Re: [lock validator] drivers/net/8139too.c: deadlock?
From: Jeff Garzik
Date: Thu Jan 26 2006 - 21:57:00 EST
Herbert Xu wrote:
You've got it.
rx_poll => rtl8139_rx => netif_receive_skb => ... => tcp_v4_rcv
In fact once we're at netif_receive_skb it's easy to see how we'll grab
xmit_lock again.
Prescription: Move TX timeout handling into a work queue.
This was my recommendation as well.
Using a work queue not only solves this locking problem, but it also
enables the possibility of sleeping, a common operation during the
hardware resets that often occur during ->tx_timeout()
Anyone who wants to make this change in a driver, don't forget that this
could race with other operations, notably dev->stop() or suspend.
Jeff
-
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/