 18:33:34 -0800
From: Andy Lutomirski <luto@...nel.org>
To: x86@...nel.org
Cc: LKML <linux-kernel@...r.kernel.org>,
Dave Hansen <dave.hansen@...ux.intel.com>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Yonghong Song <yhs@...com>,
Masami Hiramatsu <mhiramat@...nel.org>,
Andy Lutomirski <luto@...nel.org>
Subject: [PATCH v2 02/14] x86/fault: Skip the AMD erratum #91 workaround on unaffected CPUs
According to the Revision Guide for AMD Athlon™ 64 and AMD Opteron™
Processors, only early revisions of family 0xF are affected. This will
avoid unnecessarily fetching instruction bytes before sending SIGSEGV to
user programs.
Signed-off-by: Andy Lutomirski <luto@...nel.org>
---
arch/x86/mm/fault.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 441c3e9b8971..818902b08c52 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -106,6 +106,15 @@ check_prefetch_opcode(struct pt_regs *regs, unsigned char *instr,
}
}
+static bool is_amd_k8_pre_npt(void)
+{
+ struct cpuinfo_x86 *c = &boot_cpu_data;
+
+ return unlikely(IS_ENABLED(CONFIG_CPU_SUP_AMD) &&
+ c->x86_vendor == X86_VENDOR_AMD &&
+ c->x86 == 0xf && c->x86_model < 0x40);
+}
+
static int
is_prefetch(struct pt_regs *regs, unsigned long error_code, unsigned long addr)
{
@@ -113,6 +122,10 @@ is_prefetch(struct pt_regs *regs, unsigned long error_code, unsigned long addr)
unsigned char *instr;
int prefetch = 0;
+ /* Erratum #91 affects AMD K8, pre-NPT CPUs */
+ if (!is_amd_k8_pre_npt())
+ return 0;
+
/*
* If it was a exec (instruction fetch) fault on NX page, then
* do not ignore the fault:
--
2.29.2
Powered by blists - more mailing lists