From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eduardo Habkost Subject: Re: [PATCH 00/11] x86: disable virt on kdump and emergency_restart (v3) Date: Mon, 17 Nov 2008 13:04:27 -0200 Message-ID: <20081117150427.GH4481@blackpad> References: <1226598231-18311-1-git-send-email-ehabkost@redhat.com> <491FD283.60402@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: Andrew Morton , Zachary Amsden , kvm-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, kexec-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, "Rafael J. Wysocki" , Haren Myneni , Simon Horman , "Eric W. Biederman" , Ingo Molnar , Andrey Borzenkov , mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org, Vivek Goyal To: Avi Kivity Return-path: Content-Disposition: inline In-Reply-To: <491FD283.60402-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: kexec-bounces-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org Errors-To: kexec-bounces+glkk-kexec=m.gmane.org-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org List-Id: kvm.vger.kernel.org On Sun, Nov 16, 2008 at 09:57:55AM +0200, Avi Kivity wrote: > Eduardo Habkost wrote: >> Hi, >> >> This is a new spin of the series to disable vmx on kdump and on >> emergency_restart. Now we avoid doing the function pointer stuff by >> moving 4 small KVM functions to a header, as inline functions. The code >> looks much simpler now, but we have to be more careful because some >> additional code will run on kdump and reboot even when KVM is never >> loaded. >> >> I haven't tested the SVM changes on AMD CPUs. The changes are really >> simple, but some testing is welcome. >> >> This series is against tip.git#master, that already contains the >> nmi_shootdown_cpus() changes I've submitted previously. >> >> >> > > Looks good. I am slightly uneasy about moving things away from vmx.h > and svm.h; can we keep them there and #include those headers directly? I see them as bits of code that are being moved from KVM to the kernel core. I think moving those bits outside arch/x86/kvm/ makes this more clear: people can expect that code living in arch/x86/kvm/ is never used if CONFIG_KVM was not set. Because of their location, I thought svm.h and vmx.h had KVM-specific code. Now I've noticed they are independent from KVM. May I move svm.h and vmx.h to arch/x86/include/asm, then? -- Eduardo