[PATCH 10/14] drm/i915/display/vdsc: Add ultrajoiner support with DSC
Nautiyal, Ankit K
ankit.k.nautiyal at intel.com
Mon Sep 9 09:23:55 UTC 2024
On 9/6/2024 10:09 PM, Ville Syrjälä wrote:
> On Fri, Sep 06, 2024 at 07:30:16PM +0300, Ville Syrjälä wrote:
>> On Fri, Sep 06, 2024 at 06:28:03PM +0530, Ankit Nautiyal wrote:
>>> From: Stanislav Lisovskiy <stanislav.lisovskiy at intel.com>
>>>
>>> Add changes to DSC which are required for Ultrajoiner.
>>>
>>> v2:
>>> -Use correct helper for setting bits for bigjoiner secondary. (Ankit)
>>> -Use enum for joiner pipe count instead magic numbers. (Suraj)
>>> -Use primary/secondary instead of master/slave. (Suraj)
>>>
>>> Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy at intel.com>
>>> Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal at intel.com>
>>> ---
>>> drivers/gpu/drm/i915/display/intel_vdsc.c | 26 +++++++++++++++++++++--
>>> 1 file changed, 24 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/i915/display/intel_vdsc.c b/drivers/gpu/drm/i915/display/intel_vdsc.c
>>> index 8158e3702ed5..66e810c8de68 100644
>>> --- a/drivers/gpu/drm/i915/display/intel_vdsc.c
>>> +++ b/drivers/gpu/drm/i915/display/intel_vdsc.c
>>> @@ -379,9 +379,11 @@ static int intel_dsc_get_vdsc_per_pipe(const struct intel_crtc_state *crtc_state
>>> int intel_dsc_get_num_vdsc_instances(const struct intel_crtc_state *crtc_state)
>>> {
>>> int num_vdsc_instances = intel_dsc_get_vdsc_per_pipe(crtc_state);
>>> + int joined_pipes = intel_joiner_num_pipes(crtc_state);
>>>
>>> - if (crtc_state->joiner_pipes)
>>> - num_vdsc_instances *= 2;
>>> + if (joined_pipes == INTEL_BIG_JOINER_PIPES ||
>>> + joined_pipes == INTEL_ULTRA_JOINER_PIPES)
>>> + num_vdsc_instances *= joined_pipes;
>>>
>>> return num_vdsc_instances;
>>> }
>>> @@ -751,6 +753,14 @@ void intel_uncompressed_joiner_enable(const struct intel_crtc_state *crtc_state)
>>> }
>>> }
>>>
>>> +static bool intel_crtc_ultrajoiner_enable_needed(const struct intel_crtc_state *crtc_state)
>>> +{
>>> + struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
>>> +
>>> + return intel_joiner_num_pipes(crtc_state) == INTEL_ULTRA_JOINER_PIPES &&
>>> + crtc->pipe != PIPE_D;
>>> +}
>>> +
>>> void intel_dsc_enable(const struct intel_crtc_state *crtc_state)
>>> {
>>> struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
>>> @@ -770,7 +780,19 @@ void intel_dsc_enable(const struct intel_crtc_state *crtc_state)
>>> dss_ctl1_val |= JOINER_ENABLE;
>>> }
>>> if (crtc_state->joiner_pipes) {
>>> + /*
>>> + * This bit doesn't seem to follow primary/secondary logic or
>>> + * any other logic, so lets just add helper function to
>>> + * at least hide this hassle..
>>> + */
>>> + if (intel_crtc_ultrajoiner_enable_needed(crtc_state))
>> What is this crazyness? This would throw a big wrench into
>> the works, eg. the ultrajoiner readout would not work as intended.
> Hmm. I do see a note to this effect in bspec. But that doesn't make
> any real sense. I would expect that we either enable this for
> everything, or only for pipes A+C (the bigjoiner primaries).
> The latter would seem sensible, but it would also mean that
> we need to rethink the readout as well.
Yes as per bspec : The ULTRA_JOINER_ENABLE bit will be set for Pipe A,
Pipe B and Pipe C only.
Are you suggesting to write this for D as well? or just for A and C?
Also PRIMARY_ULTRA_JOINER_ENABLE is to be set for A. For other pipes
this bit will be reset.
Though only C is ultrajoiner secondary pipe, but since for other pipes
this bit is reset, how to make that distinction between C and others.
I mean readout will tell ultrajoiner primary as A and ultrajoiner
secondary as B,C,D as bit is reset for all three.
Currently we have secondary ultrajoiner pipes reading 0xE.
Regards,
Ankit
>>> + dss_ctl1_val |= ULTRA_JOINER_ENABLE;
>>> +
>>> + if (intel_crtc_is_ultrajoiner_primary(crtc_state))
>>> + dss_ctl1_val |= PRIMARY_ULTRA_JOINER_ENABLE;
>>> +
>>> dss_ctl1_val |= BIG_JOINER_ENABLE;
>>> +
>>> if (intel_crtc_is_bigjoiner_primary(crtc_state))
>>> dss_ctl1_val |= PRIMARY_BIG_JOINER_ENABLE;
>>> }
>>> --
>>> 2.45.2
>> --
>> Ville Syrjälä
>> Intel
More information about the Intel-gfx
mailing list