2013-06-05 20:50:15

by Aravind Gopalakrishnan

[permalink] [raw]
Subject: [PATCH] edac: Add code to handle new MCE signatures for new models of Fam15h

There are some newer error signatures in Family 15h models 30h-3fh.
Code has been added in the patch to support these new signatures.

Patch has been tested on Fam15h using mce_amd_inj facility and
has been verified to work correctly.

Signed-off-by: Aravind Gopalakrishnan <[email protected]>
---
drivers/edac/mce_amd.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c
index f3f0c93..9de24c7 100644
--- a/drivers/edac/mce_amd.c
+++ b/drivers/edac/mce_amd.c
@@ -134,7 +134,8 @@ static const char * const mc5_mce_desc[] = {
"Physical register file AG0 port",
"Physical register file AG1 port",
"Flag register file",
- "DE error occurred"
+ "DE error occurred",
+ "Retire Status Queue Parity"
};

static bool f12h_mc0_mce(u16 ec, u8 xec)
@@ -624,7 +625,7 @@ static void decode_mc5_mce(struct mce *m)

if (xec == 0x0 || xec == 0xc)
pr_cont("%s.\n", mc5_mce_desc[xec]);
- else if (xec < 0xd)
+ else if (xec <= 0xd)
pr_cont("%s parity error.\n", mc5_mce_desc[xec]);
else
goto wrong_mc5_mce;
--
1.7.9.5


2013-06-05 21:33:05

by Borislav Petkov

[permalink] [raw]
Subject: Re: [PATCH] edac: Add code to handle new MCE signatures for new models of Fam15h

On Wed, Jun 05, 2013 at 03:50:03PM -0500, Aravind Gopalakrishnan wrote:
> There are some newer error signatures in Family 15h models 30h-3fh.
> Code has been added in the patch to support these new signatures.
>
> Patch has been tested on Fam15h using mce_amd_inj facility and
> has been verified to work correctly.
>
> Signed-off-by: Aravind Gopalakrishnan <[email protected]>
> ---
> drivers/edac/mce_amd.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c
> index f3f0c93..9de24c7 100644
> --- a/drivers/edac/mce_amd.c
> +++ b/drivers/edac/mce_amd.c
> @@ -134,7 +134,8 @@ static const char * const mc5_mce_desc[] = {
> "Physical register file AG0 port",
> "Physical register file AG1 port",
> "Flag register file",
> - "DE error occurred"
> + "DE error occurred",
> + "Retire Status Queue Parity"

Since "parity error" is already in the printk below, I'm guessing this
string should be "Retire status queue" only?

No need to resend the patch, btw - I'll fix it up when applying.

> };
>
> static bool f12h_mc0_mce(u16 ec, u8 xec)
> @@ -624,7 +625,7 @@ static void decode_mc5_mce(struct mce *m)
>
> if (xec == 0x0 || xec == 0xc)
> pr_cont("%s.\n", mc5_mce_desc[xec]);
> - else if (xec < 0xd)
> + else if (xec <= 0xd)
> pr_cont("%s parity error.\n", mc5_mce_desc[xec]);
> else
> goto wrong_mc5_mce;

--
Regards/Gruss,
Boris.

Sent from a fat crate under my desk. Formatting is fine.
--

2013-06-05 21:55:58

by Aravind Gopalakrishnan

[permalink] [raw]
Subject: Re: [PATCH] edac: Add code to handle new MCE signatures for new models of Fam15h

On 6/5/2013 4:32 PM, Borislav Petkov wrote:
> On Wed, Jun 05, 2013 at 03:50:03PM -0500, Aravind Gopalakrishnan wrote:
>> There are some newer error signatures in Family 15h models 30h-3fh.
>> Code has been added in the patch to support these new signatures.
>>
>> Patch has been tested on Fam15h using mce_amd_inj facility and
>> has been verified to work correctly.
>>
>> Signed-off-by: Aravind Gopalakrishnan <[email protected]>
>> ---
>> drivers/edac/mce_amd.c | 5 +++--
>> 1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c
>> index f3f0c93..9de24c7 100644
>> --- a/drivers/edac/mce_amd.c
>> +++ b/drivers/edac/mce_amd.c
>> @@ -134,7 +134,8 @@ static const char * const mc5_mce_desc[] = {
>> "Physical register file AG0 port",
>> "Physical register file AG1 port",
>> "Flag register file",
>> - "DE error occurred"
>> + "DE error occurred",
>> + "Retire Status Queue Parity"
> Since "parity error" is already in the printk below, I'm guessing this
> string should be "Retire status queue" only?

Okay, that should convey the message too...

> No need to resend the patch, btw - I'll fix it up when applying.

Thanks!

>> };
>>
>> static bool f12h_mc0_mce(u16 ec, u8 xec)
>> @@ -624,7 +625,7 @@ static void decode_mc5_mce(struct mce *m)
>>
>> if (xec == 0x0 || xec == 0xc)
>> pr_cont("%s.\n", mc5_mce_desc[xec]);
>> - else if (xec < 0xd)
>> + else if (xec <= 0xd)
>> pr_cont("%s parity error.\n", mc5_mce_desc[xec]);
>> else
>> goto wrong_mc5_mce;
- Aravind.

2013-06-06 09:46:29

by Borislav Petkov

[permalink] [raw]
Subject: Re: [PATCH] edac: Add code to handle new MCE signatures for new models of Fam15h

On Wed, Jun 05, 2013 at 04:55:40PM -0500, Aravind Gopalakrishnan wrote:
> >Since "parity error" is already in the printk below, I'm guessing this
> >string should be "Retire status queue" only?
>
> Okay, that should convey the message too...

Applied, thanks.

--
Regards/Gruss,
Boris.

Sent from a fat crate under my desk. Formatting is fine.
--