> Il giorno 15 set 2021, alle ore 14:56, Jens Axboe <axboe@xxxxxxxxx> ha scritto: > > On 9/8/21 10:07 PM, Guoqing Jiang wrote: >> Hi, >> >> With latest kernel (commit 4ac6d90867a4 "Merge tag 'docs-5.15' of >> git://git.lwn.net/linux"), >> I got lots of kmemleak reports during compile kernel source code. >> >> # dmesg |grep kmemleak >> [18234.655491] kmemleak: 1 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [18890.247552] kmemleak: 2 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [19745.602271] kmemleak: 2 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [20390.965851] kmemleak: 4 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [21150.173950] kmemleak: 4 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [21929.951448] kmemleak: 15 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [23589.726859] kmemleak: 2 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [24416.441263] kmemleak: 1 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [30400.835853] kmemleak: 10 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [68673.737862] kmemleak: 2 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [72770.498898] kmemleak: 1 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> [77046.434369] kmemleak: 7 new suspected memory leaks (see >> /sys/kernel/debug/kmemleak) >> >> All of them have similar trace as follows. >> >> comm "sh", pid 27054, jiffies 4302241171 (age 47562.964s) >> hex dump (first 32 bytes): >> 01 00 00 00 00 00 00 00 00 b0 20 02 80 88 ff ff .......... ..... >> 04 00 02 00 04 00 02 00 00 00 00 00 00 00 00 00 ................ >> backtrace: >> [<000000004fa2550b>] bfq_get_queue+0x2a8/0xfd0 >> [<00000000b1757a70>] bfq_get_bfqq_handle_split+0xa4/0x240 >> [<00000000ac263274>] bfq_init_rq+0x1f7/0x1d10 >> [<00000000110283e1>] bfq_insert_requests+0xf7/0x5f0 >> [<000000002ed06e79>] blk_mq_sched_insert_requests+0xfe/0x350 >> [<000000000ebf38ac>] blk_mq_flush_plug_list+0x256/0x3e0 >> [<00000000bc647b2b>] blk_flush_plug_list+0x1ff/0x240 >> [<000000004e7e49f8>] blk_finish_plug+0x3c/0x60 >> [<000000008802f1e4>] read_pages+0x28f/0x580 >> [<000000009986d1f4>] page_cache_ra_unbounded+0x266/0x3a0 >> [<00000000492c494f>] filemap_fault+0x8a8/0xfe0 >> [<000000009cbd8d38>] __do_fault+0x70/0x150 >> [<000000002740a35f>] do_fault+0x112/0x670 >> [<00000000a74facab>] __handle_mm_fault+0x57e/0xcc0 >> [<0000000024009667>] handle_mm_fault+0xd6/0x330 >> [<00000000fb1e0780>] do_user_addr_fault+0x2a9/0x8b0 >> unreferenced object 0xffff888021e5dd90 (size 560): > > Paolo, are you on top of this one? > Thanks Guoqing for reporting this. I'm overloaded at the moment, please give me a little time and I'll have a look. Paolo > -- > Jens Axboe