On 12/8/21 7:51 PM, Josef Bacik wrote:

On Wed, Dec 08, 2021 at 02:12:09PM -0800, syzbot wrote:

quoted

syzbot has bisected this issue to:

commit 741ec653ab58f5f263f2b6df38157997661c7a50
Author: Qu Wenruo [off-list ref]
Date:   Mon Sep 27 07:22:01 2021 +0000

    btrfs: subpage: make end_compressed_bio_writeback() compatible

This isn't the right patch, this is x86 so sectorsize == pagesize, so this patch
didn't change anything at all.  Also unless the local fs is btrfs with
compression enabled I'm not entirely sure how this would even matter, the
reproducer seems to be purely io_uring related.  Thanks,

Yeah, it's not btrfs, it's just one of those "bisect gone awry" cases. btrfs
guys can ignore this one.

-- 
Jens Axboe