Posted this last week (10/21) but haven't seen any response.
Would you consider this for the next -mm? Also attached in
case kmail mangles the whitespace.
[PATCH] radeonfb: If no video memory, exit with error
Nothing good will happen if we try to ioremap and use a zero-sized
frame buffer. I observed this problem on an ia64 sx1000 box, where
the BIOS doesn't run the option ROM. If we try to continue, radeonfb
just gets hopelessly confused because the card isn't initialized
correctly.
Signed-off-by: Bjorn Helgaas <[email protected]>
===== drivers/video/aty/radeon_base.c 1.32 vs edited =====
--- 1.32/drivers/video/aty/radeon_base.c 2004-10-19 03:40:34 -06:00
+++ edited/drivers/video/aty/radeon_base.c 2004-10-21 11:50:51 -06:00
@@ -2186,7 +2186,9 @@
rinfo->video_ram = 8192 * 1024;
break;
default:
- break;
+ printk (KERN_ERR "radeonfb: no video RAM reported\n");
+ ret = -ENXIO;
+ goto err_unmap_rom;
}
}
Andrew, looks ok.
On Tue, 2004-10-26 at 16:03 -0600, Bjorn Helgaas wrote:
> Posted this last week (10/21) but haven't seen any response.
> Would you consider this for the next -mm? Also attached in
> case kmail mangles the whitespace.
>
>
> [PATCH] radeonfb: If no video memory, exit with error
>
> Nothing good will happen if we try to ioremap and use a zero-sized
> frame buffer. I observed this problem on an ia64 sx1000 box, where
> the BIOS doesn't run the option ROM. If we try to continue, radeonfb
> just gets hopelessly confused because the card isn't initialized
> correctly.
>
> Signed-off-by: Bjorn Helgaas <[email protected]>
>
> ===== drivers/video/aty/radeon_base.c 1.32 vs edited =====
> --- 1.32/drivers/video/aty/radeon_base.c 2004-10-19 03:40:34 -06:00
> +++ edited/drivers/video/aty/radeon_base.c 2004-10-21 11:50:51 -06:00
> @@ -2186,7 +2186,9 @@
> rinfo->video_ram = 8192 * 1024;
> break;
> default:
> - break;
> + printk (KERN_ERR "radeonfb: no video RAM reported\n");
> + ret = -ENXIO;
> + goto err_unmap_rom;
> }
> }
>
--
Benjamin Herrenschmidt <[email protected]>
On Tue, 2004-10-26 at 16:03 -0600, Bjorn Helgaas wrote:
> Posted this last week (10/21) but haven't seen any response.
> Would you consider this for the next -mm? Also attached in
> case kmail mangles the whitespace.
I did tell Andrew it was OK this morning.
Ben.