[PATCH 6/7] accel/qaic: Create a function to initialize BO

[Jeffrey Hugo]

On 9/17/2023 2:48 AM, Stanislaw Gruszka wrote:
> On Fri, Sep 01, 2023 at 11:22:46AM -0600, Jeffrey Hugo wrote:
>> From: Pranjal Ramajor Asha Kanojiya <quic_pkanojiy at quicinc.com>
>>
>> This makes sure that we have a single place to initialize and
>> re-initialize BO.
>>
>> Use this new API to cleanup release_dbc()
>>
>> We will need this for next patch to detach slicing to a BO.
>>
>> Signed-off-by: Pranjal Ramajor Asha Kanojiya <quic_pkanojiy at quicinc.com>
>> Reviewed-by: Jeffrey Hugo <quic_jhugo at quicinc.com>
>> Signed-off-by: Jeffrey Hugo <quic_jhugo at quicinc.com>
>> ---
>>   drivers/accel/qaic/qaic_data.c | 20 ++++++++++++++------
>>   1 file changed, 14 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/accel/qaic/qaic_data.c b/drivers/accel/qaic/qaic_data.c
>> index 6e44e00937af..2acb9dbac88b 100644
>> --- a/drivers/accel/qaic/qaic_data.c
>> +++ b/drivers/accel/qaic/qaic_data.c
>> @@ -635,6 +635,18 @@ static const struct drm_gem_object_funcs qaic_gem_funcs = {
>>   	.vm_ops = &drm_vm_ops,
>>   };
>>   
>> +static void qaic_init_bo(struct qaic_bo *bo, bool reinit)
>> +{
>> +	if (reinit) {
>> +		bo->sliced = false;
>> +		reinit_completion(&bo->xfer_done);
>> +	} else {
>> +		init_completion(&bo->xfer_done);
>> +	}
>> +	complete_all(&bo->xfer_done);
> Why do you need complete_all() here ?

This is moved from qaic_alloc_init_bo().

This puts the BO in a state where the wait_exec ioctl will fall through 
and return if userspace immediately calls it after allocating the BO, 
prior to submitting the BO to hardware.  Otherwise we need a special, 
one off check to see that the BO was never submitted to the hardware and 
handle that edge case.

-Jeff