Re: [PATCH v2] mmc: sdhci-s3c: fix the card detection in runtime-pm

["Heiko Stübner" <heiko@sntech.de>]

Hi Chris,

Am Montag, 29. Oktober 2012, 22:16:47 schrieb Chris Ball:
> Hi,
> 
> On Fri, Oct 19 2012, Heiko Stübner wrote:
> > Am Freitag, 19. Oktober 2012, 10:04:14 schrieb Seungwon Jeon:
> >> On Thursday, October 18, 2012, Heiko Stuebner <heiko@sntech.de>
> >> 
> >> > If host clock is disabled, host cannot detect a card
> >> > in case of using the internal or gpio card-detect for detection.
> >> > 
> >> > Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
> >> > Tested-by: Heiko Stuebner <heiko@sntech.de>
> >> > ---
> >> > I've added the SDHCI_CD_GPIO to the conditional. With this change it
> >> > works on my machine. But I'm not sure if this would also be necessary
> >> > for the external card detect.
> >> 
> >> Card detection will be asserted out of host in case of using
> >> SDHCI_CD_GPIO. Adding SDHCI_CD_GPIO to the conditional seems
> >> unnecessary.
> > 
> > But it _was_ necessary :-) . Because only then did the problem go away.
> > You might be right, that there exists a better solution for this, but
> > something in the original patch is at least still missing to fix the
> > problem.
> 
> Seungwon, what would you like to do here?  (I'm unhappy taking a patch
> with your name on it that's been modified in a way you disagree with.)

I agree with Seungwon, that the solution in my v2 is not the correct one :-) - 
should've called it RFC or so. The gpio-card-detect code _should_ be able to 
handle the resume of the host itself in the card detect case, but it seems 
it's not able to.

The patch in its v1 form does not fix the problem commit 2abeb5c5ded2 (mmc: 
sdhci-s3c: Add clk_(enable/disable) in runtime suspend/resume) has caused for 
Samsung devices using the gpio card-detect - it's broken in 3.7 currently.


So for 3.7 the easiest way would of course be to revert the commit above and 
find a correct solution for 3.8.


Heiko