From: Daniel Jordan <daniel.m.jordan@oracle.com> To: Tim Chen <tim.c.chen@linux.intel.com>, linux-mm@kvack.org, linux-kernel@vger.kernel.org Cc: aaron.lu@intel.com, ak@linux.intel.com, akpm@linux-foundation.org, Dave.Dice@oracle.com, dave@stgolabs.net, khandual@linux.vnet.ibm.com, ldufour@linux.vnet.ibm.com, mgorman@suse.de, mhocko@kernel.org, pasha.tatashin@oracle.com, steven.sistare@oracle.com, yossi.lev@oracle.com Subject: Re: [RFC PATCH v1 03/13] mm: add lock array to pgdat and batch fields to struct page Date: Thu, 1 Feb 2018 23:29:30 -0500 [thread overview] Message-ID: <02b757a9-5b06-51a5-a3e1-5cbc06a79996@oracle.com> (raw) In-Reply-To: <64330116-13ef-af3c-a445-f0c1b5bc1728@linux.intel.com> On 02/01/2018 05:50 PM, Tim Chen wrote: > On 01/31/2018 03:04 PM, daniel.m.jordan@oracle.com wrote: >> This patch simply adds the array of locks and struct page fields. >> Ignore for now where the struct page fields are: we need to find a place >> to put them that doesn't enlarge the struct. >> >> Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com> >> --- >> include/linux/mm_types.h | 5 +++++ >> include/linux/mmzone.h | 7 +++++++ >> mm/page_alloc.c | 3 +++ >> 3 files changed, 15 insertions(+) >> >> diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h >> index cfd0ac4e5e0e..6e9d26f0cecf 100644 >> --- a/include/linux/mm_types.h >> +++ b/include/linux/mm_types.h >> @@ -190,6 +190,11 @@ struct page { >> struct kmem_cache *slab_cache; /* SL[AU]B: Pointer to slab */ >> }; >> >> + struct { >> + unsigned lru_batch; >> + bool lru_sentinel; > > The above declaration adds at least 5 bytes to struct page. > It adds a lot of extra memory overhead when multiplied > by the number of pages in the system. Yes, I completely agree, enlarging struct page won't cut it for the final solution. > We can move sentinel bool to page flag, at least for 64 bit system. There did seem to be room for one more bit the way my kernel was configured (without losing a component in page->flags), but I'd have to look again. > And 8 bit is probably enough for lru_batch id to give a max > lru_batch number of 256 to break the locks into 256 smaller ones. > The max used in the patchset is 32 and that is already giving > pretty good spread of the locking. > It will be better if we can find some unused space in struct page > to squeeze it in. One idea we'd had was to store the batch id in the lower bits of the mem_cgroup pointer. CONFIG_MEMCG seems to be pretty ubiquitous these days, and it's a large enough struct (1048 bytes on one machine) to have room in the lower bits. Another way might be to encode the previous and next lru page pointers as pfn's instead of struct list_head *'s, shrinking the footprint of struct page's lru field to allow room for the batch id.
WARNING: multiple messages have this Message-ID (diff)
From: Daniel Jordan <daniel.m.jordan@oracle.com> To: Tim Chen <tim.c.chen@linux.intel.com>, linux-mm@kvack.org, linux-kernel@vger.kernel.org Cc: aaron.lu@intel.com, ak@linux.intel.com, akpm@linux-foundation.org, Dave.Dice@oracle.com, dave@stgolabs.net, khandual@linux.vnet.ibm.com, ldufour@linux.vnet.ibm.com, mgorman@suse.de, mhocko@kernel.org, pasha.tatashin@oracle.com, steven.sistare@oracle.com, yossi.lev@oracle.com Subject: Re: [RFC PATCH v1 03/13] mm: add lock array to pgdat and batch fields to struct page Date: Thu, 1 Feb 2018 23:29:30 -0500 [thread overview] Message-ID: <02b757a9-5b06-51a5-a3e1-5cbc06a79996@oracle.com> (raw) In-Reply-To: <64330116-13ef-af3c-a445-f0c1b5bc1728@linux.intel.com> On 02/01/2018 05:50 PM, Tim Chen wrote: > On 01/31/2018 03:04 PM, daniel.m.jordan@oracle.com wrote: >> This patch simply adds the array of locks and struct page fields. >> Ignore for now where the struct page fields are: we need to find a place >> to put them that doesn't enlarge the struct. >> >> Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com> >> --- >> include/linux/mm_types.h | 5 +++++ >> include/linux/mmzone.h | 7 +++++++ >> mm/page_alloc.c | 3 +++ >> 3 files changed, 15 insertions(+) >> >> diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h >> index cfd0ac4e5e0e..6e9d26f0cecf 100644 >> --- a/include/linux/mm_types.h >> +++ b/include/linux/mm_types.h >> @@ -190,6 +190,11 @@ struct page { >> struct kmem_cache *slab_cache; /* SL[AU]B: Pointer to slab */ >> }; >> >> + struct { >> + unsigned lru_batch; >> + bool lru_sentinel; > > The above declaration adds at least 5 bytes to struct page. > It adds a lot of extra memory overhead when multiplied > by the number of pages in the system. Yes, I completely agree, enlarging struct page won't cut it for the final solution. > We can move sentinel bool to page flag, at least for 64 bit system. There did seem to be room for one more bit the way my kernel was configured (without losing a component in page->flags), but I'd have to look again. > And 8 bit is probably enough for lru_batch id to give a max > lru_batch number of 256 to break the locks into 256 smaller ones. > The max used in the patchset is 32 and that is already giving > pretty good spread of the locking. > It will be better if we can find some unused space in struct page > to squeeze it in. One idea we'd had was to store the batch id in the lower bits of the mem_cgroup pointer. CONFIG_MEMCG seems to be pretty ubiquitous these days, and it's a large enough struct (1048 bytes on one machine) to have room in the lower bits. Another way might be to encode the previous and next lru page pointers as pfn's instead of struct list_head *'s, shrinking the footprint of struct page's lru field to allow room for the batch id. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2018-02-02 4:29 UTC|newest] Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-01-31 23:04 [RFC PATCH v1 00/13] lru_lock scalability daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 01/13] mm: add a percpu_pagelist_batch sysctl interface daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 02/13] mm: allow compaction to be disabled daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 03/13] mm: add lock array to pgdat and batch fields to struct page daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-02-01 22:50 ` Tim Chen 2018-02-01 22:50 ` Tim Chen 2018-02-02 4:29 ` Daniel Jordan [this message] 2018-02-02 4:29 ` Daniel Jordan 2018-01-31 23:04 ` [RFC PATCH v1 04/13] mm: introduce struct lru_list_head in lruvec to hold per-LRU batch info daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 05/13] mm: add batching logic to add/delete/move API's daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 06/13] mm: add lru_[un]lock_all APIs daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 07/13] mm: convert to-be-refactored lru_lock callsites to lock-all API daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 08/13] mm: temporarily convert " daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 09/13] mm: introduce add-only version of pagevec_lru_move_fn daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 10/13] mm: add LRU batch lock API's daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 11/13] mm: use lru_batch locking in release_pages daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-01-31 23:04 ` [RFC PATCH v1 12/13] mm: split up release_pages into non-sentinel and sentinel passes daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-02-02 14:40 ` Laurent Dufour 2018-02-02 14:40 ` Laurent Dufour 2018-02-02 17:00 ` Laurent Dufour 2018-02-02 17:00 ` Laurent Dufour 2018-02-06 17:47 ` Daniel Jordan 2018-02-06 17:47 ` Daniel Jordan 2018-02-05 4:58 ` [lkp-robot] [mm] 44b163e12f: kernel_BUG_at_mm/swap.c kernel test robot 2018-02-05 4:58 ` kernel test robot 2018-01-31 23:04 ` [RFC PATCH v1 13/13] mm: splice local lists onto the front of the LRU daniel.m.jordan 2018-01-31 23:04 ` daniel.m.jordan 2018-02-01 23:30 ` Tim Chen 2018-02-01 23:30 ` Tim Chen 2018-02-02 5:17 ` Daniel Jordan 2018-02-02 5:17 ` Daniel Jordan 2018-02-02 5:21 ` Aaron Lu 2018-02-02 5:21 ` Aaron Lu 2018-02-06 17:38 ` Daniel Jordan 2018-02-06 17:38 ` Daniel Jordan 2018-02-02 15:22 ` Laurent Dufour 2018-02-02 15:22 ` Laurent Dufour 2018-02-06 18:18 ` Daniel Jordan 2018-02-06 18:18 ` Daniel Jordan 2018-02-01 15:54 ` [RFC PATCH v1 00/13] lru_lock scalability Steven Whitehouse 2018-02-01 15:54 ` Steven Whitehouse 2018-02-02 4:18 ` Daniel Jordan 2018-02-02 4:18 ` Daniel Jordan 2018-02-02 10:50 ` Steven Whitehouse 2018-02-02 10:50 ` Steven Whitehouse 2018-02-08 23:36 ` Andrew Morton 2018-02-08 23:36 ` Andrew Morton 2018-02-13 21:07 ` Daniel Jordan 2018-02-13 21:07 ` Daniel Jordan
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=02b757a9-5b06-51a5-a3e1-5cbc06a79996@oracle.com \ --to=daniel.m.jordan@oracle.com \ --cc=Dave.Dice@oracle.com \ --cc=aaron.lu@intel.com \ --cc=ak@linux.intel.com \ --cc=akpm@linux-foundation.org \ --cc=dave@stgolabs.net \ --cc=khandual@linux.vnet.ibm.com \ --cc=ldufour@linux.vnet.ibm.com \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=mgorman@suse.de \ --cc=mhocko@kernel.org \ --cc=pasha.tatashin@oracle.com \ --cc=steven.sistare@oracle.com \ --cc=tim.c.chen@linux.intel.com \ --cc=yossi.lev@oracle.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.