[coreboot] more on the cs5536 and 2.6.25

[Jordan Crouse]

On 05/05/08 23:03 -0700, ron minnich wrote:
> I think 2.6.25 is hanging on the probe of ide port1.
> 
> This is 2.6.24
> 
> AMD5536: IDE controller (0x1022:0x209a rev 0x01) at  PCI slot 0000:00:0f.2
> AMD5536: not 100% native mode: will probe irqs later
> AMD5536: 0000:00:0f.2 (rev 01) UDMA100 controller
>     ide0: BM-DMA at 0x1ca0-0x1ca7, BIOS settings: hda:pio, hdb:pio
> AMD5536: IDE port disabled
> hda: CF 1GB, ATA DISK drive
> hda: MWDMA2 mode selected
> ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
> hda: max request size: 128KiB
> hda: 2030112 sectors (1039 MB) w/1KiB Cache, CHS=2014/16/63
>  hda: hda1
> 
> and 2.6.25
> 
> AMD5536: 0000:00:0f.2 (rev 01) UDMA100 controller
> AMD5536: IDE controller (0x1022:0x209a rev 0x01) at  PCI slot 0000:00:0f.2
> AMD5536: not 100% native mode: will probe irqs later
> AMD5536: IDE port disabled
>     ide0: BM-DMA at 0x1ca0-0x1ca7, BIOS settings: hda:PIO, hdb:PIO
> 
> and on one boot it did this:
> Probing IDE interface ide1...
> 
> I think it's going after ide1, even though it is not there. Not sure
> yet. Something is pretty wrong here, and it broke from 2.6.24 to
> 2.6.25.

Of course its looking for ide1, but as soon as it sees the missing bit
in the PCI config space its going to give up.

I haven't been able to reproduce hangs with the legacy IDE driver running
with v2 and 2.6.25 on norwich - everything is running fine:

AMD5536: 0000:00:0f.2 (rev 01) UDMA100 controller                          
AMD5536: IDE controller (0x1022:0x209a rev 0x01) at  PCI slot 0000:00:0f.2 
AMD5536: not 100% native mode: will probe irqs later                       
AMD5536: IDE port disabled                                                 
    ide0: BM-DMA at 0x1ca0-0x1ca7, BIOS settings: hda:PIO, hdb:PIO              
Probing IDE interface ide0...                                              
hda: TOSHIBA THNCF512MDG, CFA DISK drive                                        
hda: host max PIO5 wanted PIO255(auto-tune) selected PIO4                  
hda: host max PIO5 wanted PIO255(auto-tune) selected PIO4                  
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14                                        
Probing IDE interface ide1...  

pata_cs5536 is not healthy, however:

PCI: Unable to reserve I/O region #1:8 at 1f0 for device 0000:00:0f.2
pata_cs5536 0000:00:0f.2: failed to request/iomap BARs for port 0 (errno=-16)
pata_cs5536 0000:00:0f.2: no available native port

Again, this is with v2, so your mileage will vary, but as far as the amd74xx
driver is concerned, I don't think we're seeing a huge kernel regression.
I'm going to take some swings at the pata_cs5536 driver and see if we can't
make it better.

Jordan

PS:  In my config, at least, following the IDE setup is the USB setup - 
possibly you are hanging _there_ and just not seeing any intermediate 
kernel messages.

-- 
Jordan Crouse
Systems Software Development Engineer 
Advanced Micro Devices, Inc.