[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 09 Sep 2014 08:20:22 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: John Fastabend <john.fastabend@...il.com>
Cc: xiyou.wangcong@...il.com, davem@...emloft.net, jhs@...atatu.com,
netdev@...r.kernel.org, paulmck@...ux.vnet.ibm.com,
brouer@...hat.com
Subject: Re: [net-next PATCH v3 02/15] net: rcu-ify tcf_proto
On Tue, 2014-09-09 at 08:15 -0700, John Fastabend wrote:
>
> You mean RCU_INIT_POINTER() correct?
Right, I got this correct for other reviews ;)
>
> And to be clear which special case in rcupdate.h:999 applies here,
>
> * 1. This use of RCU_INIT_POINTER() is NULLing out the pointer -or-
> * 2. The caller has taken whatever steps are required to prevent
> * RCU readers from concurrently accessing this pointer -or-
> * 3. The referenced data structure has already been exposed to
> * readers either at compile time or via rcu_assign_pointer() -and-
> * a. You have not made -any- reader-visible changes to
> * this structure since then -or-
> * b. It is OK for readers accessing this structure from its
> * new location to see the old state of the structure. (For
> * example, the changes were to statistical counters or to
> * other state where exact synchronization is not required.)
>
In fact, everytime we remove an item from rcu protected list, its fine
to use RCU_INIT_POINTER
RCU_INIT_POINTER(*pprev, item->next);
Because we do not touch chain starting from item->next
>
> Its only NULLing out the pointer in special cases (case 1), with the
> qdisc lock dropped like in the last patch for ingress qdisc we may have
> concurrent readers (case 2) doing classification which will be walking
> the filter list, and with the tcf_ops change() its not clear to me that
> we can guarantee the structure has not been updated as stated in (case
> 3a).
>
> So the case here is 3b we are just updating the list structure to skip
> an entry and any old state is correct.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists