Re: [PATCH RFC 00/15] mm: memory book keeping and lru_lock splitting
From: Konstantin Khlebnikov
Date: Wed Feb 15 2012 - 23:51:53 EST
Hugh Dickins wrote:
On Thu, 16 Feb 2012, Konstantin Khlebnikov wrote:
There should be no logic changes in this patchset, this is only tossing bits around.
[ This patchset is on top some memcg cleanup/rework patches,
which I sent to linux-mm@ today/yesterday ]
Most of things in this patchset are self-descriptive, so here brief plan:
* Transmute struct lruvec into struct book. Like real book this struct will
store set of pages for one zone. It will be working unit for reclaimer code.
[ If memcg is disabled in config there will only one book embedded into struct zone ]
* move page-lru counters to struct book
[ this adds extra overhead in add_page_to_lru_list()/del_page_from_lru_list() for
non-memcg case, but I believe it will be invisible, only one non-atomic add/sub
in the same cacheline with lru list ]
* unify inactive_list_is_low_global() and cleanup reclaimer code
* replace struct mem_cgroup_zone with single pointer to struct book
* optimize page to book translations, move it upper in the call stack,
replace some struct zone arguments with struct book pointer.
page -> book dereference become main operation, page (even free) will always
points to one book in its zone. so page->flags bits may contains direct reference to book.
Maybe we can replace page_zone(page) with book_zone(page_book(page)), without mem cgroups
book -> zone dereference will be simple container_of().
Finally, there appears some new locking primitives for decorating lru_lock splitting logic.
Final patch actually splits zone->lru_lock into small per-book pieces.
Well done, it looks like you've beaten me by a few days: my per-memcg
per-zone locking patches are now split up and ready, except for the
commit comments and performance numbers to support them.
Heh, nice. How do you organize link from page to "book"?
My patchset still uses page_cgroup, and I afraid something broken in locking around it.
Or perhaps what we've been doing is orthogonal: I've not glanced beyond
your Subjects yet, but those do look very familiar from my own work -
though we're still using "lruvec"s rather than "book"s.
I think, good name is always good. "book" sounds much better than "mem_cgroup_zone".
Plus I want to split mm and mem-cg as much as possible, like this done in sched-grouping.
After my patchset there no links from "book" to "memcg", so now there can exist books of
different nature. It would be nice to use this memory management not only for cgroups.
For example, what do you think about per-struct-user memory controller,
"rt" prio memory group, which never pruned by "normal" prio memory group users.
or per-session-id memory auto-grouping =)
Anyway, I should be well-placed to review what you've done, and had
better switch away from my own patches to testing and reviewing yours
now, checking if we've caught anything that you're missing. Or maybe
it'll be worth posting mine anyway, we'll see: I'll look to yours first.
Competition is good. It would be nice to see your patches too, if they are as ready as my.
I already found two bugs in my patchset:
in "mm: unify inactive_list_is_low()" NR_ACTIVE_* instead of LRU_ACTIVE_*
and in "mm: handle book relocks on lumpy reclaim" I forget locking in isolate_lru_pages()
[ but with CONFIG_COMPACTION=y this branch never used, as I guess ]
All this code currently *completely untested*, but seems like it already can work.
Oh, perhaps I'm ahead of you after all :)
Not exactly, I already wrote the same code more than half-year ago.
So, it wasn't absolutely fair competition from the start. =)
After that, there two options how manage struct book on mem-cgroup create/destroy:
a) [ currently implemented ] allocate and release by rcu.
Thus lock_page_book() will be protected with rcu_read_lock().
b) allocate and never release struct book, reuse them after rcu grace period.
It allows to avoid some rcu_read_lock()/rcu_read_unlock() calls on hot paths.
Motivation:
I wrote the similar memory controller for our rhel6-based openvz/virtuozzo kernel,
including splitted lru-locks and some other [patented LOL] cool stuff.
[ common descrioption without techical details: http://wiki.openvz.org/VSwap ]
That kernel already in production and rather stable for a long time.
---
Konstantin Khlebnikov (15):
mm: rename struct lruvec into struct book
mm: memory bookkeeping core
mm: add book->pages_count
mm: unify inactive_list_is_low()
mm: add book->reclaim_stat
mm: kill struct mem_cgroup_zone
mm: move page-to-book translation upper
mm: introduce book locking primitives
mm: handle book relocks on lumpy reclaim
mm: handle book relocks in compaction
mm: handle book relock in memory controller
mm: optimize books in update_page_reclaim_stat()
mm: optimize books in pagevec_lru_move_fn()
mm: optimize putback for 0-order reclaim
mm: split zone->lru_lock
include/linux/memcontrol.h | 52 -------
include/linux/mm_inline.h | 222 ++++++++++++++++++++++++++++-
include/linux/mmzone.h | 26 ++-
include/linux/swap.h | 2
init/Kconfig | 4 +
mm/compaction.c | 35 +++--
mm/huge_memory.c | 10 +
mm/memcontrol.c | 238 ++++++++++---------------------
mm/page_alloc.c | 20 ++-
mm/swap.c | 128 ++++++-----------
mm/vmscan.c | 334 +++++++++++++++++++-------------------------
11 files changed, 554 insertions(+), 517 deletions(-)
That's a very familiar list of files to me!
Hugh
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/