From: Dave Hansen <dave.hansen@linux.intel.com> To: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org, Dave Hansen <dave.hansen@linux.intel.com>, yang.shi@linux.alibaba.com, rientjes@google.com, ying.huang@intel.com, dan.j.williams@intel.com, david@redhat.com, osalvador@suse.de Subject: [RFC][PATCH 00/13] [v5] Migrate Pages in lieu of discard Date: Mon, 25 Jan 2021 16:34:11 -0800 [thread overview] Message-ID: <20210126003411.2AC51464@viggo.jf.intel.com> (raw) The full series is also available here: https://github.com/hansendc/linux/tree/automigrate-20210122 The meat of this patch is in: [PATCH 08/13] mm/migrate: demote pages during reclaim Which also has the most changes since the last post. This version is mostly to address review comments from Yang Shi and Oscar Salvador. Review comments are documented in the individual patch changelogs. This also contains a few prerequisite patches that fix up an issue with the vm.zone_reclaim_mode sysctl ABI. -- We're starting to see systems with more and more kinds of memory such as Intel's implementation of persistent memory. Let's say you have a system with some DRAM and some persistent memory. Today, once DRAM fills up, reclaim will start and some of the DRAM contents will be thrown out. Allocations will, at some point, start falling over to the slower persistent memory. That has two nasty properties. First, the newer allocations can end up in the slower persistent memory. Second, reclaimed data in DRAM are just discarded even if there are gobs of space in persistent memory that could be used. This set implements a solution to these problems. At the end of the reclaim process in shrink_page_list() just before the last page refcount is dropped, the page is migrated to persistent memory instead of being dropped. While I've talked about a DRAM/PMEM pairing, this approach would function in any environment where memory tiers exist. This is not perfect. It "strands" pages in slower memory and never brings them back to fast DRAM. Other things need to be built to promote hot pages back to DRAM. This is also all based on an upstream mechanism that allows persistent memory to be onlined and used as if it were volatile: http://lkml.kernel.org/r/20190124231441.37A4A305@viggo.jf.intel.com == Open Issues == * For cpusets and memory policies that restrict allocations to PMEM, is it OK to demote to PMEM? Do we need a cgroup- level API to opt-in or opt-out of these migrations? -- Changes since (automigrate-20200818): * Fall back to normal reclaim when demotion fails * Fix some compile issues, when page migration and NUMA are off Changes since (automigrate-20201007): * separate out checks for "can scan anon LRU" from "can actually swap anon pages right now". Previous series conflated them and may have been overly aggressive scanning LRU * add MR_DEMOTION to tracepoint header * remove unnecessary hugetlb page check Changes since (https://lwn.net/Articles/824830/): * Use higher-level migrate_pages() API approach from Yang Shi's earlier patches. * made sure to actually check node_reclaim_mode's new bit * disabled migration entirely before introducing RECLAIM_MIGRATE * Replace GFP_NOWAIT with explicit __GFP_KSWAPD_RECLAIM and comment why we want that. * Comment on effects of that keep multiple source nodes from sharing target nodes Cc: Yang Shi <yang.shi@linux.alibaba.com> Cc: David Rientjes <rientjes@google.com> Cc: Huang Ying <ying.huang@intel.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: David Hildenbrand <david@redhat.com> Cc: osalvador <osalvador@suse.de> Cc: Huang Ying <ying.huang@intel.com>
WARNING: multiple messages have this Message-ID (diff)
From: Dave Hansen <dave.hansen@linux.intel.com> To: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org,Dave Hansen <dave.hansen@linux.intel.com>,yang.shi@linux.alibaba.com,rientjes@google.com,ying.huang@intel.com,dan.j.williams@intel.com,david@redhat.com,osalvador@suse.de Subject: [RFC][PATCH 00/13] [v5] Migrate Pages in lieu of discard Date: Mon, 25 Jan 2021 16:34:11 -0800 [thread overview] Message-ID: <20210126003411.2AC51464@viggo.jf.intel.com> (raw) The full series is also available here: https://github.com/hansendc/linux/tree/automigrate-20210122 The meat of this patch is in: [PATCH 08/13] mm/migrate: demote pages during reclaim Which also has the most changes since the last post. This version is mostly to address review comments from Yang Shi and Oscar Salvador. Review comments are documented in the individual patch changelogs. This also contains a few prerequisite patches that fix up an issue with the vm.zone_reclaim_mode sysctl ABI. -- We're starting to see systems with more and more kinds of memory such as Intel's implementation of persistent memory. Let's say you have a system with some DRAM and some persistent memory. Today, once DRAM fills up, reclaim will start and some of the DRAM contents will be thrown out. Allocations will, at some point, start falling over to the slower persistent memory. That has two nasty properties. First, the newer allocations can end up in the slower persistent memory. Second, reclaimed data in DRAM are just discarded even if there are gobs of space in persistent memory that could be used. This set implements a solution to these problems. At the end of the reclaim process in shrink_page_list() just before the last page refcount is dropped, the page is migrated to persistent memory instead of being dropped. While I've talked about a DRAM/PMEM pairing, this approach would function in any environment where memory tiers exist. This is not perfect. It "strands" pages in slower memory and never brings them back to fast DRAM. Other things need to be built to promote hot pages back to DRAM. This is also all based on an upstream mechanism that allows persistent memory to be onlined and used as if it were volatile: http://lkml.kernel.org/r/20190124231441.37A4A305@viggo.jf.intel.com == Open Issues == * For cpusets and memory policies that restrict allocations to PMEM, is it OK to demote to PMEM? Do we need a cgroup- level API to opt-in or opt-out of these migrations? -- Changes since (automigrate-20200818): * Fall back to normal reclaim when demotion fails * Fix some compile issues, when page migration and NUMA are off Changes since (automigrate-20201007): * separate out checks for "can scan anon LRU" from "can actually swap anon pages right now". Previous series conflated them and may have been overly aggressive scanning LRU * add MR_DEMOTION to tracepoint header * remove unnecessary hugetlb page check Changes since (https://lwn.net/Articles/824830/): * Use higher-level migrate_pages() API approach from Yang Shi's earlier patches. * made sure to actually check node_reclaim_mode's new bit * disabled migration entirely before introducing RECLAIM_MIGRATE * Replace GFP_NOWAIT with explicit __GFP_KSWAPD_RECLAIM and comment why we want that. * Comment on effects of that keep multiple source nodes from sharing target nodes Cc: Yang Shi <yang.shi@linux.alibaba.com> Cc: David Rientjes <rientjes@google.com> Cc: Huang Ying <ying.huang@intel.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: David Hildenbrand <david@redhat.com> Cc: osalvador <osalvador@suse.de> Cc: Huang Ying <ying.huang@intel.com>
next reply other threads:[~2021-01-26 10:26 UTC|newest] Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-01-26 0:34 Dave Hansen [this message] 2021-01-26 0:34 ` [RFC][PATCH 00/13] [v5] Migrate Pages in lieu of discard Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 01/13] mm/vmscan: restore zone_reclaim_mode ABI Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-02-10 9:42 ` Oscar Salvador 2021-01-26 0:34 ` [RFC][PATCH 02/13] mm/vmscan: move RECLAIM* bits to uapi header Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-02-10 9:44 ` Oscar Salvador 2021-01-26 0:34 ` [RFC][PATCH 03/13] mm/vmscan: replace implicit RECLAIM_ZONE checks with explicit checks Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-31 1:10 ` David Rientjes 2021-01-31 1:10 ` David Rientjes 2021-02-10 9:54 ` Oscar Salvador 2021-01-26 0:34 ` [RFC][PATCH 04/13] mm/numa: node demotion data structure and lookup Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-31 1:19 ` David Rientjes 2021-01-31 1:19 ` David Rientjes 2021-02-01 17:49 ` Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 05/13] mm/numa: automatically generate node migration order Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-29 20:46 ` Yang Shi 2021-01-29 20:46 ` Yang Shi 2021-02-01 19:13 ` Dave Hansen 2021-02-02 11:43 ` Oscar Salvador 2021-02-02 17:46 ` Yang Shi 2021-02-02 17:46 ` Yang Shi 2021-02-03 0:43 ` Dave Hansen 2021-02-04 0:26 ` Yang Shi 2021-02-04 0:26 ` Yang Shi 2021-01-26 0:34 ` [RFC][PATCH 06/13] mm/migrate: update migration order during on hotplug events Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-29 20:59 ` Yang Shi 2021-01-29 20:59 ` Yang Shi 2021-02-02 11:42 ` Oscar Salvador 2021-02-09 23:45 ` Dave Hansen 2021-02-10 8:55 ` Oscar Salvador 2021-01-26 0:34 ` [RFC][PATCH 07/13] mm/migrate: make migrate_pages() return nr_succeeded Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-29 21:04 ` Yang Shi 2021-01-29 21:04 ` Yang Shi 2021-02-09 23:41 ` Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 08/13] mm/migrate: demote pages during reclaim Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-02-02 11:55 ` Oscar Salvador 2021-02-02 22:45 ` Yang Shi 2021-02-02 22:45 ` Yang Shi 2021-02-02 22:56 ` Dave Hansen 2021-02-02 18:22 ` Yang Shi 2021-02-02 18:22 ` Yang Shi 2021-02-02 18:34 ` Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 09/13] mm/vmscan: add page demotion counter Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 10/13] mm/vmscan: add helper for querying ability to age anonymous pages Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 11/13] mm/vmscan: Consider anonymous pages without swap Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-02-02 18:56 ` Yang Shi 2021-02-02 18:56 ` Yang Shi 2021-02-02 21:35 ` Dave Hansen 2021-02-02 22:35 ` Yang Shi 2021-02-02 22:35 ` Yang Shi 2021-01-26 0:34 ` [RFC][PATCH 12/13] mm/vmscan: never demote for memcg reclaim Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-26 0:34 ` [RFC][PATCH 13/13] mm/migrate: new zone_reclaim_mode to enable reclaim migration Dave Hansen 2021-01-26 0:34 ` Dave Hansen 2021-01-31 1:13 ` [RFC][PATCH 00/13] [v5] Migrate Pages in lieu of discard David Rientjes 2021-01-31 1:13 ` David Rientjes
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20210126003411.2AC51464@viggo.jf.intel.com \ --to=dave.hansen@linux.intel.com \ --cc=dan.j.williams@intel.com \ --cc=david@redhat.com \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=osalvador@suse.de \ --cc=rientjes@google.com \ --cc=yang.shi@linux.alibaba.com \ --cc=ying.huang@intel.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.