On 8/11/22 7:28 AM, Christoph Hellwig wrote: > On Wed, Aug 10, 2022 at 12:38:08PM -0500, Mike Christie wrote: >> It sounded really nice at first, but to do that I would have a bunch of >> cases that might be specific to pr_ops, alua or scanning, etc. Then I >> also had cases where user1 does not want to retry but user2 did, so I >> have to add more SCMD bits. So, in the end it will get messier than >> you initially think. > > Do you have an overview of the different cases you've spot so far? I'll finish up what I was working on for this type of design and I'll post it as a rough RFC, so it should be easier to discuss in code format. I'm going to break out the first patch, handle your comment on it, and send that separately since it's a regression fix and not related to this discussion.
