[ELDK] ppc boot fails at PCI detection.

[Larry Johnson]

Wolfgang Denk wrote:
> Dear Renjith,
> 
> in message <a8474bd70903081839j3520e7e8y7dc506e73a6a0456 at mail.gmail.com> you wrote:
>>       [ Sequoia - AMCC PPC440EPx Evaluation Board, Rev. F, PCI=33 MHz  ].
> 
> OK.
> 
>>       U-Boot 1.2.0-gc0c292b2 (Jun  5 2007 - 07:16:12)
> 
> That's very old...
> 
>> I have also tried , latest u-boot from git (U-Boot 2009.03-rc1-00000-g00cc55
>> ) with a nand-boot option(u-boot-nand.bin) and found thing are same.
> 
> ... so the U-Boot version is probably not the problem.
> 
>> "Some times' hanging at
>>         PCI:   Bus Dev VenId DevId Class Int
>> or at
>>        Uncompressing Kernel Image ... OK
>> while Linux boot Up.
> 
> You did not mention if the board reliably comes up without PCI cards
> inserted?
> 
>> I used a PCI card (DVR-video capture card v2.2). It got detected and not
>> hanging.
> 
> You mean, with this PCI card there is never any problem?
> 
>> If I use a PCI-SATA (PCI v2.3) card (SIL3114, SATA1 from Silicon Image), it
>> has more frequent problem.
>> So I am also doubting, even it could be a Hardware trouble (at the interface
>> level).
> 
> Did you test this SIL3114 card in another 3.3V only system?
> 
>> To nullify the SATA BIOS influence , I tried Removing the Device ROM chip
>> and tried , but result same.
> 
> An x86 BIOS usually does not do much on a PowerPC system ;-)
> 
>> (2)
>> Also at this occasion, I would like to know, is there any special support
>> needed to be done at BootLoader , so that it will support PCI-SATA card.
>> Because, In normal desktop , we needed to enable SATA support specially in
>> BIOS, to detect the card and its function.
> 
> Do you want to use SATA in U-Boot?
> 
> Best regards,
> 
> Wolfgang Denk
> 

Hi Renjith and Wolfgang,

I found the thread starting at
"http://ozlabs.org/pipermail/linuxppc-dev/2009-January/067738.html"
("Broken PCI on Sequoia") after I had tracked down the same problem on
the Korat PPC440EPx card.  I don't see how PCI on the Sequoia is going
to work until this bug is fixed.

BTW, did you have a chance to try the temporary patch I posted last
week?  That fixes the problem for me.  Using the previous version of
"ppc4xx_pci.c" also works.

Best regards,
Larry