svn commit: r262351 - head/sys/netinet6

[Bjoern A. Zeeb]

On 23 Feb 2014, at 01:27 , Craig Rodrigues <rodrigc at FreeBSD.org> wrote:

> Author: rodrigc
> Date: Sun Feb 23 01:27:22 2014
> New Revision: 262351
> URL: http://svnweb.freebsd.org/changeset/base/262351
> 
> Log:
>  Remove KASSERT from in6p_lookup_mcast_ifp().
> 
>  When the devel/jenkins port, version 1.551 was started,
>  the kernel would panic if INVARIANTS was enabled in the kernel config.
> 
>  Suggested by: bms

This reads to me “we hit the assert, bms suggested to remove it”.  Great!  Caught a bug!  Where is it?  In no way the commit message tells me why?   Why did we not hit the assert in the last n years? What has changed that we hit it now? Was the assert wrong from the beginning?  Did we hit a real bug elsewhere and now lost the tracking for it?

Could you please at least for the archives explain?



> 
> Modified:
>  head/sys/netinet6/in6_mcast.c
> 
> Modified: head/sys/netinet6/in6_mcast.c
> ==============================================================================
> --- head/sys/netinet6/in6_mcast.c	Sun Feb 23 00:46:05 2014	(r262350)
> +++ head/sys/netinet6/in6_mcast.c	Sun Feb 23 01:27:22 2014	(r262351)
> @@ -1781,8 +1781,6 @@ in6p_lookup_mcast_ifp(const struct inpcb
> 	    ("%s: not INP_IPV6 inpcb", __func__));
> 	KASSERT(gsin6->sin6_family == AF_INET6,
> 	    ("%s: not AF_INET6 group", __func__));
> -	KASSERT(IN6_IS_ADDR_MULTICAST(&gsin6->sin6_addr),
> -	    ("%s: not multicast", __func__));
> 
> 	ifp = NULL;
> 	memset(&ro6, 0, sizeof(struct route_in6));

— 
Bjoern A. Zeeb                             ????????? ??? ??????? ??????:
'??? ??? ???? ??????  ??????? ?? ?? ??????? ??????? ??? ????? ????? ????
?????? ?? ????? ????',  ????????? ?????????, "??? ????? ?? ?????", ?.???