qemu gl changes are broken.

qemu gl changes are broken.

[Scott Garman]

Hello,

I tried building some images overnight with the latest master, and could 
not get them to run this morning with qemu - I got an odd error that a 
VNC server was started instead?

Just on a hunch, I reverted the following commits:

qemu: Fix broken accidental path move
qemu-git: Drop GL passthrough support
qemu-0.15: Drop GL passthrough support

rebuilt qemu-native, and now qemu works correctly.

I have no idea why qemu is trying to fallback on a VNC server, as I did 
not have one running on my machine. I know we use VNC for our 
autobuilder sanity tests, but those haven't been integrated into the 
runqemu scripts as far as I can tell.

Scott

-- 
Scott Garman
Embedded Linux Engineer - Yocto Project
Intel Open Source Technology Center

Re: qemu gl changes are broken.

[Scott Garman]

On 08/21/2012 09:21 AM, Scott Garman wrote:
> Hello,
>
> I tried building some images overnight with the latest master, and could
> not get them to run this morning with qemu - I got an odd error that a
> VNC server was started instead?
>
> Just on a hunch, I reverted the following commits:
>
> qemu: Fix broken accidental path move
> qemu-git: Drop GL passthrough support
> qemu-0.15: Drop GL passthrough support
>
> rebuilt qemu-native, and now qemu works correctly.
>
> I have no idea why qemu is trying to fallback on a VNC server, as I did
> not have one running on my machine. I know we use VNC for our
> autobuilder sanity tests, but those haven't been integrated into the
> runqemu scripts as far as I can tell.

This may need further research, I seem to have it working on another 
machine. I'll report back once I've learned more.

Scott

-- 
Scott Garman
Embedded Linux Engineer - Yocto Project
Intel Open Source Technology Center

Re: qemu gl changes are broken.

[Scott Garman]

On 08/21/2012 10:45 AM, Scott Garman wrote:
> On 08/21/2012 09:21 AM, Scott Garman wrote:
>> Hello,
>>
>> I tried building some images overnight with the latest master, and could
>> not get them to run this morning with qemu - I got an odd error that a
>> VNC server was started instead?
>>
>> Just on a hunch, I reverted the following commits:
>>
>> qemu: Fix broken accidental path move
>> qemu-git: Drop GL passthrough support
>> qemu-0.15: Drop GL passthrough support
>>
>> rebuilt qemu-native, and now qemu works correctly.
>>
>> I have no idea why qemu is trying to fallback on a VNC server, as I did
>> not have one running on my machine. I know we use VNC for our
>> autobuilder sanity tests, but those haven't been integrated into the
>> runqemu scripts as far as I can tell.
>
> This may need further research, I seem to have it working on another
> machine. I'll report back once I've learned more.

Ok, I can no longer reproduce this issue anymore, even on the original 
machine. Figures. :-/

Scott

-- 
Scott Garman
Embedded Linux Engineer - Yocto Project
Intel Open Source Technology Center

Re: qemu gl changes are broken.

[Paul Eggleton]

On Tuesday 21 August 2012 16:46:31 Scott Garman wrote:
> On 08/21/2012 10:45 AM, Scott Garman wrote:
> > On 08/21/2012 09:21 AM, Scott Garman wrote:
> >> Hello,
> >> 
> >> I tried building some images overnight with the latest master, and could
> >> not get them to run this morning with qemu - I got an odd error that a
> >> VNC server was started instead?
> >> 
> >> Just on a hunch, I reverted the following commits:
> >> 
> >> qemu: Fix broken accidental path move
> >> qemu-git: Drop GL passthrough support
> >> qemu-0.15: Drop GL passthrough support
> >> 
> >> rebuilt qemu-native, and now qemu works correctly.
> >> 
> >> I have no idea why qemu is trying to fallback on a VNC server, as I did
> >> not have one running on my machine. I know we use VNC for our
> >> autobuilder sanity tests, but those haven't been integrated into the
> >> runqemu scripts as far as I can tell.
> > 
> > This may need further research, I seem to have it working on another
> > machine. I'll report back once I've learned more.
> 
> Ok, I can no longer reproduce this issue anymore, even on the original
> machine. Figures. :-/

Is this perhaps what happens when DISPLAY is not set? Just a guess...

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre