On Wed, Mar 20, 2019 at 02:27:45PM +0800, Qu Wenruo wrote: > Do proper cleanup if we hit any error in extent_write_locked_range(), > and check the return value of flush_write_bio().
Yes that's what the code does, but the changelog should explain why this is correct. Same for "btrfs: extent_io: Handle error better in extent_writepages()". You do that in other patches, why not in this one too? If the reason is same for several patches, then copy it, eg. btrfs: "extent_io: Handle errors better in btree_write_cache_pages()" is ok. But don't resend, I'll fix it here as I made some other fixups to changelogs and am finishing the whole series.
