On 01/10/2017 12:10 PM, Vincent ABRIOU wrote:
>
>
> On 01/09/2017 04:10 PM, Hans Verkuil wrote:
>> On 09/12/2016 10:47 AM, Vincent Abriou wrote:
>>> It allows to simulate the behavior of hardware with such limitations or
>>> to connect vivid to real hardware with such limitations.
>>>
>>> Add the "allocators" module parameter option to let vivid use the
>>> dma-contig instead of vmalloc.
>>>
>>> Signed-off-by: Philipp Zabel <[email protected]>
>>> Signed-off-by: Hans Verkuil <[email protected]>
>>> Signed-off-by: Vincent Abriou <[email protected]>
>>>
>>> Cc: Philipp Zabel <[email protected]>
>>> Cc: Hans Verkuil <[email protected]>
>>> ---
>>> Documentation/media/v4l-drivers/vivid.rst | 8 ++++++++
>>> drivers/media/platform/vivid/Kconfig | 2 ++
>>> drivers/media/platform/vivid/vivid-core.c | 32
>>> ++++++++++++++++++++++++++-----
>>> 3 files changed, 37 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/Documentation/media/v4l-drivers/vivid.rst
>>> b/Documentation/media/v4l-drivers/vivid.rst
>>> index c8cf371..3e44b22 100644
>>> --- a/Documentation/media/v4l-drivers/vivid.rst
>>> +++ b/Documentation/media/v4l-drivers/vivid.rst
>>> @@ -263,6 +263,14 @@ all configurable using the following module options:
>>> removed. Unless overridden by ccs_cap_mode and/or ccs_out_mode the
>>> will default to enabling crop, compose and scaling.
>>>
>>> +- allocators:
>>> +
>>> + memory allocator selection, default is 0. It specifies the way buffers
>>> + will be allocated.
>>> +
>>> + - 0: vmalloc
>>> + - 1: dma-contig
>>
>> Could you add support for dma-sg as well? I think that would be fairly
>> trivial (unless
>> I missed something).
>>
>> Once that's added (or it's clear dma-sg won't work for some reason), then
>> I'll merge this.
>>
>> Regards,
>>
>> Hans
>>
>
> Hi Hans,
>
> What is the difference between a vmalloc allocation exported in DMABUF
> that will populate the sg and dma-sg allocation?
True. That wouldn't matter.
I'm merging this for 4.12.
Thanks!
Hans
