Re: [RFC][PATCH 2/9] deadlock prevention core
From: Peter Zijlstra <hidden>
Date: 2006-08-09 16:24:26
Also in:
linux-mm, lkml
From: Peter Zijlstra <hidden>
Date: 2006-08-09 16:24:26
Also in:
linux-mm, lkml
On Wed, 2006-08-09 at 18:18 +0200, Thomas Graf wrote:
* Peter Zijlstra [off-list ref] 2006-08-09 16:07quoted
I think Daniel was thinking of adding struct net_device * sk_buff::alloc_dev, I know I was after reading the first few mails. However if adding a field there is strict no-no.... /me takes a look at struct sk_buff Hmm, what does sk_buff::input_dev do? That seems to store the initial device?No, skb->input_dev is used when redirecting packets around in the stack and may change. Even if it would keep its value the reference to the netdevice is not valid anymore when you free the skb as the skb was queued and the refcnt acquired in __netifx_rx_schedule() has been released again thus making it possible for the netdevice to disappear.
Bah, tricky stuff that. disregards this part from -v2 then :-(