disabling bulk status

[Jack Neely]

Just aplied the patch and recompiled and restated arla.  So far this seems
to have fixed the problem.

Thnaks,
Jack Neely


--
Jack Neely  jjneely at eos.ncsu.edu
Student in Computer Science
   at North Carolina State University

Current Project: The NC State Linux Video Project
   http://prisant.ncsu.edu/~neely
WWW Pages:
   http://www.ac.net/~hneely
   http://quackmaster.net
PGP Public Key:
   http://quackmaster.net/pgp.html

On 26 Jan 2000, Love wrote:

> Jack Neely <jjneely at eos.ncsu.edu> writes:
> 
> > Disabling bulkstatus seems to have stoped the out pouring of panic
> > messages.
> 
> Then this patch might be what you want. Can you all please try it out ?
> 
> Love
> 
> PS, don't forget to turn on bulkstatus
> 
> Index: xfs_message.c
> ===================================================================
> RCS file: /afs/stacken.kth.se/src/SourceRepository/arla/xfs/linux/xfs_message.c,v
> retrieving revision 1.61
> diff -u -w -u -w -r1.61 xfs_message.c
> --- xfs_message.c	2000/01/07 10:36:11	1.61
> +++ xfs_message.c	2000/01/26 06:16:43
> @@ -228,7 +228,20 @@
>  	}
>  	t->anonrights = message->node.anonrights;
>      } else {
> +	struct xfs_message_inactivenode msg;
> +	
>  	XFSDEB(XDEBMSG, ("xfs_message_installattr: no such node\n"));
> +
> +	/*
> +	 * Eich, we got a installattr on a node we didn't have.  It
> +	 * might have been gc in a large message.  Make sure the
> +	 * user-land daemon knows we doesn't have it.
> +	 */
> +	
> +	msg.header.opcode = XFS_MSG_INACTIVENODE;
> +	msg.handle = message->node.handle;
> +	msg.flag   = XFS_NOREFS | XFS_DELETE;
> +	xfs_message_send(fd, &msg.header, sizeof(msg));
>      }
>  
>      return error;
>