v3 for alix: add ide support

[ron minnich]

This code now lets filo find IDE.

but it locks up :-(

oh well. attached.

ron

[Corey Osgood]

On Feb 7, 2008 1:15 AM, ron minnich  wrote:
> This code now lets filo find IDE.
>
> but it locks up :-(
>
> oh well. attached.
>
> ron
> printk(BIOS_DEBUG, "cs5536_ide: %s\n", __func__);

not __FUNCTION__ any more? I'm not sure what the current standards are, but
either way:
Acked-by: Corey Osgood

[Peter Stuge]

On Thu, Feb 07, 2008 at 01:21:15AM -0500, Corey Osgood wrote:
> > printk(BIOS_DEBUG, "cs5536_ide: %s\n", __func__);
>
> not __FUNCTION__ any more? I'm not sure what the current standards
> are,

__func__ is good.

http://gcc.gnu.org/onlinedocs/gcc-4.1.2/gcc/Function-Names.html

//Peter

[ron minnich]

On Feb 6, 2008 10:21 PM, Corey Osgood  wrote:
> not __FUNCTION__ any more? I'm not sure what the current standards are,

__FUNCTION__ is gcc and acts kind of like a macro. __func__ is c99 and
acts like it is a const variable. You can start looking around at
postings about why __func__ is supposed to be wonderful; either way,
__FUNCTION__ is supposed to be deprecated. So I am trying to catch up
with the 20th century, now that I am in the 21st :-)

__FUNCTION__ is deprecated in newer linux kernel code, but will not be
replaced in older code, according to
http://lkml.org/lkml/2007/5/2/470. We can probably follow the same
rule.

thanks

ron

[ron minnich]

Commited rev 575

[Marc Jones]

ron minnich wrote:
> This code now lets filo find IDE.
>
> but it locks up :-(
>
> oh well. attached.
>
> ron

I preferred the v2 way of doing this. If the AMD IDE device was found it 
got initialized. I think it needs a DTS and I think other southbridges 
with multiple PCI headers will need this too.

Back to the 5536, You can have either NAND or IDE PCI header so the 
enable_ide flag is redundant. If there is a PCI IDE card all of it's 
memory and I/O will be set in the PCI BAR. They can't use the legacy 
range so there is no real reason to not init the AMD IDE. If you wanted 
to make the AMD IDE go away it would be better to remove it via 
unwanted_vpci.

Marc

[ron minnich]

On Feb 7, 2008 11:16 AM, Marc Jones  wrote:
> I preferred the v2 way of doing this. If the AMD IDE device was found it
> got initialized. I think it needs a DTS and I think other southbridges
> with multiple PCI headers will need this too.

So it's more like the northbridge part as well.

OK, I'll think on this.

thanks

ron