[Skiboot] [PATCH 00/16] assorted MCE and SRESET handling and reporting

[Stewart Smith]

Nicholas Piggin <npiggin at gmail.com> writes:
> There is a motley assortment of precursor patches 1-8 that I ended
> up writing while implementing and testing the others, each one of
> these is independent and could be cherry-picked. The only real bug
> fix is patch 1.
>
> The rest of the series basically tries to improve handling of the non
> maskable interrupts MCE and SRESET, basically reducing cases where
> MSR[ME]=1, properly handle MCE from power saving mode, reporting SRESET
> interrupts with a register dump, reducing windows where an NMI will
> crash due to vectors or SPRs being out of synch.
>
> By the end of the series, we have the capability to handle and return
> from interrupts, and that is actually done in the case of SRESET
> interrupts. All others are still fatal, however this opens the
> possibility to contiue after a recoverable MCE, take virtual memory
> faults, implement a decrementer watchdog, etc.
>
> Thanks,
> Nick
>
> Nicholas Piggin (16):
>   core/cpu: HID update race
>   external/mambo: allow CPU targeting for most debug utils
>   core/init: rename setup_reset_vector
>   core/cpu: do not inline cpu_relax
>   core/exceptions: save current MSR in exception frame
>   asm/head: use HSRR exception registers in FIXUP_ENDIAN
>   core: hardcode P9 large decrementer width
>   core/init: rearrange final boot steps
>   core/fast-reboot: improve NMI handling during fast reset
>   core/init: enable machine check on secondaries
>   asm/head: provide asm support for interrupts to be returned from
>   core/fast-reboot: fast reboot specific sreset patch
>   asm/head: sreset handler remove FIXUP_ENDIAN
>   core/exceptions: implement an exception handler for non-powersave
>     sresets
>   core/exceptions: allow recoverable sreset exceptions
>   core/exceptions: implement support for MCE interrupts in powersave

Thanks!

I've merged all except the hardcode P9 large decementer width patch to
master as of 5d86cebd53683d055f039dbecf3f74fb0084ea83, and the HID
update race I've backported to 6.0.x and 6.2.x as of 59d9ab63b397 and
c51c7a5df601 respectively.

-- 
Stewart Smith
OPAL Architect, IBM.