On 07/07/2014 06:49 AM, Joonsoo Kim wrote:
Ccing Lisa, because there was bug report it may be related this
topic last Saturday.
http://www.spinics.net/lists/linux-mm/msg75741.html
On Fri, Jul 04, 2014 at 05:33:27PM +0200, Vlastimil Babka wrote:
On 07/04/2014 09:57 AM, Joonsoo Kim wrote:
Hello,
Hi Joonsoo,
please CC me on further updates, this is relevant to me.
Hello, Vlastimil.
Sorry for missing you. :)
This patchset aims at fixing problems due to memory isolation found by
testing my patchset [1].
These are really subtle problems so I can be wrong. If you find what I am
missing, please let me know.
Before describing bugs itself, I first explain definition of freepage.
1. pages on buddy list are counted as freepage.
2. pages on isolate migratetype buddy list are *not* counted as freepage.
I think the second point is causing us a lot of trouble. And I wonder if it's really
justified! We already have some is_migrate_isolate() checks in the fast path and now you
would add more, mostly just to keep this accounting correct.
It's not just for keeping accouting correct. There is another
purpose for is_migrate_isolate(). It forces freed pages to go into
isolate buddy list. Without it, freed pages would go into other
buddy list and will be used soon. So memory isolation can't work well
without is_migrate_isolate() checks and success rate could decrease.
Well I think that could be solved also by doing a lru/pcplists drain
right after marking pageblock as MIGRATETYPE_ISOLATE. After that moment,
anything newly put on pcplists should observe the new migratetype and go
to the correct pcplists. As putting stuff on pcplists is done with
disabled interrupts, and draining is done by IPI, I think it should work
correctly if we put the migratetype determination under the disabled irq
part in free_hot_cold_page().
And, I just added three more is_migrate_isolate() in the fast
path, but, two checks are in same *unlikely* branch and I can remove
another one easily. Therefore it's not quite problem I guess. (It even
does no-op if MEMORY_ISOLATION is disabled.)
Moreover, I removed one unconditional get_pageblock_migratetype() in
free_pcppages_bulk() so, in performance point or view, freepath would
be improved.
I haven't checked the individual patches yet, but I'll try.
So the question is, does it have to be correct? And (admiteddly not after a completely
exhaustive analysis) I think the answer is, surprisingly, that it doesn't :)
Well I of course don't mean that the freepage accounts could go random completely, but
what if we allowed them to drift a bit, limiting both the max error and the timeframe
where errors are possible? After all, watermarks checking is already racy so I don't think
it would be hurt that much.
I understand your suggestion. I once thought like as you, but give up
that idea. Watermark checking is already racy, but, it's *only*
protection to prevent memory allocation. After passing that check,
there is no mean to prevent us from allocating memory. So it should
be accurate as much as possible. If we allow someone to get the
memory without considering memory isolation, free memory could be in
really low state and system could be broken occasionally.
It would definitely require more thorough review, but I think with the
pcplists draining changes outlined above, there shouldn't be any more
inaccuracy happening.
Now if we look at what both CMA and memory hot-remove does is:
1. Mark a MAX_ORDER-aligned buch of pageblocks as MIGRATE_ISOLATE through
start_isolate_page_range(). As part of this, all free pages in that area are
moved on the isolate freelist through move_freepages_block().
2. Try to migrate away all non-free pages in the range. Also drain pcplists and lru_add
caches.
3. Check if everything was successfully isolated by test_pages_isolated(). Restart and/or
undo pageblock isolation if not.
So my idea is to throw away all special-casing of is_migrate_isolate() in the buddy
allocator, which would therefore account free pages on the isolate freelist as normal free
pages.
The accounting of isolated pages would be instead done only on the top level of CMA/hot
remove in the three steps outlined above, which would be modified as follows:
1. Calculate N as the target number of pages to be isolated. Perform the actions of step 1
as usual. Calculate X as the number of pages that move_freepages_block() has moved.
Subtract X from freepages (this is the same as it is done now), but also *remember the
value of X*
2. Migrate and drain pcplists as usual. The new free pages will either end up correctly on
isolate freelist, or not. We don't care, they will be accounted as freepages either way.
This is where some inaccuracy in accounted freepages would build up.
3. If test_pages_isolated() checks pass, subtract (N - X) from freepages. The result is
that we have a isolated range of N pages that nobody can steal now as everything is on
isolate freelist and is MAX_ORDER aligned. And we have in total subtracted N pages (first
X, then N-X). So the accounting matches reality.
If we have to undo, we undo the isolation and as part of this, we use
move_freepages_block() to move pages from isolate freelist to the normal ones. But we
don't care how many pages were moved. We simply add the remembered value of X to the
number of freepages, undoing the change from step 1. Again, the accounting matches reality.
The final point is that if we do this per MAX_ORDER blocks, the error in accounting cannot
be ever larger than 4MB and will be visible only during time a single MAX_ORDER block is
handled.
The 4MB error in accounting looks serious for me. min_free_kbytes is
4MB in 1GB system. So this 4MB error would makes all things broken in
such systems. Moreover, there are some ARCH having larger
pageblock_order than MAX_ORDER. In this case, the error will be larger
than 4MB.
In addition, I have a plan to extend CMA to work in parallel. It means
that there could be parallel memory isolation users rather than just
one user at the same time, so, we cannot easily bound the error under
some degree.
OK. I had thought that the error would be much smaller in practice, but
probably due to how buddy merging works, it would be enough if just the
very last freed page in the MAX_ORDER block was misplaced, and that
would trigger a cascading merge that will end with single page at
MAX_ORDER size becoming misplaced. So let's probably forget this approach.
As a possible improvement, we can assume during phase 2 that every page freed by migration
will end up correctly on isolate free list. So we create M free pages by migration, and
subtract M from freepage account. Then in phase 3 we either subtract (N - X - M), or add X
+ M in the undo case. (Ideally, if we succeed, X + M should be equal to N, but due to
pages on pcplists and the possible races it will be less). I think with this improvement,
any error would be negligible.
Thoughts?
Thanks for suggestion. :)
It is really good topic to think deeply.
For now, I'd like to fix these problems without side-effect as you
suggested. Your suggestion changes the meaning of freepage that
isolated pages are included in nr_freepage and there could be possible
regression in success rate of memory hotplug and CMA. Possibly, it
is the way we have to go, but, IMHO, it isn't the time to go. Before
going that way, we should fix current implementation first so that
fixes can be backported to old kernel if someone needs.
Adding the extra pcplist drains and reordering
get_pageblock_migratetype() under disabled IRQ's should be small enough
to be backportable and not bring any regressions to success rates. There
will be some cost for the extra drains, but paid only when the memory
isolation actually happens. Extending the scope of disabled irq's would
affect fast path somewhat, but I guess less than extra checks?
Interestingly,
on last Saturday, Lisa Du reported CMA accounting bugs.
http://www.spinics.net/lists/linux-mm/msg75741.html
I don't look at it in detail, but, maybe it is related to these
problems and we should fix it without side-effect.
So, in conclusion, I think that your suggestion is beyond the scope of
this patchset because of following two reasons.
1. I'd like to fix these problems without side-effect(possible
regression in success rate of memory hotplug and CMA, and nr_freepage
meanging change) due to backport possibility.
2. nr_freepage without considering memory isolation is somewhat dangerous
and not suitable for some systems.
If you have any objection, please let me know. But, I will go on
a vacation for a week so I can't answer your further comments
for a week. I will reply them next week. :)
Thanks.
3. pages on cma buddy list are counted as CMA freepage, too.
4. pages for guard are *not* counted as freepage.
Now, I describe problems and related patch.
1. Patch 2: If guard page are cleared and merged into isolate buddy list,
we should not add freepage count.
2. Patch 3: When the page return back from pcp to buddy, we should
account it to freepage counter. In this case, we should check the
pageblock migratetype of the page and should insert the page into
appropriate buddy list. Although we checked it in current code, we
didn't insert the page into appropriate buddy list so that freepage
counting can be wrong.
3. Patch 4: There is race condition so that some freepages could be
on isolate buddy list. If so, we can't use this page until next isolation
attempt on this pageblock.
4. Patch 5: There is race condition that page on isolate pageblock
can go into non-isolate buddy list. If so, buddy allocator would
merge pages on non-isolate buddy list and isolate buddy list, respectively,
and freepage count will be wrong.
5. Patch 9: move_freepages(_block) returns *not* number of moved pages.
Instead, it returns number of pages linked in that migratetype buddy list.
So accouting with this return value makes freepage count wrong.
6. Patch 10: buddy allocator would merge pages on non-isolate buddy list
and isolate buddy list, respectively. This leads to freepage counting
problem so fix it by stopping merging in this case.
Without patchset [1], above problem doesn't happens on my CMA allocation
test, because CMA reserved pages aren't used at all. So there is no
chance for above race.
With patchset [1], I did simple CMA allocation test and get below result.
- Virtual machine, 4 cpus, 1024 MB memory, 256 MB CMA reservation
- run kernel build (make -j16) on background
- 30 times CMA allocation(8MB * 30 = 240MB) attempts in 5 sec interval
- Result: more than 5000 freepage count are missed
With patchset [1] and this patchset, I found that no freepage count are
missed so that I conclude that problems are solved.
These problems can be possible on memory hot remove users, although
I didn't check it further.
Other patches are either for the base to fix these problems or for
simple clean-up. Please see individual patches for more information.
This patchset is based on linux-next-20140703.
Thanks.
[1]: Aggressively allocate the pages on cma reserved memory
https://lkml.org/lkml/2014/5/30/291
Joonsoo Kim (10):
mm/page_alloc: remove unlikely macro on free_one_page()
mm/page_alloc: correct to clear guard attribute in DEBUG_PAGEALLOC
mm/page_alloc: handle page on pcp correctly if it's pageblock is
isolated
mm/page_alloc: carefully free the page on isolate pageblock
mm/page_alloc: optimize and unify pageblock migratetype check in free
path
mm/page_alloc: separate freepage migratetype interface
mm/page_alloc: store migratetype of the buddy list into freepage
correctly
mm/page_alloc: use get_onbuddy_migratetype() to get buddy list type
mm/page_alloc: fix possible wrongly calculated freepage counter
mm/page_alloc: Stop merging pages on non-isolate and isolate buddy
list
include/linux/mm.h | 30 +++++++--
include/linux/mmzone.h | 5 ++
include/linux/page-isolation.h | 8 +++
mm/page_alloc.c | 138 +++++++++++++++++++++++++++++-----------
mm/page_isolation.c | 18 ++----
5 files changed, 147 insertions(+), 52 deletions(-)
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>