Re: [Qemu-devel] [PATCH 0/4] A few cleanups of qdev users

[Markus Armbruster]

Amit Shah <address@hidden> writes:

> On (Mon) 27 Jun 2011 [14:36:11], Markus Armbruster wrote:
>> Amit Shah <address@hidden> writes:
>> 
>> > On (Fri) 24 Jun 2011 [13:57:28], Markus Armbruster wrote:
>> >> Ping?
>> >
>> > There were a couple of things:
>> >
>> >>             port 0, guest on, host off, throttle off
>> >
>> > guest on/off, host on/off doesn't convey much -- what's on/off?
>> >
>> > Also, 'throttle' could be 'thottled'?
>> 
>> Discussion petered out with my message[*]:
>> 
>>     I chose on/off to stay consistent with how qdev shows bool
>>     properties (print_bit() in qdev-properties.c).  May be misguided.
>>     Like you, I'm having difficulties coming up with a better version
>>     that is still consise.
>>     
>>     But: should "info qtree" show such device state?  It's about
>>     configuration of the device tree, isn't it?  Connection status is
>>     useful to know, but it's not device configuration.  Other
>>     print_dev() methods may cross that line, too.  For instance,
>>     usb_bus_dev_print() prints attached, which looks suspicious (commit
>>     66a6593a).
>> 
>> Should info qtree continue to show this information?  If yes, care to
>> suggest a better format?
>
> Don't know.  I'm fine with anything the qdev guys decide.  I agree
> this isn't device state.

Unfortunately, there's no qdev maintainer making descisions.

What shall we do now?

1. Commit as is.  Need an ACK then.

2. Respin with virtser_bus_dev_print() printing the same stuff prettier.
Need ideas on a prettier format.

3. Respin with virtser_bus_dev_print() printing less stuff, but
prettier.  Need ideas on what exactly to print, and how.