2009-03-19 18:25:46

by Joe Korty

[permalink] [raw]
Subject: [PATCH 2.6.27-stable] Fix misreporting of #cores as #hyperthreads for Q9550

Fix misreporting of #cores for the Intel Quad Core Q9550.

[ for 2.6.27-stable ]

For the Q9550, in x86_64 mode, /proc/cpuinfo mistakenly
reports the #cores present as the #hyperthreads present.
i386 mode was not examined but is assumed to have the
same problem.

A backport of the following three 2.6.29-rc1 patches
fixes the problem:

[PATCH] x86: unmask CPUID levels on Intel CPUs
[PATCH] x86: unmask CPUID levels on Intel CPUs, fix
[PATCH] x86: add MSR_IA32_MISC_ENABLE bits to <asm/msr-index.h>

>From the first patch: "If the CPUID limit bit in
MSR_IA32_MISC_ENABLE is set, clear it to make all CPUID
information available. This is required for some features
to work, in particular XSAVE."

Originally-Developed-by: H. Peter Anvin <[email protected]>
Backported-by: Joe Korty <[email protected]>
Signed-off-by: Joe Korty <[email protected]>

Index: 2.6.27/include/asm-x86/msr-index.h
===================================================================
--- 2.6.27.orig/include/asm-x86/msr-index.h 2009-03-19 11:07:59.000000000 -0400
+++ 2.6.27/include/asm-x86/msr-index.h 2009-03-19 11:08:27.000000000 -0400
@@ -196,6 +196,35 @@
#define MSR_IA32_THERM_STATUS 0x0000019c
#define MSR_IA32_MISC_ENABLE 0x000001a0

+/* MISC_ENABLE bits: architectural */
+#define MSR_IA32_MISC_ENABLE_FAST_STRING (1ULL << 0)
+#define MSR_IA32_MISC_ENABLE_TCC (1ULL << 1)
+#define MSR_IA32_MISC_ENABLE_EMON (1ULL << 7)
+#define MSR_IA32_MISC_ENABLE_BTS_UNAVAIL (1ULL << 11)
+#define MSR_IA32_MISC_ENABLE_PEBS_UNAVAIL (1ULL << 12)
+#define MSR_IA32_MISC_ENABLE_ENHANCED_SPEEDSTEP (1ULL << 16)
+#define MSR_IA32_MISC_ENABLE_MWAIT (1ULL << 18)
+#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID (1ULL << 22)
+#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE (1ULL << 23)
+#define MSR_IA32_MISC_ENABLE_XD_DISABLE (1ULL << 34)
+
+/* MISC_ENABLE bits: model-specific, meaning may vary from core to core */
+#define MSR_IA32_MISC_ENABLE_X87_COMPAT (1ULL << 2)
+#define MSR_IA32_MISC_ENABLE_TM1 (1ULL << 3)
+#define MSR_IA32_MISC_ENABLE_SPLIT_LOCK_DISABLE (1ULL << 4)
+#define MSR_IA32_MISC_ENABLE_L3CACHE_DISABLE (1ULL << 6)
+#define MSR_IA32_MISC_ENABLE_SUPPRESS_LOCK (1ULL << 8)
+#define MSR_IA32_MISC_ENABLE_PREFETCH_DISABLE (1ULL << 9)
+#define MSR_IA32_MISC_ENABLE_FERR (1ULL << 10)
+#define MSR_IA32_MISC_ENABLE_FERR_MULTIPLEX (1ULL << 10)
+#define MSR_IA32_MISC_ENABLE_TM2 (1ULL << 13)
+#define MSR_IA32_MISC_ENABLE_ADJ_PREF_DISABLE (1ULL << 19)
+#define MSR_IA32_MISC_ENABLE_SPEEDSTEP_LOCK (1ULL << 20)
+#define MSR_IA32_MISC_ENABLE_L1D_CONTEXT (1ULL << 24)
+#define MSR_IA32_MISC_ENABLE_DCU_PREF_DISABLE (1ULL << 37)
+#define MSR_IA32_MISC_ENABLE_TURBO_DISABLE (1ULL << 38)
+#define MSR_IA32_MISC_ENABLE_IP_PREF_DISABLE (1ULL << 39)
+
/* Intel Model 6 */
#define MSR_P6_EVNTSEL0 0x00000186
#define MSR_P6_EVNTSEL1 0x00000187
Index: 2.6.27/arch/x86/kernel/cpu/intel.c
===================================================================
--- 2.6.27.orig/arch/x86/kernel/cpu/intel.c 2009-03-19 11:08:26.000000000 -0400
+++ 2.6.27/arch/x86/kernel/cpu/intel.c 2009-03-19 11:08:27.000000000 -0400
@@ -32,6 +32,19 @@

static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
{
+ /* Unmask CPUID levels if masked: */
+ if (c->x86 == 6 && c->x86_model >= 15) {
+ u64 misc_enable;
+
+ rdmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
+
+ if (misc_enable & MSR_IA32_MISC_ENABLE_LIMIT_CPUID) {
+ misc_enable &= ~MSR_IA32_MISC_ENABLE_LIMIT_CPUID;
+ wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
+ c->cpuid_level = cpuid_eax(0);
+ }
+ }
+
/* Netburst reports 64 bytes clflush size, but does IO in 128 bytes */
if (c->x86 == 15 && c->x86_cache_alignment == 64)
c->x86_cache_alignment = 128;
Index: 2.6.27/arch/x86/kernel/cpu/intel_64.c
===================================================================
--- 2.6.27.orig/arch/x86/kernel/cpu/intel_64.c 2009-03-19 11:08:26.000000000 -0400
+++ 2.6.27/arch/x86/kernel/cpu/intel_64.c 2009-03-19 11:08:27.000000000 -0400
@@ -9,6 +9,19 @@

static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
{
+ /* Unmask CPUID levels if masked: */
+ if (c->x86 == 6 && c->x86_model >= 15) {
+ u64 misc_enable;
+
+ rdmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
+
+ if (misc_enable & MSR_IA32_MISC_ENABLE_LIMIT_CPUID) {
+ misc_enable &= ~MSR_IA32_MISC_ENABLE_LIMIT_CPUID;
+ wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
+ c->cpuid_level = cpuid_eax(0);
+ }
+ }
+
if ((c->x86 == 0xf && c->x86_model >= 0x03) ||
(c->x86 == 0x6 && c->x86_model >= 0x0e))
set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC);


2009-03-20 00:46:34

by Greg KH

[permalink] [raw]
Subject: Re: [stable] [PATCH 2.6.27-stable] Fix misreporting of #cores as #hyperthreads for Q9550

On Thu, Mar 19, 2009 at 01:27:44PM -0400, Joe Korty wrote:
> Fix misreporting of #cores for the Intel Quad Core Q9550.
>
> [ for 2.6.27-stable ]
>
> For the Q9550, in x86_64 mode, /proc/cpuinfo mistakenly
> reports the #cores present as the #hyperthreads present.
> i386 mode was not examined but is assumed to have the
> same problem.
>
> A backport of the following three 2.6.29-rc1 patches
> fixes the problem:
>
> [PATCH] x86: unmask CPUID levels on Intel CPUs
> [PATCH] x86: unmask CPUID levels on Intel CPUs, fix
> [PATCH] x86: add MSR_IA32_MISC_ENABLE bits to <asm/msr-index.h>

Can you provide the git commit ids for these patches?

thanks,

greg k-h

>
> From the first patch: "If the CPUID limit bit in
> MSR_IA32_MISC_ENABLE is set, clear it to make all CPUID
> information available. This is required for some features
> to work, in particular XSAVE."
>
> Originally-Developed-by: H. Peter Anvin <[email protected]>
> Backported-by: Joe Korty <[email protected]>
> Signed-off-by: Joe Korty <[email protected]>
>
> Index: 2.6.27/include/asm-x86/msr-index.h
> ===================================================================
> --- 2.6.27.orig/include/asm-x86/msr-index.h 2009-03-19 11:07:59.000000000 -0400
> +++ 2.6.27/include/asm-x86/msr-index.h 2009-03-19 11:08:27.000000000 -0400
> @@ -196,6 +196,35 @@
> #define MSR_IA32_THERM_STATUS 0x0000019c
> #define MSR_IA32_MISC_ENABLE 0x000001a0
>
> +/* MISC_ENABLE bits: architectural */
> +#define MSR_IA32_MISC_ENABLE_FAST_STRING (1ULL << 0)
> +#define MSR_IA32_MISC_ENABLE_TCC (1ULL << 1)
> +#define MSR_IA32_MISC_ENABLE_EMON (1ULL << 7)
> +#define MSR_IA32_MISC_ENABLE_BTS_UNAVAIL (1ULL << 11)
> +#define MSR_IA32_MISC_ENABLE_PEBS_UNAVAIL (1ULL << 12)
> +#define MSR_IA32_MISC_ENABLE_ENHANCED_SPEEDSTEP (1ULL << 16)
> +#define MSR_IA32_MISC_ENABLE_MWAIT (1ULL << 18)
> +#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID (1ULL << 22)
> +#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE (1ULL << 23)
> +#define MSR_IA32_MISC_ENABLE_XD_DISABLE (1ULL << 34)
> +
> +/* MISC_ENABLE bits: model-specific, meaning may vary from core to core */
> +#define MSR_IA32_MISC_ENABLE_X87_COMPAT (1ULL << 2)
> +#define MSR_IA32_MISC_ENABLE_TM1 (1ULL << 3)
> +#define MSR_IA32_MISC_ENABLE_SPLIT_LOCK_DISABLE (1ULL << 4)
> +#define MSR_IA32_MISC_ENABLE_L3CACHE_DISABLE (1ULL << 6)
> +#define MSR_IA32_MISC_ENABLE_SUPPRESS_LOCK (1ULL << 8)
> +#define MSR_IA32_MISC_ENABLE_PREFETCH_DISABLE (1ULL << 9)
> +#define MSR_IA32_MISC_ENABLE_FERR (1ULL << 10)
> +#define MSR_IA32_MISC_ENABLE_FERR_MULTIPLEX (1ULL << 10)
> +#define MSR_IA32_MISC_ENABLE_TM2 (1ULL << 13)
> +#define MSR_IA32_MISC_ENABLE_ADJ_PREF_DISABLE (1ULL << 19)
> +#define MSR_IA32_MISC_ENABLE_SPEEDSTEP_LOCK (1ULL << 20)
> +#define MSR_IA32_MISC_ENABLE_L1D_CONTEXT (1ULL << 24)
> +#define MSR_IA32_MISC_ENABLE_DCU_PREF_DISABLE (1ULL << 37)
> +#define MSR_IA32_MISC_ENABLE_TURBO_DISABLE (1ULL << 38)
> +#define MSR_IA32_MISC_ENABLE_IP_PREF_DISABLE (1ULL << 39)
> +
> /* Intel Model 6 */
> #define MSR_P6_EVNTSEL0 0x00000186
> #define MSR_P6_EVNTSEL1 0x00000187
> Index: 2.6.27/arch/x86/kernel/cpu/intel.c
> ===================================================================
> --- 2.6.27.orig/arch/x86/kernel/cpu/intel.c 2009-03-19 11:08:26.000000000 -0400
> +++ 2.6.27/arch/x86/kernel/cpu/intel.c 2009-03-19 11:08:27.000000000 -0400
> @@ -32,6 +32,19 @@
>
> static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
> {
> + /* Unmask CPUID levels if masked: */
> + if (c->x86 == 6 && c->x86_model >= 15) {
> + u64 misc_enable;
> +
> + rdmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
> +
> + if (misc_enable & MSR_IA32_MISC_ENABLE_LIMIT_CPUID) {
> + misc_enable &= ~MSR_IA32_MISC_ENABLE_LIMIT_CPUID;
> + wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
> + c->cpuid_level = cpuid_eax(0);
> + }
> + }
> +
> /* Netburst reports 64 bytes clflush size, but does IO in 128 bytes */
> if (c->x86 == 15 && c->x86_cache_alignment == 64)
> c->x86_cache_alignment = 128;
> Index: 2.6.27/arch/x86/kernel/cpu/intel_64.c
> ===================================================================
> --- 2.6.27.orig/arch/x86/kernel/cpu/intel_64.c 2009-03-19 11:08:26.000000000 -0400
> +++ 2.6.27/arch/x86/kernel/cpu/intel_64.c 2009-03-19 11:08:27.000000000 -0400
> @@ -9,6 +9,19 @@
>
> static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
> {
> + /* Unmask CPUID levels if masked: */
> + if (c->x86 == 6 && c->x86_model >= 15) {
> + u64 misc_enable;
> +
> + rdmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
> +
> + if (misc_enable & MSR_IA32_MISC_ENABLE_LIMIT_CPUID) {
> + misc_enable &= ~MSR_IA32_MISC_ENABLE_LIMIT_CPUID;
> + wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
> + c->cpuid_level = cpuid_eax(0);
> + }
> + }
> +
> if ((c->x86 == 0xf && c->x86_model >= 0x03) ||
> (c->x86 == 0x6 && c->x86_model >= 0x0e))
> set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC);
>
> _______________________________________________
> stable mailing list
> [email protected]
> http://linux.kernel.org/mailman/listinfo/stable

2009-03-20 14:16:37

by Joe Korty

[permalink] [raw]
Subject: Re: [stable] [PATCH 2.6.27-stable] Fix misreporting of #cores as #hyperthreads for Q9550

On Thu, Mar 19, 2009 at 05:32:33PM -0700, Greg KH wrote:
> On Thu, Mar 19, 2009 at 01:27:44PM -0400, Joe Korty wrote:
> > Fix misreporting of #cores for the Intel Quad Core Q9550.
> >
> > [ for 2.6.27-stable ]
> >
> > For the Q9550, in x86_64 mode, /proc/cpuinfo mistakenly
> > reports the #cores present as the #hyperthreads present.
> > i386 mode was not examined but is assumed to have the
> > same problem.
> >
> > A backport of the following three 2.6.29-rc1 patches
> > fixes the problem:
> >
> > [PATCH] x86: unmask CPUID levels on Intel CPUs
> > [PATCH] x86: unmask CPUID levels on Intel CPUs, fix
> > [PATCH] x86: add MSR_IA32_MISC_ENABLE bits to <asm/msr-index.h>
>
> Can you provide the git commit ids for these patches?


066941bd4eeb159307a5d7d795100d0887c00442 [PATCH] x86: unmask CPUID levels on Intel CPUs
99fb4d349db7e7dacb2099c5cc320a9e2d31c1ef [PATCH] x86: unmask CPUID levels on Intel CPUs, fix
bdf21a49bab28f0d9613e8d8724ef9c9168b61b9 [PATCH] x86: add MSR_IA32_MISC_ENABLE bits to <asm/msr-index.h>