Re: port-xen/50659: Xen4.5 crash with vnds

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

The following reply was made to PR kern/50659; it has been noted by GNATS.

From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: kern-bug-people%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost,
        sebastian%xenoserver.net@localhost
Subject: Re: port-xen/50659: Xen4.5 crash with vnds
Date: Wed, 23 Mar 2016 09:23:22 +0100

 On Wed, Mar 23, 2016 at 02:05:01AM +0000, Robert Elz wrote:
 > The following reply was made to PR kern/50659; it has been noted by GNATS.
 > 
 > From: Robert Elz <kre%munnari.OZ.AU@localhost>
 > To: gnats-bugs%NetBSD.org@localhost
 > Cc: 
 > Subject: Re: port-xen/50659: Xen4.5 crash with vnds
 > Date: Wed, 23 Mar 2016 09:01:42 +0700
 > 
 >      Date:        Wed, 23 Mar 2016 00:35:01 +0000 (UTC)
 >      From:        mlelstv%serpens.de@localhost (Michael van Elst)
 >      Message-ID:  <20160323003501.606657ACB7%mollari.NetBSD.org@localhost>
 >  
 >    |  The EBUSY comes from specfs that prevents opening a block device twice.
 >  
 >  Ah, OK, thanks - so this is not a problem with vnd at all, and is a Xen
 >  script (lack of locking or something) or user error?
 
 I think it is Xen's problem.
 I noticed that xen 4.5 starts domUs concurently (/etc/rc.d/xendomains start)
 which prevented be from upgrading to 4.5 because of concurency problems
 in my network scripts (local change).
 
 Now if it runs scripts for each block device (or, I guess, each
 network device) in parallel, then we have a problem.
 Adding a simple locking in the shell script may not be enough:
 I tried this for my network script, but then xl start times out
 on the script :(
 We may need something better than shlock(1)
 
 -- 
 Manuel Bouyer <bouyer%antioche.eu.org@localhost>
      NetBSD: 26 ans d'experience feront toujours la difference
 --