Problem with dev_kfree_skb_any() in 2.6.0

[Benjamin Herrenschmidt]

I'm getting Ooops reports with sungem. The problem is that in the PM
code (and possibly other exceptional code path), it will clean the rings
at non interrupt time but with a spinlock_irq held.

it uses dev_kfree_skb_any() which does:

static inline void dev_kfree_skb_any(struct sk_buff *skb)
{
        if (in_irq())
                dev_kfree_skb_irq(skb);
        else
                dev_kfree_skb(skb);
}

However, in_irq() seem to only catch real IRQ time, so I end up calling
dev_kfree_skb(), which triggers the following BUG() in local_bh_enable()
        WARN_ON(irqs_disabled());

We should probably fix dev_kfree_skb_any() ? Still ugly imho though...

-        if (in_irq())
+        if (in_irq() || irqs_disabled())


Ben.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/