Re: [PATCH v5 1/3] writeback: clear the old wb's dirty IO state after switching inodes

[Tejun Heo]

Hello, Liz.

On Thu, Oct 01, 2026 at 11:50:20PM +0000, Liz Fong-Jones wrote:
> Test: a later patch in this series switches all of a replaced cgwb's
> inodes, dirty ones included, to its successor. After that, with every
> writer stopped and the filesystem synced, BdiWriteBandwidth in
> /sys/kernel/debug/bdi/*/stats stayed at 411700 and 407820 kBps with
> b_dirty, b_io and b_more_io all empty, and the replaced wb was released
> with WB_has_dirty_io set. With this patch it read 0 kBps and the wb was
> released clean, in 3 of 3 runs.

Maybe trim this to a sentence?

Acked-by: Tejun Heo <tj@xxxxxxxxxx>

Thanks.

--
tejun