kswapd0 consuming 50% of cpu

[Bruno Wolff III]

On Mon, Nov 26, 2012 at 12:19:11 -0700,
   JD <jd1008 at gmail.com> wrote:
>
>The 50% cpu consumption by kswapd appeared first in 3.7

Thanks for that clarification.

It looks like there may be multiple things causing similar problems here.
3.7-rc7 should have another patch to help this, but there are still some 
patches being tested that aren't in rc7. So you might see some improvement 
in the 3.7-rc1 kernel. One hasn't been built for Fedora yet, but I'd expect 
to see one today.

>But Bruno, the "not sending a hard reset to sleeping drives" is in 
>all 3.6.X kernels.
>You would think they would have caught it rather than propagate it.

One probably would, but it may be that the people testing this didn't have 
their drives set to go to sleep when idle and no one ran into the problem.
One would hope a test case will get added to someone's testing list to make 
sure new kernels don't break drives which go to sleep, but that doesn't 
mean it will happen.