| Message ID | 20241120103802.134295-4-ankit.k.nautiyal@intel.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | DP DSC min/max src bpc fixes | expand |
On Wed, 20 Nov 2024, Ankit Nautiyal <ankit.k.nautiyal@intel.com> wrote: > Make a separate function for setting fec_enable in crtc_state. > Drop the check for FEC support as its already checked while checking for > DSC support. That's two changes that generally shouldn't be bundled together. Aim for separating non-functional refactoring and functional changes. (Well, arguably dropping the FEC check should also be non-functional, but you know what I mean.) > Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com> > --- > drivers/gpu/drm/i915/display/intel_dp.c | 30 +++++++++++++++++-------- > 1 file changed, 21 insertions(+), 9 deletions(-) > > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c > index dee15a05e7fd..d82e25d0dc5a 100644 > --- a/drivers/gpu/drm/i915/display/intel_dp.c > +++ b/drivers/gpu/drm/i915/display/intel_dp.c > @@ -2352,6 +2352,26 @@ static int intel_edp_dsc_compute_pipe_bpp(struct intel_dp *intel_dp, > return 0; > } > > +static void intel_dp_compute_fec_config(struct intel_dp *intel_dp, > + struct intel_crtc_state *pipe_config) I think all encoder->callback_name hooks should be named something_something_callback_name(), and the same goes for helpers specifically aimed at this. This would make the function intel_dp_fec_compute_config(). Yes, in many ways "compute fec config" reads better, but there's value in being able to search for "_compute_config", and to know this is only for he ->compute_config path. BR, Jani. > +{ > + if (pipe_config->fec_enable) > + return; > + > + /* > + * Though eDP v1.5 supports FEC with DSC, unlike DP, it is optional. > + * Since, FEC is a bandwidth overhead, continue to not enable it for > + * eDP. Until, there is a good reason to do so. > + */ > + if (intel_dp_is_edp(intel_dp)) > + return; > + > + if (intel_dp_is_uhbr(pipe_config)) > + return; > + > + pipe_config->fec_enable = true; > +} > + > int intel_dp_dsc_compute_config(struct intel_dp *intel_dp, > struct intel_crtc_state *pipe_config, > struct drm_connector_state *conn_state, > @@ -2368,15 +2388,7 @@ int intel_dp_dsc_compute_config(struct intel_dp *intel_dp, > int num_joined_pipes = intel_crtc_num_joined_pipes(pipe_config); > int ret; > > - /* > - * Though eDP v1.5 supports FEC with DSC, unlike DP, it is optional. > - * Since, FEC is a bandwidth overhead, continue to not enable it for > - * eDP. Until, there is a good reason to do so. > - */ > - pipe_config->fec_enable = pipe_config->fec_enable || > - (!intel_dp_is_edp(intel_dp) && > - intel_dp_supports_fec(intel_dp, connector, pipe_config) && > - !intel_dp_is_uhbr(pipe_config)); > + intel_dp_compute_fec_config(intel_dp, pipe_config); > > if (!intel_dp_dsc_supports_format(connector, pipe_config->output_format)) > return -EINVAL;
On 11/20/2024 5:22 PM, Jani Nikula wrote: > On Wed, 20 Nov 2024, Ankit Nautiyal <ankit.k.nautiyal@intel.com> wrote: >> Make a separate function for setting fec_enable in crtc_state. >> Drop the check for FEC support as its already checked while checking for >> DSC support. > That's two changes that generally shouldn't be bundled together. > > Aim for separating non-functional refactoring and functional > changes. (Well, arguably dropping the FEC check should also be > non-functional, but you know what I mean.) Moving to separate helper should indeed have non functional change and dropping the check can be another patch. Initially I was going with a separate patch for dropping the FEC check, but couldn't make up my mind, and merged the two things. :) Will do as suggested. > >> Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com> >> --- >> drivers/gpu/drm/i915/display/intel_dp.c | 30 +++++++++++++++++-------- >> 1 file changed, 21 insertions(+), 9 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c >> index dee15a05e7fd..d82e25d0dc5a 100644 >> --- a/drivers/gpu/drm/i915/display/intel_dp.c >> +++ b/drivers/gpu/drm/i915/display/intel_dp.c >> @@ -2352,6 +2352,26 @@ static int intel_edp_dsc_compute_pipe_bpp(struct intel_dp *intel_dp, >> return 0; >> } >> >> +static void intel_dp_compute_fec_config(struct intel_dp *intel_dp, >> + struct intel_crtc_state *pipe_config) > I think all encoder->callback_name hooks should be named > something_something_callback_name(), and the same goes for helpers > specifically aimed at this. > > This would make the function intel_dp_fec_compute_config(). > > Yes, in many ways "compute fec config" reads better, but there's value > in being able to search for "_compute_config", and to know this is only > for he ->compute_config path. Makes sense to have _fec_compute_config. Will change this. Thanks Jani, for the comments and suggestions. Regards, Ankit > > BR, > Jani. > > >> +{ >> + if (pipe_config->fec_enable) >> + return; >> + >> + /* >> + * Though eDP v1.5 supports FEC with DSC, unlike DP, it is optional. >> + * Since, FEC is a bandwidth overhead, continue to not enable it for >> + * eDP. Until, there is a good reason to do so. >> + */ >> + if (intel_dp_is_edp(intel_dp)) >> + return; >> + >> + if (intel_dp_is_uhbr(pipe_config)) >> + return; >> + >> + pipe_config->fec_enable = true; >> +} >> + >> int intel_dp_dsc_compute_config(struct intel_dp *intel_dp, >> struct intel_crtc_state *pipe_config, >> struct drm_connector_state *conn_state, >> @@ -2368,15 +2388,7 @@ int intel_dp_dsc_compute_config(struct intel_dp *intel_dp, >> int num_joined_pipes = intel_crtc_num_joined_pipes(pipe_config); >> int ret; >> >> - /* >> - * Though eDP v1.5 supports FEC with DSC, unlike DP, it is optional. >> - * Since, FEC is a bandwidth overhead, continue to not enable it for >> - * eDP. Until, there is a good reason to do so. >> - */ >> - pipe_config->fec_enable = pipe_config->fec_enable || >> - (!intel_dp_is_edp(intel_dp) && >> - intel_dp_supports_fec(intel_dp, connector, pipe_config) && >> - !intel_dp_is_uhbr(pipe_config)); >> + intel_dp_compute_fec_config(intel_dp, pipe_config); >> >> if (!intel_dp_dsc_supports_format(connector, pipe_config->output_format)) >> return -EINVAL;
On Wed, 20 Nov 2024, "Nautiyal, Ankit K" <ankit.k.nautiyal@intel.com> wrote: >>> +static void intel_dp_compute_fec_config(struct intel_dp *intel_dp, >>> + struct intel_crtc_state *pipe_config) PS. For new stuff please go with crtc_state instead of pipe_config.
diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c index dee15a05e7fd..d82e25d0dc5a 100644 --- a/drivers/gpu/drm/i915/display/intel_dp.c +++ b/drivers/gpu/drm/i915/display/intel_dp.c @@ -2352,6 +2352,26 @@ static int intel_edp_dsc_compute_pipe_bpp(struct intel_dp *intel_dp, return 0; } +static void intel_dp_compute_fec_config(struct intel_dp *intel_dp, + struct intel_crtc_state *pipe_config) +{ + if (pipe_config->fec_enable) + return; + + /* + * Though eDP v1.5 supports FEC with DSC, unlike DP, it is optional. + * Since, FEC is a bandwidth overhead, continue to not enable it for + * eDP. Until, there is a good reason to do so. + */ + if (intel_dp_is_edp(intel_dp)) + return; + + if (intel_dp_is_uhbr(pipe_config)) + return; + + pipe_config->fec_enable = true; +} + int intel_dp_dsc_compute_config(struct intel_dp *intel_dp, struct intel_crtc_state *pipe_config, struct drm_connector_state *conn_state, @@ -2368,15 +2388,7 @@ int intel_dp_dsc_compute_config(struct intel_dp *intel_dp, int num_joined_pipes = intel_crtc_num_joined_pipes(pipe_config); int ret; - /* - * Though eDP v1.5 supports FEC with DSC, unlike DP, it is optional. - * Since, FEC is a bandwidth overhead, continue to not enable it for - * eDP. Until, there is a good reason to do so. - */ - pipe_config->fec_enable = pipe_config->fec_enable || - (!intel_dp_is_edp(intel_dp) && - intel_dp_supports_fec(intel_dp, connector, pipe_config) && - !intel_dp_is_uhbr(pipe_config)); + intel_dp_compute_fec_config(intel_dp, pipe_config); if (!intel_dp_dsc_supports_format(connector, pipe_config->output_format)) return -EINVAL;
Make a separate function for setting fec_enable in crtc_state. Drop the check for FEC support as its already checked while checking for DSC support. Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com> --- drivers/gpu/drm/i915/display/intel_dp.c | 30 +++++++++++++++++-------- 1 file changed, 21 insertions(+), 9 deletions(-)