On Thu, Aug 11, 2022 at 11:46:09AM +0200, Alexander Grund wrote:
On 08.08.22 15:28, Greg KH wrote:
But, we only take patches that actually do something. This one doesn't do anything at all, and has no measurable performance or bugfix that I can determine at all.
Isn't "doing less" also worth the patch?
Not if it doesn't actually fix something that a user sees.
I mean this patch removes a superflous pointer of the superblock struct making the kernel use less memory. It also saves a code line and operation during init and removes the (somewhat hidden in syntax) superflous indirect access (and hence memory read) of a pointer already available (likely even in a register) during get/set_mnt_opts.
Of course the effect here is small but I think cleanups are always good to avoid a "death by a thousand cuts" scenario, i.e. that even small things help.
We do not take "cleanup patches" in stable trees without it being a requirement for a real fix. Please read the stable kernel rules document again for what is actually allowed.
thanks,
greg k-h