On 7/5/23 4:25?PM, Michael Schmitz wrote: > Hi Jens, > > On 6/07/23 10:13, Jens Axboe wrote: >>>> should be enough for it to go into stable from 5.2 and onwards. >>> OK - I wasn't certain whether you wanted the Fixes or stable tag dropped. >>>>> (Greg didn't seem to object to the Fixes: as such, just to the >>>>> incorrect version prereq) >>>> I think it's really confusing... A patch should only have a Fixes tag if >>>> it's fixing a specific bug in that patch. Either it is, in which case >>>> you would not need Cc stable at all since it's only in 6.5-rc, or it >>> It is fixing a bug in b6f3f28f60. I should have checked whether the >>> patch series had already gone to release, not just -rc, instead of >>> just adding the stable tag out of caution. >> But this is the confusion - if it's fixing a bug in b6f3f28f60, then why >> is it marked as needing to get backported much further back, predating >> that commit? > > I see - it doesn't need to be backported that far back _alone_. It > only needs to be applied after b6f3f28f60 once that one has been > backported. OK I see - I think there's some serious misunderstandings here then :-) It sounds like it fixes a bug in b6f3f28f60 alone, and it has no business going into stable. The commit should _just_ be marked with it fixing that. If someone were to backport that previous series, then their tooling or diligence should notice this dependency and this current commit should be picked as well. There should be no Cc: stable on this patch at all, I'll fix it up. -- Jens Axboe
