On Tue, 13 Aug 2019 14:08:06 -0700 > Hello, > > syzbot found the following crash on: > > HEAD commit: d45331b0 Linux 5.3-rc4 > git tree: upstream > console output: https://syzkaller.appspot.com/x/log.txt?x=1651e6d2600000 > kernel config: https://syzkaller.appspot.com/x/.config?x=6c5e70dcab57c6af > dashboard link: https://syzkaller.appspot.com/bug?extid=0265846a0cb9a0547905 > compiler: gcc (GCC) 9.0.0 20181231 (experimental) > syz repro: https://syzkaller.appspot.com/x/repro.syz?x=12c9c336600000 > C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1766156a600000 > > IMPORTANT: if you fix the bug, please add the following tag to the commit: > Reported-by: syzbot+0265846a0cb9a0547905@xxxxxxxxxxxxxxxxxxxxxxxxx > > executing program > executing program > executing program > executing program > BUG: memory leak > unreferenced object 0xffff8881226da6c0 (size 192): > comm "syz-executor332", pid 6977, jiffies 4294941214 (age 15.840s) > hex dump (first 32 bytes): > 00 00 00 00 00 00 00 00 00 f0 bc 28 81 88 ff ff ...........(.... > 01 c8 60 00 02 0b 00 00 00 00 00 00 00 00 00 00 ..`............. > backtrace: > [<00000000b06a638e>] kmemleak_alloc_recursive include/linux/kmemleak.h:43 [inline] > [<00000000b06a638e>] slab_post_alloc_hook mm/slab.h:522 [inline] > [<00000000b06a638e>] slab_alloc mm/slab.c:3319 [inline] > [<00000000b06a638e>] kmem_cache_alloc+0x13f/0x2c0 mm/slab.c:3483 > [<00000000950a289d>] mempool_alloc_slab+0x1e/0x30 mm/mempool.c:513 > [<00000000bfcc27e2>] mempool_alloc+0x64/0x1b0 mm/mempool.c:393 > [<00000000cdf95a4a>] bio_alloc_bioset+0x180/0x2c0 block/bio.c:477 > [<00000000b239bb68>] bio_clone_fast+0x25/0x90 block/bio.c:609 > [<000000005d58c2dc>] bio_split+0x4a/0xd0 block/bio.c:1856 > [<00000000ab943734>] blk_bio_segment_split block/blk-merge.c:250 [inline] > [<00000000ab943734>] __blk_queue_split+0x355/0x730 block/blk-merge.c:272 > [<00000000e702c0ac>] blk_mq_make_request+0xb0/0x890 block/blk-mq.c:1943 > [<000000003c89773a>] generic_make_request block/blk-core.c:1052 [inline] > [<000000003c89773a>] generic_make_request+0xf6/0x4a0 block/blk-core.c:994 > [<00000000a4dcaf78>] submit_bio+0x5a/0x1e0 block/blk-core.c:1163 > [<000000003e1ce7f8>] __blkdev_direct_IO fs/block_dev.c:459 [inline] > [<000000003e1ce7f8>] blkdev_direct_IO+0x2b3/0x6d0 fs/block_dev.c:515 > [<0000000087ec76a4>] generic_file_direct_write+0xb0/0x1a0 mm/filemap.c:3230 > [<00000000ff259b44>] __generic_file_write_iter+0xec/0x230 mm/filemap.c:3413 > [<00000000223d9b6c>] blkdev_write_iter fs/block_dev.c:2026 [inline] > [<00000000223d9b6c>] blkdev_write_iter+0xbe/0x160 fs/block_dev.c:2003 > [<000000003c4a5c94>] call_write_iter include/linux/fs.h:1870 [inline] > [<000000003c4a5c94>] aio_write+0x10b/0x1d0 fs/aio.c:1583 > [<00000000581f0c84>] __io_submit_one fs/aio.c:1815 [inline] > [<00000000581f0c84>] io_submit_one+0x59b/0xe50 fs/aio.c:1862 1, add BLK_QC_T_EAGAIN --- a/block/blk-core.c +++ b/block/blk-core.c @@ -1004,7 +1004,8 @@ blk_qc_t generic_make_request(struct bio blk_qc_t ret = BLK_QC_T_NONE; if (!generic_make_request_checks(bio)) - goto out; + /* feed error back to __blkdev_direct_IO fs/block_dev.c:459 */ + return BLK_QC_T_EAGAIN; /* * We only want one ->make_request_fn to be active at a time, else -- 2, plan-B, check status if BLK_QC_T_EAGAIN is bad --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -461,7 +461,7 @@ __blkdev_direct_IO(struct kiocb *iocb, s } qc = submit_bio(bio); - if (qc == BLK_QC_T_EAGAIN) { + if (qc == BLK_QC_T_EAGAIN || bio->bi_status) { if (!ret) ret = -EAGAIN; goto error; --
