[OpenPower-Firmware] poor correctable MC errors logging

[Sergey Kachkin]

Hi Mahesh,

thanks for your reply.

>We can improve it to print CPU pir number. Do you also want location code
info there ?

Yes, I would prefer as much info as possible that may help to distinguish
one MC problem from another and isolate the root cause. So adding more
details would be beneficial.
Am I correct that CPU numbers etc  will be printed for other similar
recoverable errors also?  I have never seen anything except ERAT but
wondering if it could be also SLB / TLB multihit etc.


regards,
Sergey
YADRO


On Fri, Mar 16, 2018 at 6:52 PM, Mahesh Jagannath Salgaonkar <
mahesh at linux.vnet.ibm.com> wrote:

> On 03/14/2018 07:05 PM, Sergey Kachkin wrote:
> > Hi,
> >
> > recently there was a number of HMI logging improvements which may help to
> > isolate the source of HMI errors, but troubleshooting MCs like below is
> > also challenging.
> > Can we have additional logging for MCs also?
>
> We can improve it to print CPU pir number. Do you also want location
> code info there ?
>
>


> >
> >
> >    1. Feb 15 02:56:33 host kernel: Severe Machine check interrupt
> >    [Recovered]
> >    2. Feb 15 02:56:33 host kernel:   Initiator: CPU
> >    3. Feb 15 02:56:33 host kernel:   Error type: ERAT [Multihit]
> >    4. Feb 15 02:56:33 host kernel:     Effective address:
> c00003eefc12f018
> >    5. Feb 15 03:04:19 host kernel: Severe Machine check interrupt
> >    [Recovered]
> >    6. Feb 15 03:04:19 host kernel:   Initiator: CPU
> >    7. Feb 15 03:04:19 host kernel:   Error type: ERAT [Multihit]
> >    8. Feb 15 03:04:19 host kernel:     Effective address:
> c00003eefc12f018
> >    9.
> >
> >
> >
> > * [282d5fee5c4f](https://github.com/open-power/skiboot/commit/
> 282d5fee5c4f)
> > core/hmi: Use pr_fmt macro for tagging log messages
> > * [c531ff957669](https://github.com/open-power/skiboot/commit/
> c531ff957669)
> > opal/hmi: HMI logging with location code info.
> > * [b33ed1e6b6b0](https://github.com/open-power/skiboot/commit/
> b33ed1e6b6b0)
> > core/hmi: Do not display FIR details if none of the bits are set.
> > * [45a961515be6](https://github.com/open-power/skiboot/commit/
> 45a961515be6)
> > core/hmi: Display chip location code while displaying core FIR.
> >
> >
> > thanks,
> >
> > regards,
> > Sergey
> > YADRO
> >
> >
> >
> > _______________________________________________
> > OpenPower-Firmware mailing list
> > OpenPower-Firmware at lists.ozlabs.org
> > https://lists.ozlabs.org/listinfo/openpower-firmware
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ozlabs.org/pipermail/openpower-firmware/attachments/20180316/a3273a98/attachment.html>