We cache isolate_start_pfn before entering isolate_migratepages(). If
pageblock is skipped in isolate_migratepages() due to whatever reason,
cc->migrate_pfn can be far from isolate_start_pfn hence we flush pages
that were freed. For example, the following scenario can be possible:
- assume order-9 compaction, pageblock order is 9
- start_isolate_pfn is 0x200
- isolate_migratepages()
- skip a number of pageblocks
- start to isolate from pfn 0x600
- cc->migrate_pfn = 0x620
- return
- last_migrated_pfn is set to 0x200
- check flushing condition
- current_block_start is set to 0x600
- last_migrated_pfn < current_block_start then do useless flush
This wrong flush would not help the performance and success rate so this
patch tries to fix it. One simple way to know the exact position where
we start to isolate migratable pages is that we cache it in
isolate_migratepages() before entering actual isolation. This patch
implements that and fixes the problem.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Minchan Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
There is a problem that trace events are not properly enabled with
boot cmdline. The problem is that if we pass "trace_event=kmem:mm_page_alloc"
to the boot cmdline, it enables all kmem trace events, and not just
the page_alloc event.
This is caused by the parsing mechanism. When we parse the cmdline, the buffer
contents is modified due to tokenization. And, if we use this buffer
again, we will get the wrong result.
Unfortunately, this buffer is be accessed three times to set trace events
properly at boot time. So, we need to handle this situation.
There is already code handling ",", but we need another for ":".
This patch adds it.
Link: http://lkml.kernel.org/r/[email protected]
Cc: [email protected] # 3.19+
Signed-off-by: Joonsoo Kim <[email protected]>
[ added missing return ret; ]
Signed-off-by: Steven Rostedt <[email protected]>
Compaction has anti fragmentation algorithm. It is that freepage should
be more than pageblock order to finish the compaction if we don't find any
freepage in requested migratetype buddy list. This is for mitigating
fragmentation, but, there is a lack of migratetype consideration and it is
too excessive compared to page allocator's anti fragmentation algorithm.
Not considering migratetype would cause premature finish of compaction.
For example, if allocation request is for unmovable migratetype, freepage
with CMA migratetype doesn't help that allocation and compaction should
not be stopped. But, current logic regards this situation as compaction
is no longer needed, so finish the compaction.
Secondly, condition is too excessive compared to page allocator's logic.
We can steal freepage from other migratetype and change pageblock
migratetype on more relaxed conditions in page allocator. This is
designed to prevent fragmentation and we can use it here. Imposing hard
constraint only to the compaction doesn't help much in this case since
page allocator would cause fragmentation again.
To solve these problems, this patch borrows anti fragmentation logic from
page allocator. It will reduce premature compaction finish in some cases
and reduce excessive compaction work.
stress-highalloc test in mmtests with non movable order 7 allocation shows
considerable increase of compaction success rate.
Compaction success rate (Compaction success * 100 / Compaction stalls, %)
31.82 : 42.20
I tested it on non-reboot 5 runs stress-highalloc benchmark and found that
there is no more degradation on allocation success rate than before. That
roughly means that this patch doesn't result in more fragmentations.
Vlastimil suggests additional idea that we only test for fallbacks when
migration scanner has scanned a whole pageblock. It looked good for
fragmentation because chance of stealing increase due to making more free
pages in certain pageblock. So, I tested it, but, it results in decreased
compaction success rate, roughly 38.00. I guess the reason that if system
is low memory condition, watermark check could be failed due to not enough
order 0 free page and so, sometimes, we can't reach a fallback check
although migrate_pfn is aligned to pageblock_nr_pages. I can insert code
to cope with this situation but it makes code more complicated so I don't
include his idea at this patch.
[[email protected]: fix CONFIG_CMA=n build]
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This is preparation step to use page allocator's anti fragmentation logic
in compaction. This patch just separates fallback freepage checking part
from fallback freepage management part. Therefore, there is no functional
change.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Freepage with MIGRATE_CMA can be used only for MIGRATE_MOVABLE and they
should not be expanded to other migratetype buddy list to protect them
from unmovable/reclaimable allocation. Implementing these requirements in
__rmqueue_fallback(), that is, finding largest possible block of freepage
has bad effect that high order freepage with MIGRATE_CMA are broken
continually although there are suitable order CMA freepage. Reason is
that they are not be expanded to other migratetype buddy list and next
__rmqueue_fallback() invocation try to finds another largest block of
freepage and break it again. So, MIGRATE_CMA fallback should be handled
separately. This patch introduces __rmqueue_cma_fallback(), that just
wrapper of __rmqueue_smallest() and call it before __rmqueue_fallback() if
migratetype == MIGRATE_MOVABLE.
This results in unintended behaviour change that MIGRATE_CMA freepage is
always used first rather than other migratetype as movable allocation's
fallback. But, as already mentioned above, MIGRATE_CMA can be used only
for MIGRATE_MOVABLE, so it is better to use MIGRATE_CMA freepage first as
much as possible. Otherwise, we needlessly take up precious freepages
with other migratetype and increase chance of fragmentation.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently, freepage isolation in one pageblock doesn't consider how many
freepages we isolate. When I traced flow of compaction, compaction
sometimes isolates more than 256 freepages to migrate just 32 pages.
In this patch, freepage isolation is stopped at the point that we
have more isolated freepage than isolated page for migration. This
results in slowing down free page scanner and make compaction success
rate higher.
stress-highalloc test in mmtests with non movable order 7 allocation shows
increase of compaction success rate.
Compaction success rate (Compaction success * 100 / Compaction stalls, %)
27.13 : 31.82
pfn where both scanners meets on compaction complete
(separate test due to enormous tracepoint buffer)
(zone_start=4096, zone_end=1048576)
586034 : 654378
In fact, I didn't fully understand why this patch results in such good
result. There was a guess that not used freepages are released to pcp list
and on next compaction trial we won't isolate them again so compaction
success rate would decrease. To prevent this effect, I tested with adding
pcp drain code on release_freepages(), but, it has no good effect.
Anyway, this patch reduces waste time to isolate unneeded freepages so
seems reasonable.
Vlastimil said:
: I briefly tried it on top of the pivot-changing series and with order-9
: allocations it reduced free page scanned counter by almost 10%. No effect
: on success rates (maybe because pivot changing already took care of the
: scanners meeting problem) but the scanning reduction is good on its own.
:
: It also explains why e14c720efd ("mm, compaction: remember position
: within pageblock in free pages scanner") had less than expected
: improvements. It would only actually stop within pageblock in case of
: async compaction detecting contention. I guess that's also why the
: infinite loop problem fixed by 1d5bfe1ffb affected so relatively few
: people.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Tested-by: Vlastimil Babka <[email protected]>
Reviewed-by: Zhang Yanfei <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Rik van Riel <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
What we want to check here is whether there is highorder freepage in buddy
list of other migratetype in order to steal it without fragmentation.
But, current code just checks cc->order which means allocation request
order. So, this is wrong.
Without this fix, non-movable synchronous compaction below pageblock order
would not stopped until compaction is complete, because migratetype of
most pageblocks are movable and high order freepage made by compaction is
usually on movable type buddy list.
There is some report related to this bug. See below link.
http://www.spinics.net/lists/linux-mm/msg81666.html
Although the issued system still has load spike comes from compaction,
this makes that system completely stable and responsive according to his
report.
stress-highalloc test in mmtests with non movable order 7 allocation
doesn't show any notable difference in allocation success rate, but, it
shows more compaction success rate.
Compaction success rate (Compaction success * 100 / Compaction stalls, %)
18.47 : 28.94
Fixes: 1fb3f8ca0e ("mm: compaction: capture a suitable high-order page immediately when it is made available")
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Reviewed-by: Zhang Yanfei <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: <[email protected]> [3.7+]
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Compaction deferring logic is heavy hammer that block the way to the
compaction. It doesn't consider overall system state, so it could prevent
user from doing compaction falsely. In other words, even if system has
enough range of memory to compact, compaction would be skipped due to
compaction deferring logic. This patch add new tracepoint to understand
work of deferring logic. This will also help to check compaction success
and fail.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
It'd be useful to know current range where compaction work for detailed
analysis. With it, we can know pageblock where we actually scan and
isolate, and, how much pages we try in that pageblock and can guess why it
doesn't become freepage with pageblock order roughly.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We now have tracepoint for begin event of compaction and it prints start
position of both scanners, but, tracepoint for end event of compaction
doesn't print finish position of both scanners. It'd be also useful to
know finish position of both scanners so this patch add it. It will help
to find odd behavior or problem on compaction internal logic.
And mode is added to both begin/end tracepoint output, since according to
mode, compaction behavior is quite different.
And lastly, status format is changed to string rather than status number
for readability.
[[email protected]: fix sparse warning]
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Dan Carpenter <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
To check the range that compaction is working, tracepoint print
start/end pfn of zone and start pfn of both scanner with decimal format.
Since we manage all pages in order of 2 and it is well represented by
hexadecimal, this patch change the tracepoint format from decimal to
hexadecimal. This would improve readability. For example, it makes us
easily notice whether current scanner try to compact previously
attempted pageblock or not.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
compound_head() is implemented with assumption that there would be race
condition when checking tail flag. This assumption is only true when we
try to access arbitrary positioned struct page.
The situation that virt_to_head_page() is called is different case. We
call virt_to_head_page() only in the range of allocated pages, so there
is no race condition on tail flag. In this case, we don't need to
handle race condition and we can reduce overhead slightly. This patch
implements compound_head_fast() which is similar with compound_head()
except tail flag race handling. And then, virt_to_head_page() uses this
optimized function to improve performance.
I saw 1.8% win in a fast-path loop over kmem_cache_alloc/free, (14.063
ns -> 13.810 ns) if target object is on tail page.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Cc: Jesper Dangaard Brouer <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We had to insert a preempt enable/disable in the fastpath a while ago in
order to guarantee that tid and kmem_cache_cpu are retrieved on the same
cpu. It is the problem only for CONFIG_PREEMPT in which scheduler can
move the process to other cpu during retrieving data.
Now, I reach the solution to remove preempt enable/disable in the
fastpath. If tid is matched with kmem_cache_cpu's tid after tid and
kmem_cache_cpu are retrieved by separate this_cpu operation, it means
that they are retrieved on the same cpu. If not matched, we just have
to retry it.
With this guarantee, preemption enable/disable isn't need at all even if
CONFIG_PREEMPT, so this patch removes it.
I saw roughly 5% win in a fast-path loop over kmem_cache_alloc/free in
CONFIG_PREEMPT. (14.821 ns -> 14.049 ns)
Below is the result of Christoph's slab_test reported by Jesper Dangaard
Brouer.
* Before
Single thread testing
=====================
1. Kmalloc: Repeatedly allocate then free test
10000 times kmalloc(8) -> 49 cycles kfree -> 62 cycles
10000 times kmalloc(16) -> 48 cycles kfree -> 64 cycles
10000 times kmalloc(32) -> 53 cycles kfree -> 70 cycles
10000 times kmalloc(64) -> 64 cycles kfree -> 77 cycles
10000 times kmalloc(128) -> 74 cycles kfree -> 84 cycles
10000 times kmalloc(256) -> 84 cycles kfree -> 114 cycles
10000 times kmalloc(512) -> 83 cycles kfree -> 116 cycles
10000 times kmalloc(1024) -> 81 cycles kfree -> 120 cycles
10000 times kmalloc(2048) -> 104 cycles kfree -> 136 cycles
10000 times kmalloc(4096) -> 142 cycles kfree -> 165 cycles
10000 times kmalloc(8192) -> 238 cycles kfree -> 226 cycles
10000 times kmalloc(16384) -> 403 cycles kfree -> 264 cycles
2. Kmalloc: alloc/free test
10000 times kmalloc(8)/kfree -> 68 cycles
10000 times kmalloc(16)/kfree -> 68 cycles
10000 times kmalloc(32)/kfree -> 69 cycles
10000 times kmalloc(64)/kfree -> 68 cycles
10000 times kmalloc(128)/kfree -> 68 cycles
10000 times kmalloc(256)/kfree -> 68 cycles
10000 times kmalloc(512)/kfree -> 74 cycles
10000 times kmalloc(1024)/kfree -> 75 cycles
10000 times kmalloc(2048)/kfree -> 74 cycles
10000 times kmalloc(4096)/kfree -> 74 cycles
10000 times kmalloc(8192)/kfree -> 75 cycles
10000 times kmalloc(16384)/kfree -> 510 cycles
* After
Single thread testing
=====================
1. Kmalloc: Repeatedly allocate then free test
10000 times kmalloc(8) -> 46 cycles kfree -> 61 cycles
10000 times kmalloc(16) -> 46 cycles kfree -> 63 cycles
10000 times kmalloc(32) -> 49 cycles kfree -> 69 cycles
10000 times kmalloc(64) -> 57 cycles kfree -> 76 cycles
10000 times kmalloc(128) -> 66 cycles kfree -> 83 cycles
10000 times kmalloc(256) -> 84 cycles kfree -> 110 cycles
10000 times kmalloc(512) -> 77 cycles kfree -> 114 cycles
10000 times kmalloc(1024) -> 80 cycles kfree -> 116 cycles
10000 times kmalloc(2048) -> 102 cycles kfree -> 131 cycles
10000 times kmalloc(4096) -> 135 cycles kfree -> 163 cycles
10000 times kmalloc(8192) -> 238 cycles kfree -> 218 cycles
10000 times kmalloc(16384) -> 399 cycles kfree -> 262 cycles
2. Kmalloc: alloc/free test
10000 times kmalloc(8)/kfree -> 65 cycles
10000 times kmalloc(16)/kfree -> 66 cycles
10000 times kmalloc(32)/kfree -> 65 cycles
10000 times kmalloc(64)/kfree -> 66 cycles
10000 times kmalloc(128)/kfree -> 66 cycles
10000 times kmalloc(256)/kfree -> 71 cycles
10000 times kmalloc(512)/kfree -> 72 cycles
10000 times kmalloc(1024)/kfree -> 71 cycles
10000 times kmalloc(2048)/kfree -> 71 cycles
10000 times kmalloc(4096)/kfree -> 71 cycles
10000 times kmalloc(8192)/kfree -> 65 cycles
10000 times kmalloc(16384)/kfree -> 511 cycles
Most of the results are better than before.
Note that this change slightly worses performance in !CONFIG_PREEMPT,
roughly 0.3%. Implementing each case separately would help performance,
but, since it's so marginal, I didn't do that. This would help
maintanance since we have same code for all cases.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Tested-by: Jesper Dangaard Brouer <[email protected]>
Acked-by: Jesper Dangaard Brouer <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Kim Phillips reported following build failure.
LD init/built-in.o
mm/built-in.o: In function `free_pages_prepare':
mm/page_alloc.c:770: undefined reference to `.kernel_map_pages'
mm/built-in.o: In function `prep_new_page':
mm/page_alloc.c:933: undefined reference to `.kernel_map_pages'
mm/built-in.o: In function `map_pages':
mm/compaction.c:61: undefined reference to `.kernel_map_pages'
make: *** [vmlinux] Error 1
Reason for this problem is that commit 031bc5743f
("mm/debug-pagealloc: make debug-pagealloc boottime configurable")
forgot to remove the old declaration of kernel_map_pages() for some
architectures. This patch removes them to fix build failure.
Reported-by: Kim Phillips <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Heiko Carstens <[email protected]>
Cc: Martin Schwidefsky <[email protected]>
Cc: David Miller <[email protected]>
Cc: David Howells <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
zsmalloc has many size_classes to reduce fragmentation and they are in 16
bytes unit, for example, 16, 32, 48, etc., if PAGE_SIZE is 4096. And,
zsmalloc has constraint that each zspage has 4 pages at maximum.
In this situation, we can see interesting aspect. Let's think about
size_class for 1488, 1472, ..., 1376. To prevent external fragmentation,
they uses 4 pages per zspage and so all they can contain 11 objects at
maximum.
16384 (4096 * 4) = 1488 * 11 + remains
16384 (4096 * 4) = 1472 * 11 + remains
16384 (4096 * 4) = ...
16384 (4096 * 4) = 1376 * 11 + remains
It means that they have same characteristics and classification between
them isn't needed. If we use one size_class for them, we can reduce
fragementation and save some memory since both the 1488 and 1472 sized
classes can only fit 11 objects into 4 pages, and an object that's 1472
bytes can fit into an object that's 1488 bytes, merging these classes to
always use objects that are 1488 bytes will reduce the total number of
size classes. And reducing the total number of size classes reduces
overall fragmentation, because a wider range of compressed pages can fit
into a single size class, leaving less unused objects in each size class.
For this purpose, this patch implement size_class merging. If there is
size_class that have same pages_per_zspage and same number of objects per
zspage with previous size_class, we don't create new size_class. Instead,
we use previous, same characteristic size_class. With this way, above
example sizes (1488, 1472, ..., 1376) use just one size_class so we can
get much more memory utilization.
Below is result of my simple test.
TEST ENV: EXT4 on zram, mount with discard option WORKLOAD: untar kernel
source code, remove directory in descending order in size. (drivers arch
fs sound include net Documentation firmware kernel tools)
Each line represents orig_data_size, compr_data_size, mem_used_total,
fragmentation overhead (mem_used - compr_data_size) and overhead ratio
(overhead to compr_data_size), respectively, after untar and remove
operation is executed.
* untar-nomerge.out
orig_size compr_size used_size overhead overhead_ratio
525.88MB 199.16MB 210.23MB 11.08MB 5.56%
288.32MB 97.43MB 105.63MB 8.20MB 8.41%
177.32MB 61.12MB 69.40MB 8.28MB 13.55%
146.47MB 47.32MB 56.10MB 8.78MB 18.55%
124.16MB 38.85MB 48.41MB 9.55MB 24.58%
103.93MB 31.68MB 40.93MB 9.25MB 29.21%
84.34MB 22.86MB 32.72MB 9.86MB 43.13%
66.87MB 14.83MB 23.83MB 9.00MB 60.70%
60.67MB 11.11MB 18.60MB 7.49MB 67.48%
55.86MB 8.83MB 16.61MB 7.77MB 88.03%
53.32MB 8.01MB 15.32MB 7.31MB 91.24%
* untar-merge.out
orig_size compr_size used_size overhead overhead_ratio
526.23MB 199.18MB 209.81MB 10.64MB 5.34%
288.68MB 97.45MB 104.08MB 6.63MB 6.80%
177.68MB 61.14MB 66.93MB 5.79MB 9.47%
146.83MB 47.34MB 52.79MB 5.45MB 11.51%
124.52MB 38.87MB 44.30MB 5.43MB 13.96%
104.29MB 31.70MB 36.83MB 5.13MB 16.19%
84.70MB 22.88MB 27.92MB 5.04MB 22.04%
67.11MB 14.83MB 19.26MB 4.43MB 29.86%
60.82MB 11.10MB 14.90MB 3.79MB 34.17%
55.90MB 8.82MB 12.61MB 3.79MB 42.97%
53.32MB 8.01MB 11.73MB 3.73MB 46.53%
As you can see above result, merged one has better utilization (overhead
ratio, 5th column) and uses less memory (mem_used_total, 3rd column).
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Nitin Gupta <[email protected]>
Cc: Jerome Marchand <[email protected]>
Cc: Sergey Senozhatsky <[email protected]>
Reviewed-by: Dan Streetman <[email protected]>
Cc: Luigi Semenzato <[email protected]>
Cc: <[email protected]>
Cc: "seungho1.park" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Extended memory to store page owner information is initialized some time
later than that page allocator starts. Until initialization, many pages
can be allocated and they have no owner information. This make debugging
using page owner harder, so some fixup will be helpful.
This patch fixes up this situation by setting fake owner information
immediately after page extension is initialized. Information doesn't tell
the right owner, but, at least, it can tell whether page is allocated or
not, more correctly.
On my testing, this patch catches 13343 early allocated pages, although
they are mostly allocated from page extension feature. Anyway, after
then, there is no page left that it is allocated and has no page owner
flag.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Jungsoo Son <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This is the page owner tracking code which is introduced so far ago. It
is resident on Andrew's tree, though, nobody tried to upstream so it
remain as is. Our company uses this feature actively to debug memory leak
or to find a memory hogger so I decide to upstream this feature.
This functionality help us to know who allocates the page. When
allocating a page, we store some information about allocation in extra
memory. Later, if we need to know status of all pages, we can get and
analyze it from this stored information.
In previous version of this feature, extra memory is statically defined in
struct page, but, in this version, extra memory is allocated outside of
struct page. It enables us to turn on/off this feature at boottime
without considerable memory waste.
Although we already have tracepoint for tracing page allocation/free,
using it to analyze page owner is rather complex. We need to enlarge the
trace buffer for preventing overlapping until userspace program launched.
And, launched program continually dump out the trace buffer for later
analysis and it would change system behaviour with more possibility rather
than just keeping it in memory, so bad for debug.
Moreover, we can use page_owner feature further for various purposes. For
example, we can use it for fragmentation statistics implemented in this
patch. And, I also plan to implement some CMA failure debugging feature
using this interface.
I'd like to give the credit for all developers contributed this feature,
but, it's not easy because I don't know exact history. Sorry about that.
Below is people who has "Signed-off-by" in the patches in Andrew's tree.
Contributor:
Alexander Nyberg <[email protected]>
Mel Gorman <[email protected]>
Dave Hansen <[email protected]>
Minchan Kim <[email protected]>
Michal Nazarewicz <[email protected]>
Andrew Morton <[email protected]>
Jungsoo Son <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Jungsoo Son <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_mmap_private() in nommu.c try to allocate physically contiguous pages
with arbitrary size in some cases and we now have good abstract function
to do exactly same thing, alloc_pages_exact(). So, change to use it.
There is no functional change. This is the preparation step for support
page owner feature accurately.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Jungsoo Son <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we have prepared to avoid using debug-pagealloc in boottime. So
introduce new kernel-parameter to disable debug-pagealloc in boottime, and
makes related functions to be disabled in this case.
Only non-intuitive part is change of guard page functions. Because guard
page is effective only if debug-pagealloc is enabled, turning off
according to debug-pagealloc is reasonable thing to do.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Jungsoo Son <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Until now, debug-pagealloc needs extra flags in struct page, so we need to
recompile whole source code when we decide to use it. This is really
painful, because it takes some time to recompile and sometimes rebuild is
not possible due to third party module depending on struct page. So, we
can't use this good feature in many cases.
Now, we have the page extension feature that allows us to insert extra
flags to outside of struct page. This gets rid of third party module
issue mentioned above. And, this allows us to determine if we need extra
memory for this page extension in boottime. With these property, we can
avoid using debug-pagealloc in boottime with low computational overhead in
the kernel built with CONFIG_DEBUG_PAGEALLOC. This will help our
development process greatly.
This patch is the preparation step to achive above goal. debug-pagealloc
originally uses extra field of struct page, but, after this patch, it will
use field of struct page_ext. Because memory for page_ext is allocated
later than initialization of page allocator in CONFIG_SPARSEMEM, we should
disable debug-pagealloc feature temporarily until initialization of
page_ext. This patch implements this.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Jungsoo Son <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
When we debug something, we'd like to insert some information to every
page. For this purpose, we sometimes modify struct page itself. But,
this has drawbacks. First, it requires re-compile. This makes us
hesitate to use the powerful debug feature so development process is
slowed down. And, second, sometimes it is impossible to rebuild the
kernel due to third party module dependency. At third, system behaviour
would be largely different after re-compile, because it changes size of
struct page greatly and this structure is accessed by every part of
kernel. Keeping this as it is would be better to reproduce errornous
situation.
This feature is intended to overcome above mentioned problems. This
feature allocates memory for extended data per page in certain place
rather than the struct page itself. This memory can be accessed by the
accessor functions provided by this code. During the boot process, it
checks whether allocation of huge chunk of memory is needed or not. If
not, it avoids allocating memory at all. With this advantage, we can
include this feature into the kernel in default and can avoid rebuild and
solve related problems.
Until now, memcg uses this technique. But, now, memcg decides to embed
their variable to struct page itself and it's code to extend struct page
has been removed. I'd like to use this code to develop debug feature, so
this patch resurrect it.
To help these things to work well, this patch introduces two callbacks for
clients. One is the need callback which is mandatory if user wants to
avoid useless memory allocation at boot-time. The other is optional, init
callback, which is used to do proper initialization after memory is
allocated. Detailed explanation about purpose of these functions is in
code comment. Please refer it.
Others are completely same with previous extension code in memcg.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Jungsoo Son <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Page guard is used by debug-pagealloc feature. Currently, it is
open-coded, but, I think that more abstraction of it makes core page
allocator code more readable.
There is no functional difference.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Gioh Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Unlike SLUB, sometimes, object isn't started at the beginning of the slab
in the SLAB. This causes the unalignment problem when after slab merging
is supported by commit 12220dea07 ("mm/slab: support slab merge").
Alignment mismatch check is introduced ("mm/slab: fix unalignment problem
on Malta with EVA due to slab merge") to prevent merge in this case.
This causes undesirable result that merging happens between infrequently
used kmem_caches if there are kmem_caches with same size and is 256 bytes,
are merged into pool_workqueue rather than kmalloc-256, because
kmem_caches for kmalloc are at the tail of the list.
To prevent this situation, this patch reverses iteration order in
find_mergeable() to find frequently used kmem_caches. This change helps
to merge kmem_cache to frequently used kmem_caches, such as kmalloc
kmem_caches.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Joonsoo Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
high_memory isn't direct mapped memory so retrieving it's physical address
isn't appropriate. But, it would be useful to check physical address of
highmem boundary so it's justfiable to get physical address from it. In
x86, there is a validation check if CONFIG_DEBUG_VIRTUAL and it triggers
following boot failure reported by Ingo.
...
BUG: Int 6: CR2 00f06f53
...
Call Trace:
dump_stack+0x41/0x52
early_idt_handler+0x6b/0x6b
cma_declare_contiguous+0x33/0x212
dma_contiguous_reserve_area+0x31/0x4e
dma_contiguous_reserve+0x11d/0x125
setup_arch+0x7b5/0xb63
start_kernel+0xb8/0x3e6
i386_start_kernel+0x79/0x7d
To fix boot regression, this patch implements workaround to avoid
validation check in x86 when retrieving physical address of high_memory.
__pa_nodebug() used by this patch is implemented only in x86 so there is
no choice but to use dirty #ifdef.
[[email protected]: tweak comment]
Signed-off-by: Joonsoo Kim <[email protected]>
Reported-by: Ingo Molnar <[email protected]>
Tested-by: Ingo Molnar <[email protected]>
Cc: Marek Szyprowski <[email protected]>
Cc: Russell King <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I'm stuck into panic that too litte free memory is left when I boot with
trace_buf_size parameter. After digging into the problem, I found that
trace_buf_size is the size of trace buffer on each cpu rather than total
size of trace buffer. To prevent victim like me, change description of
trace_buf_size parameter more accurately.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
One thing I did in this patch is fixing freepage accounting. If we
clear guard page and link it onto isolate buddy list, we should not
increase freepage count. This patch adds conditional branch to skip
counting in this case. Without this patch, this overcounting happens
frequently if guard order is set and CMA is used.
Another thing fixed in this patch is the target to reset order. In
__free_one_page(), we check the buddy page whether it is a guard page or
not. And, if so, we should clear guard attribute on the buddy page and
reset order of it to 0. But, current code resets original page's order
rather than buddy one's. Maybe, this doesn't have any problem, because
whole merged page's order will be re-assigned soon. But, it is better
to correct code.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Gioh Kim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Unlike SLUB, sometimes, object isn't started at the beginning of the
slab in SLAB. This causes the unalignment problem after slab merging is
supported by commit 12220dea07 ("mm/slab: support slab merge").
Following is the report from Markos that fail to boot on Malta with EVA.
Calibrating delay loop... 19.86 BogoMIPS (lpj=99328)
pid_max: default: 32768 minimum: 301
Mount-cache hash table entries: 4096 (order: 0, 16384 bytes)
Mountpoint-cache hash table entries: 4096 (order: 0, 16384 bytes)
Kernel bug detected[#1]:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.17.0-05639-g12220dea07f1 #1631
task: 1f04f5d8 ti: 1f050000 task.ti: 1f050000
epc : 80141190 alloc_unbound_pwq+0x234/0x304
Not tainted
ra : 80141184 alloc_unbound_pwq+0x228/0x304
Process swapper/0 (pid: 1, threadinfo=1f050000, task=1f04f5d8, tls=00000000)
Call Trace:
alloc_unbound_pwq+0x234/0x304
apply_workqueue_attrs+0x11c/0x294
__alloc_workqueue_key+0x23c/0x470
init_workqueues+0x320/0x400
do_one_initcall+0xe8/0x23c
kernel_init_freeable+0x9c/0x224
kernel_init+0x10/0x100
ret_from_kernel_thread+0x14/0x1c
[ end trace cb88537fdc8fa200 ]
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
alloc_unbound_pwq() allocates slab object from pool_workqueue. This
kmem_cache requires 256 bytes alignment, but, current merging code
doesn't honor that, and merge it with kmalloc-256. kmalloc-256 requires
only cacheline size alignment so that above failure occurs. However, in
x86, kmalloc-256 is luckily aligned in 256 bytes, so the problem didn't
happen on it.
To fix this problem, this patch introduces alignment mismatch check in
find_mergeable(). This will fix the problem.
Signed-off-by: Joonsoo Kim <[email protected]>
Reported-by: Markos Chandras <[email protected]>
Tested-by: Markos Chandras <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Current pageblock isolation logic could isolate each pageblock
individually. This causes freepage accounting problem if freepage with
pageblock order on isolate pageblock is merged with other freepage on
normal pageblock. We can prevent merging by restricting max order of
merging to pageblock order if freepage is on isolate pageblock.
A side-effect of this change is that there could be non-merged buddy
freepage even if finishing pageblock isolation, because undoing
pageblock isolation is just to move freepage from isolate buddy list to
normal buddy list rather than to consider merging. So, the patch also
makes undoing pageblock isolation consider freepage merge. When
un-isolation, freepage with more than pageblock order and it's buddy are
checked. If they are on normal pageblock, instead of just moving, we
isolate the freepage and free it in order to get merged.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: "Kirill A. Shutemov" <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Yasuaki Ishimatsu <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Cc: Tang Chen <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Bartlomiej Zolnierkiewicz <[email protected]>
Cc: Wen Congyang <[email protected]>
Cc: Marek Szyprowski <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Laura Abbott <[email protected]>
Cc: Heesub Shin <[email protected]>
Cc: "Aneesh Kumar K.V" <[email protected]>
Cc: Ritesh Harjani <[email protected]>
Cc: Gioh Kim <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
In free_pcppages_bulk(), we use cached migratetype of freepage to
determine type of buddy list where freepage will be added. This
information is stored when freepage is added to pcp list, so if
isolation of pageblock of this freepage begins after storing, this
cached information could be stale. In other words, it has original
migratetype rather than MIGRATE_ISOLATE.
There are two problems caused by this stale information.
One is that we can't keep these freepages from being allocated.
Although this pageblock is isolated, freepage will be added to normal
buddy list so that it could be allocated without any restriction. And
the other problem is incorrect freepage accounting. Freepages on
isolate pageblock should not be counted for number of freepage.
Following is the code snippet in free_pcppages_bulk().
/* MIGRATE_MOVABLE list may include MIGRATE_RESERVEs */
__free_one_page(page, page_to_pfn(page), zone, 0, mt);
trace_mm_page_pcpu_drain(page, 0, mt);
if (likely(!is_migrate_isolate_page(page))) {
__mod_zone_page_state(zone, NR_FREE_PAGES, 1);
if (is_migrate_cma(mt))
__mod_zone_page_state(zone, NR_FREE_CMA_PAGES, 1);
}
As you can see above snippet, current code already handle second
problem, incorrect freepage accounting, by re-fetching pageblock
migratetype through is_migrate_isolate_page(page).
But, because this re-fetched information isn't used for
__free_one_page(), first problem would not be solved. This patch try to
solve this situation to re-fetch pageblock migratetype before
__free_one_page() and to use it for __free_one_page().
In addition to move up position of this re-fetch, this patch use
optimization technique, re-fetching migratetype only if there is isolate
pageblock. Pageblock isolation is rare event, so we can avoid
re-fetching in common case with this optimization.
This patch also correct migratetype of the tracepoint output.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Minchan Kim <[email protected]>
Acked-by: Michal Nazarewicz <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: "Kirill A. Shutemov" <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Yasuaki Ishimatsu <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Cc: Tang Chen <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Bartlomiej Zolnierkiewicz <[email protected]>
Cc: Wen Congyang <[email protected]>
Cc: Marek Szyprowski <[email protected]>
Cc: Laura Abbott <[email protected]>
Cc: Heesub Shin <[email protected]>
Cc: "Aneesh Kumar K.V" <[email protected]>
Cc: Ritesh Harjani <[email protected]>
Cc: Gioh Kim <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
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.
3. pages on cma buddy list are counted as CMA freepage, too.
Now, I describe problems and related patch.
Patch 1: There is race conditions on getting pageblock migratetype that
it results in misplacement of freepages on buddy list, incorrect
freepage count and un-availability of freepage.
Patch 2: Freepages on pcp list could have stale cached information to
determine migratetype of buddy list to go. This causes misplacement of
freepages on buddy list and incorrect freepage count.
Patch 4: Merging between freepages on different migratetype of
pageblocks will cause freepages accouting problem. This patch fixes it.
Without patchset [3], 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 [3], 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 [3] and this patchset, I found that no freepage count are
missed so that I conclude that problems are solved.
On my simple memory offlining test, these problems also occur on that
environment, too.
This patch (of 4):
There are two paths to reach core free function of buddy allocator,
__free_one_page(), one is free_one_page()->__free_one_page() and the
other is free_hot_cold_page()->free_pcppages_bulk()->__free_one_page().
Each paths has race condition causing serious problems. At first, this
patch is focused on first type of freepath. And then, following patch
will solve the problem in second type of freepath.
In the first type of freepath, we got migratetype of freeing page
without holding the zone lock, so it could be racy. There are two cases
of this race.
1. pages are added to isolate buddy list after restoring orignal
migratetype
CPU1 CPU2
get migratetype => return MIGRATE_ISOLATE
call free_one_page() with MIGRATE_ISOLATE
grab the zone lock
unisolate pageblock
release the zone lock
grab the zone lock
call __free_one_page() with MIGRATE_ISOLATE
freepage go into isolate buddy list,
although pageblock is already unisolated
This may cause two problems. One is that we can't use this page anymore
until next isolation attempt of this pageblock, because freepage is on
isolate buddy list. The other is that freepage accouting could be wrong
due to merging between different buddy list. Freepages on isolate buddy
list aren't counted as freepage, but ones on normal buddy list are
counted as freepage. If merge happens, buddy freepage on normal buddy
list is inevitably moved to isolate buddy list without any consideration
of freepage accouting so it could be incorrect.
2. pages are added to normal buddy list while pageblock is isolated.
It is similar with above case.
This also may cause two problems. One is that we can't keep these
freepages from being allocated. Although this pageblock is isolated,
freepage would be added to normal buddy list so that it could be
allocated without any restriction. And the other problem is same as
case 1, that it, incorrect freepage accouting.
This race condition would be prevented by checking migratetype again
with holding the zone lock. Because it is somewhat heavy operation and
it isn't needed in common case, we want to avoid rechecking as much as
possible. So this patch introduce new variable, nr_isolate_pageblock in
struct zone to check if there is isolated pageblock. With this, we can
avoid to re-check migratetype in common case and do it only if there is
isolated pageblock or migratetype is MIGRATE_ISOLATE. This solve above
mentioned problems.
Changes from v3:
Add one more check in free_one_page() that checks whether migratetype is
MIGRATE_ISOLATE or not. Without this, abovementioned case 1 could happens.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Minchan Kim <[email protected]>
Acked-by: Michal Nazarewicz <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: "Kirill A. Shutemov" <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Yasuaki Ishimatsu <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Cc: Tang Chen <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Bartlomiej Zolnierkiewicz <[email protected]>
Cc: Wen Congyang <[email protected]>
Cc: Marek Szyprowski <[email protected]>
Cc: Laura Abbott <[email protected]>
Cc: Heesub Shin <[email protected]>
Cc: "Aneesh Kumar K.V" <[email protected]>
Cc: Ritesh Harjani <[email protected]>
Cc: Gioh Kim <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Commit 7d49d88683 ("mm, compaction: reduce zone checking frequency in
the migration scanner") has a side-effect that changes the iteration
range calculation. Before the change, block_end_pfn is calculated using
start_pfn, but now it blindly adds pageblock_nr_pages to the previous
value.
This causes the problem that isolation_start_pfn is larger than
block_end_pfn when we isolate the page with more than pageblock order.
In this case, isolation would fail due to an invalid range parameter.
To prevent this, this patch implements skipping the range until a proper
target pageblock is met. Without this patch, CMA with more than
pageblock order always fails but with this patch it will succeed.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Commit edc2ca6124 ("mm, compaction: move pageblock checks up from
isolate_migratepages_range()") commonizes isolate_migratepages variants
and make them use isolate_migratepages_block().
isolate_migratepages_block() could stop the execution when enough pages
are isolated, but, there is no code in isolate_migratepages_range() to
handle this case. In the result, even if isolate_migratepages_block()
returns prematurely without checking all pages in the range,
isolate_migratepages_block() is called repeately on the following
pageblock and some pages in the previous range are skipped to check.
Then, CMA is failed frequently due to this fact.
To fix this problem, this patch let isolate_migratepages_range() know
the situation that enough pages are isolated and stop the isolation in
that case.
Note that isolate_migratepages() has no such problem, because, it always
stops the isolation after just one call of isolate_migratepages_block().
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Michal Nazarewicz <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Commit bf0dea23a9 ("mm/slab: use percpu allocator for cpu cache")
changed the allocation method for cpu cache array from slab allocator to
percpu allocator. Alignment should be provided for aligned memory in
percpu allocator case, but, that commit mistakenly set this alignment to
0. So, percpu allocator returns unaligned memory address. It doesn't
cause any problem on x86 which permits unaligned access, but, it causes
the problem on sparc64 which needs strong guarantee of alignment.
Following bug report is reported from David Miller.
I'm getting tons of the following on sparc64:
[603965.383447] Kernel unaligned access at TPC[546b58] free_block+0x98/0x1a0
[603965.396987] Kernel unaligned access at TPC[546b60] free_block+0xa0/0x1a0
...
[603970.554394] log_unaligned: 333 callbacks suppressed
...
This patch provides a proper alignment parameter when allocating cpu
cache to fix this unaligned memory access problem on sparc64.
Reported-by: David Miller <[email protected]>
Tested-by: David Miller <[email protected]>
Tested-by: Meelis Roos <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Because of chicken and egg problem, initialization of SLAB is really
complicated. We need to allocate cpu cache through SLAB to make the
kmem_cache work, but before initialization of kmem_cache, allocation
through SLAB is impossible.
On the other hand, SLUB does initialization in a more simple way. It uses
percpu allocator to allocate cpu cache so there is no chicken and egg
problem.
So, this patch try to use percpu allocator in SLAB. This simplifies the
initialization step in SLAB so that we could maintain SLAB code more
easily.
In my testing there is no performance difference.
This implementation relies on percpu allocator. Because percpu allocator
uses vmalloc address space, vmalloc address space could be exhausted by
this change on many cpu system with *32 bit* kernel. This implementation
can cover 1024 cpus in worst case by following calculation.
Worst: 1024 cpus * 4 bytes for pointer * 300 kmem_caches *
120 objects per cpu_cache = 140 MB
Normal: 1024 cpus * 4 bytes for pointer * 150 kmem_caches(slab merge) *
80 objects per cpu_cache = 46 MB
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Jeremiah Mahler <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Slab merge is good feature to reduce fragmentation. If new creating slab
have similar size and property with exsitent slab, this feature reuse it
rather than creating new one. As a result, objects are packed into fewer
slabs so that fragmentation is reduced.
Below is result of my testing.
* After boot, sleep 20; cat /proc/meminfo | grep Slab
<Before>
Slab: 25136 kB
<After>
Slab: 24364 kB
We can save 3% memory used by slab.
For supporting this feature in SLAB, we need to implement SLAB specific
kmem_cache_flag() and __kmem_cache_alias(), because SLUB implements some
SLUB specific processing related to debug flag and object size change on
these functions.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Slab merge is good feature to reduce fragmentation. Now, it is only
applied to SLUB, but, it would be good to apply it to SLAB. This patch is
preparation step to apply slab merge to SLAB by commonizing slab merge
logic.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Anton noticed (http://www.spinics.net/lists/linux-mm/msg67489.html) that
on ppc LPARs with memoryless nodes, a large amount of memory was consumed
by slabs and was marked unreclaimable. He tracked it down to slab
deactivations in the SLUB core when we allocate remotely, leading to poor
efficiency always when memoryless nodes are present.
After much discussion, Joonsoo provided a few patches that help
significantly. They don't resolve the problem altogether:
- memory hotplug still needs testing, that is when a memoryless node
becomes memory-ful, we want to dtrt
- there are other reasons for going off-node than memoryless nodes,
e.g., fully exhausted local nodes
Neither case is resolved with this series, but I don't think that should
block their acceptance, as they can be explored/resolved with follow-on
patches.
The series consists of:
[1/3] topology: add support for node_to_mem_node() to determine the
fallback node
[2/3] slub: fallback to node_to_mem_node() node if allocating on
memoryless node
- Joonsoo's patches to cache the nearest node with memory for each
NUMA node
[3/3] Partial revert of 81c98869fa (""kthread: ensure locality of
task_struct allocations")
- At Tejun's request, keep the knowledge of memoryless node fallback
to the allocator core.
This patch (of 3):
We need to determine the fallback node in slub allocator if the allocation
target node is memoryless node. Without it, the SLUB wrongly select the
node which has no memory and can't use a partial slab, because of node
mismatch. Introduced function, node_to_mem_node(X), will return a node Y
with memory that has the nearest distance. If X is memoryless node, it
will return nearest distance node, but, if X is normal node, it will
return itself.
We will use this function in following patch to determine the fallback
node.
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Nishanth Aravamudan <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Han Pingtian <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Benjamin Herrenschmidt <[email protected]>
Cc: Michael Ellerman <[email protected]>
Cc: Anton Blanchard <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Wanpeng Li <[email protected]>
Cc: Tejun Heo <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
cache_free_alien() is rarely used function when node mismatch. But, it is
defined with inline attribute so it is inlined to __cache_free() which is
core free function of slab allocator. It uselessly makes
kmem_cache_free()/kfree() functions large. What we really need to inline
is just checking node match so this patch factor out other parts of
cache_free_alien() to reduce code size of kmem_cache_free()/ kfree().
<Before>
nm -S mm/slab.o | grep -e "T kfree" -e "T kmem_cache_free"
00000000000011e0 0000000000000228 T kfree
0000000000000670 0000000000000216 T kmem_cache_free
<After>
nm -S mm/slab.o | grep -e "T kfree" -e "T kmem_cache_free"
0000000000001110 00000000000001b5 T kfree
0000000000000750 0000000000000181 T kmem_cache_free
You can see slightly reduced size of text: 0x228->0x1b5, 0x216->0x181.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Our intention of __ac_put_obj() is that it doesn't affect anything if
sk_memalloc_socks() is disabled. But, because __ac_put_obj() is too
small, compiler inline it to ac_put_obj() and affect code size of free
path. This patch add noinline keyword for __ac_put_obj() not to distrupt
normal free path at all.
<Before>
nm -S slab-orig.o |
grep -e "t cache_alloc_refill" -e "T kfree" -e "T kmem_cache_free"
0000000000001e80 00000000000002f5 t cache_alloc_refill
0000000000001230 0000000000000258 T kfree
0000000000000690 000000000000024c T kmem_cache_free
<After>
nm -S slab-patched.o |
grep -e "t cache_alloc_refill" -e "T kfree" -e "T kmem_cache_free"
0000000000001e00 00000000000002e5 t cache_alloc_refill
00000000000011e0 0000000000000228 T kfree
0000000000000670 0000000000000216 T kmem_cache_free
cache_alloc_refill: 0x2f5->0x2e5
kfree: 0x256->0x228
kmem_cache_free: 0x24c->0x216
code size of each function is reduced slightly.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, due to likely keyword, compiled code of cache_flusharray() is on
unlikely.text section. Although it is uncommon case compared to free to
cpu cache case, it is common case than free_block(). But, free_block() is
on normal text section. This patch fix this odd situation to remove
likely keyword.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we track caller if tracing or slab debugging is enabled. If they are
disabled, we could save one argument passing overhead by calling
__kmalloc(_node)(). But, I think that it would be marginal. Furthermore,
default slab allocator, SLUB, doesn't use this technique so I think that
it's okay to change this situation.
After this change, we can turn on/off CONFIG_DEBUG_SLAB without full
kernel build and remove some complicated '#if' defintion. It looks more
benefitial to me.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We don't need to keep kmem_cache definition in include/linux/slab.h if we
don't need to inline kmem_cache_size(). According to my code inspection,
this function is only called at lc_create() in lib/lru_cache.c which may
be called at initialization phase of something, so we don't need to inline
it. Therfore, move it to slab_common.c and move kmem_cache definition to
internal header.
After this change, we can change kmem_cache definition easily without full
kernel build. For instance, we can turn on/off CONFIG_SLUB_STATS without
full kernel build.
[[email protected]: export kmem_cache_size() to modules]
[[email protected]: add header files to fix kmemcheck.c build errors]
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Randy Dunlap <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This reverts commit a640616822 ("slab: remove BAD_ALIEN_MAGIC").
commit a640616822 ("slab: remove BAD_ALIEN_MAGIC") assumes that the
system with !CONFIG_NUMA has only one memory node. But, it turns out to
be false by the report from Geert. His system, m68k, has many memory
nodes and is configured in !CONFIG_NUMA. So it couldn't boot with above
change.
Here goes his failure report.
With latest mainline, I'm getting a crash during bootup on m68k/ARAnyM:
enable_cpucache failed for radix_tree_node, error 12.
kernel BUG at /scratch/geert/linux/linux-m68k/mm/slab.c:1522!
*** TRAP #7 *** FORMAT=0
Current process id is 0
BAD KERNEL TRAP: 00000000
Modules linked in:
PC: [<0039c92c>] kmem_cache_init_late+0x70/0x8c
SR: 2200 SP: 00345f90 a2: 0034c2e8
d0: 0000003d d1: 00000000 d2: 00000000 d3: 003ac942
d4: 00000000 d5: 00000000 a0: 0034f686 a1: 0034f682
Process swapper (pid: 0, task=0034c2e8)
Frame format=0
Stack from 00345fc4:
002f69ef 002ff7e5 000005f2 000360fa 0017d806 003921d4 00000000
00000000 00000000 00000000 00000000 00000000 003ac942 00000000
003912d6
Call Trace: [<000360fa>] parse_args+0x0/0x2ca
[<0017d806>] strlen+0x0/0x1a
[<003921d4>] start_kernel+0x23c/0x428
[<003912d6>] _sinittext+0x2d6/0x95e
Code: f7e5 4879 002f 69ef 61ff ffca 462a 4e47 <4879> 0035 4b1c 61ff
fff0 0cc4 7005 23c0 0037 fd20 588f 265f 285f 4e75 48e7 301c
Disabling lock debugging due to kernel taint
Kernel panic - not syncing: Attempted to kill the idle task!
Although there is a alternative way to fix this issue such as disabling
use of alien cache on !CONFIG_NUMA, but, reverting issued commit is better
to me in this time.
Signed-off-by: Joonsoo Kim <[email protected]>
Reported-by: Geert Uytterhoeven <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Vladimir Davydov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently, there are two users on CMA functionality, one is the DMA
subsystem and the other is the KVM on powerpc. They have their own code
to manage CMA reserved area even if they looks really similar. From my
guess, it is caused by some needs on bitmap management. KVM side wants
to maintain bitmap not for 1 page, but for more size. Eventually it use
bitmap where one bit represents 64 pages.
When I implement CMA related patches, I should change those two places
to apply my change and it seem to be painful to me. I want to change
this situation and reduce future code management overhead through this
patch.
This change could also help developer who want to use CMA in their new
feature development, since they can use CMA easily without copying &
pasting this reserved area management code.
In previous patches, we have prepared some features to generalize CMA
reserved area management and now it's time to do it. This patch moves
core functions to mm/cma.c and change DMA APIs to use these functions.
There is no functional change in DMA APIs.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Michal Nazarewicz <[email protected]>
Acked-by: Zhang Yanfei <[email protected]>
Acked-by: Minchan Kim <[email protected]>
Reviewed-by: Aneesh Kumar K.V <[email protected]>
Cc: Alexander Graf <[email protected]>
Cc: Aneesh Kumar K.V <[email protected]>
Cc: Gleb Natapov <[email protected]>
Acked-by: Marek Szyprowski <[email protected]>
Tested-by: Marek Szyprowski <[email protected]>
Cc: Paolo Bonzini <[email protected]>
Cc: Benjamin Herrenschmidt <[email protected]>
Cc: Paul Mackerras <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Richard Yao reported a month ago that his system have a trouble with
vmap_area_lock contention during performance analysis by /proc/meminfo.
Andrew asked why his analysis checks /proc/meminfo stressfully, but he
didn't answer it.
https://lkml.org/lkml/2014/4/10/416
Although I'm not sure that this is right usage or not, there is a
solution reducing vmap_area_lock contention with no side-effect. That
is just to use rcu list iterator in get_vmalloc_info().
rcu can be used in this function because all RCU protocol is already
respected by writers, since Nick Piggin commit db64fe0225 ("mm:
rewrite vmap layer") back in linux-2.6.28
Specifically :
insertions use list_add_rcu(),
deletions use list_del_rcu() and kfree_rcu().
Note the rb tree is not used from rcu reader (it would not be safe),
only the vmap_area_list has full RCU protection.
Note that __purge_vmap_area_lazy() already uses this rcu protection.
rcu_read_lock();
list_for_each_entry_rcu(va, &vmap_area_list, list) {
if (va->flags & VM_LAZY_FREE) {
if (va->va_start < *start)
*start = va->va_start;
if (va->va_end > *end)
*end = va->va_end;
nr += (va->va_end - va->va_start) >> PAGE_SHIFT;
list_add_tail(&va->purge_list, &valist);
va->flags |= VM_LAZY_FREEING;
va->flags &= ~VM_LAZY_FREE;
}
}
rcu_read_unlock();
Peter:
: While rcu list traversal over the vmap_area_list is safe, this may
: arrive at different results than the spinlocked version. The rcu list
: traversal version will not be a 'snapshot' of a single, valid instant
: of the entire vmap_area_list, but rather a potential amalgam of
: different list states.
Joonsoo:
: Yes, you are right, but I don't think that we should be strict here.
: Meminfo is already not a 'snapshot' at specific time. While we try to get
: certain stats, the other stats can change. And, although we may arrive at
: different results than the spinlocked version, the difference would not be
: large and would not make serious side-effect.
[[email protected]: add more commit description]
Signed-off-by: Joonsoo Kim <[email protected]>
Reported-by: Richard Yao <[email protected]>
Acked-by: Eric Dumazet <[email protected]>
Cc: Peter Hurley <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Andi Kleen <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, there is no code to hold two lock simultaneously, since we don't
call slab_destroy() with holding any lock. So, lockdep annotation is
useless now. Remove it.
v2: don't remove BAD_ALIEN_MAGIC in this patch. It will be removed
in the following patch.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I haven't heard that this alien cache lock is contended, but to reduce
chance of contention would be better generally. And with this change,
we can simplify complex lockdep annotation in slab code. In the
following patch, it will be implemented.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we have separate alien_cache structure, so it'd be better to hold
the lock on alien_cache while manipulating alien_cache. After that, we
don't need the lock on array_cache, so remove it.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently, we use array_cache for alien_cache. Although they are mostly
similar, there is one difference, that is, need for spinlock. We don't
need spinlock for array_cache itself, but to use array_cache for
alien_cache, array_cache structure should have spinlock. This is
needless overhead, so removing it would be better. This patch prepare
it by introducing alien_cache and using it. In the following patch, we
remove spinlock in array_cache.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
In free_block(), if freeing object makes new free slab and number of
free_objects exceeds free_limit, we start to destroy this new free slab
with holding the kmem_cache node lock. Holding the lock is useless and,
generally, holding a lock as least as possible is good thing. I never
measure performance effect of this, but we'd be better not to hold the
lock as much as possible.
Commented by Christoph:
This is also good because kmem_cache_free is no longer called while
holding the node lock. So we avoid one case of recursion.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Acked-by: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
node isn't changed, so we don't need to retreive this structure
everytime we move the object. Maybe compiler do this optimization, but
making it explicitly is better.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Acked-by: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patchset does some cleanup and tries to remove lockdep annotation.
Patches 1~2 are just for really really minor improvement.
Patches 3~9 are for clean-up and removing lockdep annotation.
There are two cases that lockdep annotation is needed in SLAB.
1) holding two node locks
2) holding two array cache(alien cache) locks
I looked at the code and found that we can avoid these cases without any
negative effect.
1) occurs if freeing object makes new free slab and we decide to
destroy it. Although we don't need to hold the lock during destroying
a slab, current code do that. Destroying a slab without holding the
lock would help the reduction of the lock contention. To do it, I
change the implementation that new free slab is destroyed after
releasing the lock.
2) occurs on similar situation. When we free object from non-local
node, we put this object to alien cache with holding the alien cache
lock. If alien cache is full, we try to flush alien cache to proper
node cache, and, in this time, new free slab could be made. Destroying
it would be started and we will free metadata object which comes from
another node. In this case, we need another node's alien cache lock to
free object. This forces us to hold two array cache locks and then we
need lockdep annotation although they are always different locks and
deadlock cannot be possible. To prevent this situation, I use same way
as 1).
In this way, we can avoid 1) and 2) cases, and then, can remove lockdep
annotation. As short stat noted, this makes SLAB code much simpler.
This patch (of 9):
slab_should_failslab() is called on every allocation, so to optimize it
is reasonable. We normally don't allocate from kmem_cache. It is just
used when new kmem_cache is created, so it's very rare case. Therefore,
add unlikely macro to help compiler optimization.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: David Rientjes <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
min_partial means minimum number of slab cached in node partial list.
So, if nr_partial is less than it, we keep newly empty slab on node
partial list rather than freeing it. But if nr_partial is equal or
greater than it, it means that we have enough partial slabs so should
free newly empty slab. Current implementation missed the equal case so
if we set min_partial is 0, then, at least one slab could be cached.
This is critical problem to kmemcg destroying logic because it doesn't
works properly if some slabs is cached. This patch fixes this problem.
Fixes 91cb69620284 ("slub: make dead memcg caches discard free slabs
immediately").
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vladimir Davydov <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Acked-by: David Rientjes <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Commit b1cb0982bd ("change the management method of free objects of
the slab") introduced a bug on slab leak detector
('/proc/slab_allocators'). This detector works like as following
decription.
1. traverse all objects on all the slabs.
2. determine whether it is active or not.
3. if active, print who allocate this object.
but that commit changed the way how to manage free objects, so the logic
determining whether it is active or not is also changed. In before, we
regard object in cpu caches as inactive one, but, with this commit, we
mistakenly regard object in cpu caches as active one.
This intoduces kernel oops if DEBUG_PAGEALLOC is enabled. If
DEBUG_PAGEALLOC is enabled, kernel_map_pages() is used to detect who
corrupt free memory in the slab. It unmaps page table mapping if object
is free and map it if object is active. When slab leak detector check
object in cpu caches, it mistakenly think this object active so try to
access object memory to retrieve caller of allocation. At this point,
page table mapping to this object doesn't exist, so oops occurs.
Following is oops message reported from Dave.
It blew up when something tried to read /proc/slab_allocators
(Just cat it, and you should see the oops below)
Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
Modules linked in:
[snip...]
CPU: 1 PID: 9386 Comm: trinity-c33 Not tainted 3.14.0-rc5+ #131
task: ffff8801aa46e890 ti: ffff880076924000 task.ti: ffff880076924000
RIP: 0010:[<ffffffffaa1a8f4a>] [<ffffffffaa1a8f4a>] handle_slab+0x8a/0x180
RSP: 0018:ffff880076925de0 EFLAGS: 00010002
RAX: 0000000000001000 RBX: 0000000000000000 RCX: 000000005ce85ce7
RDX: ffffea00079be100 RSI: 0000000000001000 RDI: ffff880107458000
RBP: ffff880076925e18 R08: 0000000000000001 R09: 0000000000000000
R10: 0000000000000000 R11: 000000000000000f R12: ffff8801e6f84000
R13: ffffea00079be100 R14: ffff880107458000 R15: ffff88022bb8d2c0
FS: 00007fb769e45740(0000) GS:ffff88024d040000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffff8801e6f84ff8 CR3: 00000000a22db000 CR4: 00000000001407e0
DR0: 0000000002695000 DR1: 0000000002695000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000070602
Call Trace:
leaks_show+0xce/0x240
seq_read+0x28e/0x490
proc_reg_read+0x3d/0x80
vfs_read+0x9b/0x160
SyS_read+0x58/0xb0
tracesys+0xd4/0xd9
Code: f5 00 00 00 0f 1f 44 00 00 48 63 c8 44 3b 0c 8a 0f 84 e3 00 00 00 83 c0 01 44 39 c0 72 eb 41 f6 47 1a 01 0f 84 e9 00 00 00 89 f0 <4d> 8b 4c 04 f8 4d 85 c9 0f 84 88 00 00 00 49 8b 7e 08 4d 8d 46
RIP handle_slab+0x8a/0x180
To fix the problem, I introduce an object status buffer on each slab.
With this, we can track object status precisely, so slab leak detector
would not access active object and no kernel oops would occur. Memory
overhead caused by this fix is only imposed to CONFIG_DEBUG_SLAB_LEAK
which is mainly used for debugging, so memory overhead isn't big
problem.
Signed-off-by: Joonsoo Kim <[email protected]>
Reported-by: Dave Jones <[email protected]>
Reported-by: Tetsuo Handa <[email protected]>
Reviewed-by: Vladimir Davydov <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently, if allocation constraint to node is NUMA_NO_NODE, we search a
partial slab on numa_node_id() node. This doesn't work properly on a
system having memoryless nodes, since it can have no memory on that node
so there must be no partial slab on that node.
On that node, page allocation always falls back to numa_mem_id() first.
So searching a partial slab on numa_node_id() in that case is the proper
solution for the memoryless node case.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Nishanth Aravamudan <[email protected]>
Acked-by: David Rientjes <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Cc: Pekka Enberg <[email protected]>
Cc: Wanpeng Li <[email protected]>
Cc: Han Pingtian <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
'cma: Remove potential deadlock situation' introduces per cma area mutex
for bitmap management. It is good, but there is one mistake. When we
can't find appropriate area in bitmap, we release cma_mutex global lock
rather than cma->lock and this is a bug. So fix it.
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
Commit a41adfaa23 ("slab: introduce byte sized index for the freelist
of a slab") changes the size of freelist index and also changes
prototype of accessor function to freelist index. And there was a
mistake.
The mistake is that although it changes the size of freelist index
correctly, it changes the size of the index of freelist index
incorrectly. With patch, freelist index can be 1 byte or 2 bytes, that
means that num of object on on a slab can be more than 255. So we need
more than 1 byte for the index to find the index of free object on
freelist. But, above patch makes this index type 1 byte, so slab which
have more than 255 objects cannot work properly and in consequence of
it, the system cannot boot.
This issue was reported by Steven King on m68knommu which would use
2 bytes freelist index:
https://lkml.org/lkml/2014/4/16/433
To fix is easy. To change the type of the index of freelist index on
accessor functions is enough to fix this bug. Although 2 bytes is
enough, I use 4 bytes since it have no bad effect and make things more
easier. This fix was suggested and tested by Steven in his original
report.
Signed-off-by: Joonsoo Kim <[email protected]>
Reported-and-acked-by: Steven King <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Tested-by: James Hogan <[email protected]>
Tested-by: David Miller <[email protected]>
Cc: Pekka Enberg <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
zram is ram based block device and can be used by backend of filesystem.
When filesystem deletes a file, it normally doesn't do anything on data
block of that file. It just marks on metadata of that file. This
behavior has no problem on disk based block device, but has problems on
ram based block device, since we can't free memory used for data block.
To overcome this disadvantage, there is REQ_DISCARD functionality. If
block device support REQ_DISCARD and filesystem is mounted with discard
option, filesystem sends REQ_DISCARD to block device whenever some data
blocks are discarded. All we have to do is to handle this request.
This patch implements to flag up QUEUE_FLAG_DISCARD and handle this
REQ_DISCARD request. With it, we can free memory used by zram if it isn't
used.
[[email protected]: tweak comments]
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Nitin Gupta <[email protected]>
Cc: Sergey Senozhatsky <[email protected]>
Cc: Jerome Marchand <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
isolation_suitable() and migrate_async_suitable() is used to be sure
that this pageblock range is fine to be migragted. It isn't needed to
call it on every page. Current code do well if not suitable, but, don't
do well when suitable.
1) It re-checks isolation_suitable() on each page of a pageblock that was
already estabilished as suitable.
2) It re-checks migrate_async_suitable() on each page of a pageblock that
was not entered through the next_pageblock: label, because
last_pageblock_nr is not otherwise updated.
This patch fixes situation by 1) calling isolation_suitable() only once
per pageblock and 2) always updating last_pageblock_nr to the pageblock
that was just checked.
Additionally, move PageBuddy() check after pageblock unit check, since
pageblock check is the first thing we should do and makes things more
simple.
[[email protected]: rephrase commit description]
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Rik van Riel <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
It is odd to drop the spinlock when we scan (SWAP_CLUSTER_MAX - 1) th
pfn page. This may results in below situation while isolating
migratepage.
1. try isolate 0x0 ~ 0x200 pfn pages.
2. When low_pfn is 0x1ff, ((low_pfn+1) % SWAP_CLUSTER_MAX) == 0, so drop
the spinlock.
3. Then, to complete isolating, retry to aquire the lock.
I think that it is better to use SWAP_CLUSTER_MAX th pfn for checking the
criteria about dropping the lock. This has no harm 0x0 pfn, because, at
this time, locked variable would be false.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Rik van Riel <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
suitable_migration_target() checks that pageblock is suitable for
migration target. In isolate_freepages_block(), it is called on every
page and this is inefficient. So make it called once per pageblock.
suitable_migration_target() also checks if page is highorder or not, but
it's criteria for highorder is pageblock order. So calling it once
within pageblock range has no problem.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Rik van Riel <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Purpose of compaction is to get a high order page. Currently, if we
find high-order page while searching migration target page, we break it
to order-0 pages and use them as migration target. It is contrary to
purpose of compaction, so disallow high-order page to be used for
migration target.
Additionally, clean-up logic in suitable_migration_target() to simplify
the code. There is no functional changes from this clean-up.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Rik van Riel <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This is a preparation patch to unify the use of vma_resv_map()
regardless of the map type. This patch prepares it by removing
resv_map_put(), which only works for HPAGE_RESV_OWNER's resv_map, not
for all resv_maps.
[[email protected]: update changelog]
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Davidlohr Bueso <[email protected]>
Reviewed-by: Aneesh Kumar K.V <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: David Gibson <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
To change a protection method for region tracking to find grained one,
we pass the resv_map, instead of list_head, to region manipulation
functions.
This doesn't introduce any functional change, and it is just for
preparing a next step.
[[email protected]: update changelog]
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Davidlohr Bueso <[email protected]>
Reviewed-by: Aneesh Kumar K.V <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: David Gibson <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently, to track reserved and allocated regions, we use two different
ways, depending on the mapping. For MAP_SHARED, we use
address_mapping's private_list and, while for MAP_PRIVATE, we use a
resv_map.
Now, we are preparing to change a coarse grained lock which protect a
region structure to fine grained lock, and this difference hinder it.
So, before changing it, unify region structure handling, consistently
using a resv_map regardless of the kind of mapping.
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Davidlohr Bueso <[email protected]>
Reviewed-by: Aneesh Kumar K.V <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: David Gibson <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
commit 'slab: restrict the number of objects in a slab' uses
__builtin_constant_p() on #if macro. It is wrong usage of builtin
function, but it is compiled on x86 without any problem, so I can't
find it before 0 day build system find it.
This commit fixes the situation by using KMALLOC_MIN_SIZE, instead of
KMALLOC_SHIFT_LOW. KMALLOC_SHIFT_LOW is parsed to ilog2() on some
architecture and this ilog2() uses __builtin_constant_p() and results in
the problem. This problem would disappear by using KMALLOC_MIN_SIZE,
since it is just constant.
Tested-by: David Rientjes <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Pekka Enberg <[email protected]>
Now, the size of the freelist for the slab management diminish,
so that the on-slab management structure can waste large space
if the object of the slab is large.
Consider a 128 byte sized slab. If on-slab is used, 31 objects can be
in the slab. The size of the freelist for this case would be 31 bytes
so that 97 bytes, that is, more than 75% of object size, are wasted.
In a 64 byte sized slab case, no space is wasted if we use on-slab.
So set off-slab determining constraint to 128 bytes.
Acked-by: Christoph Lameter <[email protected]>
Acked-by: David Rientjes <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Pekka Enberg <[email protected]>
Currently, the freelist of a slab consist of unsigned int sized indexes.
Since most of slabs have less number of objects than 256, large sized
indexes is needless. For example, consider the minimum kmalloc slab. It's
object size is 32 byte and it would consist of one page, so 256 indexes
through byte sized index are enough to contain all possible indexes.
There can be some slabs whose object size is 8 byte. We cannot handle
this case with byte sized index, so we need to restrict minimum
object size. Since these slabs are not major, wasted memory from these
slabs would be negligible.
Some architectures' page size isn't 4096 bytes and rather larger than
4096 bytes (One example is 64KB page size on PPC or IA64) so that
byte sized index doesn't fit to them. In this case, we will use
two bytes sized index.
Below is some number for this patch.
* Before *
kmalloc-512 525 640 512 8 1 : tunables 54 27 0 : slabdata 80 80 0
kmalloc-256 210 210 256 15 1 : tunables 120 60 0 : slabdata 14 14 0
kmalloc-192 1016 1040 192 20 1 : tunables 120 60 0 : slabdata 52 52 0
kmalloc-96 560 620 128 31 1 : tunables 120 60 0 : slabdata 20 20 0
kmalloc-64 2148 2280 64 60 1 : tunables 120 60 0 : slabdata 38 38 0
kmalloc-128 647 682 128 31 1 : tunables 120 60 0 : slabdata 22 22 0
kmalloc-32 11360 11413 32 113 1 : tunables 120 60 0 : slabdata 101 101 0
kmem_cache 197 200 192 20 1 : tunables 120 60 0 : slabdata 10 10 0
* After *
kmalloc-512 521 648 512 8 1 : tunables 54 27 0 : slabdata 81 81 0
kmalloc-256 208 208 256 16 1 : tunables 120 60 0 : slabdata 13 13 0
kmalloc-192 1029 1029 192 21 1 : tunables 120 60 0 : slabdata 49 49 0
kmalloc-96 529 589 128 31 1 : tunables 120 60 0 : slabdata 19 19 0
kmalloc-64 2142 2142 64 63 1 : tunables 120 60 0 : slabdata 34 34 0
kmalloc-128 660 682 128 31 1 : tunables 120 60 0 : slabdata 22 22 0
kmalloc-32 11716 11780 32 124 1 : tunables 120 60 0 : slabdata 95 95 0
kmem_cache 197 210 192 21 1 : tunables 120 60 0 : slabdata 10 10 0
kmem_caches consisting of objects less than or equal to 256 byte have
one or more objects than before. In the case of kmalloc-32, we have 11 more
objects, so 352 bytes (11 * 32) are saved and this is roughly 9% saving of
memory. Of couse, this percentage decreases as the number of objects
in a slab decreases.
Here are the performance results on my 4 cpus machine.
* Before *
Performance counter stats for 'perf bench sched messaging -g 50 -l 1000' (10 runs):
229,945,138 cache-misses ( +- 0.23% )
11.627897174 seconds time elapsed ( +- 0.14% )
* After *
Performance counter stats for 'perf bench sched messaging -g 50 -l 1000' (10 runs):
218,640,472 cache-misses ( +- 0.42% )
11.504999837 seconds time elapsed ( +- 0.21% )
cache-misses are reduced by this patchset, roughly 5%.
And elapsed times are improved by 1%.
Acked-by: Christoph Lameter <[email protected]>
Acked-by: David Rientjes <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Pekka Enberg <[email protected]>
To prepare to implement byte sized index for managing the freelist
of a slab, we should restrict the number of objects in a slab to be less
or equal to 256, since byte only represent 256 different values.
Setting the size of object to value equal or more than newly introduced
SLAB_OBJ_MIN_SIZE ensures that the number of objects in a slab is less or
equal to 256 for a slab with 1 page.
If page size is rather larger than 4096, above assumption would be wrong.
In this case, we would fall back on 2 bytes sized index.
If minimum size of kmalloc is less than 16, we use it as minimum object
size and give up this optimization.
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Pekka Enberg <[email protected]>
In the following patches, to get/set free objects from the freelist
is changed so that simple casting doesn't work for it. Therefore,
introduce helper functions.
Acked-by: Christoph Lameter <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Pekka Enberg <[email protected]>
This logic is not simple to understand so that making separate function
helping readability. Additionally, we can use this change in the
following patch which implement for freelist to have another sized index
in according to nr objects.
Acked-by: Christoph Lameter <[email protected]>
Signed-off-by: Joonsoo Kim <[email protected]>
Signed-off-by: Pekka Enberg <[email protected]>
We should remove the page from the list if we fail with ENOSYS, since
migrate_pages() consider error cases except -ENOMEM and -EAGAIN as
permanent failure and it assumes that the page would be removed from the
list. Without this patch, we could overcount number of failure.
In addition, we should put back the new hugepage if
!hugepage_migration_support(). If not, we would leak hugepage memory.
Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Christoph Lameter <[email protected]>
Reviewed-by: Wanpeng Li <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: Rafael Aquini <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Wanpeng Li <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we have an infrastructure in rmap_walk() to handle difference from
variants of rmap traversing functions.
So, just use it in page_mkclean().
In this patch, I change following things.
1. remove some variants of rmap traversing functions.
cf> page_mkclean_file
2. mechanical change to use rmap_walk() in page_mkclean().
Signed-off-by: Joonsoo Kim <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Hillf Danton <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we have an infrastructure in rmap_walk() to handle difference from
variants of rmap traversing functions.
So, just use it in page_referenced().
In this patch, I change following things.
1. remove some variants of rmap traversing functions.
cf> page_referenced_ksm, page_referenced_anon,
page_referenced_file
2. introduce new struct page_referenced_arg and pass it to
page_referenced_one(), main function of rmap_walk, in order to count
reference, to store vm_flags and to check finish condition.
3. mechanical change to use rmap_walk() in page_referenced().
[[email protected]: fix BUG at rmap_walk]
Signed-off-by: Joonsoo Kim <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Hillf Danton <[email protected]>
Signed-off-by: Wanpeng Li <[email protected]>
Cc: Sasha Levin <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we have an infrastructure in rmap_walk() to handle difference from
variants of rmap traversing functions.
So, just use it in try_to_munlock().
In this patch, I change following things.
1. remove some variants of rmap traversing functions.
cf> try_to_unmap_ksm, try_to_unmap_anon, try_to_unmap_file
2. mechanical change to use rmap_walk() in try_to_munlock().
3. copy and paste comments.
Signed-off-by: Joonsoo Kim <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Hillf Danton <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Now, we have an infrastructure in rmap_walk() to handle difference from
variants of rmap traversing functions.
So, just use it in try_to_unmap().
In this patch, I change following things.
1. enable rmap_walk() if !CONFIG_MIGRATION.
2. mechanical change to use rmap_walk() in try_to_unmap().
Signed-off-by: Joonsoo Kim <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Hillf Danton <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
There are a lot of common parts in traversing functions, but there are
also a little of uncommon parts in it. By assigning proper function
pointer on each rmap_walker_control, we can handle these difference
correctly.
Following are differences we should handle.
1. difference of lock function in anon mapping case
2. nonlinear handling in file mapping case
3. prechecked condition:
checking memcg in page_referenced(),
checking VM_SHARE in page_mkclean()
checking temporary vma in try_to_unmap()
4. exit condition:
checking page_mapped() in try_to_unmap()
So, in this patch, I introduce 4 function pointers to handle above
differences.
Signed-off-by: Joonsoo Kim <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Hillf Danton <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
In each rmap traverse case, there is some difference so that we need
function pointers and arguments to them in order to handle these
For this purpose, struct rmap_walk_control is introduced in this patch,
and will be extended in following patch. Introducing and extending are
separate, because it clarify changes.
Signed-off-by: Joonsoo Kim <[email protected]>
Reviewed-by: Naoya Horiguchi <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Hillf Danton <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Joonsoo Kim