commit ac3f3b0a55 upstream.
__alloc_pages_direct_reclaim() is called from slowpath allocation where
high atomic reserves can be unreserved after there is a progress in
reclaim and yet no suitable page is found. Later should_reclaim_retry()
gets called from slow path allocation to decide if the reclaim needs to be
retried before OOM kill path is taken.
should_reclaim_retry() checks the available(reclaimable + free pages)
memory against the min wmark levels of a zone and returns:
a) true, if it is above the min wmark so that slow path allocation will
do the reclaim retries.
b) false, thus slowpath allocation takes oom kill path.
should_reclaim_retry() can also unreserves the high atomic reserves **but
only after all the reclaim retries are exhausted.**
In a case where there are almost none reclaimable memory and free pages
contains mostly the high atomic reserves but allocation context can't use
these high atomic reserves, makes the available memory below min wmark
levels hence false is returned from should_reclaim_retry() leading the
allocation request to take OOM kill path. This can turn into a early oom
kill if high atomic reserves are holding lot of free memory and
unreserving of them is not attempted.
(early)OOM is encountered on a VM with the below state:
[ 295.998653] Normal free:7728kB boost:0kB min:804kB low:1004kB
high:1204kB reserved_highatomic:8192KB active_anon:4kB inactive_anon:0kB
active_file:24kB inactive_file:24kB unevictable:1220kB writepending:0kB
present:70732kB managed:49224kB mlocked:0kB bounce:0kB free_pcp:688kB
local_pcp:492kB free_cma:0kB
[ 295.998656] lowmem_reserve[]: 0 32
[ 295.998659] Normal: 508*4kB (UMEH) 241*8kB (UMEH) 143*16kB (UMEH)
33*32kB (UH) 7*64kB (UH) 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB
0*4096kB = 7752kB
Per above log, the free memory of ~7MB exist in the high atomic reserves
is not freed up before falling back to oom kill path.
Fix it by trying to unreserve the high atomic reserves in
should_reclaim_retry() before __alloc_pages_direct_reclaim() can fallback
to oom kill path.
Link: https://lkml.kernel.org/r/1700823445-27531-1-git-send-email-quic_charante@quicinc.com
Fixes: 0aaa29a56e ("mm, page_alloc: reserve pageblocks for high-order atomic allocations on demand")
Signed-off-by: Charan Teja Kalla <quic_charante@quicinc.com>
Reported-by: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
Suggested-by: Michal Hocko <mhocko@suse.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Pavankumar Kondeti <quic_pkondeti@quicinc.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Joakim Tjernlund <Joakim.Tjernlund@infinera.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For compound pages, the head sets the PG_head flag and the tail sets the
compound_head to indicate the head page. If a user allocates a compound
page and frees it with a different order, the compound page information
will not be properly initialized. To detect this problem,
compound_order(page) and the order argument are compared, but this is not
checked when the order argument is zero. That error should be checked
regardless of the order.
Link: https://lkml.kernel.org/r/20231023083217.1866451-1-hyesoo.yu@samsung.com
Signed-off-by: Hyesoo Yu <hyesoo.yu@samsung.com>
Reviewed-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In current PCP auto-tuning design, if the number of pages allocated is
much more than that of pages freed on a CPU, the PCP high may become the
maximal value even if the allocating/freeing depth is small, for example,
in the sender of network workloads. If a CPU was used as sender
originally, then it is used as receiver after context switching, we need
to fill the whole PCP with maximal high before triggering PCP draining for
consecutive high order freeing. This will hurt the performance of some
network workloads.
To solve the issue, in this patch, we will track the consecutive page
freeing with a counter in stead of relying on PCP draining. So, we can
detect consecutive page freeing much earlier.
On a 2-socket Intel server with 128 logical CPU, we tested
SCTP_STREAM_MANY test case of netperf test suite with 64-pair processes.
With the patch, the network bandwidth improves 5.0%. This restores the
performance drop caused by PCP auto-tuning.
Link: https://lkml.kernel.org/r/20231016053002.756205-10-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
One target of PCP is to minimize pages in PCP if the system free pages is
too few. To reach that target, when page reclaiming is active for the
zone (ZONE_RECLAIM_ACTIVE), we will stop increasing PCP high in allocating
path, decrease PCP high and free some pages in freeing path. But this may
be too late because the background page reclaiming may introduce latency
for some workloads. So, in this patch, during page allocation we will
detect whether the number of free pages of the zone is below high
watermark. If so, we will stop increasing PCP high in allocating path,
decrease PCP high and free some pages in freeing path. With this, we can
reduce the possibility of the premature background page reclaiming caused
by too large PCP.
The high watermark checking is done in allocating path to reduce the
overhead in hotter freeing path.
Link: https://lkml.kernel.org/r/20231016053002.756205-9-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The target to tune PCP high automatically is as follows,
- Minimize allocation/freeing from/to shared zone
- Minimize idle pages in PCP
- Minimize pages in PCP if the system free pages is too few
To reach these target, a tuning algorithm as follows is designed,
- When we refill PCP via allocating from the zone, increase PCP high.
Because if we had larger PCP, we could avoid to allocate from the
zone.
- In periodic vmstat updating kworker (via refresh_cpu_vm_stats()),
decrease PCP high to try to free possible idle PCP pages.
- When page reclaiming is active for the zone, stop increasing PCP
high in allocating path, decrease PCP high and free some pages in
freeing path.
So, the PCP high can be tuned to the page allocating/freeing depth of
workloads eventually.
One issue of the algorithm is that if the number of pages allocated is
much more than that of pages freed on a CPU, the PCP high may become the
maximal value even if the allocating/freeing depth is small. But this
isn't a severe issue, because there are no idle pages in this case.
One alternative choice is to increase PCP high when we drain PCP via
trying to free pages to the zone, but don't increase PCP high during PCP
refilling. This can avoid the issue above. But if the number of pages
allocated is much less than that of pages freed on a CPU, there will be
many idle pages in PCP and it is hard to free these idle pages.
1/8 (>> 3) of PCP high will be decreased periodically. The value 1/8 is
kind of arbitrary. Just to make sure that the idle PCP pages will be
freed eventually.
On a 2-socket Intel server with 224 logical CPU, we run 8 kbuild instances
in parallel (each with `make -j 28`) in 8 cgroup. This simulates the
kbuild server that is used by 0-Day kbuild service. With the patch, the
build time decreases 3.5%. The cycles% of the spinlock contention (mostly
for zone lock) decreases from 11.0% to 0.5%. The number of PCP draining
for high order pages freeing (free_high) decreases 65.6%. The number of
pages allocated from zone (instead of from PCP) decreases 83.9%.
Link: https://lkml.kernel.org/r/20231016053002.756205-8-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Suggested-by: Mel Gorman <mgorman@techsingularity.net>
Suggested-by: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The page allocation performance requirements of different workloads are
usually different. So, we need to tune PCP (per-CPU pageset) high to
optimize the workload page allocation performance. Now, we have a system
wide sysctl knob (percpu_pagelist_high_fraction) to tune PCP high by hand.
But, it's hard to find out the best value by hand. And one global
configuration may not work best for the different workloads that run on
the same system. One solution to these issues is to tune PCP high of each
CPU automatically.
This patch adds the framework for PCP high auto-tuning. With it,
pcp->high of each CPU will be changed automatically by tuning algorithm at
runtime. The minimal high (pcp->high_min) is the original PCP high value
calculated based on the low watermark pages. While the maximal high
(pcp->high_max) is the PCP high value when percpu_pagelist_high_fraction
sysctl knob is set to MIN_PERCPU_PAGELIST_HIGH_FRACTION. That is, the
maximal pcp->high that can be set via sysctl knob by hand.
It's possible that PCP high auto-tuning doesn't work well for some
workloads. So, when PCP high is tuned by hand via the sysctl knob, the
auto-tuning will be disabled. The PCP high set by hand will be used
instead.
This patch only adds the framework, so pcp->high will be set to
pcp->high_min (original default) always. We will add actual auto-tuning
algorithm in the following patches in the series.
Link: https://lkml.kernel.org/r/20231016053002.756205-7-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In page allocator, PCP (Per-CPU Pageset) is refilled and drained in
batches to increase page allocation throughput, reduce page
allocation/freeing latency per page, and reduce zone lock contention. But
too large batch size will cause too long maximal allocation/freeing
latency, which may punish arbitrary users. So the default batch size is
chosen carefully (in zone_batchsize(), the value is 63 for zone > 1GB) to
avoid that.
In commit 3b12e7e979 ("mm/page_alloc: scale the number of pages that are
batch freed"), the batch size will be scaled for large number of page
freeing to improve page freeing performance and reduce zone lock
contention. Similar optimization can be used for large number of pages
allocation too.
To find out a suitable max batch scale factor (that is, max effective
batch size), some tests and measurement on some machines were done as
follows.
A set of debug patches are implemented as follows,
- Set PCP high to be 2 * batch to reduce the effect of PCP high
- Disable free batch size scaling to get the raw performance.
- The code with zone lock held is extracted from rmqueue_bulk() and
free_pcppages_bulk() to 2 separate functions to make it easy to
measure the function run time with ftrace function_graph tracer.
- The batch size is hard coded to be 63 (default), 127, 255, 511,
1023, 2047, 4095.
Then will-it-scale/page_fault1 is used to generate the page
allocation/freeing workload. The page allocation/freeing throughput
(page/s) is measured via will-it-scale. The page allocation/freeing
average latency (alloc/free latency avg, in us) and allocation/freeing
latency at 99 percentile (alloc/free latency 99%, in us) are measured with
ftrace function_graph tracer.
The test results are as follows,
Sapphire Rapids Server
======================
Batch throughput free latency free latency alloc latency alloc latency
page/s avg / us 99% / us avg / us 99% / us
----- ---------- ------------ ------------ ------------- -------------
63 513633.4 2.33 3.57 2.67 6.83
127 517616.7 4.35 6.65 4.22 13.03
255 520822.8 8.29 13.32 7.52 25.24
511 524122.0 15.79 23.42 14.02 49.35
1023 525980.5 30.25 44.19 25.36 94.88
2047 526793.6 59.39 84.50 45.22 140.81
Ice Lake Server
===============
Batch throughput free latency free latency alloc latency alloc latency
page/s avg / us 99% / us avg / us 99% / us
----- ---------- ------------ ------------ ------------- -------------
63 620210.3 2.21 3.68 2.02 4.35
127 627003.0 4.09 6.86 3.51 8.28
255 630777.5 7.70 13.50 6.17 15.97
511 633651.5 14.85 22.62 11.66 31.08
1023 637071.1 28.55 42.02 20.81 54.36
2047 638089.7 56.54 84.06 39.28 91.68
Cascade Lake Server
===================
Batch throughput free latency free latency alloc latency alloc latency
page/s avg / us 99% / us avg / us 99% / us
----- ---------- ------------ ------------ ------------- -------------
63 404706.7 3.29 5.03 3.53 4.75
127 422475.2 6.12 9.09 6.36 8.76
255 411522.2 11.68 16.97 10.90 16.39
511 428124.1 22.54 31.28 19.86 32.25
1023 414718.4 43.39 62.52 40.00 66.33
2047 429848.7 86.64 120.34 71.14 106.08
Commet Lake Desktop
===================
Batch throughput free latency free latency alloc latency alloc latency
page/s avg / us 99% / us avg / us 99% / us
----- ---------- ------------ ------------ ------------- -------------
63 795183.13 2.18 3.55 2.03 3.05
127 803067.85 3.91 6.56 3.85 5.52
255 812771.10 7.35 10.80 7.14 10.20
511 817723.48 14.17 27.54 13.43 30.31
1023 818870.19 27.72 40.10 27.89 46.28
Coffee Lake Desktop
===================
Batch throughput free latency free latency alloc latency alloc latency
page/s avg / us 99% / us avg / us 99% / us
----- ---------- ------------ ------------ ------------- -------------
63 510542.8 3.13 4.40 2.48 3.43
127 514288.6 5.97 7.89 4.65 6.04
255 516889.7 11.86 15.58 8.96 12.55
511 519802.4 23.10 28.81 16.95 26.19
1023 520802.7 45.30 52.51 33.19 45.95
2047 519997.1 90.63 104.00 65.26 81.74
From the above data, to restrict the allocation/freeing latency to be less
than 100 us in most times, the max batch scale factor needs to be less
than or equal to 5.
Although it is reasonable to use 5 as max batch scale factor for the
systems tested, there are also slower systems. Where smaller value should
be used to constrain the page allocation/freeing latency.
So, in this patch, a new kconfig option (PCP_BATCH_SCALE_MAX) is added to
set the max batch scale factor. Whose default value is 5, and users can
reduce it when necessary.
Link: https://lkml.kernel.org/r/20231016053002.756205-5-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In commit f26b3fa046 ("mm/page_alloc: limit number of high-order pages
on PCP during bulk free"), the PCP (Per-CPU Pageset) will be drained when
PCP is mostly used for high-order pages freeing to improve the cache-hot
pages reusing between page allocating and freeing CPUs.
On system with small per-CPU data cache slice, pages shouldn't be cached
before draining to guarantee cache-hot. But on a system with large
per-CPU data cache slice, some pages can be cached before draining to
reduce zone lock contention.
So, in this patch, instead of draining without any caching, "pcp->batch"
pages will be cached in PCP before draining if the size of the per-CPU
data cache slice is more than "3 * batch".
In theory, if the size of per-CPU data cache slice is more than "2 *
batch", we can reuse cache-hot pages between CPUs. But considering the
other usage of cache (code, other data accessing, etc.), "3 * batch" is
used.
Note: "3 * batch" is chosen to make sure the optimization works on recent
x86_64 server CPUs. If you want to increase it, please check whether it
breaks the optimization.
On a 2-socket Intel server with 128 logical CPU, with the patch, the
network bandwidth of the UNIX (AF_UNIX) test case of lmbench test suite
with 16-pair processes increase 70.5%. The cycles% of the spinlock
contention (mostly for zone lock) decreases from 46.1% to 21.3%. The
number of PCP draining for high order pages freeing (free_high) decreases
89.9%. The cache miss rate keeps 0.2%.
Link: https://lkml.kernel.org/r/20231016053002.756205-4-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "mm: PCP high auto-tuning", v3.
The page allocation performance requirements of different workloads are
often different. So, we need to tune the PCP (Per-CPU Pageset) high on
each CPU automatically to optimize the page allocation performance.
The list of patches in series is as follows,
[1/9] mm, pcp: avoid to drain PCP when process exit
[2/9] cacheinfo: calculate per-CPU data cache size
[3/9] mm, pcp: reduce lock contention for draining high-order pages
[4/9] mm: restrict the pcp batch scale factor to avoid too long latency
[5/9] mm, page_alloc: scale the number of pages that are batch allocated
[6/9] mm: add framework for PCP high auto-tuning
[7/9] mm: tune PCP high automatically
[8/9] mm, pcp: decrease PCP high if free pages < high watermark
[9/9] mm, pcp: reduce detecting time of consecutive high order page freeing
Patch [1/9], [2/9], [3/9] optimize the PCP draining for consecutive
high-order pages freeing.
Patch [4/9], [5/9] optimize batch freeing and allocating.
Patch [6/9], [7/9], [8/9] implement and optimize a PCP high
auto-tuning method.
Patch [9/9] optimize the PCP draining for consecutive high order page
freeing based on PCP high auto-tuning.
The test results for patches with performance impact are as follows,
kbuild
======
On a 2-socket Intel server with 224 logical CPU, we run 8 kbuild instances
in parallel (each with `make -j 28`) in 8 cgroup. This simulates the
kbuild server that is used by 0-Day kbuild service.
build time lock contend% free_high alloc_zone
---------- ---------- --------- ----------
base 100.0 14.0 100.0 100.0
patch1 99.5 12.8 19.5 95.6
patch3 99.4 12.6 7.1 95.6
patch5 98.6 11.0 8.1 97.1
patch7 95.1 0.5 2.8 15.6
patch9 95.0 1.0 8.8 20.0
The PCP draining optimization (patch [1/9], [3/9]) and PCP batch
allocation optimization (patch [5/9]) reduces zone lock contention a
little. The PCP high auto-tuning (patch [7/9], [9/9]) reduces build time
visibly. Where the tuning target: the number of pages allocated from zone
reduces greatly. So, the zone contention cycles% reduces greatly.
With PCP tuning patches (patch [7/9], [9/9]), the average used memory
during test increases up to 18.4% because more pages are cached in PCP.
But at the end of the test, the number of the used memory decreases to the
same level as that of the base patch. That is, the pages cached in PCP
will be released to zone after not being used actively.
netperf SCTP_STREAM_MANY
========================
On a 2-socket Intel server with 128 logical CPU, we tested
SCTP_STREAM_MANY test case of netperf test suite with 64-pair processes.
score lock contend% free_high alloc_zone cache miss rate%
----- ---------- --------- ---------- ----------------
base 100.0 2.1 100.0 100.0 1.3
patch1 99.4 2.1 99.4 99.4 1.3
patch3 106.4 1.3 13.3 106.3 1.3
patch5 106.0 1.2 13.2 105.9 1.3
patch7 103.4 1.9 6.7 90.3 7.6
patch9 108.6 1.3 13.7 108.6 1.3
The PCP draining optimization (patch [1/9]+[3/9]) improves performance.
The PCP high auto-tuning (patch [7/9]) reduces performance a little
because PCP draining cannot be triggered in time sometimes. So, the cache
miss rate% increases. The further PCP draining optimization (patch [9/9])
based on PCP tuning restore the performance.
lmbench3 UNIX (AF_UNIX)
=======================
On a 2-socket Intel server with 128 logical CPU, we tested UNIX
(AF_UNIX socket) test case of lmbench3 test suite with 16-pair
processes.
score lock contend% free_high alloc_zone cache miss rate%
----- ---------- --------- ---------- ----------------
base 100.0 51.4 100.0 100.0 0.2
patch1 116.8 46.1 69.5 104.3 0.2
patch3 199.1 21.3 7.0 104.9 0.2
patch5 200.0 20.8 7.1 106.9 0.3
patch7 191.6 19.9 6.8 103.8 2.8
patch9 193.4 21.7 7.0 104.7 2.1
The PCP draining optimization (patch [1/9], [3/9]) improves performance
much. The PCP tuning (patch [7/9]) reduces performance a little because
PCP draining cannot be triggered in time sometimes. The further PCP
draining optimization (patch [9/9]) based on PCP tuning restores the
performance partly.
The patchset adds several fields in struct per_cpu_pages. The struct
layout before/after the patchset is as follows,
base
====
struct per_cpu_pages {
spinlock_t lock; /* 0 4 */
int count; /* 4 4 */
int high; /* 8 4 */
int batch; /* 12 4 */
short int free_factor; /* 16 2 */
short int expire; /* 18 2 */
/* XXX 4 bytes hole, try to pack */
struct list_head lists[13]; /* 24 208 */
/* size: 256, cachelines: 4, members: 7 */
/* sum members: 228, holes: 1, sum holes: 4 */
/* padding: 24 */
} __attribute__((__aligned__(64)));
patched
=======
struct per_cpu_pages {
spinlock_t lock; /* 0 4 */
int count; /* 4 4 */
int high; /* 8 4 */
int high_min; /* 12 4 */
int high_max; /* 16 4 */
int batch; /* 20 4 */
u8 flags; /* 24 1 */
u8 alloc_factor; /* 25 1 */
u8 expire; /* 26 1 */
/* XXX 1 byte hole, try to pack */
short int free_count; /* 28 2 */
/* XXX 2 bytes hole, try to pack */
struct list_head lists[13]; /* 32 208 */
/* size: 256, cachelines: 4, members: 11 */
/* sum members: 237, holes: 2, sum holes: 3 */
/* padding: 16 */
} __attribute__((__aligned__(64)));
The size of the struct doesn't changed with the patchset.
This patch (of 9):
In commit f26b3fa046 ("mm/page_alloc: limit number of high-order pages
on PCP during bulk free"), the PCP (Per-CPU Pageset) will be drained when
PCP is mostly used for high-order pages freeing to improve the cache-hot
pages reusing between page allocation and freeing CPUs.
But, the PCP draining mechanism may be triggered unexpectedly when process
exits. With some customized trace point, it was found that PCP draining
(free_high == true) was triggered with the order-1 page freeing with the
following call stack,
=> free_unref_page_commit
=> free_unref_page
=> __mmdrop
=> exit_mm
=> do_exit
=> do_group_exit
=> __x64_sys_exit_group
=> do_syscall_64
Checking the source code, this is the page table PGD freeing
(mm_free_pgd()). It's a order-1 page freeing if
CONFIG_PAGE_TABLE_ISOLATION=y. Which is a common configuration for
security.
Just before that, page freeing with the following call stack was found,
=> free_unref_page_commit
=> free_unref_page_list
=> release_pages
=> tlb_batch_pages_flush
=> tlb_finish_mmu
=> exit_mmap
=> __mmput
=> exit_mm
=> do_exit
=> do_group_exit
=> __x64_sys_exit_group
=> do_syscall_64
So, when a process exits,
- a large number of user pages of the process will be freed without
page allocation, it's highly possible that pcp->free_factor becomes >
0. In fact, this is expected behavior to improve process exit
performance.
- after freeing all user pages, the PGD will be freed, which is a
order-1 page freeing, PCP will be drained.
All in all, when a process exits, it's high possible that the PCP will be
drained. This is an unexpected behavior.
To avoid this, in the patch, the PCP draining will only be triggered for 2
consecutive high-order page freeing.
On a 2-socket Intel server with 224 logical CPU, we run 8 kbuild instances
in parallel (each with `make -j 28`) in 8 cgroup. This simulates the
kbuild server that is used by 0-Day kbuild service. With the patch, the
cycles% of the spinlock contention (mostly for zone lock) decreases from
14.0% to 12.8% (with PCP size == 367). The number of PCP draining for
high order pages freeing (free_high) decreases 80.5%.
This helps network workload too for reduced zone lock contention. On a
2-socket Intel server with 128 logical CPU, with the patch, the network
bandwidth of the UNIX (AF_UNIX) test case of lmbench test suite with
16-pair processes increase 16.8%. The cycles% of the spinlock contention
(mostly for zone lock) decreases from 51.4% to 46.1%. The number of PCP
draining for high order pages freeing (free_high) decreases 30.5%. The
cache miss rate keeps 0.2%.
Link: https://lkml.kernel.org/r/20231016053002.756205-1-ying.huang@intel.com
Link: https://lkml.kernel.org/r/20231016053002.756205-2-ying.huang@intel.com
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
When guard page debug is enabled and set_page_guard returns success, we
miss to forward page to point to start of next split range and we will do
split unexpectedly in page range without target page. Move start page
update before set_page_guard to fix this.
As we split to wrong target page, then splited pages are not able to merge
back to original order when target page is put back and splited pages
except target page is not usable. To be specific:
Consider target page is the third page in buddy page with order 2.
| buddy-2 | Page | Target | Page |
After break down to target page, we will only set first page to Guard
because of bug.
| Guard | Page | Target | Page |
When we try put_page_back_buddy with target page, the buddy page of target
if neither guard nor buddy, Then it's not able to construct original page
with order 2
| Guard | Page | buddy-0 | Page |
All pages except target page is not in free list and is not usable.
Link: https://lkml.kernel.org/r/20230927094401.68205-1-shikemeng@huaweicloud.com
Fixes: 06be6ff3d2 ("mm,hwpoison: rework soft offline for free pages")
Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Acked-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Commit 4b23a68f95 ("mm/page_alloc: protect PCP lists with a spinlock")
bypasses the pcplist on lock contention and returns the page directly to
the buddy list of the page's migratetype.
For pages that don't have their own pcplist, such as CMA and HIGHATOMIC,
the migratetype is temporarily updated such that the page can hitch a ride
on the MOVABLE pcplist. Their true type is later reassessed when flushing
in free_pcppages_bulk(). However, when lock contention is detected after
the type was already overridden, the bypass will then put the page on the
wrong buddy list.
Once on the MOVABLE buddy list, the page becomes eligible for fallbacks
and even stealing. In the case of HIGHATOMIC, otherwise ineligible
allocations can dip into the highatomic reserves. In the case of CMA, the
page can be lost from the CMA region permanently.
Use a separate pcpmigratetype variable for the pcplist override. Use the
original migratetype when going directly to the buddy. This fixes the bug
and should make the intentions more obvious in the code.
Originally sent here to address the HIGHATOMIC case:
https://lore.kernel.org/lkml/20230821183733.106619-4-hannes@cmpxchg.org/
Changelog updated in response to the CMA-specific bug report.
[mgorman@techsingularity.net: updated changelog]
Link: https://lkml.kernel.org/r/20230911181108.GA104295@cmpxchg.org
Fixes: 4b23a68f95 ("mm/page_alloc: protect PCP lists with a spinlock")
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Joe Liu <joe.liu@mediatek.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In the past, movable allocations could be disallowed from CMA through
PF_MEMALLOC_PIN. As CMA pages are funneled through the MOVABLE pcplist,
this required filtering that cornercase during allocations, such that
pinnable allocations wouldn't accidentally get a CMA page.
However, since 8e3560d963 ("mm: honor PF_MEMALLOC_PIN for all movable
pages"), PF_MEMALLOC_PIN automatically excludes __GFP_MOVABLE. Once
again, MOVABLE implies CMA is allowed.
Remove the stale filtering code. Also remove a stale comment that was
introduced as part of the filtering code, because the filtering let
order-0 pages fall through to the buddy allocator. See 1d91df85f3
("mm/page_alloc: handle a missing case for memalloc_nocma_{save/restore}
APIs") for context. The comment's been obsolete since the introduction of
the explicit ALLOC_HIGHATOMIC flag in eb2e2b425c ("mm/page_alloc:
explicitly record high-order atomic allocations in alloc_flags").
Link: https://lkml.kernel.org/r/20230824153821.243148-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: David Hildenbrand <david@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Pasha Tatashin <pasha.tatashin@soleen.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
If pfn is outside zone boundaries in the first round, ret will be set to
1. But if pfn is changed to inside the zone boundaries in zone span
seqretry path, ret is still set to 1 leading to false page outside zone
error info.
This is from code inspection. The race window should be really small thus
hard to trigger in real world.
[akpm@linux-foundation.org: code simplification, per Matthew]
Link: https://lkml.kernel.org/r/20230704111823.940331-1-linmiaohe@huawei.com
Fixes: bdc8cb9845 ("[PATCH] memory hotplug locking: zone span seqlock")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
__build_all_zonelists() acquires zonelist_update_seq by first disabling
interrupts via local_irq_save() and then acquiring the seqlock with
write_seqlock(). This is troublesome and leads to problems on PREEMPT_RT.
The problem is that the inner spinlock_t becomes a sleeping lock on
PREEMPT_RT and must not be acquired with disabled interrupts.
The API provides write_seqlock_irqsave() which does the right thing in one
step. printk_deferred_enter() has to be invoked in non-migrate-able
context to ensure that deferred printing is enabled and disabled on the
same CPU. This is the case after zonelist_update_seq has been acquired.
There was discussion on the first submission that the order should be:
local_irq_disable();
printk_deferred_enter();
write_seqlock();
to avoid pitfalls like having an unaccounted printk() coming from
write_seqlock_irqsave() before printk_deferred_enter() is invoked. The
only origin of such a printk() can be a lockdep splat because the lockdep
annotation happens after the sequence count is incremented. This is
exceptional and subject to change.
It was also pointed that PREEMPT_RT can be affected by the printk problem
since its write_seqlock_irqsave() does not really disable interrupts.
This isn't the case because PREEMPT_RT's printk implementation differs
from the mainline implementation in two important aspects:
- Printing happens in a dedicated threads and not at during the
invocation of printk().
- In emergency cases where synchronous printing is used, a different
driver is used which does not use tty_port::lock.
Acquire zonelist_update_seq with write_seqlock_irqsave() and then defer
printk output.
Link: https://lkml.kernel.org/r/20230623201517.yw286Knb@linutronix.de
Fixes: 1007843a91 ("mm/page_alloc: fix potential deadlock on zonelist_update_seq seqlock")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: John Ogness <john.ogness@linutronix.de>
Cc: Luis Claudio R. Goncalves <lgoncalv@redhat.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Waiman Long <longman@redhat.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The current calculation of min_free_kbytes only uses ZONE_DMA and
ZONE_NORMAL pages,but the ZONE_MOVABLE zone->_watermark[WMARK_MIN] will
also divide part of min_free_kbytes.This will cause the min watermark of
ZONE_NORMAL to be too small in the presence of ZONE_MOVEABLE.
__GFP_HIGH and PF_MEMALLOC allocations usually don't need movable zone
pages, so just like ZONE_HIGHMEM, cap pages_min to a small value in
__setup_per_zone_wmarks().
On my testing machine with 16GB of memory (transparent hugepage is turned
off by default, and movablecore=12G is configured) The following is a
comparative test data of watermark_min
no patch add patch
ZONE_DMA 1 8
ZONE_DMA32 151 709
ZONE_NORMAL 233 1113
ZONE_MOVABLE 1434 128
min_free_kbytes 7288 7326
Link: https://lkml.kernel.org/r/20230625031656.23941-1-liuq131@chinatelecom.cn
Signed-off-by: liuq <liuq131@chinatelecom.cn>
Reviewed-by: "Huang, Ying" <ying.huang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Pull mm updates from Andrew Morton:
- Yosry Ahmed brought back some cgroup v1 stats in OOM logs
- Yosry has also eliminated cgroup's atomic rstat flushing
- Nhat Pham adds the new cachestat() syscall. It provides userspace
with the ability to query pagecache status - a similar concept to
mincore() but more powerful and with improved usability
- Mel Gorman provides more optimizations for compaction, reducing the
prevalence of page rescanning
- Lorenzo Stoakes has done some maintanance work on the
get_user_pages() interface
- Liam Howlett continues with cleanups and maintenance work to the
maple tree code. Peng Zhang also does some work on maple tree
- Johannes Weiner has done some cleanup work on the compaction code
- David Hildenbrand has contributed additional selftests for
get_user_pages()
- Thomas Gleixner has contributed some maintenance and optimization
work for the vmalloc code
- Baolin Wang has provided some compaction cleanups,
- SeongJae Park continues maintenance work on the DAMON code
- Huang Ying has done some maintenance on the swap code's usage of
device refcounting
- Christoph Hellwig has some cleanups for the filemap/directio code
- Ryan Roberts provides two patch series which yield some
rationalization of the kernel's access to pte entries - use the
provided APIs rather than open-coding accesses
- Lorenzo Stoakes has some fixes to the interaction between pagecache
and directio access to file mappings
- John Hubbard has a series of fixes to the MM selftesting code
- ZhangPeng continues the folio conversion campaign
- Hugh Dickins has been working on the pagetable handling code, mainly
with a view to reducing the load on the mmap_lock
- Catalin Marinas has reduced the arm64 kmalloc() minimum alignment
from 128 to 8
- Domenico Cerasuolo has improved the zswap reclaim mechanism by
reorganizing the LRU management
- Matthew Wilcox provides some fixups to make gfs2 work better with the
buffer_head code
- Vishal Moola also has done some folio conversion work
- Matthew Wilcox has removed the remnants of the pagevec code - their
functionality is migrated over to struct folio_batch
* tag 'mm-stable-2023-06-24-19-15' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (380 commits)
mm/hugetlb: remove hugetlb_set_page_subpool()
mm: nommu: correct the range of mmap_sem_read_lock in task_mem()
hugetlb: revert use of page_cache_next_miss()
Revert "page cache: fix page_cache_next/prev_miss off by one"
mm/vmscan: fix root proactive reclaim unthrottling unbalanced node
mm: memcg: rename and document global_reclaim()
mm: kill [add|del]_page_to_lru_list()
mm: compaction: convert to use a folio in isolate_migratepages_block()
mm: zswap: fix double invalidate with exclusive loads
mm: remove unnecessary pagevec includes
mm: remove references to pagevec
mm: rename invalidate_mapping_pagevec to mapping_try_invalidate
mm: remove struct pagevec
net: convert sunrpc from pagevec to folio_batch
i915: convert i915_gpu_error to use a folio_batch
pagevec: rename fbatch_count()
mm: remove check_move_unevictable_pages()
drm: convert drm_gem_put_pages() to use a folio_batch
i915: convert shmem_sg_free_table() to use a folio_batch
scatterlist: add sg_set_folio()
...
Commit 73444bc4d8 ("mm, page_alloc: do not wake kswapd with zone lock
held") moved wakeup_kswapd() from steal_suitable_fallback() to rmqueue()
using ZONE_BOOSTED_WATERMARK flag.
Only allocation contexts that include ALLOC_KSWAPD (which corresponds to
__GFP_KSWAPD_RECLAIM) should wake kswapd, for callers are supposed to
remove __GFP_KSWAPD_RECLAIM if trying to hold pgdat->kswapd_wait has a
risk of deadlock. But since zone->flags is a shared variable, a thread
doing !__GFP_KSWAPD_RECLAIM allocation request might observe this flag
being set immediately after another thread doing __GFP_KSWAPD_RECLAIM
allocation request set this flag, causing possibility of deadlock.
Link: https://lkml.kernel.org/r/c3c3dacf-dd3b-77c9-f96a-d0982b4b2a4f@I-love.SAKURA.ne.jp
Fixes: 73444bc4d8 ("mm, page_alloc: do not wake kswapd with zone lock held")
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: "Huang, Ying" <ying.huang@intel.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
