>This is not recalled from the network layer, this is the raw irq handler.
This patch is wrong agreed. (excuse me)
The reason is that it's enforced by the design that there can't be bh
handlers run in the middle of irq handlers... (I just thought about it
some days ago, but now I was searching for a fix and I forgot it).
So I return to the "works but I don't know why" mode. You can find my
patch that is reported working in the first email of this thread in
attachment.
But I think that if nicholas want he could try also this start_bh_atomic()
bogus patch, maybe the bug is that some bh handler is run by mistake in
irq handler?
I discovered that the cpu lockups exactly at the first spin_lock_irqsave
of ei_start_xmit, and this made me to think (again) that the ei_start_xmit
happened in the middle of ei_interrupt.
Andrea Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/