At Thu, 18 Sep 2014 14:51:26 +0530,
Sudip Mukherjee wrote:
>
> On Mon, Sep 15, 2014 at 08:17:52PM +0200, Takashi Iwai wrote:
> > At Mon, 15 Sep 2014 20:55:54 +0530,
> > Sudip Mukherjee wrote:
> > >
> > > On Mon, Sep 15, 2014 at 04:29:48PM +0200, Takashi Iwai wrote:
> > > > At Mon, 15 Sep 2014 19:39:41 +0530,
> > > > Sudip Mukherjee wrote:
> <snip>
> > > > >
> > > > > diff --git a/sound/pci/ctxfi/ctresource.c
> > > > > b/sound/pci/ctxfi/ctresource.c
> > > > > index e49d2be..80beecb 100644
> > > > > --- a/sound/pci/ctxfi/ctresource.c
> > > > > +++ b/sound/pci/ctxfi/ctresource.c
> > > > (snip)
> > > > > @@ -282,8 +287,9 @@ int rsc_mgr_uninit(struct rsc_mgr *mgr)
> > > > > case SUM:
> > > > > break;
> > > > > default:
> > > > > - pr_err("ctxfi: Invalid resource type value
> > > > > %d!\n",
> > > > > - mgr->type);
> > > > > + dev_err(&(((struct hw *)mgr->hw)->pci->dev),
> > > > > + "ctxfi: Invalid resource type value
> > > > > %d!\n",
> > > > > + mgr->type);
> > > >
> > > > Did you really conclude that this is the best way?
> > > > Also, is it good to mix up the usages of both card->dev and &pci->dev?
> > > > Think again.
> > > >
> > >
> > > i have a doubt regarding this :-
> > > in the snd_card_new() card->dev is being assigned with &pci->dev ,
> > > then are not they the same ?
> >
> > Yes, but how can it be guaranteed in future? We may avoid the
> > problems in future by keeping the consistency at this moment. It's
> > one of the good points of keeping code consistent, in addition to:
> > increased readability / understandability and making the bug easier to
> > be spotted.
>
> understood the point.
> >
> > > i was trying to get some way of finding out the reference of card->dev
> > > from the resource manager ,
> > > but ... :(
> > > i will try again and if i cant find any way i will ask for some hint from
> > > you.
> >
> > Good! A hint is that there is no 100% perfect way to achieve this.
> > It's always compromise, and you'll have to choose which one is better
> > than others. For that, you'll have to evaluate multiple
> > implementations, and it's a really good exercise for coding.
> >
>
> i can find the rsc_mgr from the snd_card or ct_atc
> but if i want to find the reference to the device from the managers , then
> either i can go with
> 1) &(((struct hw *)mgr->hw)->pci->dev)
> or
> 2) use pci_get_drvdata with ((struct hw *)mgr->hw)->pci
> or
> 3) using container_of with &(((struct hw *)mgr->hw)->pci->dev) to find the
> pointer to snd_card via card_dev , but it becomes too complicated :(
>
> we can get to all the managers (rsc_mgr, amixer_mgr, src_mgr ...) using
> atc->rsc_mgrs[] , then there should be some simple way to go
> the opposite way (reaching atc from the managers) . container_of will not
> work .. :(
> or am i missing something??
Adding a new field pointing to card to each struct.
Takashi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
