Skip to contentSkip to conversation

Pci multi bridge bug

U-Boot-Users Philippe Simal 3 participants 3 messages Feb 5, 2004

lore
Share & tools
lore
#0 lore raw

Hello,

I'm running u-boot-1.0.0 on a IBM PPC440GP.
I have a custom made board with 2 pci-to-pci bridge Intel 21154BE and several devices behind it. (* PCI bus layout)
On boot-up, u-boot immediately recognizes the 2 bridges and 1 device on bus 0.
It also has no problem for enumirating all the devices on bus 1, but u-boot can't find any device on bus 2.
When I look at the bridge configuration using the 'pci 0 long' command, I see that the secondary bus address field of bridge 1 is set to 1. This should be 2. I also see that on bridge 1 but also on bridge 0 the subordinate bus address field always is filled with a 0.

First of all I already implemented the type1 configuration cycle bug fix at (..\drivers\Pci_indirect.c).

To work this kind of PCI configuration, I changed the following lines:
1. The problem with the secondary bus address is fixed by changing pci_hose_write_config_byte(hose, dev, PCI_SECONDARY_BUS, sub_bus + 1); to
pci_hose_write_config_byte(hose, dev, PCI_SECONDARY_BUS,
hose->current_busno);
in
(file: '..\drivers\Pci_auto.c'; function: 'static void
pciauto_prescan_setup_bridge())
2. The problem with the subordinate bus address is solved by swapping the code lines from

     pciauto_postscan_setup_bridge(hose, dev, sub_bus);
     sub_bus = hose->current_busno;
     to
     sub_bus = hose->current_busno;
     pciauto_postscan_setup_bridge(hose, dev, sub_bus);
    (file: '..\drivers\Pci_auto.c';   function: int pciauto_config_device())

I'm not sure if this fix is fully failed proof but it works fine for this design.

I hope my findings can be a contribution to the u-boot project.

* PCI bus layout.

BUS0

         |
         |
         |     /-------\  BUS1
         |---|            |--------       -> 7 PCI devices
         |     \-------/
         |   Bridge 0
         |
         |
         |     /-------\  BUS2
         |---|            |--------     -> 7 PCI devices
         |     \-------/
         |   Bridge 1
         |
         |
         |--- 1 PCI device

Greetings,
Philippe

#1 lore raw
Quoted 3 lines · In message <001e01c3ebc6$e8e9e340$10c8c8c8@pc005> you wrote:

First of all I already implemented the type1 configuration cycle bug fix at (..\drivers\Pci_indirect.c).

...

Quoted 1 line

I hope my findings can be a contribution to the u-boot project.

Can you please submit a proper patch?

Best regards,

Wolfgang Denk

Signature · 3 lines
See us @ Embedded World, Nuremberg, Feb 17 - 19,  Hall 12.0 Booth 440
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd at denx.de
Overflow on /dev/null, please empty the bit bucket.
#2 lore raw
In reply to Philippe Simal

Philippe:

I sent in a patch for this a couple of weeks ago (2 weeks?).

Search the archives for:
sibling p2p bridge

I'm sure its in Wolfgang's Q for cvs...

-travis

Quoted 75 lines · Philippe Simal

Hello,

I'm running u-boot-1.0.0 on a IBM PPC440GP.
I have a custom made board with 2 pci-to-pci bridge Intel 21154BE and several devices behind it. (* PCI bus layout)
On boot-up, u-boot immediately recognizes the 2 bridges and 1 device on bus 0.
It also has no problem for enumirating all the devices on bus 1, but u-boot can't find any device on bus 2.
When I look at the bridge configuration using the 'pci 0 long' command, I see that the secondary bus address field of bridge 1 is set to 1. This should be 2. I also see that on bridge 1 but also on bridge 0 the subordinate bus address field always is filled with a 0.

First of all I already implemented the type1 configuration cycle bug fix at (..\drivers\Pci_indirect.c).

To work this kind of PCI configuration, I changed the following lines:
1. The problem with the secondary bus address is fixed by changing pci_hose_write_config_byte(hose, dev, PCI_SECONDARY_BUS, sub_bus + 1); to
pci_hose_write_config_byte(hose, dev, PCI_SECONDARY_BUS,
hose->current_busno);
in
(file: '..\drivers\Pci_auto.c'; function: 'static void
pciauto_prescan_setup_bridge())
2. The problem with the subordinate bus address is solved by swapping the code lines from

     pciauto_postscan_setup_bridge(hose, dev, sub_bus);
     sub_bus = hose->current_busno;
     to
     sub_bus = hose->current_busno;
     pciauto_postscan_setup_bridge(hose, dev, sub_bus);
    (file: '..\drivers\Pci_auto.c';   function: int pciauto_config_device())

I'm not sure if this fix is fully failed proof but it works fine for this design.

I hope my findings can be a contribution to the u-boot project.

* PCI bus layout.

   BUS0
         |
         |
         |     /-------\  BUS1
         |---|            |--------       -> 7 PCI devices
         |     \-------/
         |   Bridge 0
         |
         |
         |     /-------\  BUS2
         |---|            |--------     -> 7 PCI devices
         |     \-------/
         |   Bridge 1
         |
         |
         |--- 1 PCI device

Greetings,
Philippe

-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
U-Boot-Users mailing list
U-Boot-Users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/u-boot-users