Messages in this thread | | | Date | Wed, 2 Dec 2015 11:10:17 +0000 | From | Mel Gorman <> | Subject | Re: [lkp] [mm, page_alloc] d0164adc89: -100.0% fsmark.app_overhead |
| |
On Tue, Dec 01, 2015 at 03:04:31PM +0100, Michal Hocko wrote: > On Tue 01-12-15 12:23:41, Will Deacon wrote: > > Hi Michal, > > > > On Mon, Nov 30, 2015 at 02:02:00PM +0100, Michal Hocko wrote: > > > [Let's CC Will - see the question at the end of the email please] > > > > [...] > > > > > > > There is no reference to OOM possibility in the email that I can see. Can > > > > > you give examples of the OOM messages that shows the problem sites? It was > > > > > suspected that there may be some callers that were accidentally depending > > > > > on access to emergency reserves. If so, either they need to be fixed (if > > > > > the case is extremely rare) or a small reserve will have to be created > > > > > for callers that are not high priority but still cannot reclaim. > > > > > > __virtblk_add_req calls > > > virtqueue_add_sgs(vq, sgs, num_out, num_in, vbr, GFP_ATOMIC) > > > alloc_indirect(gfp) > > > gfp &= ~(__GFP_HIGHMEM | __GFP_HIGH) > > > > > > So this is true __GFP_ATOMIC, we just drop __GFP_HIGH so it doesn't get > > > access to more reserves. It still does ALLOC_HARDER. So I think the real > > > issue is somewhere else when something should have triggered kswapd and > > > it doesn't do that anymore. I have tried to find that offender the last > > > time but didn't manage to find any. > > > > > > Btw. I completely miss why b92b1b89a33c ("virtio: force vring > > > descriptors to be allocated from lowmem") had to clear __GFP_HIGH. Will > > > do you remember why you have dropped that flag as well? > > > > Right, that looks unnecessary, but it could be that we were masking a > > bug somewhere else. > > OK, I will send a patch to remove __GFP_HIGH because it is clearly > misleding and doesn't have anything to do with the highmem zone. >
Thanks for looking into this. I just sent a patch that includes a changelog explaining why this bug triggers now and would have been hidden before.
> > > Also I do not seem to find any user of alloc_indirect which would do > > > __GFP_HIGHMEM. All of them are either GFP_KERNEL or GFP_ATOMIC. So > > > either I am missing something or this is not really needed. Maybe the > > > situation was different back in 2012. > > > > I tried to revisit the thread leading to that patch, but it doesn't make > > a whole lot of sense: > > > > https://lkml.org/lkml/2012/10/17/143 > > > > I certainly remember debugging the failure (i.e. it wasn't theoretical), > > and we were ending up with highmem addresses being passed in the virtio > > ring (due to the zero-copy stuff in 9p) and also for the descriptors > > themselves. The discussion at the time makes it sound like GFP_ATOMIC > > was giving us those... > > Hmm, unless I am missing something GFP_ATOMIC resp. GFP_KERNEL cannot > fallback to the highmem zone - see GFP_ZONE_TABLE. Maybe the highmem > pointer got there from a different path than alloc_indirect?
GFP_ATOMIC should not be returning highmem addresses ever. I checked briefly but did not spot where 9p is getting highmem pages from but it wasn't via GFP_ATOMIC.
-- Mel Gorman SUSE Labs
| |