On Wed, May 26, 2021 at 11:51 AM Sean Christopherson <seanjc@xxxxxxxxxx> wrote: > > On Wed, May 26, 2021, David Matlack wrote: > > On Wed, May 26, 2021 at 10:29 AM Sean Christopherson <seanjc@xxxxxxxxxx> wrote: > > > > > > Put version information in the subject, otherwise it's not always obvious which > > > patch you want to be accepted, e.g. > > > > > > [PATCH v2] KVM: x86/mmu: Fix comment mentioning skip_4k > > > > Got it. My thinking was that I changed the title of the patch so > > should omit the v2, but that doesn't really make sense. > > Ha, yeah, the version should get bumped even if a patch/series gets heavily > rewritten. There are exceptions (though I'm struggling to think of a good > example), but even then it's helpful to describe the relationship to any > previous series. > > It's also customery to describe the changes between versions in the cover letter, > or in the case of a one-off patch, in the part of the patch that git ignores. > > And my own personal preference is to also include lore links to previous versions, > e.g. in this case I would do something like: > > v2: Reword comment to document min_level. [sean] > > v1: https://lkml.kernel.org/r/20210526163227.3113557-1-dmatlack@xxxxxxxxxx > > Providing the explicit link in addition to the delta summaray makes it easy for > reviewers to see the history and understand the context of _why_ changes were > made. That's especially helpful for reviewers that didn't read/review earlier > versions. Great advice. Thanks Sean!
