netdev - Re: yield() in netlink_broadcast

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Fri, 6 Apr 2012 08:05:34 -0700
From:	Stephen Hemminger <shemminger@...tta.com>
To:	Fredrick <fjohnber@...o.com>
Cc:	netdev@...r.kernel.org, eric.dumazet@...il.com
Subject: Re: yield() in netlink_broadcast_filtered

On Fri, 06 Apr 2012 07:11:56 -0700
Fredrick <fjohnber@...o.com> wrote:

> 
> 
> I see there is a yield being called from
> netlink_broadcast_filtered.
> .....
> int netlink_broadcast_filtered(...)
> {
> ....
>          if (info.delivered) {
>                  if (info.congested && (allocation & __GFP_WAIT))
>                          yield();
>                  return 0;
>          }
>          return -ESRCH;
> }
> .....
> 
> But I don't see the point of calling it.
> After the yield, there is nothing being done.
> It just returns. So why yield ?
> Why can't it simply return?
> 

Because without that yield it is easily possible for one process
to generate lots of netlink messages and overrun the consumers.
The yield allows the now ready listening processes to run.
There is no good mechanism to totally prevent overrunning the
socket of processes reading for netlink events, but this yield()
is good enough to avoid the problem with the typical case.
The example we are familar with is a full route table (1M routes)
and a link change causing a routing flap.
--
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