platforms/44x/warp-nand.c bogosity

arch/powerpc/platforms/44x/warp-nand.c was added this year and updated 
quite recently.

warp-nand.c is empty unless CONFIG_MTD_NAND_NDFC=y.

MTD_NAND_NDFC depends on !PPC_MERGE.

!PPC_MERGE can never be true on PPC now that arch/ppc/ got removed.

So warp-nand.c is dead code, but that does not seem to be intentionally?

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

On Fri, 1 Aug 2008 18:14:01 +0300
Adrian Bunk [off-list ref] wrote:

arch/powerpc/platforms/44x/warp-nand.c was added this year and updated 
quite recently.

warp-nand.c is empty unless CONFIG_MTD_NAND_NDFC=y.

MTD_NAND_NDFC depends on !PPC_MERGE.

!PPC_MERGE can never be true on PPC now that arch/ppc/ got removed.

So warp-nand.c is dead code, but that does not seem to be intentionally?

I don't believe it is intentional, no.  Sean can you look at this and
see what's up?  I have a couple of fixes to queue up from you already
and if you can get this one done in the next few days I'll included it
as well.

josh

Josh Boyer wrote:

On Fri, 1 Aug 2008 18:14:01 +0300
Adrian Bunk [off-list ref] wrote:

quoted

arch/powerpc/platforms/44x/warp-nand.c was added this year and updated quite recently.

warp-nand.c is empty unless CONFIG_MTD_NAND_NDFC=y.

MTD_NAND_NDFC depends on !PPC_MERGE.

!PPC_MERGE can never be true on PPC now that arch/ppc/ got removed.

So warp-nand.c is dead code, but that does not seem to be intentionally?

I don't believe it is intentional, no.  Sean can you look at this and
see what's up?  I have a couple of fixes to queue up from you already
and if you can get this one done in the next few days I'll included it
as well.

josh

The current community ndfc driver won't build for arch/powerpc.
It still uses ioremap64 and includes <asm/ibm4xx.h> stuff.

Valentine.

On Fri, 1 Aug 2008 18:14:01 +0300
"Adrian Bunk" [off-list ref] wrote:

arch/powerpc/platforms/44x/warp-nand.c was added this year and
updated quite recently.

warp-nand.c is empty unless CONFIG_MTD_NAND_NDFC=y.

MTD_NAND_NDFC depends on !PPC_MERGE.

!PPC_MERGE can never be true on PPC now that arch/ppc/ got removed.

So warp-nand.c is dead code, but that does not seem to be
intentionally?

I use a local copy of ndfc.c that was modified to work with
arch/powerpc. I submitted the patch, but it was never accepted into the
kernel, and to be honest I don't remember why. I will look into this.

So the code is definitely not dead. You can't realistically run a warp
without the NAND.

Cheers,
   Sean

On Fri, 1 Aug 2008 11:26:28 -0400
"Josh Boyer" [off-list ref] wrote:

I don't believe it is intentional, no.  Sean can you look at this and
see what's up?  I have a couple of fixes to queue up from you already
and if you can get this one done in the next few days I'll included it
as well.

So I went back in the archives. Somebody else, Thomas Gleixner I
believe, was already working on the ndfc port. So, since I *had* to
have a working ndfc, I put a local copy in my kernel and forgot about
it.

It looks like Thomas has dropped the port. I should probably resubmit
the patch with the arch/ppc code removed. However, I don't think this
would be for 2.6.27 timeframe since there will probably be discussion.
For one, my patch just gets it working... the warp-nand.c file shows
that it is not a complete solution ;) Second, there are some byte order
issues.

So could we just leave warp-nand.c "there but disabled" for now?

Cheers,
   Sean

For those who are interested, I have posted a patch for the ndfc
driver. This is the driver we use un-bogusitize the warp-nand ;)

I removed the CONFIG_PPC_MERGE defines, but other than that this is the
driver we use in production.

Stefan: This has the byte order reversing that I believe you where
going to look into wrt u-boot. This may need to be a config option.

Cheers,
   Sean