On Fri, Feb 4, 2022 at 8:28 PM Ulf Hansson <ulf.hansson@xxxxxxxxxx> wrote:
>
> On Tue, 25 Jan 2022 at 06:50, Kai-Heng Feng <kai.heng.feng@xxxxxxxxxxxxx> wrote:
> >
> > Commit 5b4258f6721f ("misc: rtsx: rts5249 support runtime PM") doesn't
> > use pm_runtime_{get,put}() helpers when it should, so the RPM refcount
> > keeps at zero, hence its parent driver, rtsx_pci, has to do lots of
> > weird tricks to keep it from runtime suspending.
> >
> > So use those helpers at right places to properly manage runtime PM.
> >
> > Fixes: 5b4258f6721f ("misc: rtsx: rts5249 support runtime PM")
> > Cc: Ricky WU <ricky_wu@xxxxxxxxxxx>
> > Signed-off-by: Kai-Heng Feng <kai.heng.feng@xxxxxxxxxxxxx>
>
> The runtime PM reference counting for the mmc host device is managed
> by the mmc core. Have a look at __mmc_claim_host() and
> mmc_release_host().
>
> In other words, the runtime PM reference counting should not be needed
> in the mmc host driver, unless there are some specific cases, like for
> example during ->probe|remove().
>
> So perhaps it's only the changes in the ->probe|remove() functions
> that you need to fix the problems? No?
Yes you are right. Let me send a patch to remove redundant PM helpers
from this patch.
Kai-Heng
>
> [...]
>
> Kind regards
> Uffe
