Re: [PATCH v2] dm-verity: support block number limits for different ioprio classes

[Mikulas Patocka]

On Thu, 27 Mar 2025, Mikulas Patocka wrote:

> There is also a bug that prefetch may race with suspend, sending I/Os even 
> when the dm-verity device is suspended - but that is fixable by adding 
> flush_workqueue to the postsuspend hook.
> 
> Mikulas

flush_workqueue won't make it - it seems that I'll have to add a new 
function dm_bufio_quisce that will wait until all prefetches terminate and 
use this function in the postsuspend hook of targets that do prefetch.

Mikulas