Date: Wed, 23 Nov 2011 14:53:37 +0100 From: Marko Zec <zec@fer.hr> To: freebsd-net@freebsd.org Cc: Nikos Vassiliadis <nvass@gmx.com> Subject: Re: arprequest triggered panic Message-ID: <201111231453.37900.zec@fer.hr> In-Reply-To: <4ECCF257.9090801@gmx.com> References: <4EC63D37.4050105@gmx.com> <20111123112204.GY96616@glebius.int.ru> <4ECCF257.9090801@gmx.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 23 November 2011 14:17:11 Nikos Vassiliadis wrote:
> On 11/23/2011 1:22 PM, Gleb Smirnoff wrote:
> > I'd suspect VIMAGE. Can you please try w/o it and if it appears to be
> > VIMAGE-related, then please file a PR.
>
> It seems VIMAGE related. I'll ask at virtualization@.
From the backtrace it looks like the curvnet context is not properly set in a
timer-driven call graph originating at lagg_port_setlladdr().
Perhaps this (untested) patch could help:
--- //depot/user/zec/vimage_8/src/sys/net/if_lagg.c 2011-09-06
05:45:07.000000000 0000
+++ /u/marko/p4/zec/vimage_8/src/sys/net/if_lagg.c 2011-09-06
05:45:07.000000000 0000
@@ -468,7 +468,9 @@
ifp = llq->llq_ifp;
/* Set the link layer address */
+ CURVNET_SET(ifp->if_vnet);
error = if_setlladdr(ifp, llq->llq_lladdr, ETHER_ADDR_LEN);
+ CURVNET_RESTORE();
if (error)
printf("%s: setlladdr failed on %s\n", __func__,
ifp->if_xname);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201111231453.37900.zec>