[v1] x86/cpu/amd: Correct the microcode table for Zenbleed

[PATCH] x86/cpu/amd: Correct the microcode table for Zenbleed

Posted by Andrew Cooper 5 days, 10 hours ago

The good revisions are tied to exact steppings, meaning it's not valid to
match on model number alone, let alone a range.

This is probably only a latent issue.  From public microcode archives, the
followin CPUs exist 17-30-00, 17-60-00, 17-70-00 and would be captured by the
model ranges.  They're likely pre-production steppings, and likely didn't get
Zenbleed microcode, but it's still incorrect to compare them to a different
stepping's revision.

Either way, convert the logic to use x86_match_min_microcode_rev(), which is
the preferred mechanism.

Fixes: 522b1d69219d ("x86/cpu/amd: Add a Zenbleed fix")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Borislav Petkov <bp@alien8.de>
CC: Mario Limonciello <mario.limonciello@amd.com>
CC: x86@kernel.org
CC: linux-kernel@vger.kernel.org
---
 arch/x86/kernel/cpu/amd.c | 29 ++++++++---------------------
 1 file changed, 8 insertions(+), 21 deletions(-)

diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 5d46709c58d0..9721d24727e9 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -951,26 +951,13 @@ static void init_amd_zen1(struct cpuinfo_x86 *c)
 	}
 }
 
-static bool cpu_has_zenbleed_microcode(void)
-{
-	u32 good_rev = 0;
-
-	switch (boot_cpu_data.x86_model) {
-	case 0x30 ... 0x3f: good_rev = 0x0830107b; break;
-	case 0x60 ... 0x67: good_rev = 0x0860010c; break;
-	case 0x68 ... 0x6f: good_rev = 0x08608107; break;
-	case 0x70 ... 0x7f: good_rev = 0x08701033; break;
-	case 0xa0 ... 0xaf: good_rev = 0x08a00009; break;
-
-	default:
-		return false;
-	}
-
-	if (boot_cpu_data.microcode < good_rev)
-		return false;
-
-	return true;
-}
+static const struct x86_cpu_id amd_zenbleed_microcode[] = {
+	ZEN_MODEL_STEP_UCODE(0x17, 0x31, 0x0, 0x0830107b),
+	ZEN_MODEL_STEP_UCODE(0x17, 0x60, 0x1, 0x0860010c),
+	ZEN_MODEL_STEP_UCODE(0x17, 0x68, 0x1, 0x08608107),
+	ZEN_MODEL_STEP_UCODE(0x17, 0x71, 0x0, 0x08701033),
+	ZEN_MODEL_STEP_UCODE(0x17, 0xa0, 0x0, 0x08a00009),
+};
 
 static void zen2_zenbleed_check(struct cpuinfo_x86 *c)
 {
@@ -980,7 +967,7 @@ static void zen2_zenbleed_check(struct cpuinfo_x86 *c)
 	if (!cpu_has(c, X86_FEATURE_AVX))
 		return;
 
-	if (!cpu_has_zenbleed_microcode()) {
+	if (!x86_match_min_microcode_rev(amd_zenbleed_microcode)) {
 		pr_notice_once("Zenbleed: please update your microcode for the most optimal fix\n");
 		msr_set_bit(MSR_AMD64_DE_CFG, MSR_AMD64_DE_CFG_ZEN2_FP_BACKUP_FIX_BIT);
 	} else {

base-commit: ac3fd01e4c1efce8f2c054cdeb2ddd2fc0fb150d
-- 
2.39.5

Re: [PATCH] x86/cpu/amd: Correct the microcode table for Zenbleed

Posted by Andrew Cooper 5 days, 10 hours ago

On 26/11/2025 11:26 am, Andrew Cooper wrote:
> The good revisions are tied to exact steppings, meaning it's not valid to
> match on model number alone, let alone a range.
>
> This is probably only a latent issue.  From public microcode archives, the
> followin CPUs exist 17-30-00, 17-60-00, 17-70-00 and would be captured by the
> model ranges.  They're likely pre-production steppings, and likely didn't get
> Zenbleed microcode, but it's still incorrect to compare them to a different
> stepping's revision.
>
> Either way, convert the logic to use x86_match_min_microcode_rev(), which is
> the preferred mechanism.
>
> Fixes: 522b1d69219d ("x86/cpu/amd: Add a Zenbleed fix")
> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> ---
> CC: Borislav Petkov <bp@alien8.de>
> CC: Mario Limonciello <mario.limonciello@amd.com>
> CC: x86@kernel.org
> CC: linux-kernel@vger.kernel.org
> ---
>  arch/x86/kernel/cpu/amd.c | 29 ++++++++---------------------
>  1 file changed, 8 insertions(+), 21 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
> index 5d46709c58d0..9721d24727e9 100644
> --- a/arch/x86/kernel/cpu/amd.c
> +++ b/arch/x86/kernel/cpu/amd.c
> @@ -951,26 +951,13 @@ static void init_amd_zen1(struct cpuinfo_x86 *c)
>  	}
>  }
>  
> -static bool cpu_has_zenbleed_microcode(void)
> -{
> -	u32 good_rev = 0;
> -
> -	switch (boot_cpu_data.x86_model) {
> -	case 0x30 ... 0x3f: good_rev = 0x0830107b; break;
> -	case 0x60 ... 0x67: good_rev = 0x0860010c; break;
> -	case 0x68 ... 0x6f: good_rev = 0x08608107; break;
> -	case 0x70 ... 0x7f: good_rev = 0x08701033; break;
> -	case 0xa0 ... 0xaf: good_rev = 0x08a00009; break;
> -
> -	default:
> -		return false;
> -	}
> -
> -	if (boot_cpu_data.microcode < good_rev)
> -		return false;
> -
> -	return true;
> -}
> +static const struct x86_cpu_id amd_zenbleed_microcode[] = {
> +	ZEN_MODEL_STEP_UCODE(0x17, 0x31, 0x0, 0x0830107b),
> +	ZEN_MODEL_STEP_UCODE(0x17, 0x60, 0x1, 0x0860010c),
> +	ZEN_MODEL_STEP_UCODE(0x17, 0x68, 0x1, 0x08608107),
> +	ZEN_MODEL_STEP_UCODE(0x17, 0x71, 0x0, 0x08701033),
> +	ZEN_MODEL_STEP_UCODE(0x17, 0xa0, 0x0, 0x08a00009),
> +};

Sorry, this needs a {} terminator to be correct.  I'll send out a v2 in
a bit, assuming there are no other comments.

~Andrew

Re: [PATCH] x86/cpu/amd: Correct the microcode table for Zenbleed

Posted by Borislav Petkov 5 days, 9 hours ago

On Wed, Nov 26, 2025 at 11:33:43AM +0000, Andrew Cooper wrote:
> Sorry, this needs a {} terminator to be correct.  I'll send out a v2 in
> a bit, assuming there are no other comments.

I'm just glad that I'm not the only one to fall for this no-terminator shit.

:-P

In any case, makes sense. It'll have to wait for after the merge window along
with your other patch.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [PATCH] x86/cpu/amd: Correct the microcode table for Zenbleed

Posted by david laight 5 days, 7 hours ago

On Wed, 26 Nov 2025 13:44:53 +0100
Borislav Petkov <bp@alien8.de> wrote:

> On Wed, Nov 26, 2025 at 11:33:43AM +0000, Andrew Cooper wrote:
> > Sorry, this needs a {} terminator to be correct.  I'll send out a v2 in
> > a bit, assuming there are no other comments.  
> 
> I'm just glad that I'm not the only one to fall for this no-terminator shit.

Could you add an #define wrapper, eg

#define x86_match_min_microcode_rev(table) \
	x86_match_min_microcode_rev(table, ARRAY_COUNT(table))

to pass the count into the function?
Then the terminator can be made optional.

The data will be smaller, the code slightly larger.
So a net size gain since the performance can't be critical.

	David

Re: [PATCH] x86/cpu/amd: Correct the microcode table for Zenbleed

Posted by Andrew Cooper 5 days, 9 hours ago

On 26/11/2025 12:44 pm, Borislav Petkov wrote:
> On Wed, Nov 26, 2025 at 11:33:43AM +0000, Andrew Cooper wrote:
>> Sorry, this needs a {} terminator to be correct.  I'll send out a v2 in
>> a bit, assuming there are no other comments.
> I'm just glad that I'm not the only one to fall for this no-terminator shit.
>
> :-P
>
> In any case, makes sense. It'll have to wait for after the merge window along
> with your other patch.

There's no rush.  I'm fixing Xen, and at least ensuring that I have sent
a patch in Linux's direction.

~Andrew

[PATCH v2] x86/cpu/amd: Correct the microcode table for Zenbleed

Posted by Andrew Cooper 5 days, 9 hours ago

The good revisions are tied to exact steppings, meaning it's not valid to
match on model number alone, let alone a range.

This is probably only a latent issue.  From public microcode archives, the
following CPUs exist 17-30-00, 17-60-00, 17-70-00 and would be captured by the
model ranges.  They're likely pre-production steppings, and likely didn't get
Zenbleed microcode, but it's still incorrect to compare them to a different
steppings revision.

Either way, convert the logic to use x86_match_min_microcode_rev(), which is
the preferred mechanism.

Fixes: 522b1d69219d ("x86/cpu/amd: Add a Zenbleed fix")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Borislav Petkov <bp@alien8.de>
CC: Mario Limonciello <mario.limonciello@amd.com>
CC: x86@kernel.org
CC: linux-kernel@vger.kernel.org

v2:
 * Terminate the list with {}.
---
 arch/x86/kernel/cpu/amd.c | 30 +++++++++---------------------
 1 file changed, 9 insertions(+), 21 deletions(-)

diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 5d46709c58d0..a92750f3079a 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -951,26 +951,14 @@ static void init_amd_zen1(struct cpuinfo_x86 *c)
 	}
 }
 
-static bool cpu_has_zenbleed_microcode(void)
-{
-	u32 good_rev = 0;
-
-	switch (boot_cpu_data.x86_model) {
-	case 0x30 ... 0x3f: good_rev = 0x0830107b; break;
-	case 0x60 ... 0x67: good_rev = 0x0860010c; break;
-	case 0x68 ... 0x6f: good_rev = 0x08608107; break;
-	case 0x70 ... 0x7f: good_rev = 0x08701033; break;
-	case 0xa0 ... 0xaf: good_rev = 0x08a00009; break;
-
-	default:
-		return false;
-	}
-
-	if (boot_cpu_data.microcode < good_rev)
-		return false;
-
-	return true;
-}
+static const struct x86_cpu_id amd_zenbleed_microcode[] = {
+	ZEN_MODEL_STEP_UCODE(0x17, 0x31, 0x0, 0x0830107b),
+	ZEN_MODEL_STEP_UCODE(0x17, 0x60, 0x1, 0x0860010c),
+	ZEN_MODEL_STEP_UCODE(0x17, 0x68, 0x1, 0x08608107),
+	ZEN_MODEL_STEP_UCODE(0x17, 0x71, 0x0, 0x08701033),
+	ZEN_MODEL_STEP_UCODE(0x17, 0xa0, 0x0, 0x08a00009),
+	{}
+};
 
 static void zen2_zenbleed_check(struct cpuinfo_x86 *c)
 {
@@ -980,7 +968,7 @@ static void zen2_zenbleed_check(struct cpuinfo_x86 *c)
 	if (!cpu_has(c, X86_FEATURE_AVX))
 		return;
 
-	if (!cpu_has_zenbleed_microcode()) {
+	if (!x86_match_min_microcode_rev(amd_zenbleed_microcode)) {
 		pr_notice_once("Zenbleed: please update your microcode for the most optimal fix\n");
 		msr_set_bit(MSR_AMD64_DE_CFG, MSR_AMD64_DE_CFG_ZEN2_FP_BACKUP_FIX_BIT);
 	} else {
-- 
2.39.5