On 02/14/2013 01:05 PM, Manuel Huber wrote:
> Hello!
>
> We use an AMD FX-61000 CPU and we are having trouble using kernel versions
> 2.6.38.8 and 3.5.7 (newest) and some more... The only version that
> worked is 2.6.35.9 (so far). But let's focus on 3.5.7:
>
> I finally managed to get error messages from the 3.5.7 kernel (Thanks to
> Philippe Gerum, who was so kind to direct me to the mailing list and also
> gave me some hints about debugging,... ). I've seen that there is a
> kernel BUG on line 139 and 170. Maybe thats the reason for the problem?
> I included the kernel config and the full log from startup. Of course it can
> very well be a wrong configuration...
>
> However, it seems like I could avoid the problemes with the 2.6.35.9
> kernel (It was related to RTnet + rt8169 sometimes getting stuck;
> it doesn't happen on the 3.5.7 kernel, as far as I can tell
> [tested on an AMD Athlon 64 X2]; now I'm using the rt8139C card which
> works like a charm), so I just wanted to submit the problem,
> maybe it helps to further improve this project...
>
> Of coure I would love to use the current version and I can test
> configurations and boot the system with different options if that
> helps with finding the error...
>
> Best regards, Manuel
> -------------- next part --------------
[ 1.201759] kernel BUG at arch/x86/kernel/ipipe.c:589!
Could you try the following patch?
diff --git a/arch/x86/kernel/ipipe.c b/arch/x86/kernel/ipipe.c
index f833663..6d0fd72 100644
--- a/arch/x86/kernel/ipipe.c
+++ b/arch/x86/kernel/ipipe.c
@@ -586,6 +586,8 @@ int __ipipe_handle_irq(struct pt_regs *regs)
if (likely(vector < 0)) {
irq = __this_cpu_read(vector_irq[~vector]);
+ if (irq < 0)
+ printk("invalid irq for vector 0x%08x\n", vector);
BUG_ON(irq < 0);
} else { /* Software-generated. */
irq = vector;
--
Gilles.
_______________________________________________
Xenomai mailing list
[email protected]
http://www.xenomai.org/mailman/listinfo/xenomai
