Re: [Qemu-devel] [PATCH 00/13] cuda: misc fixes and cleanups

[Mark Cave-Ayland]

On 06/02/16 14:30, Hervé Poussineau wrote:

> Hi Mark,
> 
> Le 06/02/2016 14:32, Mark Cave-Ayland a écrit :
>> On 23/01/16 20:39, Hervé Poussineau wrote:
>>
>>> Hi,
>>>
>>> This patchset cleans up a little bit the Apple CUDA emulation:
>>> - correctly reject unknown commands
>>> - correctly reject commands with wrong parameters
>>> - support changing the frequency of auto-polling
>>> - support changing device list probed in auto-poll
>>> - add logs when using FILE_SERVER_FLAG/SET_POWER_MESSAGE
>>> - remove unused commands (GET/SET_6805_ADDR)
>>> - remove unimplemented GET_SET_IIC/COMBINED_FORMAT_IIC
>>>
>>> GET_SET_IIC/COMBINED_FORMAT_IIC commands should be added again once
>>> we implement the I2C bus provided by CUDA.
>>>
>>> This also fixes MacBugs hanging at startup in the absence of
>>> ADB mouse input.
>>>
>>> Hervé
>>
>> Hi Hervé,
>>
>> I've just tried to rebase your git remote onto master to run some tests
>> and it looks like due to recent changes then I get a number of
>> conflicts. Any chance that you could rebase this patchset onto master
>> and update your remote accordingly?
> 
> I've rebased this patchset. This is available at
> http://repo.or.cz/qemu/hpoussin.git branch cuda
> 
> As a bonus, you have an implementation of CUDA_GET_SET_IIC command and
> the I2C bus.
> Unfortunately, I was unable to test if it really works.

I've posted a couple of minor comments on a couple of patches, but in
general I think this is a good improvement, especially given that there
is active work fixing up more CUDA functionality.

Other than the comments mentioned above, there is also the compile fix
for patch 4 posted to the list plus I still see some IIC debugging
enabled from the git repository which outputs to stdout, e.g.

CUDA: XXX receive data from I2C [addr 48] => ff
CUDA: XXX send data to I2C [addr 40]: 05 c0

Perhaps for the moment it is best to leave out the IIC functionality
simply to get the majority of the patchset included in preparation for
more detailed work later?

With all these issues addressed in a revised v2 patchset, I'm fairly
confident I can give the Tested-by requested by David before inclusion
in ppc-next since I didn't observe any regressions on my test images
with the patchset in its current form.


ATB,

Mark.