Skip to contentSkip to conversation

x86/microcode/intel: Check microcode revision before updating sibling threads

PATCH In review Ashok Raj 2 participants 4 messages Feb 16, 2018 – Feb 18, 2018

Subsystems (2)
lore
Share & tools
lore

Ashok Raj

author
#0 lore raw

PATCH 1/1 x86/microcode/intel: Check microcode revision before updating sibling threads

After updating microcode on one of the threads in the core, the thread sibling automatically gets the update since the microcode resources are shared. Check the ucode revision on the cpu before performing a ucode update.

---
 arch/x86/kernel/cpu/microcode/intel.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
arch/x86/kernel/cpu/microcode/intel.c +13 -3
index 09b95a7..036d1db 100644
@@ -776,7 +776,7 @@ static enum ucode_state apply_microcode_intel(int cpu)
 {
 	struct microcode_intel *mc;
 	struct ucode_cpu_info *uci;
-	struct cpuinfo_x86 *c;
+	struct cpuinfo_x86 *c = &cpu_data(cpu);
 	static int prev_rev;
 	u32 rev;
 
@@ -793,6 +793,18 @@ static enum ucode_state apply_microcode_intel(int cpu)
 			return UCODE_NFOUND;
 	}
 
+	rev = intel_get_microcode_revision();
+	/*
+	 * Its possible the microcode got udpated
+	 * because its sibling update was done earlier.
+	 * Skip the udpate in that case.
+	 */
+	if (rev >= mc->hdr.rev) {
+		uci->cpu_sig.rev = rev;
+		c->microcode = rev;
+		return UCODE_OK;
+	}
+
 	/* write microcode via MSR 0x79 */
 	wrmsrl(MSR_IA32_UCODE_WRITE, (unsigned long)mc->bits);
 
@@ -813,8 +825,6 @@ static enum ucode_state apply_microcode_intel(int cpu)
 		prev_rev = rev;
 	}
 
-	c = &cpu_data(cpu);
-
 	uci->cpu_sig.rev = rev;
 	c->microcode = rev;
 
Signature · 1 line
2.7.4
#1 lore raw
Quoted 4 lines · Ashok Raj

After updating microcode on one of the threads in the core, the thread sibling automatically gets the update since the microcode resources are shared. Check the ucode revision on the cpu before performing a ucode update.

s/cpu/CPU

Quoted 26 lines
---
 arch/x86/kernel/cpu/microcode/intel.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
arch/x86/kernel/cpu/microcode/intel.c +2 -1
index 09b95a7..036d1db 100644
@@ -776,7 +776,7 @@ static enum ucode_state apply_microcode_intel(int cpu)
 {
 	struct microcode_intel *mc;
 	struct ucode_cpu_info *uci;
-	struct cpuinfo_x86 *c;
+	struct cpuinfo_x86 *c = &cpu_data(cpu);
 	static int prev_rev;
 	u32 rev;
 
@@ -793,6 +793,18 @@ static enum ucode_state apply_microcode_intel(int cpu)
 			return UCODE_NFOUND;
 	}
 
+	rev = intel_get_microcode_revision();

arch/x86/kernel/cpu/microcode/intel.c static enum ucode_state apply_microcode_intel(int cpu) · line 806

Quoted 10 lines
arch/x86/kernel/cpu/microcode/intel.c +10 -0
+	/*
+	 * Its possible the microcode got udpated
+	 * because its sibling update was done earlier.
+	 * Skip the udpate in that case.
+	 */
+	if (rev >= mc->hdr.rev) {
+		uci->cpu_sig.rev = rev;
+		c->microcode = rev;
+		return UCODE_OK;
+	}

s/udpate
/update

Also, more fundamentally, during microcode early testing, isn't it possible for internal iterations of the microcode to have the same revision, but be different?

This patch would prevent re-loading it - for a seemingly minimal benefit.

Thanks,

Ingo

author
#2 lore raw

Hi Ingo

Quoted 3 lines · Ingo Molnar

Also, more fundamentally, during microcode early testing, isn't it possible for internal iterations of the microcode to have the same revision, but be different?

Atleast we don't do that here. Such debug microcode isn't OS loadable. Anything that comes out ready for OS loading, the version will keep moving.

Quoted 1 line

This patch would prevent re-loading it - for a seemingly minimal benefit.

Cheers,
Ashok

#3 lore raw
Quoted 9 lines · Raj, Ashok

Hi Ingo

Quoted 3 lines · Ingo Molnar

Also, more fundamentally, during microcode early testing, isn't it possible for internal iterations of the microcode to have the same revision, but be different?

Atleast we don't do that here. Such debug microcode isn't OS loadable. Anything that comes out ready for OS loading, the version will keep moving.

Ok.

Thanks,

Ingo