qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCH 18/18] usb: add ehci adapter


From: David Ahern
Subject: Re: [Qemu-devel] [PATCH 18/18] usb: add ehci adapter
Date: Tue, 17 May 2011 08:19:50 -0600
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.15) Gecko/20110421 Fedora/3.1.9-2.fc14 Thunderbird/3.1.9


On 05/17/11 07:46, Gerd Hoffmann wrote:
> 
>   Hi,
> 
>>>> As someone who spent a significant amount of time working on the EHCI
>>>> code last year I am absolutely not ok with this. The entire
>>>> contribution
>>>> history for EHCI lost - and for no reason.
>>>
>>> There is a reason.  I've tried to keep the history, but it was a big
>>> mess with conflicts and build errors due to ehci being out-of-tree for a
>>> loooooong time.
>>
>> Not true. Back in March it took me less 2-1/2 hours to see a request for
>> a 0.14 version, update my git repo, merge master onto ehci, fix merges,
>> test and send out:
>>
>> http://comments.gmane.org/gmane.comp.emulators.kvm.devel/68898
>>
>> Again, most of the changes are to the EHCI code. The rest are
>> sprinklings to add the adapter and adhere to USB API.
> 
> And it totally ignores the usb changes which went into 0.14 to prepare
> the qemu usb subsystem for higher speeds.  There is no such thing like a
> v1 or v2 usb bus.
> 
> You've mailed out a single monster patch instead of a patch series. That
> is the problem with merges:  You can trivially get a full diff against
> master, but you can't easily get a nicely review-able patch series.

Come on Gerd. That was not an attempt to get it included. Someone asked
for a patch and I took the existing tree, merged with latest and through
out the patch. The v1/v2 is not in Jan's tree. That's a hack I have
locally to have mixed devices. What is shows is that is not a big deal
to move to latest code from Jan's last update in November.

> 
> I've tried to rebase the ehci tree.  First, to fix the patch series
> issue.  Second to adapt patches to qemu changes, to make the patch
> series bisect-able.  Also to drop obsolete bits.  That didn't work out
> as described.  You didn't even try.

Right, you did not even try to work with Jan's tree. There is no reason
the EHCI code could not have been brought in to qemu that way.


David


> 
> I'm not impressed.
> 
> cheers,
>   Gerd
> 



reply via email to

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