Re: [PATCH 0/6] treewide: Add and use dev_fmt similar to pr_fmt

[Corey Minyard]

On 05/09/2018 12:04 PM, Joe Perches wrote:
> On Wed, 2018-05-09 at 11:47 -0500, Corey Minyard wrote:
> > On 05/09/2018 10:15 AM, Joe Perches wrote:
> > > The pr_fmt mechanism exists for pr_<level> logging message prefixing,
> > > but no similar capability exists for dev_<level> message prefixing.
> > >
> > > Many uses of dev_<level> have an embedded prefix for logging output.
> > >
> > > So add a similar dev_fmt macro that can automatically prefix the
> > > dev_<level> logging output.
> > >
> > > Rename the existing dev_<level> functions to _dev_<level> and add new
> > > macros that call _dev_<level> with the desired prefix if defined.
> > >
> > > The new default #define for dev_fmt is blank.
> > >
> > > Convert ipmi and infiniband to use this mechanism.
> > The IPMI changes look good to me.
> Oh good.
>
> > There are some conflicts with a patch I have pulling out the proc
> > interface that is destined for 3.18.
> I'm sure you mean 4.18.

Oops, yes :). I was just looking at a 3.x kernel and it stuck in my brain.

> > I can take the IPMI changes into my tree, if you want.
> These patches are not at all urgent and were done
> on top of next-20180509.
>
> As there are dependencies between the patch that
> introduces dev_fmt and the reset of the patches,
> I think it makes sense to take these as a single
> patchset rather than take parts into various trees.

The dependency isn't hard, the changes work without dev_fmt,
it just won't print the prefix. But I'm fine with you keeping
them.

> Respinning the IPMI patches is trivial and can be
> done whenever appropriate.
>
> When do you expect your IPMI patches to hit -next?
I went ahead and pulled it in now, it's been tested well enough
in my tree.

For patches 3, 4, and 5:

Acked-by: Corey Minyard <cminyard@xxxxxxxxxx>