On Tue, Jul 16, 2019 at 03:57:30AM +0000, howaboutsynergy@xxxxxxxxxxxxxx wrote: > > PID: 21120 TASK: ffff8fd7f10ddac0 CPU: 9 COMMAND: "stress" > > START: __schedule at ffffffff987f5a0c > > [ffff9cb9a9f57928] isolate_migratepages_block at ffffffff981b35c6 > > [ffff9cb9a9f57950] isolate_migratepages_block at ffffffff981b37bd > > [ffff9cb9a9f57a10] compact_zone at ffffffff981b4c07 > > [ffff9cb9a9f57ab8] compact_zone_order at ffffffff981b51de > > [ffff9cb9a9f57b78] try_to_compact_pages at ffffffff981b5a17 > > [ffff9cb9a9f57bd8] __alloc_pages_direct_compact at ffffffff981d95a7 > > [ffff9cb9a9f57c30] __alloc_pages_slowpath at ffffffff981d99ee > > [ffff9cb9a9f57cd0] release_pages at ffffffff9819aa08 > > [ffff9cb9a9f57d00] __pagevec_lru_add_fn at ffffffff9819ada9 > > [ffff9cb9a9f57d40] __alloc_pages_nodemask at ffffffff981da668 > > [ffff9cb9a9f57da0] do_huge_pmd_anonymous_page at ffffffff98202c01 > > [ffff9cb9a9f57df0] __handle_mm_fault at ffffffff981bfa4c > > [ffff9cb9a9f57ea0] handle_mm_fault at ffffffff981c01f9 > > [ffff9cb9a9f57ec8] __do_page_fault at ffffffff9803d5c7 > > [ffff9cb9a9f57f28] do_page_fault at ffffffff9803d85d > > [ffff9cb9a9f57f48] page_fault at ffffffff98800de8 > > [ffff9cb9a9f57f50] page_fault at ffffffff98800dfe > > RIP: 00005879fc8c4c10 RSP: 00007ffd393aa0d0 RFLAGS: 00010206 > > RAX: 0000000010b8a000 RBX: 0000714104876010 RCX: 0000714358a5a3db > > RDX: 0000000000000000 RSI: 00000002540bf000 RDI: 0000714104876000 > > RBP: 00005879fc8c5a54 R8: 0000714104876010 R9: 0000000000000000 > > R10: 0000000000000022 R11: 00000002540be400 R12: ffffffffffffffff > > R13: 0000000000000002 R14: 0000000000001000 R15: 00000002540be400 > > ORIG_RAX: ffffffffffffffff CS: 0033 SS: 002b > > crash> High CPU usage in this path is not something I've observed recently. When it happens and CPU usage is high, can you run the following commands please? trace-cmd record -e compaction:* sleep 10 trace-cmd report > trace.log and send me the resulting trace.log please? -- Mel Gorman SUSE Labs
