Re: [Qemu-devel] [PULL v2 12/27] target/mips: Convert to CPUClass::tlb_fill

[Peter Maydell]

On Tue, 14 May 2019 at 17:05, Aleksandar Markovic
<address@hidden> wrote:
> I am not complaining about myself looking at another commit message, but have 
> future maintainers and future developers in mind. Their effort needed for 
> deciphering commit messages like this one is multiple times larger than 
> putting together a clear, full, and right-on-the-money message by the 
> submitter. The commit messages should be made convenient for their readers, 
> not writters, shouldn't they?

Yeah, good commit messages are important; in the end there
is a judgement call to be made about how much detail is useful.

I think one thing that might have affected our differing
views here is that you were only cc'd on the MIPS-related patch,
so will have been looking at it in isolation. I read the whole
series starting with the initial patches which introduced the
API, so had a fuller context for looking at this patch.
(I think future readers will also be able to acquire that
fuller context easily because they can just look through
the git history for the nearby commit that adds the tlb_fill
method if they want the rationale for the refactoring.)

thanks
-- PMM