netdev - Re: [PATCH] ipv4: synchronize bind() with RTM

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

Date:	Thu, 21 Oct 2010 13:12:57 +0200
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Timo Teräs <timo.teras@....fi>
Cc:	David Miller <davem@...emloft.net>, netdev@...r.kernel.org
Subject: Re: [PATCH] ipv4: synchronize bind() with RTM_NEWADDR notifications

Le jeudi 21 octobre 2010 à 13:58 +0300, Timo Teräs a écrit :
> On 10/21/2010 01:50 PM, David Miller wrote:
> > From: Timo Teräs <timo.teras@....fi>
> > Date: Thu, 21 Oct 2010 13:41:37 +0300
> > 
> >> Is inet_bind() called from non-userland context? If yes, then this is a
> >> bad idea. Otherwise I don't think it's that hot path...
> > 
> > It is.
> 
> Yet, almost immediately after that there is lock_sock() which can also
> sleep. How does that work then?
> 

I am not sure I understand your question. Maybe my answer wont be clear.

rtnl_lock() can take ages on some setups, because its using one single
and shared mutex. Its use should be restricted to administrative
purposes.

bind() is a pretty hot path on many workloads, this is hardly what we
call an administrative task.

lock_sock() uses a per socket lock, and it is a _bit_ more scalable, you
can have 4096 cpus all using bind()/recv()/send()/... at the same time,
it just works.

--
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