From: "Andreas Färber" <afaerber@suse.de>
To: Igor Mammedov <imammedo@redhat.com>
Cc: ehabkost@redhat.com, qemu-devel@nongnu.org, armbru@redhat.com,
lcapitulino@redhat.com, anthony@codemonkey.ws,
pbonzini@redhat.com
Subject: Re: [Qemu-devel] [PATCH] qdev: Fix device_add bus assumptions
Date: Mon, 22 Apr 2013 15:35:49 +0200 [thread overview]
Message-ID: <51753CB5.80607@suse.de> (raw)
In-Reply-To: <20130418104156.7a5349f0@nial.usersys.redhat.com>
Am 18.04.2013 10:41, schrieb Igor Mammedov:
> On Tue, 16 Apr 2013 03:50:21 +0200
> Andreas Färber <afaerber@suse.de> wrote:
>
>> Drop an unreachable fallback bus assignment to SysBus.
>>
>> If no ,bus= is specified, only search busses recursively for bus type if
>> the DeviceClass has a bus_type specified. Handle resulting NULL cases.
>>
>> Signed-off-by: Andreas Färber <afaerber@suse.de>
>> ---
>> qdev-monitor.c | 16 ++++++++--------
>> 1 file changed, 8 insertions(+), 8 deletions(-)
>>
>> diff --git a/qdev-monitor.c b/qdev-monitor.c
>> index 9a78ccf..73d7946 100644
>> --- a/qdev-monitor.c
>> +++ b/qdev-monitor.c
>> @@ -18,6 +18,7 @@
>> */
>>
>> #include "hw/qdev.h"
>> +#include "hw/sysbus.h"
>> #include "monitor/monitor.h"
>> #include "monitor/qdev.h"
>> #include "qmp-commands.h"
>> @@ -415,7 +416,7 @@ DeviceState *qdev_device_add(QemuOpts *opts)
>> DeviceClass *k;
>> const char *driver, *path, *id;
>> DeviceState *qdev;
>> - BusState *bus;
>> + BusState *bus = NULL;
>>
>> driver = qemu_opt_get(opts, "driver");
>> if (!driver) {
>> @@ -453,7 +454,7 @@ DeviceState *qdev_device_add(QemuOpts *opts)
>> driver, object_get_typename(OBJECT(bus)));
>> return NULL;
>> }
>> - } else {
>> + } else if (k->bus_type != NULL) {
>> bus = qbus_find_recursive(sysbus_get_default(), NULL, k->bus_type);
>> if (!bus) {
>> qerror_report(QERR_NO_BUS_FOR_DEVICE,
>> @@ -461,18 +462,17 @@ DeviceState *qdev_device_add(QemuOpts *opts)
>> return NULL;
>> }
>> }
>> - if (qdev_hotplug && !bus->allow_hotplug) {
>> + if (qdev_hotplug && bus && !bus->allow_hotplug) {
>> qerror_report(QERR_BUS_NO_HOTPLUG, bus->name);
>> return NULL;
>> }
>>
>> - if (!bus) {
>> - bus = sysbus_get_default();
>> - }
>> -
> I've checked all direct childs of TYPE_DEVICE and they all set k->bus_type,
> with only one exception of TYPE_CPU.
> So it should be safe to remove fallback
> from qdev_device_add POV.
Sorry, don't understand what exactly you are suggesting?
> However TYPE_CPU breaks assumption that device always has parent_bus set to not
> NULL in qdev_unplug() and qdev_print()
>
> It would be better to add something like this:
> // untested
>
> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> index 4eb0134..45009ba 100644
> --- a/hw/core/qdev.c
> +++ b/hw/core/qdev.c
> @@ -207,7 +207,7 @@ void qdev_unplug(DeviceState *dev, Error **errp)
> {
> DeviceClass *dc = DEVICE_GET_CLASS(dev);
>
> - if (!dev->parent_bus->allow_hotplug) {
> + if (dev->parent_bus && !dev->parent_bus->allow_hotplug) {
> error_set(errp, QERR_BUS_NO_HOTPLUG, dev->parent_bus->name);
> return;
> }
This part looks good, can you resend as a proper patch? Even if we can't
test the no-parent_bus unplug path ATM, it shouldn't hurt to fix
assumptions about it.
Andreas
> diff --git a/qdev-monitor.c b/qdev-monitor.c
> index 9a78ccf..2476e4e 100644
> --- a/qdev-monitor.c
> +++ b/qdev-monitor.c
> @@ -557,7 +557,9 @@ static void qdev_print(Monitor *mon, DeviceState *dev, int indent)
> qdev_print_props(mon, dev, DEVICE_CLASS(class)->props, indent);
> class = object_class_get_parent(class);
> } while (class != object_class_by_name(TYPE_DEVICE));
> - bus_print_dev(dev->parent_bus, mon, dev, indent);
> + if (dev->parent_bus) {
> + bus_print_dev(dev->parent_bus, mon, dev, indent);
> + }
> QLIST_FOREACH(child, &dev->child_bus, sibling) {
> qbus_print(mon, child, indent);
> }
>
>> /* create device, set properties */
>> qdev = DEVICE(object_new(driver));
>> - qdev_set_parent_bus(qdev, bus);
>> +
>> + if (bus) {
>> + qdev_set_parent_bus(qdev, bus);
>> + }
>>
>> id = qemu_opts_id(opts);
>> if (id) {
>
>
--
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
next prev parent reply other threads:[~2013-04-22 13:35 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-04-16 1:50 [Qemu-devel] [PATCH] qdev: Fix device_add bus assumptions Andreas Färber
2013-04-18 8:41 ` Igor Mammedov
2013-04-18 9:01 ` Igor Mammedov
2013-04-22 11:51 ` Libaiqing
2013-04-22 13:27 ` Andreas Färber
2013-04-22 13:54 ` KONRAD Frédéric
2013-04-22 14:02 ` Andreas Färber
2013-04-22 14:15 ` KONRAD Frédéric
2013-04-22 14:22 ` Andreas Färber
2013-04-22 15:11 ` KONRAD Frédéric
2013-04-22 13:35 ` Andreas Färber [this message]
2013-04-22 18:36 ` Anthony Liguori
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=51753CB5.80607@suse.de \
--to=afaerber@suse.de \
--cc=anthony@codemonkey.ws \
--cc=armbru@redhat.com \
--cc=ehabkost@redhat.com \
--cc=imammedo@redhat.com \
--cc=lcapitulino@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).