* L!TF Bulletin: The state of the horrors
@ 2018-06-21 15:33 Thomas Gleixner
2018-06-22 12:24 ` [MODERATED] " Borislav Petkov
0 siblings, 1 reply; 3+ messages in thread
From: Thomas Gleixner @ 2018-06-21 15:33 UTC (permalink / raw)
To: speck
[-- Attachment #1: Type: text/plain, Size: 452 bytes --]
Hi!
I've updated the repository (master branch and the 4.17/16/14.y stable
branches) with the following items:
- The 1/6 patch from Konrad fixing the pr_fmt issue in bugs.c
- The swap size fix from Vlastimil
- The SMT control knob series with a few tweaks as discussed on the
list
- The 2/6 patch from Konrad adding the L1D_FLUSH cpu feature
Find the latest L!TF combo bundle based on 4.18-rc1 attached.
Thanks,
tglx
[-- Attachment #2: Type: application/octet-stream, Size: 37162 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
* [MODERATED] Re: L!TF Bulletin: The state of the horrors
2018-06-21 15:33 L!TF Bulletin: The state of the horrors Thomas Gleixner
@ 2018-06-22 12:24 ` Borislav Petkov
2018-06-22 13:37 ` Thomas Gleixner
0 siblings, 1 reply; 3+ messages in thread
From: Borislav Petkov @ 2018-06-22 12:24 UTC (permalink / raw)
To: speck
On Thu, Jun 21, 2018 at 05:33:28PM +0200, speck for Thomas Gleixner wrote:
> Hi!
>
> I've updated the repository (master branch and the 4.17/16/14.y stable
> branches) with the following items:
>
> - The 1/6 patch from Konrad fixing the pr_fmt issue in bugs.c
>
> - The swap size fix from Vlastimil
>
> - The SMT control knob series with a few tweaks as discussed on the
> list
>
> - The 2/6 patch from Konrad adding the L1D_FLUSH cpu feature
>
> Find the latest L!TF combo bundle based on 4.18-rc1 attached.
One more ontop, as we talked on IRC:
---
From: Borislav Petkov <bp@suse.de>
Date: Fri, 22 Jun 2018 11:34:11 +0200
Subject: [PATCH] x86/CPU/AMD: Move TOPOEXT reenablement before reading smp_num_siblings
The TOPOEXT reenablement is a workaround for broken BIOSen which didn't
enable the CPUID bit. amd_get_topology_early(), however, relies on
that bit being set so that it can read out the CPUID leaf and set
smp_num_siblings properly.
Move the reenablement up to early_init_amd(). While at it, simplify
amd_get_topology_early().
Signed-off-by: Borislav Petkov <bp@suse.de>
---
arch/x86/kernel/cpu/amd.c | 37 +++++++++++++++++--------------------
1 file changed, 17 insertions(+), 20 deletions(-)
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index b20dee982a53..74061e421e62 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -318,12 +318,8 @@ static void legacy_fixup_core_id(struct cpuinfo_x86 *c)
static void amd_get_topology_early(struct cpuinfo_x86 *c)
{
- if (boot_cpu_has(X86_FEATURE_TOPOEXT)) {
- u32 eax, ebx, ecx, edx;
-
- cpuid(0x8000001e, &eax, &ebx, &ecx, &edx);
- smp_num_siblings = ((ebx >> 8) & 0xff) + 1;
- }
+ if (cpu_has(c, X86_FEATURE_TOPOEXT))
+ smp_num_siblings = ((cpuid_ebx(0x8000001e) >> 8) & 0xff) + 1;
}
/*
@@ -345,7 +341,6 @@ static void amd_get_topology(struct cpuinfo_x86 *c)
cpuid(0x8000001e, &eax, &ebx, &ecx, &edx);
node_id = ecx & 0xff;
- smp_num_siblings = ((ebx >> 8) & 0xff) + 1;
if (c->x86 == 0x15)
c->cu_id = ebx & 0xff;
@@ -622,6 +617,7 @@ static void early_detect_mem_encrypt(struct cpuinfo_x86 *c)
static void early_init_amd(struct cpuinfo_x86 *c)
{
+ u64 value;
u32 dummy;
early_init_amd_mc(c);
@@ -693,6 +689,20 @@ static void early_init_amd(struct cpuinfo_x86 *c)
early_detect_mem_encrypt(c);
+ /* Re-enable TopologyExtensions if switched off by BIOS */
+ if (c->x86 == 0x15 &&
+ (c->x86_model >= 0x10 && c->x86_model <= 0x6f) &&
+ !cpu_has(c, X86_FEATURE_TOPOEXT)) {
+
+ if (msr_set_bit(0xc0011005, 54) > 0) {
+ rdmsrl(0xc0011005, value);
+ if (value & BIT_64(54)) {
+ set_cpu_cap(c, X86_FEATURE_TOPOEXT);
+ pr_info_once(FW_INFO "CPU: Re-enabling disabled Topology Extensions Support.\n");
+ }
+ }
+ }
+
amd_get_topology_early(c);
}
@@ -785,19 +795,6 @@ static void init_amd_bd(struct cpuinfo_x86 *c)
{
u64 value;
- /* re-enable TopologyExtensions if switched off by BIOS */
- if ((c->x86_model >= 0x10) && (c->x86_model <= 0x6f) &&
- !cpu_has(c, X86_FEATURE_TOPOEXT)) {
-
- if (msr_set_bit(0xc0011005, 54) > 0) {
- rdmsrl(0xc0011005, value);
- if (value & BIT_64(54)) {
- set_cpu_cap(c, X86_FEATURE_TOPOEXT);
- pr_info_once(FW_INFO "CPU: Re-enabling disabled Topology Extensions Support.\n");
- }
- }
- }
-
/*
* The way access filter has a performance penalty on some workloads.
* Disable it on the affected CPUs.
--
2.17.0.582.gccdcbd54c
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
--
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: L!TF Bulletin: The state of the horrors
2018-06-22 12:24 ` [MODERATED] " Borislav Petkov
@ 2018-06-22 13:37 ` Thomas Gleixner
0 siblings, 0 replies; 3+ messages in thread
From: Thomas Gleixner @ 2018-06-22 13:37 UTC (permalink / raw)
To: speck
On Fri, 22 Jun 2018, speck for Borislav Petkov wrote:
> One more ontop, as we talked on IRC:
>
> ---
> From: Borislav Petkov <bp@suse.de>
> Date: Fri, 22 Jun 2018 11:34:11 +0200
> Subject: [PATCH] x86/CPU/AMD: Move TOPOEXT reenablement before reading smp_num_siblings
>
> The TOPOEXT reenablement is a workaround for broken BIOSen which didn't
> enable the CPUID bit. amd_get_topology_early(), however, relies on
> that bit being set so that it can read out the CPUID leaf and set
> smp_num_siblings properly.
>
> Move the reenablement up to early_init_amd(). While at it, simplify
> amd_get_topology_early().
Applied to all branches and pushed out.
Thanks
tglx
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2018-06-22 13:37 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-06-21 15:33 L!TF Bulletin: The state of the horrors Thomas Gleixner
2018-06-22 12:24 ` [MODERATED] " Borislav Petkov
2018-06-22 13:37 ` Thomas Gleixner
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.