Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-10 Thread David Hildenbrand
On 10.09.20 13:05, Vlastimil Babka wrote: > On 9/10/20 12:29 PM, David Hildenbrand wrote: >> On 09.09.20 13:55, Vlastimil Babka wrote: >>> On 9/9/20 1:36 PM, Michal Hocko wrote: On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: > Here's a version that will apply on top of next-20200908.

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-10 Thread Vlastimil Babka
On 9/10/20 12:29 PM, David Hildenbrand wrote: > On 09.09.20 13:55, Vlastimil Babka wrote: >> On 9/9/20 1:36 PM, Michal Hocko wrote: >>> On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: Here's a version that will apply on top of next-20200908. The first 4 patches need no change. For

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-10 Thread David Hildenbrand
On 09.09.20 13:55, Vlastimil Babka wrote: > On 9/9/20 1:36 PM, Michal Hocko wrote: >> On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: >>> Here's a version that will apply on top of next-20200908. The first 4 >>> patches need no change. >>> >>> 8< >>> >From

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-09 Thread David Hildenbrand
On 09.09.20 13:44, David Hildenbrand wrote: > On 09.09.20 13:36, Michal Hocko wrote: >> On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: >>> Here's a version that will apply on top of next-20200908. The first 4 >>> patches need no change. >>> >>> 8< >>> >From

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-09 Thread Vlastimil Babka
On 9/9/20 1:36 PM, Michal Hocko wrote: > On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: >> Here's a version that will apply on top of next-20200908. The first 4 >> patches need no change. >> >> 8< >> >From 8febc17272b8e8b378e2e5ea5e76b2616f029c5b Mon Sep 17 00:00:00 2001 >> From:

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-09 Thread Pavel Tatashin
> because we do not want to call out pcp free lists). I strongly suspect > that only the memory hotplug really cares for this hard guanrantee. > alloc_contig_range simply goes with EBUSY. > You are right, the strong requirement is for hotplug case only, but if PCP disable/enable functionality is

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-09 Thread David Hildenbrand
On 09.09.20 13:36, Michal Hocko wrote: > On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: >> Here's a version that will apply on top of next-20200908. The first 4 >> patches need no change. >> >> 8< >> >From 8febc17272b8e8b378e2e5ea5e76b2616f029c5b Mon Sep 17 00:00:00 2001 >> From:

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-09 Thread Michal Hocko
On Wed 09-09-20 12:48:54, Vlastimil Babka wrote: > Here's a version that will apply on top of next-20200908. The first 4 patches > need no change. > > 8< > >From 8febc17272b8e8b378e2e5ea5e76b2616f029c5b Mon Sep 17 00:00:00 2001 > From: Vlastimil Babka > Date: Mon, 7 Sep 2020 17:20:39

Re: [RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-09 Thread Vlastimil Babka
Here's a version that will apply on top of next-20200908. The first 4 patches need no change. 8< >From 8febc17272b8e8b378e2e5ea5e76b2616f029c5b Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Mon, 7 Sep 2020 17:20:39 +0200 Subject: [PATCH] mm, page_alloc: disable pcplists during

[RFC 5/5] mm, page_alloc: disable pcplists during page isolation

2020-09-07 Thread Vlastimil Babka
Page isolation can race with process freeing pages to pcplists in a way that a page from isolated pageblock can end up on pcplist. This can be fixed by repeated draining of pcplists, as done by patch "mm/memory_hotplug: drain per-cpu pages again during memory offline" in [1]. David and Michal