From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: [BUG?] Interrupts enabled after xen_acpi_processor_resume+0x0/0x34 [xen_acpi_processor] Date: Fri, 14 Mar 2014 13:14:15 -0400 Message-ID: <20140314171415.GF30560@phenom.dumpdata.com> References: <20140128150848.GA1428@redhat.com> <20140128160404.GA5732@phenom.dumpdata.com> <20140129082521.GA1362@redhat.com> <20140131160140.GC23648@phenom.dumpdata.com> <20140203101215.GA1725@redhat.com> <20140203141429.GD3400@phenom.dumpdata.com> <20140210143727.GA15771@phenom.dumpdata.com> <20140226102620.GA4686@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20140226102620.GA4686@redhat.com> Sender: linux-kernel-owner@vger.kernel.org To: Stanislaw Gruszka Cc: David Rientjes , boris.ostrovsky@oracle.com, david.vrabel@citrix.com, Ben Guthro , linux-kernel@vger.kernel.org, "Rafael J. Wysocki" , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org On Wed, Feb 26, 2014 at 11:26:21AM +0100, Stanislaw Gruszka wrote: > On Mon, Feb 10, 2014 at 09:37:27AM -0500, Konrad Rzeszutek Wilk wrote: > > > > But I'm not sure if that is good solution. It crate some not necessery > > > > sysfs directories and files. Additionaly it can restore CPU C-states > > > > after some other drivers resume, which prehaps require proper C-states. > > > > > > Yes. > > > > > > > > Hence maybe adding direct notify from xen core resume will be better > > > > idea (proposed patch below). Plese let me know what you think, I'll > > > > provide solution which you choose to bug reporters for testing. > > > > > > Let me think about it for a day or so. > > > > Sorry for the delay. I think this is fine. > > I'm sorry for delay too. I provided test kernel with the patch to bug > reporter, but have not get any answer since two weeks now. I tested it and it works great. > > I'll post the patch in the next email, I hope someone can test it (I'm > not xen user). I'm pretty sure it fixes the problem, but not sure if it > does not cause some crash. To test is enough to perform suspend/resume > cycle (thought I'm not sure if suspended should be just guest or whole > system with hypervisor). I put your put in the queue for 3.15. Thanks! > > Thanks > Stanislaw