AMCC 440SP PCI Issue

I am having an issue with my PCI configuration on a
custom board based closely off of the AMCC Luan board
(440SP processor).  When I type "pci 1" or "pci 2"
from the U-Boot prompt, I get a bunch of
"pci_bus_to_hose() failed" messages ("pci 0" returns
fine).  What could cause this.  I pretty much kept the
PCI configuration settings (mmap, config.h, board.c)
from the Luan port.  I checked the bootstrap settings
and all of the settings are identical.  Does anyone
have any idea what could cause this?  I looked at the
hoses structure, and it's empty.

Thanks,
   Jeff Stevens

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: (no subject)

I made the change that Marc suggested and still no luck. If both dimms
are in use the system hangs and if only the second Dimm slot is in use I
get:

U-Boot 1.1.4 (Sep 30 2006 - 10:08:43)

CPU:   AMCC PowerPC 440GX Rev. F at 800 MHz (PLB=160, OPB=80, EBC=40
MHz)
       I2C boot EEPROM enabled
       Internal PCI arbiter enabled
       32 kB I-Cache 32 kB D-Cache
Board: Ocotea - AMCC PPC440GX Evaluation Board
I2C:   ready
DRAM:  ERROR: Cannot determine a common read delay.
### ERROR ### Please RESET the board ###

Any Ideals?

 Gerald

-----Original Message-----
From: Gerald Jackson 
Sent: Saturday, September 30, 2006 10:13 AM
To: Marc.Howard <at> KLA-Tencor.com; sr <at> denx.de;
u-boot-users <at> lists.sourceforge.net
Subject: RE: [U-Boot-Users] (no subject)

Here is the error I get if I use the second DIMMS

U-Boot 1.1.4 (Jul 18 2006 - 07:36:20)

Re: Linux Hangs After Uncompressing Kernel

wenwen ha scritto:
> i download a 6M Byte kernel(28M bytes uncompressed) to
SDRAM(0x4000000), it hangs after uncompressing kernel.
>
> when i use standalone downlod the kernel, i found the text section
which offset is 8M bytes is not the same as the objdump file of the elf
file. is there any limited size about kernel ?
> i used amcc ppc440 sequoia board

Seems definitely huge,
what are you doing exactly ?

--

-- 
Marco Cavallini
Koan s.a.s. - Bergamo - ITALIA
Embedded and Real-Time Software Engineering
 - Atmel Third Party Consultant
Tel. +39-(0)35-255.235 - Fax +39-178-223.9748
www.KoanSoftware.com   |    www.KaeilOS.com

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
U-Boot-Users mailing list

Re: Linux Hangs After Uncompressing Kernel

Hi Wenwen,

On Saturday 30 September 2006 18:33, wenwen wrote:
> i download a 6M Byte kernel(28M bytes uncompressed) to SDRAM(0x4000000), it
> hangs after uncompressing kernel.

I have never seen such a big kernel image. I can't believe that it should be 
possible to configure such a huge kernel, even with all options enabled. You 
must be doing something wrong here.

> when i use standalone downlod the kernel, i found the text section which
> offset is 8M bytes is not the same as the objdump file of the elf file. is
> there any limited size about kernel ? i used amcc ppc440 sequoia board

Where did you get the linux kernel? What version are you using?

Best regards,
Stefan

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: (no subject)

Hi Gerald,

On Sunday 01 October 2006 05:04, Gerald Jackson wrote:
> I made the change that Marc suggested and still no luck. If both dimms
> are in use the system hangs

Hangs where? Also upon SDRAM initialization?

> and if only the second Dimm slot is in use I 
> get:
>
> U-Boot 1.1.4 (Sep 30 2006 - 10:08:43)
>
> CPU:   AMCC PowerPC 440GX Rev. F at 800 MHz (PLB=160, OPB=80, EBC=40
> MHz)
>        I2C boot EEPROM enabled
>        Internal PCI arbiter enabled
>        32 kB I-Cache 32 kB D-Cache
> Board: Ocotea - AMCC PPC440GX Evaluation Board
> I2C:   ready
> DRAM:  ERROR: Cannot determine a common read delay.
> ### ERROR ### Please RESET the board ###
>
> Any Ideals?

Ideals? Not really. But I have an idea! 

How did you get the Board? Directly from AMCC? Did it never work at all?

Your PLB frequency seems strange to me (160MHz). Please try to configure your 

Re: AMCC 440SP PCI Issue

Hi Jeff,

On Sunday 01 October 2006 04:35, Jeff Stevens wrote:
> I am having an issue with my PCI configuration on a
> custom board based closely off of the AMCC Luan board
> (440SP processor).  When I type "pci 1" or "pci 2"
> from the U-Boot prompt, I get a bunch of
> "pci_bus_to_hose() failed" messages ("pci 0" returns
> fine).  What could cause this.

It seems to me that only the first PCI interface of the Luan is initialized in 
U-Boot. That should be the reason for those error messages.

> I pretty much kept the 
> PCI configuration settings (mmap, config.h, board.c)
> from the Luan port.  I checked the bootstrap settings
> and all of the settings are identical.  Does anyone
> have any idea what could cause this?  I looked at the
> hoses structure, and it's empty.

Which PCI busses are you using on your board? As target or host? Which are 
accessed by your Linux kernel?

Best regards,
Stefan

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash

Re: (no subject)

Stefan,

Yes the board works and Pibs works with both Dimms slot in use.
The board came from AMCC and it is the latest board with the 800MHz
440GX chip on it.

The system hangs upon SDRAM initialization, but only with both dimms
slots used.

I didn't even know this problem exist until another company that is
writing a device driver for me told me that the second DIMMS slot
doesn't work on there board. Sure enough it doesn't work on mines
either.

Gerald

-----Original Message-----
From: Stefan Roese [mailto:sr <at> denx.de] 
Sent: Sunday, October 01, 2006 5:23 AM
To: u-boot-users <at> lists.sourceforge.net
Cc: Gerald Jackson; Marc.Howard <at> kla-tencor.com
Subject: Re: [U-Boot-Users] (no subject)

Hi Gerald,

On Sunday 01 October 2006 05:04, Gerald Jackson wrote:
> I made the change that Marc suggested and still no luck. If both dimms
> are in use the system hangs

Re: AMCC 440SP PCI Issue

I am using all three pci busses are host interfaces:

PCIX0: PMC Slot
PCIX1: Fiber Channel Controller
PCIX2: USB Controller

When linux does boot, it sees all of these devices
just fine.  However, when it doesn't, it hangs after
it calls pciauto_bus_scan.

Thanks,
   Jeff

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

AMCC 440SP PCIX Register access with BDI2000?

Is there any way to access the PCIX Registers using
the BDI2000?  I accidentally set the CPU wait bit for
PCIX1, and now it won't boot.  If I could access the
PCIX registers (located and 0x91EC80044) I could clear
the wait bit, and start U-Boot.  However, it looks
like the BDI2000 can only access 32-bit addresses, so
when I have it dump 0x91ec80044 it dumps 0x1ec80044
instead.  Is there an indirect access register set?

Thanks,
   Jeff Stevens

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: AMCC 440SP PCIX Register access with BDI2000?

I can only give you half an answer.  The BDI2000 config file lets you
setup TLB entries.  For my 440GX board I have the following in the
[INIT] section of the config file:

WTLB    0xe0000095  0x1400001B

This maps the 32-bit virtual address 0xe0000000 to the 36-bit physical
address 0x140000000.  A similar incantation should map the PCIX
registers into the 32-bit address space, but I don't remember the format
of the fields in the command...

jeff

Jeff Stevens wrote:
> Is there any way to access the PCIX Registers using
> the BDI2000?  I accidentally set the CPU wait bit for
> PCIX1, and now it won't boot.  If I could access the
> PCIX registers (located and 0x91EC80044) I could clear
> the wait bit, and start U-Boot.  However, it looks
> like the BDI2000 can only access 32-bit addresses, so
> when I have it dump 0x91ec80044 it dumps 0x1ec80044
> instead.  Is there an indirect access register set?
> 
> Thanks,
>    Jeff Stevens
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around 
> http://mail.yahoo.com