Re: Fix quilt merge error in acpi-cpufreq.c

[Paul Gortmaker <paul.gortmaker@windriver.com>]

On Wed, Apr 15, 2009 at 10:00 PM, Rusty Russell <rusty@rustcorp.com.au> wrote:
> On Thu, 16 Apr 2009 04:17:49 am H. Peter Anvin wrote:
>> Linus Torvalds wrote:
>> "build fix" is valid and proper use: it tells that it
>> fixes a compilation error, which succinctly communicates both the
>> priority of the fix and how it needs to be validated.
>
> Side note: I really prefer to see the compile error output in this case: great
> for googling.  It annoys me when people skip this.
>
> Anyway, Impact: had lead me to think harder about my messages than the
> free-form commit style did.  Perhaps it's too rigid, but it helped.
>
> Let's get concrete.  Here's the top 3 non-merge commits in gitk:
>
>    ALSA: hda - Fix the cmd cache keys for amp verbs
>
>    Fix the key value generation for get/set amp verbs.  The upper bits of
>    the parameter have to be combined with the verb value to be unique for
>    each direction/index of amp access.
>
>    This fixes the resume problem on some hardwares like Macbook after
>    the channel mode is changed.
>
> I have no idea what this patch does.  It seems to be a fix; what are the
> symptoms of the problem, and how long has it been there?
>
>    ALSA: add missing definitions(letters) to HD-Audio.txt
>
>    impact: Add missing definitions(letters).
>
> This is actually a pure documentation patch.  "Fix typos" or "Documentation
> fixes" would seem sufficient for subject, and no body needed.
>
>    ALSA: sound/pci: use memdup_user()
>
>    Remove open-coded memdup_user().
>
> Again, the body seems gratuitous.
>
> Anyone want to try to write a guide on writing good commit messages?

I'd put together some notes on making good commit messages for
internal use, trying to guide people into describing the symptoms of a
problem and why what is being submitted is the correct technical fix
-- vs. the semi-useless cop-outs of an English translation of the C
code itself (i.e. "increment foo comparison by one" or similar).

It isn't anything near a big wordy guide in its own right (and
hopefully it doesn't need to be), but I'm willing to use it as a
start, plus bits from this discussion, and once folks agree on the
content I put together being OK, we can stick it in Documentation
SubmittingPatches or wherever seems appropriate.  Assuming people in
general think there is a need for it...

Paul.

> Rusty.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>