Re: [External] Re: [PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-09 Thread Muchun Song
On Wed, Sep 9, 2020 at 5:59 PM Chris Down wrote: > > Muchun Song writes: > >1. Double the threshold currently hard coded as "10" with each iteration > >suggested by Vlastimil. It is also a good idea. > > I think this sounds reasonable, although I'd like to see what the difference > in >

Re: [External] Re: [PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-09 Thread Chris Down
Muchun Song writes: 1. Double the threshold currently hard coded as "10" with each iteration suggested by Vlastimil. It is also a good idea. I think this sounds reasonable, although I'd like to see what the difference in reclaim looks like in practice. 2. In the while loop, we can check

Re: [External] Re: [PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-08 Thread Muchun Song
Hi Chris, On Tue, Sep 8, 2020 at 11:09 PM Chris Down wrote: > > drop_caches by its very nature can be extremely performance intensive -- if > someone wants to abort after trying too long, they can just send a > TASK_KILLABLE signal, no? If exiting the loop and returning to usermode > doesn't >

Re: [PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-08 Thread Chris Down
drop_caches by its very nature can be extremely performance intensive -- if someone wants to abort after trying too long, they can just send a TASK_KILLABLE signal, no? If exiting the loop and returning to usermode doesn't reliably work when doing that, then _that's_ something to improve, but

Re: [PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-08 Thread Vlastimil Babka
On 9/8/20 5:09 PM, Chris Down wrote: > drop_caches by its very nature can be extremely performance intensive -- if > someone wants to abort after trying too long, they can just send a > TASK_KILLABLE signal, no? If exiting the loop and returning to usermode > doesn't > reliably work when doing

Re: [PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-08 Thread Chris Down
Vlastimil Babka writes: On 9/8/20 5:09 PM, Chris Down wrote: drop_caches by its very nature can be extremely performance intensive -- if someone wants to abort after trying too long, they can just send a TASK_KILLABLE signal, no? If exiting the loop and returning to usermode doesn't reliably

[PATCH] mm/vmscan: fix infinite loop in drop_slab_node

2020-09-08 Thread zangchunxin
From: Chunxin Zang On our server, there are about 10k memcg in one machine. They use memory very frequently. When I tigger drop caches，the process will infinite loop in drop_slab_node. There are two reasons: 1.We have too many memcgs, even though one object freed in one memcg, the sum of