[PATCH] Do not skip PCI Express to PCI bridge when scanning OF node

[PATCH] Do not skip PCI Express to PCI bridge when scanning OF node

[Jon Loeliger]

From: Haiying Wang <haiying.wang@freescale.com>

Signed-off-by: Zhang Wei <wei.zhang@freescale.com>
Signed-off-by: Haiying Wang <haiying.wang@freescale.com>
Signed-off-by: Jon Loeliger <jdl@freescale.com>
---
 arch/powerpc/kernel/prom_parse.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/powerpc/kernel/prom_parse.c b/arch/powerpc/kernel/prom_parse.c
index ea6fd55..0809977 100644
--- a/arch/powerpc/kernel/prom_parse.c
+++ b/arch/powerpc/kernel/prom_parse.c
@@ -267,7 +267,7 @@ int of_irq_map_pci(struct pci_dev *pdev, struct of_irq *out_irq)
 		ppdev = pdev->bus->self;
 
 		/* Ouch, it's a host bridge... */
-		if (ppdev == NULL) {
+		if (ppdev == NULL || ppdev->devfn == 0) {
 #ifdef CONFIG_PPC64
 			ppnode = pci_bus_to_OF_node(pdev->bus);
 #else
-- 
1.5.0.1.213.g509b

Re: [PATCH] Do not skip PCI Express to PCI bridge when scanning OF node

[Kumar Gala]

On Mar 7, 2007, at 12:35 PM, Jon Loeliger wrote:

> From: Haiying Wang <haiying.wang@freescale.com>
>
> Signed-off-by: Zhang Wei <wei.zhang@freescale.com>
> Signed-off-by: Haiying Wang <haiying.wang@freescale.com>
> Signed-off-by: Jon Loeliger <jdl@freescale.com>

The signed-off-by's are out of order if this patch is really from  
Haiyang.

Also, I think a better description about why this was needed should  
be provided.  I'm guessing devices behind a PCI Ex to PCI bridge end  
up pdev->bus->self set?  Seems like a bug in the generic pci code.

- k

Re: [PATCH] Do not skip PCI Express to PCI bridge when scanning OF node

[Benjamin Herrenschmidt]

On Wed, 2007-03-07 at 12:35 -0600, Jon Loeliger wrote:
> From: Haiying Wang <haiying.wang@freescale.com>
> 
> Signed-off-by: Zhang Wei <wei.zhang@freescale.com>
> Signed-off-by: Haiying Wang <haiying.wang@freescale.com>
> Signed-off-by: Jon Loeliger <jdl@freescale.com>
> ---
>  arch/powerpc/kernel/prom_parse.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/arch/powerpc/kernel/prom_parse.c b/arch/powerpc/kernel/prom_parse.c
> index ea6fd55..0809977 100644
> --- a/arch/powerpc/kernel/prom_parse.c
> +++ b/arch/powerpc/kernel/prom_parse.c
> @@ -267,7 +267,7 @@ int of_irq_map_pci(struct pci_dev *pdev, struct of_irq *out_irq)
>  		ppdev = pdev->bus->self;
>  
>  		/* Ouch, it's a host bridge... */
> -		if (ppdev == NULL) {
> +		if (ppdev == NULL || ppdev->devfn == 0) {

No way. There is nothing "special" per spec with devfn == 0.

What exactly is that supposed to do ?

Ben.