Re: 3.18-rc3: soft lockup in compact_zone, dead machine

[Vlastimil Babka]

On 11/07/2014 11:06 AM, Pavel Machek wrote:
> Hi!
>
> My main machine died completely, it seems that original failure was
> soft lockup in compact_zone().

(expanding CC)

Welcome to the club...

http://article.gmane.org/gmane.linux.kernel.mm/124451/match=isolate_freepages_block+very+high+intermittent+overhead

https://lkml.org/lkml/2014/11/4/144

https://lkml.org/lkml/2014/11/4/904

How reproducible is your case? So far it seems that git revert 
e14c720efdd73c6d69cd8d07fa894bcd11fe1973 helped one of the reporters.
I still don't know what's wrong, but suspect a free scanner 
(cc->free_pfn) position being broken (i.e. underflow), which would allow 
isolate_migratepages() to run for a loooooooong time. The code does 
cond_resched() periodically, so soft lockup looks strange, but I guess 
that can happen in some contexts?

Vlastimil



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/