From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Wray Subject: Re: [RESUBMIT] [PATCH] xen, tools: pincpu use vcpu and cpumap_t Date: Thu, 05 May 2005 16:49:50 +0100 Message-ID: <427A409E.2010306@hp.com> References: <20050426223109.GC14932@us.ibm.com> <4277AC20.7060804@hp.com> <20050503235807.GB11696@us.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20050503235807.GB11696@us.ibm.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Ryan Harper Cc: Ian Pratt , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org Ryan Harper wrote: > * Mike Wray [2005-05-03 12:01]: > >>Ryan Harper wrote: >> >>>Here is an update to the previous pincpu patches. This patch updates >>>the pincpu dom0op to use a cpumap_t pointer, rather than specifying a >>>size in the interface. cpumap_t is current typedef'ed to an unsigned >>>long. xm list -v now truncates the cpumap field to only show the bits >>>of the map that are relevant. That is, if you have four VCPUS, you will >>>only see the first four bits of the cpumap. >>> >>>This patch compiles against current unstable tree, but I could not test >>>it as the current unstable tree crashes dom0 when booting domU. I back >>>ported the patch to 20050420 snapshot and it worked there. >>> >> >>This patch causes the dom0 crash I reported on the current unstable >>(attached) so it needs fixing before applying. > > > I was not able to recreate dom0 crash with the previous patch against > current nightly snapshot (20050503). None-the-less addressing some of > your feedback. > > Attached is an updated patch. > o bumped cpumap_t up to u64, per Ian and Keir > o Fixed deref of user-space pointer without copy_from_user(), Mike Wray > o Added put_domain() in ed > MAX_VIRT_CPUS error path, Mike Wray OK, I'll give it another whirl. Mike