netdev - RE: [bnx2x] Re: Kernel 3.18.11 hangs when inserting netconsle module on a DELL M620 VRTX Blade

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]

Message-ID: <B5657A6538887040AD3A81F1008BEC63C26537@avmb3.qlogic.org>
Date:	Wed, 8 Apr 2015 14:42:08 +0000
From:	Yuval Mintz <Yuval.Mintz@...gic.com>
To:	Peter Hurley <peter@...leysoftware.com>,
	Urban Loesch <bind@...s.net>
CC:	linux-kernel <linux-kernel@...r.kernel.org>,
	Ariel Elior <Ariel.Elior@...gic.com>,
	netdev <netdev@...r.kernel.org>
Subject: RE: [bnx2x] Re: Kernel 3.18.11 hangs when inserting netconsle
 module on a DELL M620 VRTX Blade

> > I'have installed a new DELL VRTX M620 Blade with kernel 3.18.11.
> > After system startup I tried to activate the kernel netconsole with remote
> logging enabled.
> >
> > I executed the following command and the shell I issued it becomes
> unresponsive and hangs.
> >
> > #  modprobe netconsole
> netconsole="@/eth0,514@...1.10.197/00:10:db:fc:60:0c"
> >
> > The system load increases slowly and the CPU #11 uses 100% of soft
> > irq. Only a soft reset witohut loading the netconsole module after startup
> solves the issue.

I suspect this is a regression introduced by 9a2620c87745
"bnx2x: prevent WARN during driver unload".

bnx2x locks & unlocks spin_lock_bh() during the napi poll, which shouldn't
be done while interrupts are disabled. This break interoperability with netpoll,
as it disables irqs prior to sending the skb on the bnx2x's interface.

Can you please try compiling your kernel without CONFIG_NET_RX_BUSY_POLL?
I suspect that might solve your issue.

Regardless, we'll investigate this further and hopefully come up with a fix soon.

Thanks,
Yuval