netdev - Re: [PATCH] bnx2x: fix panic when TX ring is full

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 Jun 2012 18:56:22 +0300
From:	"Dmitry Kravkov" <dmitry@...adcom.com>
To:	"Eric Dumazet" <eric.dumazet@...il.com>
cc:	"Tomas Hruby" <thruby@...gle.com>,
	"David Miller" <davem@...emloft.net>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"therbert@...gle.com" <therbert@...gle.com>,
	"evansr@...gle.com" <evansr@...gle.com>,
	"Eilon Greenstein" <eilong@...adcom.com>,
	"Merav Sicron" <meravs@...adcom.com>,
	"Yaniv Rosner" <yaniv.rosner@...adcom.com>,
	"willemb@...gle.com" <willemb@...gle.com>
Subject: Re: [PATCH] bnx2x: fix panic when TX ring is full

On Thu, 2012-06-21 at 17:12 +0200, Eric Dumazet wrote:
> On Thu, 2012-06-21 at 15:19 +0300, Dmitry Kravkov wrote:
> 
> > The crash happens with default configuration since
> > [4acb41903b2f99f3dffd4c3df9acc84ca5942cb2] "net/tcp: Fix tcp memory
> > limits initialization when !CONFIG_SYSCTL", but it can be hit by
> > increasing values of tcp_wmem even earlier.
> 
> This makes no sense.
Bisected to this commit and reproduced before the commit only after:
echo "4096 16384 4194304" > /proc/sys/net/ipv4/tcp_wmem
by this max nr_frags raised from 8 to 17, when running 40 netperfs

i've decreased rx queue to 200, during the test


> > From: Dmitry Kravkov <dmitry@...adcom.com>
> > Subject: [PATCH net-next] bnx2x: reservation for NEXT tx BDs
> > 
> > Commit [4acb41903b2f99f3dffd4c3df9acc84ca5942cb2]
> > net/tcp: Fix tcp memory limits initialization when !CONFIG_SYSCTL
> > provided new default value for tcp_wmem, since heavy tcp
> > traffic may cause the TSO packet to consume 20 BDs + 1 for next page
> > descriptor.
> 
> This is completely bogus. I have no idea how you came to this.
> 
> A forwarding workload can trigger same bug, if GRO is enabled.
> 
> Remove this wrong bit, please ?
> 
> 
> 



--
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