On 05/22/2010 01:03 AM, Josef Bacik wrote: > In order for AIO to work, we need to implement aio_write. This patch converts > our btrfs_file_write to btrfs_aio_write. I've tested this with xfstests and > nothing broke, and the AIO stuff magically started working. Thanks, > > Signed-off-by: Josef Bacik <josef@xxxxxxxxxx> > Hi, Josef, I've tested your patch(May 22) with my tools, and one case triggered a bug which made writev operation hang up, more information is followed. - Steps to trigger it: # mount /dev/sda8 /home/btrfsdisk -o nodatacow # gcc direct-io.c -o direct-io # ./direct-io O_DIRECT writev /home/btrfsdisk/testrw 4M then on another tty, after "dmesg"... [snip] device fsid f44b0879c75c0e99-1d4b28f2d5c503ae devid 1 transid 11177 /dev/sda8 btrfs: setting nodatacow INFO: task direct-io:1399 blocked for more than 120 seconds. "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. direct-io D 0000000000000003 0 1399 1341 0x00000000 ffff880137c379c8 0000000000000082 ffff880137c379d8 ffffffff00000000 ffff880137c37fd8 ffff880139730000 0000000000015440 ffff880137c37fd8 0000000000015440 0000000000015440 0000000000015440 0000000000015440 Call Trace: [<ffffffffa0119d4a>] wait_extent_bit+0xe3/0x163 [btrfs] [<ffffffff8106651f>] ? autoremove_wake_function+0x0/0x39 [<ffffffffa0119e47>] lock_extent_bits+0x7d/0xa8 [btrfs] [<ffffffffa0119e88>] lock_extent+0x16/0x18 [btrfs] [<ffffffffa01025ce>] btrfs_direct_IO+0x8e/0x1be [btrfs] [<ffffffff810c7301>] generic_file_direct_write+0xed/0x16d [<ffffffffa010bb91>] btrfs_file_aio_write+0x2af/0x8d2 [btrfs] [<ffffffff81100eae>] ? try_get_mem_cgroup_from_mm+0x39/0x49 [<ffffffffa010b8e2>] ? btrfs_file_aio_write+0x0/0x8d2 [btrfs] [<ffffffff811063ed>] do_sync_readv_writev+0xc1/0x100 [<ffffffff81106120>] ? might_fault+0x21/0x23 [<ffffffff81106151>] ? copy_from_user+0x2f/0x31 [<ffffffff811c90ab>] ? security_file_permission+0x16/0x18 [<ffffffff81107145>] do_readv_writev+0xa7/0x127 [<ffffffff81107208>] vfs_writev+0x43/0x4e [<ffffffff811072f8>] sys_writev+0x4a/0x93 [<ffffffff81009c32>] system_call_fastpath+0x16/0x1b So, can you figure out if anything in your patch leads to the bug? -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
