On Thu 2020-09-03 06:55:47, Stephen Rothwell wrote: > Hi Petr, > > On Wed, 2 Sep 2020 09:26:11 +0200 Petr Mladek <pmladek@xxxxxxxx> wrote: > > > > The problem is that this commit is not in mainline. It is living > > only in printk/linux.git. > > > > Could we use the SHA1 from the maintainer tree when it would not get rebased? > > > > Or should we rather avoid Fixes: tag referencing commits that are not > > in mainline? > > > > I am sorry to bother you with this silly question. I do not see any > > hint in Documentation/process/submitting-patches.rst. > > Well, in theory, maintainers trees should not be rebased after they > have been published (except in exceptional circumstances), so using > SHA1s from them should be OK. Especially if the fixing commit is in > the same maintainers tree (which it should be, right). It does mean > that maintainers need to be a bit more careful if they do rebase their > trees to update any Fixes tags (or other commit references) that are > affected by the rebase. Thanks a lot for info. I have rebased the last 5 commits in the printk-rework branch and added the missing SHAs there. Best Regards, Petr
