Skip to contentSkip to conversation

x86, MCE, fix MSR_IA32_MCI_CTL2 CMCI threshold setup

PATCH In review Huang Ying 3 participants 4 messages Mar 29, 2010 x86

lore
Share & tools
lore

Huang Ying

author
#0 lore raw

PATCH 1/1 x86, MCE, fix MSR_IA32_MCI_CTL2 CMCI threshold setup

It is reported that CMCI is not raised when number of corrected error reaches preset threshold. After inspection, it is found that MSR_IA32_MCI_CTL2 threshold field is not setup properly. This patch fixed it.

---
 arch/x86/include/asm/mce.h             |    3 +++
 arch/x86/kernel/cpu/mcheck/mce_intel.c |    1 +
 2 files changed, 4 insertions(+), 0 deletions(-)
arch/x86/include/asm/mce.h +3 -0
index 6c3fdd6..355f298 100644
@@ -38,6 +38,9 @@
 #define MCM_ADDR_MEM	 3	/* memory address */
 #define MCM_ADDR_GENERIC 7	/* generic */
 
+/* CTL2 register defines */
+#define MCI_CTL2_THRESHOLD_MASK	0x7fff
+
 #define MCJ_CTX_MASK		3
 #define MCJ_CTX(flags)		((flags) & MCJ_CTX_MASK)
 #define MCJ_CTX_RANDOM		0    /* inject context: random */
arch/x86/kernel/cpu/mcheck/mce_intel.c +1 -0
index d15df6e..ffe730d 100644
@@ -101,6 +101,7 @@ static void cmci_discover(int banks, int boot)
 			continue;
 		}
 
+		val &= ~MCI_CTL2_THRESHOLD_MASK;
 		val |= CMCI_EN | CMCI_THRESHOLD;
 		wrmsrl(MSR_IA32_MCx_CTL2(i), val);
 		rdmsrl(MSR_IA32_MCx_CTL2(i), val);
Reviewed-by on v1 1/1
#1 lore raw
Quoted 32 lines · Huang Ying

It is reported that CMCI is not raised when number of corrected error reaches preset threshold. After inspection, it is found that MSR_IA32_MCI_CTL2 threshold field is not setup properly. This patch fixed it.

---
 arch/x86/include/asm/mce.h             |    3 +++
 arch/x86/kernel/cpu/mcheck/mce_intel.c |    1 +
 2 files changed, 4 insertions(+), 0 deletions(-)
arch/x86/include/asm/mce.h +3 -0
index 6c3fdd6..355f298 100644
@@ -38,6 +38,9 @@
 #define MCM_ADDR_MEM	 3	/* memory address */
 #define MCM_ADDR_GENERIC 7	/* generic */
 
+/* CTL2 register defines */
+#define MCI_CTL2_THRESHOLD_MASK	0x7fff
+
 #define MCJ_CTX_MASK		3
 #define MCJ_CTX(flags)		((flags) & MCJ_CTX_MASK)
 #define MCJ_CTX_RANDOM		0    /* inject context: random */
arch/x86/kernel/cpu/mcheck/mce_intel.c +0 -0
index d15df6e..ffe730d 100644

arch/x86/kernel/cpu/mcheck/mce_intel.c static void cmci_discover(int banks, int boot) · line 107

Quoted 7 lines
arch/x86/kernel/cpu/mcheck/mce_intel.c +1 -0
@@ -101,6 +101,7 @@ static void cmci_discover(int banks, int boot)
 			continue;
 		}
 
+		val &= ~MCI_CTL2_THRESHOLD_MASK;
 		val |= CMCI_EN | CMCI_THRESHOLD;
 		wrmsrl(MSR_IA32_MCx_CTL2(i), val);
 		rdmsrl(MSR_IA32_MCx_CTL2(i), val);

Hum, it seems that the CTL2 of reported environment had be initialized to have more than CMCI_THRESHOLD(=1) by BIOS etc. Could you explain more about your inspection?

Maybe we could handle this environment if kernel supports APEI's firmware-first for corrected MCE and if this is in that case.

Patch looks good. How about cc-ing stable?

Thanks,
H.Seto

author
#2 lore raw

Hi, Seto san,

Thanks for review.

Quoted 36 lines · Hidetoshi Seto
Quoted 35 lines · Huang Ying

It is reported that CMCI is not raised when number of corrected error reaches preset threshold. After inspection, it is found that MSR_IA32_MCI_CTL2 threshold field is not setup properly. This patch fixed it.

---
 arch/x86/include/asm/mce.h             |    3 +++
 arch/x86/kernel/cpu/mcheck/mce_intel.c |    1 +
 2 files changed, 4 insertions(+), 0 deletions(-)
arch/x86/include/asm/mce.h +3 -0
index 6c3fdd6..355f298 100644
@@ -38,6 +38,9 @@
 #define MCM_ADDR_MEM	 3	/* memory address */
 #define MCM_ADDR_GENERIC 7	/* generic */
 
+/* CTL2 register defines */
+#define MCI_CTL2_THRESHOLD_MASK	0x7fff
+
 #define MCJ_CTX_MASK		3
 #define MCJ_CTX(flags)		((flags) & MCJ_CTX_MASK)
 #define MCJ_CTX_RANDOM		0    /* inject context: random */
arch/x86/kernel/cpu/mcheck/mce_intel.c +0 -0
index d15df6e..ffe730d 100644
@@ -101,6 +101,7 @@ static void cmci_discover(int banks, int boot)
 			continue;
 		}
 
Quoted 8 lines

arch/x86/kernel/cpu/mcheck/mce_intel.c static void cmci_discover(int banks, int boot) · line 107

Quoted 4 lines
arch/x86/kernel/cpu/mcheck/mce_intel.c +1 -0
+		val &= ~MCI_CTL2_THRESHOLD_MASK;
 		val |= CMCI_EN | CMCI_THRESHOLD;
 		wrmsrl(MSR_IA32_MCx_CTL2(i), val);
 		rdmsrl(MSR_IA32_MCx_CTL2(i), val);

Hum, it seems that the CTL2 of reported environment had be initialized to have more than CMCI_THRESHOLD(=1) by BIOS etc. Could you explain more about your inspection?

In CTL2, threshold is initialized to 10 by BIOS.

Quoted 3 lines

Maybe we could handle this environment if kernel supports APEI's firmware-first for corrected MCE and if this is in that case.

I still don't know much about the details of the interaction between BIOS and OS here.

Best Regards,
Huang Ying

#3 lore raw
In reply to Hidetoshi Seto
Quoted 4 lines · Hidetoshi Seto

Maybe we could handle this environment if kernel supports APEI's firmware-first for corrected MCE and if this is in that case.

The problem at hand has nothing to do with firmware first or APEI.

-Andi

Signature · 1 line
[email protected] -- Speaking for myself only.