On Tue, Dec 31, 2013 at 10:01 -0800, Randy Dunlap wrote:
>
> On 12/31/13 05:48, Gerhard Sittig wrote:
> > On Tue, Dec 31, 2013 at 11:51 +0100, Krzysztof Hałasa wrote:
> >>
> >> drivers/crypto/ixp4xx_crypto.c: In function 'ixp_module_init':
> >> drivers/crypto/ixp4xx_crypto.c:1419:2: error: 'dev' undeclared (first use in this function)
> >>
> >> Now builds. Not tested on real hw.
> >>
> >> Signed-off-by: Krzysztof Hałasa <[email protected]>
> >
> > I understood that a 'Fixes: <hash> <subject>' line is suggested when
> > you can clearly identify what broke what you fix. Is it 27c1789ca6a6
> > "DMA-API: crypto: remove last references to 'static struct device *dev'"?
> > Then v3.13-rc1 would have been affected.
>
> is 'Fixes: <whatever>' documented somewhere?
Last time I checked (few days ago, 'git grep -w Fixes:') it
wasn't, at least not within the kernel source tree and its
Documentation hierarchy. Today's master still does not have it.
But a quick search in recent LAKML messages reveals that it's
been in use by e.g. Olof Johansson, Jason Cooper, Rusty Russell,
Dan Carpenter, Russell King, Mark Brown, Stephen Warren, Paul
Walmsley, and I'm not making this up but am just referring to
what repeatedly was requested in the past. The form used there
was "Fixes: <hash> (<oneline>)" though. A doc update may be due
to have a canonical format.
I do see the benefit of this tag, as it does not hide this
information in the commit message's prose, makes submitters think
about that property in more explicit ways and makes them provide
the information such that others need not do the research, and
greatly helps those involved in tests and release management to
determine affected versions and required actions that need to be
taken. The tag is very useful.
virtually yours
Gerhard Sittig
--
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: [email protected]
On Wed, Jan 01, 2014 at 01:37:46PM +0100, Gerhard Sittig wrote:
> Last time I checked (few days ago, 'git grep -w Fixes:') it
> wasn't, at least not within the kernel source tree and its
> Documentation hierarchy. Today's master still does not have it.
>
> But a quick search in recent LAKML messages reveals that it's
> been in use by e.g. Olof Johansson, Jason Cooper, Rusty Russell,
> Dan Carpenter, Russell King, Mark Brown, Stephen Warren, Paul
> Walmsley, and I'm not making this up but am just referring to
> what repeatedly was requested in the past. The form used there
> was "Fixes: <hash> (<oneline>)" though. A doc update may be due
> to have a canonical format.
It was discussed at kernel summit, and I believe the resulting format
was only published in one of the kernel summit mailing lists. The
outcome of it as I understand was that the format is:
Fixes: <hash> ("<one line summary>")
Though the addition of the quotes was (iirc) only something added after
the vocal discussion. If you mail Linus a patch with a fixes line which
doesn't match that, Linus will probably fix it up... (Linus has done that
with one of mine in the past.)
On Wed, Jan 01, 2014 at 12:46:25PM +0000, Russell King - ARM Linux wrote:
> On Wed, Jan 01, 2014 at 01:37:46PM +0100, Gerhard Sittig wrote:
> > Last time I checked (few days ago, 'git grep -w Fixes:') it
> > wasn't, at least not within the kernel source tree and its
> > Documentation hierarchy. Today's master still does not have it.
> >
> > But a quick search in recent LAKML messages reveals that it's
> > been in use by e.g. Olof Johansson, Jason Cooper, Rusty Russell,
> > Dan Carpenter, Russell King, Mark Brown, Stephen Warren, Paul
> > Walmsley, and I'm not making this up but am just referring to
> > what repeatedly was requested in the past. The form used there
> > was "Fixes: <hash> (<oneline>)" though. A doc update may be due
> > to have a canonical format.
>
> It was discussed at kernel summit, and I believe the resulting format
> was only published in one of the kernel summit mailing lists. The
> outcome of it as I understand was that the format is:
>
> Fixes: <hash> ("<one line summary>")
fwiw, Linus also recommended setting core.abbrev = 12.
thx,
Jason.