Re: xfrm_state locking regression...
From: David Miller <davem@davemloft.net>
Date: 2008-09-09 00:21:03
From: David Miller <davem@davemloft.net>
Date: 2008-09-09 00:21:03
From: Herbert Xu <herbert@gondor.apana.org.au> Date: Tue, 9 Sep 2008 10:18:28 +1000
On Mon, Sep 08, 2008 at 05:09:07PM -0700, David Miller wrote:quoted
What happens to all of the entries on the local gc_list which you will be skipping when you break out of that last loop? It looks like they will never get processed.I made gc_list static, so they'll be processed when the GC task runs again. Also whenever completed is incremented we schedule the GC task.
Thanks, I didn't catch that "static" bit :-)