[parisc-linux-cvs] Re: 2.6.10-rc1-pa4 mercury_cfg_ops and MMIO address routing

[Grant Grundler]

On Sat, Oct 30, 2004 at 09:07:38PM -0600, Grant Grundler wrote:
> Log message:
> 	2.6.10-rc1-pa4 - fix MMIO address listed in /proc/iomem
> 	
> 	o LBA now reports proper addresses routed down from SBA.
> 	Matching SBA code was accidentally committed ~4 weeks ago.
> 	
> 	o add mercury_cfg_ops. This fixes the tg3 init delay problem.
> 	PDC took ~20ms per PCI cfg write. tg3 performs ~1600 writes (~30 seconds
> 	per NIC) during init. Alternative is the relocate PDC to host ram.
> 	But that would burn ~4MB of RAM. This is cheaper.
> 	
> 	o add debug code to iosapic
> 	o add PCI_F_EXTEND to asm/pci.h - replaces private definitions
> 
> Patches:
> http://cvs.parisc-linux.org/linux-2.6/Makefile.diff?cvsroot=&r1=1.285&r2=1.286
> http://cvs.parisc-linux.org/linux-2.6/drivers/parisc/iosapic.c.diff?cvsroot=&r1=1.8&r2=1.9
> http://cvs.parisc-linux.org/linux-2.6/drivers/parisc/lba_pci.c.diff?cvsroot=&r1=1.16&r2=1.17
> http://cvs.parisc-linux.org/linux-2.6/drivers/parisc/sba_iommu.c.diff?cvsroot=&r1=1.19&r2=1.20
> http://cvs.parisc-linux.org/linux-2.6/include/asm-parisc/pci.h.diff?cvsroot=&r1=1.10&r2=1.11

Oh...I forgot about this. Nice. :^)
Thanks Randolph!

It's worth reading the comment/code I added to lba_pci.c about
MMIO address routing and how Sprockets firmware doesn't set it
up properly. Other firmware might be "cheating" the same way.

BTW, rp3440 is HPMCing on serial port initialization and it's
not obvious to me why. IRQ Routing is messed up too - so I'm
not sure what's been broken with recent merges. I'll try to fix
it this weekend though.

grant