From: Jan Kiszka Subject: Re: [PATCH] KVM: x86: Convert tsc_write_lock to raw_spinlock Date: Mon, 07 Feb 2011 16:58:16 +0100 Message-ID: <4D501698.1000507@siemens.com> References: <4D4BCB97.6000900@siemens.com> <4D4C698A.4010201@redhat.com> <4D4FD8EE.6040009@siemens.com> <4D4FFD97.6010805@redhat.com> <4D5008F0.5060200@siemens.com> <4D500C9F.2080501@redhat.com> <4D50120F.4030809@siemens.com> <4D501547.6040907@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Cc: Zachary Amsden , Marcelo Tosatti , kvm , Linux Kernel Mailing List To: Avi Kivity Return-path: In-Reply-To: <4D501547.6040907@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On 2011-02-07 16:52, Avi Kivity wrote: > On 02/07/2011 05:38 PM, Jan Kiszka wrote: >>> >>> I don't know as it is allowed to sleep, it doesn't call any sleeping >>> functions to my knowledge. What worries me in the RT case is that the >>> spinlock acquired for hardware_enable might be preempted and run on >>> another CPU, which obviously isn't what you want. >> >> I see now, there are calls to raw_smp_processor_id. >> >> I think it's best to make this a raw lock. At this chance, some >> read-only users of vm_list should be rcu'ified. Will have a look. > > vm_list is rarely used, for either read or write. I don't see the need > to rcu it. Avoid that code under this lock expands the preempt-disabled period, specifically under -rt, and specifically as the number of objects over which we loop is user-defined. Jan -- Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux