|
From: | Anthony Liguori |
Subject: | Re: [Qemu-devel] [PATCH 07/10] qtest: IRQ interception infrastructure (v2) |
Date: | Sat, 25 Feb 2012 19:34:59 -0600 |
User-agent: | Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.23) Gecko/20110922 Lightning/1.0b2 Thunderbird/3.1.15 |
On 02/25/2012 07:31 PM, Andreas Färber wrote:
Am 25.02.2012 22:16, schrieb Anthony Liguori:On 02/25/2012 02:20 PM, Paolo Bonzini wrote:On 02/25/2012 08:42 PM, Anthony Liguori wrote:+ object_property_add_child(object_resolve_path("/i440fx/piix3", NULL), + "ioapic", OBJECT(dev), NULL);Jan objected to putting this under /i440fx/piix3.I think for our model, having it as a child property makes quite a lot of sense.Weren't you discussing rearranging the tree so that devices are under a separate node from block devices etc. (e.g., /devices)? Would be a good idea to introduce a (dummy) object_resolve_device_path() to avoid having to change hardcoded absolute string paths like the above later.
No need. object_resolve_path("i440fx/piix3", NULL) would have the same effect.But I also don't think it's necessary. Direct calls to object_resolve_path almost always indicates the need to refactor code.
Regards, Anthony Liguori
Or do that change first so that new paths like this one at least don't need to be touched again. Andreas
[Prev in Thread] | Current Thread | [Next in Thread] |