From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paolo Bonzini Subject: Re: Documentation of the kvm->srcu lock Date: Mon, 09 Sep 2013 15:51:53 +0200 Message-ID: <522DD279.6000605@redhat.com> References: <20130909140032.3fb22a4a@thhw500> <522DBB6D.1050206@redhat.com> <522DBE89.1060505@linux.vnet.ibm.com> <20130909144758.6e6c0538@thhw500> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: kvm@vger.kernel.org, Xiao Guangrong To: Thomas Huth Return-path: Received: from mx1.redhat.com ([209.132.183.28]:5434 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752415Ab3IINvv (ORCPT ); Mon, 9 Sep 2013 09:51:51 -0400 In-Reply-To: <20130909144758.6e6c0538@thhw500> Sender: kvm-owner@vger.kernel.org List-ID: Il 09/09/2013 14:47, Thomas Huth ha scritto: > Am Mon, 09 Sep 2013 20:26:49 +0800 > schrieb Xiao Guangrong : > >> On 09/09/2013 08:13 PM, Paolo Bonzini wrote: >>> Il 09/09/2013 14:00, Thomas Huth ha scritto: >>>> >>>> Hi all! >>>> >>>> I am quite new to KVM, and I recently had to struggle with the >>>> kvm->srcu lock a little bit ... documentation for this lock seems >>>> to be quite sparse, I've only found some descriptions in mailing >>>> list archives (the description at the following URL is quite good: >>>> http://www.mail-archive.com/kvm@vger.kernel.org/msg90003.html ). >>>> >>>> So I wonder whether it would be helpful for others, too, if some >>>> information about this lock was added to the documentation file >>>> Documentation/virtual/kvm/locking.txt for example? I'd like to >>>> suggest the following text as a first draft (using some information >>>> from the URL above): >>>> >>>> Name: kvm->srcu >>>> Type: srcu lock >>>> Arch: any >>>> Protects: - kvm->memslots >>>> - kvm->buses >>>> Comment: The srcu read lock must be held while accessing memslots (e.g. >>>> when using gfn_to_* functions) and while accessing in-kernel >>>> MMIO/PIO address->device structure mapping (kvm->buses). >>>> The srcu index can be stored in kvm_vcpu->srcu_idx per vcpu >>>> if it is needed by multiple functions. >>>> >>>> What do you think? Since I am not an expert in this area, extension >>>> and/or corrections are very welcome! >>> >>> Yes, this looks good. Xiao, any suggestions? >> >> It's nice indeed! :) >> >> Reviewed-by: Xiao Guangrong > > Great :-) ... ok, can you then simply pick up the above text or shall I > submit a real patch for that file? Sending a patch would be appreciated. Paolo