qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH] qdev: obey no_user


From: Andreas Färber
Subject: Re: [Qemu-devel] [PATCH] qdev: obey no_user
Date: Mon, 07 Jan 2013 23:36:36 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/17.0 Thunderbird/17.0

Am 07.01.2013 23:20, schrieb Peter Maydell:
> On 7 January 2013 22:10, Andreas Färber <address@hidden> wrote:
>> Am 07.01.2013 21:16, schrieb Peter Maydell:
>>> On 7 January 2013 20:12, Anthony Liguori <address@hidden> wrote:
>>>> Peter Maydell <address@hidden> writes:
>>>>> It seems to me like arbitrarily allowing the monitor to construct
>>>>> no-user devices isn't really the right way to attack the problem
>>>>> of "allow complete machine construction by management tools"...
>>>>
>>>> There is no such thing as a 'no-user' device.  It's a silly distinction
>>>> that has never had a consistent meaning.
>>>
>>> Then let's just rip that flag out completely.
>>
>> That's a bad idea, given that we are about to make the CPU a device.
> 
> We already have a pile of devices which the user can't usefully
> use -device on...CPUs would be just another one, right?

Not sure what you're arguing for here? The CPU is the worst example of a
device I know in that it is not limited to a particular bus and messes
with global state and threads.

Also, I'm sure that there will be objects/devices that a management tool
is not supposed to mess with either, once we start using
object_initialize() more.

If we agree that "no_user = 1" is not ideal or used inconsistently, then
we should IMO talk about how to replace/amend it and not conclude to
just rip it out and leave users without sensible error messages while
praying for The Omnipotent Management Tool that AFAICS we don't have
today...

Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg



reply via email to

[Prev in Thread] Current Thread [Next in Thread]