On Mon, Nov 30, 2020 at 09:29:02PM +0100, Paolo Bonzini wrote:
On 30/11/20 20:44, Mike Christie wrote:
I have never seen a public/open-source vhost-scsi testsuite.
For patch 23 (the one that adds the lun reset support which is built on patch 22), we can't add it to stable right now if you wanted to, because it has a bug in it. Michael T, sent the fix:
https://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git/commit/?h=linu...
to Linus today.
Ok, so at least it was only a close call and anyway not for something that most people would be running on their machines. But it still seems to me that the state of CI in Linux is abysmal compared to what is needed to arbitrarily(*) pick up patches and commit them to "stable" trees.
Paolo
(*) A ML bot is an arbitrary choice as far as we are concerned since we cannot know how it makes a decision.
The choice of patches is "arbitrary", but the decision is human. The patches are reviewed coming out of the AI, sent to public mailing list(s) for review, followed by 2 reminders asking for reviews.
The process for AUTOSEL patches generally takes longer than most patches do for upstream.
It's quite easy to NAK a patch too, just reply saying "no" and it'll be dropped (just like this patch was dropped right after your first reply) so the burden on maintainers is minimal.