diff options
| author | Philip Guenther <guenther@cvs.openbsd.org> | 2019-05-18 18:11:47 +0000 |
|---|---|---|
| committer | Philip Guenther <guenther@cvs.openbsd.org> | 2019-05-18 18:11:47 +0000 |
| commit | 6b99de2a01f10e8ca1de7101f370a7a0563aeed0 (patch) | |
| tree | df3c194b8336f595fd9f029dd58b05a2d437ca31 | |
| parent | 8deba4de36f85af63dd5ad4fd0893e41ba73f1ed (diff) | |
Both FreeBSD and I appear to have been confused by intel's deep-dive doc:
the RDCL_NO bit of the ARCH_CAPA msr only indicates one of the MDS
subissues is mitigated; only the MDS_NO bit indicates they're all
mitigated.
ok jsg@ mlarkin@
| -rw-r--r-- | sys/arch/amd64/amd64/cpu.c | 5 |
1 files changed, 2 insertions, 3 deletions
diff --git a/sys/arch/amd64/amd64/cpu.c b/sys/arch/amd64/amd64/cpu.c index 9e6c7b854b7..e14c8eaa9b0 100644 --- a/sys/arch/amd64/amd64/cpu.c +++ b/sys/arch/amd64/amd64/cpu.c @@ -1,4 +1,4 @@ -/* $OpenBSD: cpu.c,v 1.135 2019/05/17 19:07:15 guenther Exp $ */ +/* $OpenBSD: cpu.c,v 1.136 2019/05/18 18:11:46 guenther Exp $ */ /* $NetBSD: cpu.c,v 1.1 2003/04/26 18:39:26 fvdl Exp $ */ /*- @@ -214,8 +214,7 @@ replacemds(void) if (strcmp(cpu_vendor, "GenuineIntel") != 0 || ((ci->ci_feature_sefflags_edx & SEFF0EDX_ARCH_CAP) && - (rdmsr(MSR_ARCH_CAPABILITIES) & - (ARCH_CAPABILITIES_RDCL_NO | ARCH_CAPABILITIES_MDS_NO)))) { + (rdmsr(MSR_ARCH_CAPABILITIES) & ARCH_CAPABILITIES_MDS_NO))) { /* Unaffected, nop out the handling code */ has_verw = 0; } else if (ci->ci_feature_sefflags_edx & SEFF0EDX_MD_CLEAR) { |