[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CALnjE+pg1ecFjisxnHyrMsBCyZ6daF=wCH+FxEc8XtCqAwyo6g@mail.gmail.com>
Date: Mon, 22 Apr 2013 16:49:05 -0700
From: Pravin Shelar <pshelar@...ira.com>
To: David Miller <davem@...emloft.net>
Cc: netdev@...r.kernel.org, dev@...nvswitch.org, jesse@...ira.com
Subject: Re: [PATCH net-next v2 1/2] genl: Allow concurrent genl callbacks.
On Mon, Apr 22, 2013 at 3:55 PM, David Miller <davem@...emloft.net> wrote:
> From: Pravin Shelar <pshelar@...ira.com>
> Date: Mon, 22 Apr 2013 13:41:41 -0700
>
>> Therefore scatchpad in SKB is protected and parallel netlink dump operations
>> can go only on different netlink sockets.
>
> What about configuration changes running in parallel with the dump?
>
> You must synchronize with those too, otherwise protecting the
> intra-dump state is meaningless since the underlying objects
> can change underneath the dump.
That is synchronized using rw-sem.
This is how locking looks like:
genl_dump() (if lockless = true)
SEM_READ_LOCK()
Per_sock_mutex()
-->netlink_dump()
Per_sock_unmutex()
SEM_READ_UNLOCK()
---
genl_conf_changes:
SEM_WRITE_LOCK()
GENL_LOCK()
conf_changes()
GENL_UNLOCK()
SEM_WRITE_UNLOCK()
---
Now I think I shld call this parallel operations rather than lockless
operations, since I do take read lock for entire genl-ops
.
Let me send updated patch.
--
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