Additionally, it would seem saner to standardize rules around when
code is expected to hit the maintainers hands for kernel
releases. Both yours and Martins deals with that, there really
shouldn't be the need to have this specified in detail per sub-system.
Yeah. There is basically nothing specific about SCSI in my write-up
outside of the branch naming.
I deliberately didn't mention coding style preferences. We have so much
20+ year old cruft in SCSI that's impossible to even entertain. But I do
request new code to follow coding-style.rst. BYOXT.
Also note that the original target audience for my document. It was
aimed at onboarding new driver contributors from hardware companies. So
people that don't live and breathe Linux development and who are not
intimately familiar with the kernel development process. It's possible
that we have this information in Documentation/ these days; I'll go
look. But it didn't exist when this doc was written. And in my
experience nobody coming to Linux development from the outside
understands what the "merge window" is. And when the appropriate time is
to submit patches and features. I think this would be a great area to
have a common set of guidelines and documentation. I'd prefer for this
to be global and then let maintainers apply their own wiggle room
instead of documenting particular rules for a given subsystem.
One pet peeve I have is that people are pretty bad at indicating the
intended target tree. I often ask for it in private mail but the
practice doesn't seem to stick. I spend a ton of time guessing whether a
patch is a fix for a new feature in the x+1 queue or a fix for the
current release. It is not always obvious.
Martin K. Petersen Oracle Linux Engineering