On Tue 27-12-16 12:48:24, Klaus Ethgen wrote: [...] > By the way, I find the following two messages often. Maybe they are > unrelated, maybe not. > [31633.189121] Purging GPU memory, 144 pages freed, 5692 pages still pinned. > [31638.530025] Unable to lock GPU to purge memory. I do not think this makes much of a difference for the oom reports. See more below: [...] > [28756.498366] Xorg invoked oom-killer: gfp_mask=0x24200d4(GFP_USER|GFP_DMA32|__GFP_RECLAIMABLE), nodemask=0, order=0, oom_score_adj=0 [...] > [28756.498761] Node 0 active_anon:409988kB inactive_anon:233592kB active_file:1272348kB inactive_file:720884kB unevictable:27744kB isolated(anon):0kB isolated(file):0kB mapped:190792kB dirty:4656kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 0kB anon_thp: 35332kB writeback_tmp:0kB unstable:0kB pages_scanned:4062806 all_unreclaimable? yes > [28756.498769] DMA free:4116kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:1568kB inactive_file:0kB unevictable:0kB writepending:0kB present:15984kB managed:15908kB mlocked:0kB slab_reclaimable:9848kB slab_unreclaimable:376kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB > lowmem_reserve[]: 0 833 3008 3008 > [28756.498782] Normal free:42404kB min:42416kB low:53020kB high:63624kB active_anon:1316kB inactive_anon:20540kB active_file:548376kB inactive_file:60kB unevictable:0kB writepending:260kB present:892920kB managed:854328kB mlocked:0kB slab_reclaimable:184192kB slab_unreclaimable:47020kB kernel_stack:2728kB pagetables:0kB bounce:0kB free_pcp:936kB local_pcp:216kB free_cma:0kB [...] > > [28757.732436] updatedb.mlocat invoked oom-killer: gfp_mask=0x2400840(GFP_NOFS|__GFP_NOFAIL), nodemask=0, order=0, oom_score_adj=0 > [28757.732649] Node 0 active_anon:124120kB inactive_anon:61840kB active_file:1280688kB inactive_file:713860kB unevictable:27744kB isolated(anon):0kB isolated(file):0kB mapped:103164kB dirty:0kB writeback:4kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 0kB anon_thp: 27352kB writeback_tmp:0kB unstable:0kB pages_scanned:3674354 all_unreclaimable? yes > [28757.732656] DMA free:4116kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:1568kB inactive_file:0kB unevictable:0kB writepending:0kB present:15984kB managed:15908kB mlocked:0kB slab_reclaimable:9848kB slab_unreclaimable:376kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB > lowmem_reserve[]: 0 833 3008 3008 > [28757.732669] Normal free:42324kB min:42416kB low:53020kB high:63624kB active_anon:1312kB inactive_anon:20408kB active_file:549628kB inactive_file:40kB unevictable:0kB writepending:0kB present:892920kB managed:854328kB mlocked:0kB slab_reclaimable:184036kB slab_unreclaimable:46960kB kernel_stack:2384kB pagetables:0kB bounce:0kB free_pcp:408kB local_pcp:124kB free_cma:0kB > lowmem_reserve[]: 0 0 17397 17397 [...] > [31617.991795] gkrellm invoked oom-killer: gfp_mask=0x25000c0(GFP_KERNEL_ACCOUNT), nodemask=0, order=0, oom_score_adj=0 > [31617.991950] Node 0 active_anon:410912kB inactive_anon:188952kB active_file:1274928kB inactive_file:579668kB unevictable:28468kB isolated(anon):0kB isolated(file):0kB mapped:184664kB dirty:68kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 0kB anon_thp: 34744kB writeback_tmp:0kB unstable:0kB pages_scanned:5362697 all_unreclaimable? yes > [31617.991957] DMA free:4116kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:1568kB inactive_file:0kB unevictable:0kB writepending:0kB present:15984kB managed:15908kB mlocked:0kB slab_reclaimable:9848kB slab_unreclaimable:376kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB > lowmem_reserve[]: 0 833 3008 3008 > [31617.991970] Normal free:42300kB min:42416kB low:53020kB high:63624kB active_anon:1380kB inactive_anon:20244kB active_file:558700kB inactive_file:16kB unevictable:0kB writepending:16kB present:892920kB managed:854328kB mlocked:0kB slab_reclaimable:173456kB slab_unreclaimable:45024kB kernel_stack:3144kB pagetables:0kB bounce:0kB free_pcp:1392kB local_pcp:612kB free_cma:0kB All of those are lowmem requests triggering the oom killer while there is a lot of page cache sitting on the active Normal list. So it smells like the same issue I was referring to and the mentioned patch should fix the issue. -- Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>