[LinuxBIOS] Re: dead loop in compute_allocate_resource

[Steve Magnani]

I ran into something similar over the weekend, with a s2735 variant (Intel
7501 Xeon development kit) I'm working on. I had assumed it was something
I did wrong in Config.lb - I set the USB and IDE devices "on" even though
my southbridge LPC setup disables them. When I set the USB and IDE devices
"off" in Config.lb my problem went away. Not a solution, I know, but it
might help track down the problem.

Steve

[Steve Magnani]

Might be the keyboard or floppy, too - I turned those off as well.

Steve
-----Original Message-----
From: "Steve Magnani" 
To: [e-mail], [e-mail], [e-mail]
Date: Tue, 05 Jul 2005 18:04:38 -0500
Subject: Re: dead loop in compute_allocate_resource
> I ran into something similar over the weekend, with a s2735 variant
> (Intel
> 7501 Xeon development kit) I'm working on. I had assumed it was
> something
> I did wrong in Config.lb - I set the USB and IDE devices "on" even
> though
> my southbridge LPC setup disables them. When I set the USB and IDE
> devices
> "off" in Config.lb my problem went away. Not a solution, I know, but it
> might help track down the problem.
>
> Steve

[yhlu]

I enable enable com1 get it through. ....

YH

PCI: 00:1f.2 20 *  [0x000030a0 - 0x000030af] io
        1: PNP: 002e.2 60 *  [0x000003f8 - 0x000003ff] io
        1: PCI: 00:1f.2 10 *  [0x000030b0 - 0x000030b7] io
        2: PCI: 00:1f.2 10 *  [0x000030b0 - 0x000030b7] io

On 7/5/05, Steve Magnani  wrote:
> Might be the keyboard or floppy, too - I turned those off as well.
>
> Steve
> -----Original Message-----
> From: "Steve Magnani" 
> To: [e-mail], [e-mail], [e-mail]
> Date: Tue, 05 Jul 2005 18:04:38 -0500
> Subject: Re: dead loop in compute_allocate_resource
> > I ran into something similar over the weekend, with a s2735 variant
> > (Intel
> > 7501 Xeon development kit) I'm working on. I had assumed it was
> > something
> > I did wrong in Config.lb - I set the USB and IDE devices "on" even
> > though
> > my southbridge LPC setup disables them. When I set the USB and IDE
> > devices
> > "off" in Config.lb my problem went away. Not a solution, I know, but it
> > might help track down the problem.
> >
> > Steve

[yhlu]

solution will be add
if (resource->flags & IORESOURCE_FIXED)  return 
in pick_largest_resource.

YH

--- device.c.orig       2005-06-22 12:54:44.000000000 -0700
+++ device.c    2005-07-05 19:32:20.000000000 -0700
@@ -180,6 +180,7 @@
                state->seen_last = 1;
                return;
        }
+       if (resource->flags & IORESOURCE_FIXED ) return; //skip it 
        if (last && (
                    (last->align < resource->align) ||
                    ((last->align == resource->align) &&

yhlu  writes:
> solution will be add
> if (resource->flags & IORESOURCE_FIXED)  return 
> in pick_largest_resource.

This solution seems suspicious.  I would expect this to have
triggered before if a proper fix was as simple as that.

Eric

[yhlu]

actually you are using

if (resource->flags & IORESOURCE_FIXED) {
         continue;
}

in the main loop

YH

On 7/5/05, Eric W. Biederman  wrote:
> yhlu  writes:
> > solution will be add
> > if (resource->flags & IORESOURCE_FIXED)  return
> > in pick_largest_resource.
>
> This solution seems suspicious.  I would expect this to have
> triggered before if a proper fix was as simple as that.
>
> Eric