On Sun, Feb 21, 2016 at 09:36:44AM +0800, Qu Wenruo wrote:
>
>
> On 02/19/2016 09:41 PM, Anand Jain wrote:
> >
> >
> > Saw below warn leading to bug when running btrfs/011, not
> > reproducible. Any idea ?
> >
> Seems like another wq_destroy race.
>
> But it's hard to locate which wq is the cause from backtrace only.
>
> What's the wq btrfs_stop_all_workers+0xcd is going to free?
You can always try to guess on your locally built sources, provided the
configs do not enable "too much" debugging. What I get and looks
reasonable:
(gdb) l *(btrfs_stop_all_workers+0xcd)
0x2bc7d is in btrfs_stop_all_workers (fs/btrfs/disk-io.c:2154).
warning: Source file is more recent than executable.
2149 btrfs_destroy_workqueue(fs_info->delayed_workers);
2150 btrfs_destroy_workqueue(fs_info->caching_workers);
2151 btrfs_destroy_workqueue(fs_info->readahead_workers);
2152 btrfs_destroy_workqueue(fs_info->flush_workers);
2153 btrfs_destroy_workqueue(fs_info->qgroup_rescan_workers);
2154 btrfs_destroy_workqueue(fs_info->extent_workers);
2155 }
2156
2157 static void free_root_extent_buffers(struct btrfs_root *root)
2158 {
> And what's the wq end_workqueue_bio+0x85 is going to add?
(gdb) l *(end_workqueue_bio+0x85)
0x2ab55 is in end_workqueue_bio (fs/btrfs/disk-io.c:726).
721
722 if (bio->bi_rw & REQ_WRITE) {
723 if (end_io_wq->metadata == BTRFS_WQ_ENDIO_METADATA) {
724 wq = fs_info->endio_meta_write_workers;
725 func = btrfs_endio_meta_write_helper;
726 } else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_FREE_SPACE) {
727 wq = fs_info->endio_freespace_worker;
728 func = btrfs_freespace_write_helper;
729 } else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56) {
730 wq = fs_info->endio_raid56_workers;
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
