[PATCH 2/4] gpu: host1x: Enable gather filter

[Mikko Perttunen]

On 08/20/2017 07:59 PM, Dmitry Osipenko wrote:
> On 20.08.2017 19:44, Dmitry Osipenko wrote:
>> On 20.08.2017 19:24, Dmitry Osipenko wrote:
>>> On 18.08.2017 19:15, Mikko Perttunen wrote:
>>>> The gather filter is a feature present on Tegra124 and newer where the
>>>> hardware prevents GATHERed command buffers from executing commands
>>>> normally reserved for the CDMA pushbuffer which is maintained by the
>>>> kernel driver.
>>>>
>>>> This commit enables the gather filter on all supporting hardware.
>>>>
>>>> Signed-off-by: Mikko Perttunen <mperttunen at nvidia.com>
>>>> ---
>>>>   drivers/gpu/host1x/hw/channel_hw.c          | 22 ++++++++++++++++++++++
>>>>   drivers/gpu/host1x/hw/hw_host1x04_channel.h | 12 ++++++++++++
>>>>   drivers/gpu/host1x/hw/hw_host1x05_channel.h | 12 ++++++++++++
>>>>   3 files changed, 46 insertions(+)
>>>>
>>>> diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c
>>>> index 0161da331702..5c0dc6bb51d1 100644
>>>> --- a/drivers/gpu/host1x/hw/channel_hw.c
>>>> +++ b/drivers/gpu/host1x/hw/channel_hw.c
>>>> @@ -181,10 +181,32 @@ static int channel_submit(struct host1x_job *job)
>>>>   	return err;
>>>>   }
>>>>   
>>>> +static void enable_gather_filter(struct host1x *host,
>>>> +				 struct host1x_channel *ch)
>>>> +{
>>>> +#if HOST1X_HW >= 6
>>>> +	u32 val;
>>>> +
>>>> +	if (!host->hv_regs)
>>>> +		return;
>>>
>>> Is it really possible that gather filter could be not present on HW without
>>> hypervisor? Maybe there is other way to enable it in that case?

The hardware may have the hypervisor but Linux may be running as a 
virtual machine without access to the hypervisor, in which case we 
cannot access the registers, and it's the responsibility of the other OS 
acting as hypervisor to enable the gather filter.

>>>
>>> Is possible at all that hypervisor could be missed?
>>
>> BTW, this is also incoherent with the 'syncpoint protection' patch which doesn't
>> check for hypervisor presence.
>>
> 
> However, I noticed that check and it's wrongly placed ;) See comment to the
> 'syncpoint protection' patch.
>