On Wed, 2012-01-04 at 16:21 -0800, Roland Dreier wrote: > On Wed, Dec 21, 2011 at 2:53 PM, Nicholas A. Bellinger > <nab@xxxxxxxxxxxxxxx> wrote: > > Anyways, since we are removing se_device TCQ depth checking all-together > > for v3.3, it should be a moot point now with the following: > > I don't see what the depth checking has to do with the soft lockup... > the issue is that as long as the processing thread finds work in its > queue it will keep running and never reschedule. > Then I must have read the original mail incorrectly. I thought that the transport_tcq_window_closed() usage with UP when the depth was exhausted was the root cause here, but after re-reading the thread you're right that depth_left does not apply here. Sorry for the confusion. So i'm still not sure what is the properly resolution here if cond_resched() on each iteration adds a performance regression.. Any ideas guys..? --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
