From mboxrd@z Thu Jan 1 00:00:00 1970 From: walimis Subject: Re: [PATCH 2/2] kvm tools: change option type of RNG from increment to boolean Date: Wed, 17 Aug 2011 23:14:05 +0800 Message-ID: <20110817151405.GA32698@walimis-desktop> References: <1313574294-23123-1-git-send-email-walimisdev@gmail.com> <1313574294-23123-2-git-send-email-walimisdev@gmail.com> <4E4BA997.7050408@kernel.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Pekka Enberg , Ingo Molnar , kvm@vger.kernel.org To: Sasha Levin Return-path: Received: from mail-qy0-f181.google.com ([209.85.216.181]:46080 "EHLO mail-qy0-f181.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753352Ab1HQPaZ (ORCPT ); Wed, 17 Aug 2011 11:30:25 -0400 Received: by qyk34 with SMTP id 34so689013qyk.19 for ; Wed, 17 Aug 2011 08:30:24 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: On Wed, Aug 17, 2011 at 03:02:55PM +0300, Sasha Levin wrote: >On Wed, Aug 17, 2011 at 2:44 PM, Pekka Enberg wrote: >> On 8/17/11 12:44 PM, Liming Wang wrote: >>> >>> Becasue virtio random generator is a single device, change its option >>> type to boolean. >>> >>> Signed-off-by: Liming Wang >> >> I suppose the idea here was to support multiple rng devices. Sasha? > >I wrote the idea behind supporting multiple rng devices in the commit >message of that patch: > >Since multiple hardware rng devices of the same type are currently >unsupported by the kernel, this serves more as an example of a basic >virtio driver under kvm tools and can be used to debug the PCI layer. > >Currently I use it mostly to easily test the virtio-pci and related >code, for example - when I added MSI-X I've tried creating a bunch of >virtio-rng devices and seeing how the kernel handles them. I see. I found this issue because kvm tools crashed when I used multiple rng devices to test. > >Since it having multiple virtio-rng devices doesn't really do anything >in the guest at the moment it could also be removed (unless we fix the >kernel to support them :) ). > >IMO this can go either way, theres no reason to keep support for >multiple devices besides making development a bit easier. OK, as you said, we can keep it for development. walimis