From mboxrd@z Thu Jan 1 00:00:00 1970 From: George Dunlap Subject: Re: RFC: Automatically making a PCI device assignable in the config file Date: Fri, 12 Jul 2013 17:08:17 +0100 Message-ID: <51E029F1.7070006@eu.citrix.com> References: <20957.26468.175816.677766@mariner.uk.xensource.com> <51DD7420.5050803@eu.citrix.com> <51DE988E.7020700@citrix.com> <20960.81.509830.418330@mariner.uk.xensource.com> <20130712134833.GB3518@localhost.localdomain> <20960.5652.606227.862830@mariner.uk.xensource.com> <20130712150138.GB3846@phenom.dumpdata.com> <51E01C29.1000600@eu.citrix.com> <20130712160214.GB8063@phenom.dumpdata.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20130712160214.GB8063@phenom.dumpdata.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Konrad Rzeszutek Wilk Cc: Ian Campbell , Ian Jackson , David Vrabel , "xen-devel@lists.xen.org" List-Id: xen-devel@lists.xenproject.org On 12/07/13 17:02, Konrad Rzeszutek Wilk wrote: > On Fri, Jul 12, 2013 at 04:09:29PM +0100, George Dunlap wrote: >> On 12/07/13 16:01, Konrad Rzeszutek Wilk wrote: >>> On Fri, Jul 12, 2013 at 03:43:32PM +0100, Ian Jackson wrote: >>>> Konrad Rzeszutek Wilk writes ("Re: [Xen-devel] RFC: Automatically making a PCI device assignable in the config file"): >>>>> On Fri, Jul 12, 2013 at 02:10:41PM +0100, Ian Jackson wrote: >>>>>> I think at the very least there should be the ability to have it as a >>>>>> global config option. >>>>> Like the seatbelt option or perhaps a better name 'expert' in the /etc/xl.conf? >>>> I was thinking "default.pci.seize" or something. >>> Why not combine this with the other - the one where you can call >>> 'xl vpcu-set ' where the number of >>> VCPUs is greater than PCPUs. And if the 'expert' option is turned >>> on it will let you. If it is not, it will notify you about this >>> being a bad idea. >> ...because you may want to have one and not the other? > Seatbelts are seatbelts. They have many parts - but nonethless > at the end of the day you describe them by just one word. > > I am not a big fan of many knobs - even thought I did work at some > point for IBM which was known for its knob-for-everything products. Well remember that my main concern here is with the *changing* of behavior, not the *existence* of behavior. (i.e., xl pci-assignable-add already behaves in the way in which xl pci-attach would behave with the seatbelt off). If we had already had a single "expert" option in xl.conf, then we would be in nearly the same situation -- add xl pci-attach to the "expert" option and have it change from safe to unsafe without any warning, or add yet another parameter. -George