|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [Xen-devel] [linux-4.1 bisection] complete test-armhf-armhf-xl-multivcpu
branch xen-unstable xenbranch xen-unstable job test-armhf-armhf-xl-multivcpu testid debian-install Tree: linux git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git Tree: linuxfirmware git://xenbits.xen.org/osstest/linux-firmware.git Tree: qemuu git://xenbits.xen.org/qemu-xen.git Tree: xen git://xenbits.xen.org/xen.git *** Found and reproduced problem changeset *** Bug is in tree: linux git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git Bug introduced: c5ad33184354260be6d05de57e46a5498692f6d6 Bug not present: c5bcec6cbcbf520f088dc7939934bbf10c20c5a5 Last fail repro: http://logs.test-lab.xenproject.org/osstest/logs/99777/ commit c5ad33184354260be6d05de57e46a5498692f6d6 Author: Lukasz Odzioba <lukasz.odzioba@xxxxxxxxx> Date: Fri Jun 24 14:50:01 2016 -0700 mm/swap.c: flush lru pvecs on compound page arrival [ Upstream commit 8f182270dfec432e93fae14f9208a6b9af01009f ] Currently we can have compound pages held on per cpu pagevecs, which leads to a lot of memory unavailable for reclaim when needed. In the systems with hundreads of processors it can be GBs of memory. On of the way of reproducing the problem is to not call munmap explicitly on all mapped regions (i.e. after receiving SIGTERM). After that some pages (with THP enabled also huge pages) may end up on lru_add_pvec, example below. void main() { #pragma omp parallel { size_t size = 55 * 1000 * 1000; // smaller than MEM/CPUS void *p = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS , -1, 0); if (p != MAP_FAILED) memset(p, 0, size); //munmap(p, size); // uncomment to make the problem go away } } When we run it with THP enabled it will leave significant amount of memory on lru_add_pvec. This memory will be not reclaimed if we hit OOM, so when we run above program in a loop: for i in `seq 100`; do ./a.out; done many processes (95% in my case) will be killed by OOM. The primary point of the LRU add cache is to save the zone lru_lock contention with a hope that more pages will belong to the same zone and so their addition can be batched. The huge page is already a form of batched addition (it will add 512 worth of memory in one go) so skipping the batching seems like a safer option when compared to a potential excess in the caching which can be quite large and much harder to fix because lru_add_drain_all is way to expensive and it is not really clear what would be a good moment to call it. Similarly we can reproduce the problem on lru_deactivate_pvec by adding: madvise(p, size, MADV_FREE); after memset. This patch flushes lru pvecs on compound page arrival making the problem less severe - after applying it kill rate of above example drops to 0%, due to reducing maximum amount of memory held on pvec from 28MB (with THP) to 56kB per CPU. Suggested-by: Michal Hocko <mhocko@xxxxxxxx> Link: http://lkml.kernel.org/r/1466180198-18854-1-git-send-email-lukasz.odzioba@xxxxxxxxx Signed-off-by: Lukasz Odzioba <lukasz.odzioba@xxxxxxxxx> Acked-by: Michal Hocko <mhocko@xxxxxxxx> Cc: Kirill Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx> Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx> Cc: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx> Cc: Ming Li <mingli199x@xxxxxx> Cc: Minchan Kim <minchan@xxxxxxxxxx> Cc: <stable@xxxxxxxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> Signed-off-by: Sasha Levin <sasha.levin@xxxxxxxxxx> For bisection revision-tuple graph see: http://logs.test-lab.xenproject.org/osstest/results/bisect/linux-4.1/test-armhf-armhf-xl-multivcpu.debian-install.html Revision IDs in each graph node refer, respectively, to the Trees above. ---------------------------------------- Running cs-bisection-step --graph-out=/home/logs/results/bisect/linux-4.1/test-armhf-armhf-xl-multivcpu.debian-install --summary-out=tmp/99777.bisection-summary --basis-template=96211 --blessings=real,real-bisect linux-4.1 test-armhf-armhf-xl-multivcpu debian-install Searching for failure / basis pass: 99714 fail [host=cubietruck-gleizes] / 96211 [host=cubietruck-metzinger] 96183 ok. Failure / basis pass flights: 99714 / 96183 Tree: linux git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git Tree: linuxfirmware git://xenbits.xen.org/osstest/linux-firmware.git Tree: qemuu git://xenbits.xen.org/qemu-xen.git Tree: xen git://xenbits.xen.org/xen.git Latest 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 Basis pass 95123c0b81d9478b8155fe15093b88f57ef7d0bd c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f c6f7d21747805b50123fc1b8d73518fea2aa9096 Generating revisions with ./adhoc-revtuple-generator git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git#95123c0b81d9478b8155fe15093b88f57ef7d0bd-5880876e94699ce010554f483ccf0009997955ca git://xenbits.xen.org/osstest/linux-firmware.git#c530a75c1e6a472b0eb9558310b518f0dfcd8860-c530a75c1e6a472b0eb9558310b518f0dfcd8860 git://xenbits.xen.org/qemu-xen.git#44a072f0de0d57c95c2212bbce02888832b7b74f-44a072f0de0d57c95c2212bbce02888832b7b74f git://xenbits.xen.org/xen.git#c6f7d21747805b50123fc1b8d73518fea2aa9096-e763268781d341fef05d461f3057e6ced5e033f2 Loaded 2001 nodes in revision graph Searching for test results: 96211 [host=cubietruck-metzinger] 96160 pass 95123c0b81d9478b8155fe15093b88f57ef7d0bd c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f c6f7d21747805b50123fc1b8d73518fea2aa9096 96183 pass 95123c0b81d9478b8155fe15093b88f57ef7d0bd c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f c6f7d21747805b50123fc1b8d73518fea2aa9096 97279 fail irrelevant 97434 fail irrelevant 97394 fail irrelevant 97496 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f b48be35ac86cd6369124cf06ca3006d086095297 97558 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f b48be35ac86cd6369124cf06ca3006d086095297 97613 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f b48be35ac86cd6369124cf06ca3006d086095297 97644 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f b48be35ac86cd6369124cf06ca3006d086095297 97692 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 97730 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 99604 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 99688 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 99664 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 99695 fail 8ca7bf099ae0e6ff096b3910895b5285a112aeb5 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99677 pass 95123c0b81d9478b8155fe15093b88f57ef7d0bd c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f c6f7d21747805b50123fc1b8d73518fea2aa9096 99767 pass c5bcec6cbcbf520f088dc7939934bbf10c20c5a5 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99733 pass a13b0f0a244b15e576f6edf4ffb9ce41ea6f3837 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99704 pass fcc5d265d134e891abd67169c77358cd5ea2fc77 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99701 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 99709 fail d4b08964d00a0b99e999a2bb1ce417e54b5c607f c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99714 fail 5880876e94699ce010554f483ccf0009997955ca c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f e763268781d341fef05d461f3057e6ced5e033f2 99739 pass 1ff20a560eba527ba652502a2da1cd431e1e2fea c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99746 pass 7f3724b8951735ef1d5ae4f2846b8af98a665d73 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99756 pass c5bcec6cbcbf520f088dc7939934bbf10c20c5a5 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99753 fail 683854270f84daa09baffe2b21d64ec88c614fa9 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99763 fail c5ad33184354260be6d05de57e46a5498692f6d6 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99770 fail c5ad33184354260be6d05de57e46a5498692f6d6 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99774 pass c5bcec6cbcbf520f088dc7939934bbf10c20c5a5 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 99777 fail c5ad33184354260be6d05de57e46a5498692f6d6 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 Searching for interesting versions Result found: flight 96160 (pass), for basis pass Result found: flight 97692 (fail), for basis failure Repro found: flight 99677 (pass), for basis pass Repro found: flight 99688 (fail), for basis failure 0 revisions at c5bcec6cbcbf520f088dc7939934bbf10c20c5a5 c530a75c1e6a472b0eb9558310b518f0dfcd8860 44a072f0de0d57c95c2212bbce02888832b7b74f 7da483b0236d8974cc97f81780dcf8e559a63175 No revisions left to test, checking graph state. Result found: flight 99756 (pass), for last pass Result found: flight 99763 (fail), for first failure Repro found: flight 99767 (pass), for last pass Repro found: flight 99770 (fail), for first failure Repro found: flight 99774 (pass), for last pass Repro found: flight 99777 (fail), for first failure *** Found and reproduced problem changeset *** Bug is in tree: linux git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git Bug introduced: c5ad33184354260be6d05de57e46a5498692f6d6 Bug not present: c5bcec6cbcbf520f088dc7939934bbf10c20c5a5 Last fail repro: http://logs.test-lab.xenproject.org/osstest/logs/99777/ commit c5ad33184354260be6d05de57e46a5498692f6d6 Author: Lukasz Odzioba <lukasz.odzioba@xxxxxxxxx> Date: Fri Jun 24 14:50:01 2016 -0700 mm/swap.c: flush lru pvecs on compound page arrival [ Upstream commit 8f182270dfec432e93fae14f9208a6b9af01009f ] Currently we can have compound pages held on per cpu pagevecs, which leads to a lot of memory unavailable for reclaim when needed. In the systems with hundreads of processors it can be GBs of memory. On of the way of reproducing the problem is to not call munmap explicitly on all mapped regions (i.e. after receiving SIGTERM). After that some pages (with THP enabled also huge pages) may end up on lru_add_pvec, example below. void main() { #pragma omp parallel { size_t size = 55 * 1000 * 1000; // smaller than MEM/CPUS void *p = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS , -1, 0); if (p != MAP_FAILED) memset(p, 0, size); //munmap(p, size); // uncomment to make the problem go away } } When we run it with THP enabled it will leave significant amount of memory on lru_add_pvec. This memory will be not reclaimed if we hit OOM, so when we run above program in a loop: for i in `seq 100`; do ./a.out; done many processes (95% in my case) will be killed by OOM. The primary point of the LRU add cache is to save the zone lru_lock contention with a hope that more pages will belong to the same zone and so their addition can be batched. The huge page is already a form of batched addition (it will add 512 worth of memory in one go) so skipping the batching seems like a safer option when compared to a potential excess in the caching which can be quite large and much harder to fix because lru_add_drain_all is way to expensive and it is not really clear what would be a good moment to call it. Similarly we can reproduce the problem on lru_deactivate_pvec by adding: madvise(p, size, MADV_FREE); after memset. This patch flushes lru pvecs on compound page arrival making the problem less severe - after applying it kill rate of above example drops to 0%, due to reducing maximum amount of memory held on pvec from 28MB (with THP) to 56kB per CPU. Suggested-by: Michal Hocko <mhocko@xxxxxxxx> Link: http://lkml.kernel.org/r/1466180198-18854-1-git-send-email-lukasz.odzioba@xxxxxxxxx Signed-off-by: Lukasz Odzioba <lukasz.odzioba@xxxxxxxxx> Acked-by: Michal Hocko <mhocko@xxxxxxxx> Cc: Kirill Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx> Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx> Cc: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx> Cc: Ming Li <mingli199x@xxxxxx> Cc: Minchan Kim <minchan@xxxxxxxxxx> Cc: <stable@xxxxxxxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> Signed-off-by: Sasha Levin <sasha.levin@xxxxxxxxxx> pnmtopng: 89 colors found Revision graph left in /home/logs/results/bisect/linux-4.1/test-armhf-armhf-xl-multivcpu.debian-install.{dot,ps,png,html,svg}. ---------------------------------------- 99777: tolerable ALL FAIL flight 99777 linux-4.1 real-bisect [real] http://logs.test-lab.xenproject.org/osstest/logs/99777/ Failures :-/ but no regressions. Tests which did not succeed, including tests which could not be run: test-armhf-armhf-xl-multivcpu 9 debian-install fail baseline untested jobs: test-armhf-armhf-xl-multivcpu fail ------------------------------------------------------------ sg-report-flight on osstest.test-lab.xenproject.org logs: /home/logs/logs images: /home/logs/images Logs, config files, etc. are available at http://logs.test-lab.xenproject.org/osstest/logs Explanation of these reports, and of osstest in general, is at http://xenbits.xen.org/gitweb/?p=osstest.git;a=blob;f=README.email;hb=master http://xenbits.xen.org/gitweb/?p=osstest.git;a=blob;f=README;hb=master Test harness code can be found at http://xenbits.xen.org/gitweb?p=osstest.git;a=summary _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxx https://lists.xen.org/xen-devel
|
 |
Lists.xenproject.org is hosted with RackSpace, monitoring our |