Re: [coreboot] pci device memory allocation problem

> Doing it this way is better: first assign align value to be at least
> of min_align then compare align with bridge->align to see if
> bridge->align needs updated.
> ... ...
>    while((dev = largest_resource(bus, &resource, type_mask, type))) {
>    	resource_t size;
>    	/* Do NOT I repeat do not ignore resources which have zero size.
>    	 * If they need to be ignored dev->read_resources should not even
>    	 * return them.   Some resources must be set even when they have
>    	 * no size.  PCI bridge resources are a good example of this.
>    	 */
>    	/* Make certain we are dealing with a good minimum size */
>    	size = resource->size;
>    	align = resource->align;
>    	if (align < min_align) {
>    		align = min_align;
>    	}
> /* Propogate the resource alignment to the bridge register  */
> if (align > bridge->align) {
> 	bridge->align = align;
> }
>
>
> if (resource->flags & IORESOURCE_FIXED) {
> 	continue;
> }
>
> ... ...

I noticed this problem, if certian devices are disabled (like onboard vga for example). The other enabled devices try to use 0xfec00000 for memory allocation, causing APIC to fail. I think the above code looks good :-)
Have you done up a patch?

pci device memory allocation problem

[aaron lwe]

> I noticed this problem, if certian devices are disabled (like onboard vga for example). The other enabled devices try to use 0xfec00000 for memory allocation, causing APIC to fail. I think the above code looks good :-)
>  Have you done up a patch?

yes, I just made a patch:

Index: src/devices/device.c
===================================================================
--- src/devices/device.c        (revision 3254)
+++ src/devices/device.c        (working copy)
@@ -305,11 +305,6 @@
                 * return them.   Some resources must be set even when they have
                 * no size.  PCI bridge resources are a good example of this.
                 */
-               /* Propogate the resource alignment to the bridge register  */
-               if (resource->align > bridge->align) {
-                       bridge->align = resource->align;
-               }
-
                /* Make certain we are dealing with a good minimum size */
                size = resource->size;
                align = resource->align;
@@ -317,6 +312,11 @@
                        align = min_align;
                }

+               /* Propogate the resource alignment to the bridge register  */
+               if (align > bridge->align) {
+                       bridge->align = align;
+               }
+
                if (resource->flags & IORESOURCE_FIXED) {
                        continue;
                }

First time to send an attachment, not sure if it works, so I pasted
the code in this mail ;-)

pci device memory allocation problem

[Paul Menzel]

Dear Aaron,

Am Mittwoch, den 23.04.2008, 22:07 +0800 schrieb aaron lwe:
> First time to send an attachment, not sure if it works, so I pasted
> the code in this mail ;-)

I received your attachment and was able to open it.

Although I cannot test or verify your patch, I think you have to add a
Signed-Off-By line [1]. Otherwise your patch cannot be acked and then
commited.

Thanks and according to the comments I read great investigation and
solution,

Paul

[1] http://www.coreboot.org/Development_Guidelines#Sign-off_Procedure

pci device memory allocation problem

[Joe]

You just need to add a Signed-Off-By as stated in the development
guidelines. In the mean time I will test it and if all goes well you'll get
my Acked-by, and then it can be commited.

Thanks,
Joseph Smith
Set-Top-Linux
www.settoplinux.org
> -----Original Message-----
> From: [e-mail] [mailto:[e-mail]]
> On Behalf Of aaron lwe
> Sent: Wednesday, April 23, 2008 10:07 AM
> To: [e-mail]
> Cc: [e-mail]; [e-mail]
> Subject: [coreboot] pci device memory allocation problem
> > I noticed this problem, if certian devices are disabled (like onboard
>
> vga for example). The other enabled devices try to use 0xfec00000 for
> memory allocation, causing APIC to fail. I think the above code looks good
> :-)
> > Have you done up a patch?
>
> yes, I just made a patch:
> Index: src/devices/device.c
>
> --- src/devices/device.c        (revision 3254)
> +++ src/devices/device.c        (working copy)
> @@ -305,11 +305,6 @@
>                  * return them.   Some resources must be set even when
> they have
>                  * no size.  PCI bridge resources are a good example of
> this.
>                  */
>           /* Propogate the resource alignment to the bridge register
>
> */
>           if (resource->align > bridge->align) {
>                   bridge->align = resource->align;
>           }
>           /* Make certain we are dealing with a good minimum size */
>           size = resource->size;
>           align = resource->align;
>
> @@ -317,6 +312,11 @@
>                         align = min_align;
>                 }
>           /* Propogate the resource alignment to the bridge register
>
> */
>           if (align > bridge->align) {
>                   bridge->align = align;
>           }
>           if (resource->flags & IORESOURCE_FIXED) {
>                   continue;
>           }
>
> First time to send an attachment, not sure if it works, so I pasted
> the code in this mail ;-)

pci device memory allocation problem

[aaron lwe]

> You just need to add a Signed-Off-By as stated in the development
>  guidelines. In the mean time I will test it and if all goes well you'll get
>  my Acked-by, and then it can be commited.

Oh, I see.

Signed-off-by: Aaron Lwe

Thanks.