[Intel-gfx] [PATCH 00/27] add pm_runtime_last_busy_and_autosuspend() helper

Rafael J. Wysocki rjw at rjwysocki.net
Thu Sep 25 21:54:36 CEST 2014


On Thursday, September 25, 2014 04:27:58 PM Wolfram Sang wrote:
> 
> --Bn2rw/3z4jIqBvZU
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: inline
> Content-Transfer-Encoding: quoted-printable
> 
> On Thu, Sep 25, 2014 at 09:22:01AM -0500, Felipe Balbi wrote:
> > On Thu, Sep 25, 2014 at 01:27:18PM +0530, Vinod Koul wrote:
> > > On Wed, Sep 24, 2014 at 03:32:19PM -0500, Felipe Balbi wrote:
> > > > > > > OK, I guess this is as good as it gets.
> > > > > > >=20
> > > > > > > What tree would you like it go through?
> > > > > >=20
> > > > > > Do we really need this new helper ? I mean, the very moment when =
> we
> > > > > > decide to implement ->runtime_idle() we will need to get rid of t=
> his
> > > > > > change. I wonder if it's really valid...
> > > > >=20
> > > > > I'm not sure I'm following?  This seems to simply implement what dr=
> ivers
> > > > > have been doing already as one function.  Why would it be invalid t=
> o reduce
> > > > > code duplication?
> > > >=20
> > > > For two reasons:
> > > >=20
> > > > 1) the helper has no inteligence whatsoever. It just calls the same
> > > > functions.
> > > >=20
> > > > 2) the duplication will vanish whenever someone implements
> > > > ->runtime_idle() and have that call pm_runtime_autosuspend() (like PCI
> > > > and USB buses are doing today). This will just be yet another line th=
> at
> > > > needs to change.
> > > >=20
> > > > Frankly though, no strong feelings, I just think it's a commit that
> > > > doesn't bring that any benefits other than looking like one line was
> > > > removed.
> > > and yes that is what it tries to do nothing more nothing less. If in fu=
> ture
> > > there are no users (today we have quite a few), then we can remove the =
> dead
> > > macro, no harm. But that is not the situation today.
> >=20
> > as I said, a commit that's bound to be useless. It's not like you're
> > saving 10 lines of code, it's only one. Replacing two simple lines with
> > a function which takes <joke> almost as many characters to type </joke>.
> >=20
> > IMO, this is pretty useless and I'd rather not see them in the drivers I
> > maintain, sorry.
> 
> It is not a NACK from me; yet from a high-level perspective I agree with
> Felipe.

OK

I'd rather not merge something that driver people don't want to use.

Vinod?

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.



More information about the Intel-gfx mailing list