From mboxrd@z Thu Jan 1 00:00:00 1970 From: Huang Ying Subject: Re: [PATCH -v3] QEMU-KVM: MCE: Relay UCR MCE to guest Date: Mon, 21 Sep 2009 10:23:47 +0800 Message-ID: <1253499827.15717.530.camel@yhuang-dev.sh.intel.com> References: <1253260739.15717.511.camel@yhuang-dev.sh.intel.com> <20090918170535.GA9100@amt.cnet> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: Avi Kivity , Andi Kleen , Anthony Liguori , "kvm@vger.kernel.org" To: Marcelo Tosatti Return-path: Received: from mga09.intel.com ([134.134.136.24]:46037 "EHLO mga09.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753566AbZIUCYH (ORCPT ); Sun, 20 Sep 2009 22:24:07 -0400 In-Reply-To: <20090918170535.GA9100@amt.cnet> Sender: kvm-owner@vger.kernel.org List-ID: On Sat, 2009-09-19 at 01:05 +0800, Marcelo Tosatti wrote: > On Fri, Sep 18, 2009 at 03:58:59PM +0800, Huang Ying wrote: > > UCR (uncorrected recovery) MCE is supported in recent Intel CPUs, > > where some hardware error such as some memory error can be reported > > without PCC (processor context corrupted). To recover from such MCE, > > the corresponding memory will be unmapped, and all processes accessing > > the memory will be killed via SIGBUS. > > > > For KVM, if QEMU/KVM is killed, all guest processes will be killed > > too. So we relay SIGBUS from host OS to guest system via a UCR MCE > > injection. Then guest OS can isolate corresponding memory and kill > > necessary guest processes only. SIGBUS sent to main thread (not VCPU > > threads) will be broadcast to all VCPU threads as UCR MCE. > > > > v3: > > > > - Re-raise SIGBUS for SIGBUS not from MCE > > - Kill itself for error in kvm_inject_x86_mce > > This is broken, non-MCE SIGBUS causes qemu-kvm to call the sigbus > handler in a loop. Sorry. I tested the wrong branch in previous development. I will fix this. > BTW, how are you testing this and what guests have been tested? I use a self-made kernel module to simulate SIGBUS from MCE handler to qemu. Until now, only Linux guest has been tested. Best Regards, Huang Ying