Problems with swapping in v4.5-rc on POWER

[Michael Ellerman]

On Fri, 2016-03-04 at 09:58 -0800, Hugh Dickins wrote:
> 
> The alternative bisection was as unsatisfactory as the first:
> again it fingered an irrelevant merge (rather than any commit
> pulled in by that merge) as the bad commit.
> 
> It seems this issue is too intermittent for bisection to be useful,
> on my load anyway.

Darn. Thanks for trying.

> The best I can do now is try v4.4 for a couple of days, to verify that
> still comes out good (rather than the machine going bad coincident with
> v4.5-rc), then try v4.5-rc7 to verify that that still comes out bad.

Thanks, that would still be helpful.

> I'll report back on those; but beyond that, I'll have to leave it to you.

I haven't had any luck here :/

Can you give us a more verbose description of your test setup?

 - G5, which exact model?
 - 4k pages, no THP.
 - how much ram & swap?
 - building linus' tree, make -j ?
 - source and output on tmpfs? (how big?)
 - what device is the swap device? (you said SSD I think?)
 - anything else I've forgotten?

Oh and can you send us your bisect logs, we can at least trust the bad results
I think.

cheers