Re: drivers/mmc/host/sdhci-of-core.c on sparc64

Re: drivers/mmc/host/sdhci-of-core.c on sparc64

[Chris Ball]

Hi,

On Thu, Jan 13, 2011 at 03:41:40PM -0800, Andrew Morton wrote:
> > > drivers/mmc/host/sdhci-of-core.c:24:25: asm/machdep.h: No such file or directory
> > > drivers/mmc/host/sdhci-of-core.c: In function `sdhci_of_wp_inverted':
> > > drivers/mmc/host/sdhci-of-core.c:115: error: implicit declaration of function `machine_is'
> > > 
> > > That code's been there for a while.  Did someone change Kconfig?
> > 
> > Can you attach a .config?  <asm/machdep.h> is arch-specific, so I'd
> > suggest that you're building on an unsupported arch.
> > 
> 
> sparc64 allmodconfig.

You're right, Andrew -- Rob (CC'd) changed the MMC Kconfig to build this
driver on Sparc.  Mainline commit 236cdc7bc71 ("of: make drivers depend
on CONFIG_OF instead of CONFIG_PPC_OF").

Rob also posted a patch to devicetree-discuss@, on top of the one above
("mmc: sdhci-of: fix build on non-powerpc platforms"), to fix up the
Sparc build by ifdef'ing for PPC inside the driver.  Grant merged the
first patch but not the second, hence sparc64 allmodconfig is broken
now.  The reason Grant didn't merge the second patch may be that Wolfram
objected to #ifdef proliferation inside the driver.

Options, as I see it:
 * revert the commit such that MMC_SDHCI_OF once again depends on PPC_OF
 * take the second patch as-is
 * come up with a less-#ifdeffy second patch

Wolfram, would appreciate your input on what we should do here.  Thanks,

-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: drivers/mmc/host/sdhci-of-core.c on sparc64

[Grant Likely]

On Thu, Jan 13, 2011 at 6:12 PM, Chris Ball <cjb@laptop.org> wrote:
> Hi,
>
> On Thu, Jan 13, 2011 at 03:41:40PM -0800, Andrew Morton wrote:
>> > > drivers/mmc/host/sdhci-of-core.c:24:25: asm/machdep.h: No such file or directory
>> > > drivers/mmc/host/sdhci-of-core.c: In function `sdhci_of_wp_inverted':
>> > > drivers/mmc/host/sdhci-of-core.c:115: error: implicit declaration of function `machine_is'
>> > >
>> > > That code's been there for a while.  Did someone change Kconfig?
>> >
>> > Can you attach a .config?  <asm/machdep.h> is arch-specific, so I'd
>> > suggest that you're building on an unsupported arch.
>> >
>>
>> sparc64 allmodconfig.
>
> You're right, Andrew -- Rob (CC'd) changed the MMC Kconfig to build this
> driver on Sparc.  Mainline commit 236cdc7bc71 ("of: make drivers depend
> on CONFIG_OF instead of CONFIG_PPC_OF").
>
> Rob also posted a patch to devicetree-discuss@, on top of the one above
> ("mmc: sdhci-of: fix build on non-powerpc platforms"), to fix up the
> Sparc build by ifdef'ing for PPC inside the driver.  Grant merged the
> first patch but not the second, hence sparc64 allmodconfig is broken
> now.  The reason Grant didn't merge the second patch may be that Wolfram
> objected to #ifdef proliferation inside the driver.
>
> Options, as I see it:
>  * revert the commit such that MMC_SDHCI_OF once again depends on PPC_OF
>  * take the second patch as-is
>  * come up with a less-#ifdeffy second patch
>
> Wolfram, would appreciate your input on what we should do here.  Thanks,

I've applied the 2nd patch.  I had applied it to my test tree, and
indeed I replied saying I did, but I had a corrupt git tree event
shortly after applying it to my test branch and evidently lost track
of it.

It's been pushed out to my next-devicetree branch:

git://git.secretlab.ca/git/linux-2.6.git next-devicetree

I'll ask Linus to pull after doing some sanity testing.

g.