[Intel-gfx] [PATCH] drm/i915/display: Enable DP Display Audio WA
Shankar, Uma
uma.shankar at intel.com
Tue Apr 7 13:04:57 UTC 2020
> -----Original Message-----
> From: Jani Nikula <jani.nikula at linux.intel.com>
> Sent: Tuesday, April 7, 2020 4:45 PM
> To: Shankar, Uma <uma.shankar at intel.com>; intel-gfx at lists.freedesktop.org
> Cc: Vehmanen, Kai <kai.vehmanen at intel.com>
> Subject: Re: [Intel-gfx] [PATCH] drm/i915/display: Enable DP Display Audio WA
>
> On Tue, 07 Apr 2020, Uma Shankar <uma.shankar at intel.com> wrote:
> > Enable Display Audio WA #1406928334 for 4k+VDSC usecase on DP
> > encoders.
>
> I didn't actually read the wa, but please describe the main points here.
Sure will add more details.
> >
> > Signed-off-by: Uma Shankar <uma.shankar at intel.com>
> > ---
> > drivers/gpu/drm/i915/display/intel_audio.c | 110 +++++++++++++++++++++
> > drivers/gpu/drm/i915/i915_reg.h | 16 +++
> > 2 files changed, 126 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_audio.c
> > b/drivers/gpu/drm/i915/display/intel_audio.c
> > index 950160f1a89f..035f2949e9c0 100644
> > --- a/drivers/gpu/drm/i915/display/intel_audio.c
> > +++ b/drivers/gpu/drm/i915/display/intel_audio.c
> > @@ -512,6 +512,112 @@ static void hsw_audio_codec_disable(struct
> intel_encoder *encoder,
> > mutex_unlock(&dev_priv->av_mutex);
> > }
> >
> > +static void enable_audio_dsc_wa(struct intel_encoder *encoder,
> > + const struct intel_crtc_state *crtc_state) {
> > + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
>
> Please name the variable i915 now that we can.
Ok.
> > + struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
> > + enum pipe pipe = crtc->pipe;
> > + u64 link_clks_available, link_clks_required, link_clk;
> > + u64 tu_data, tu_line, link_clks_active;
> > + u64 hblank_rise, hblank_early_prog, samples_room;
> > + u64 h_active, h_total, hblank_delta, pixel_clk, v_total, lanes;
> > + u64 vdsc_bpp, fec_coeff, refresh_rate, cdclk;
>
> Most of these probably don't need to be u64.
Yeah will drop u64.
> > + u64 rounding_factor = 100000;
>
> Macro.
Sure, will update this.
> > + u32 val;
> > +
> > + val = intel_de_read(dev_priv, AUD_CONFIG_BE);
> > +
> > + if (INTEL_GEN(dev_priv) == 11)
> > + val |= HBLANK_EARLY_ENABLE_ICL(pipe);
> > + else if (INTEL_GEN(dev_priv) >= 12)
> > + val |= HBLANK_EARLY_ENABLE_TGL(pipe);
> > +
>
> So if it's gen < 11 without dsc, you'll just read and write the register? Probably want
> to bail out early.
Right, will fix this.
> > + if (crtc_state->dsc.compression_enable &&
> > + (crtc_state->hw.adjusted_mode.hdisplay >= 3840 &&
> > + crtc_state->hw.adjusted_mode.vdisplay >= 2160)) {
> > + h_active = crtc_state->hw.adjusted_mode.hdisplay;
> > + h_total = crtc_state->hw.adjusted_mode.htotal;
> > + v_total = crtc_state->hw.adjusted_mode.vtotal;
> > + hblank_rise = crtc_state->hw.adjusted_mode.hsync_start;
> > + pixel_clk = crtc_state->hw.adjusted_mode.clock;
> > + link_clk = crtc_state->port_clock;
> > + lanes = crtc_state->lane_count;
> > + vdsc_bpp = crtc_state->pipe_bpp;
> > + refresh_rate = crtc_state->hw.adjusted_mode.vrefresh;
> > + cdclk = dev_priv->cdclk.hw.cdclk;
> > + /* fec= 0.972261, using rounding multiplier of 1000000 */
> > + fec_coeff = 972261;
> > +
> > + if (!(h_active && link_clk && lanes && vdsc_bpp && cdclk)) {
> > + drm_err(&dev_priv->drm, "Null Parameters received\n");
>
> What's the user going to do with that information?
>
> > + return;
>
> No need to write HBLANK_EARLY_ENABLE_* then?
Ideally we should not get Null params here, will add a WARN_ON instead.
> > + }
> > +
> > + drm_dbg_kms(&dev_priv->drm, "h_active = %llu link_clk = %llu :"
> > + "lanes = %llu vdsc_bpp = %llu cdclk = %llu\n",
> > + h_active, link_clk, lanes, vdsc_bpp, cdclk);
> > +
> > + link_clks_available = ((((h_total - h_active) *
> > + ((link_clk * rounding_factor) /
> > + pixel_clk)) / rounding_factor) - 28);
> > +
> > + link_clks_required = DIV_ROUND_UP(192000, (refresh_rate *
> > + v_total)) *
> > + ((48 / lanes) + 2);
> > +
> > + if (link_clks_available > link_clks_required)
> > + hblank_delta = 32;
> > + else
> > + hblank_delta = DIV_ROUND_UP(((((5 * rounding_factor) /
> > + link_clk) + ((5 *
> > + rounding_factor) / cdclk)) *
> > + pixel_clk), rounding_factor);
> > +
> > + tu_data = (pixel_clk * vdsc_bpp * 8) / ((link_clk *
> > + lanes * fec_coeff) /
> > + 1000000);
> > + tu_line = (((h_active * link_clk * fec_coeff) / 1000000) /
> > + (64 * pixel_clk));
> > + link_clks_active = (tu_line - 1) * 64 + tu_data;
> > + hblank_rise = ((link_clks_active + 6 *
> > + DIV_ROUND_UP(link_clks_active, 250) + 4) *
> > + ((pixel_clk * rounding_factor) /
> > + link_clk)) / rounding_factor;
> > +
> > + hblank_early_prog = h_active - hblank_rise + hblank_delta;
> > +
> > + if (hblank_early_prog < 32) {
> > + val &= ~HBLANK_START_COUNT_MASK(pipe);
> > + val |=
> HBLANK_START_COUNT(HBLANK_START_COUNT_32, pipe);
> > + } else if (hblank_early_prog < 64) {
> > + val &= ~HBLANK_START_COUNT_MASK(pipe);
> > + val |=
> HBLANK_START_COUNT(HBLANK_START_COUNT_64, pipe);
> > + } else if (hblank_early_prog < 96) {
> > + val &= ~HBLANK_START_COUNT_MASK(pipe);
> > + val |=
> HBLANK_START_COUNT(HBLANK_START_COUNT_96, pipe);
> > + } else {
> > + val &= ~HBLANK_START_COUNT_MASK(pipe);
> > + val |=
> HBLANK_START_COUNT(HBLANK_START_COUNT_128, pipe);
> > + }
> > +
> > + samples_room = ((((h_total - h_active) *
> > + ((link_clk * rounding_factor) /
> > + pixel_clk)) / rounding_factor) - 12) /
> > + ((48 / lanes) + 2);
>
> Please abstract the calculation of hblank_early_prog and samples_room to separate
> helpers, and pass in crtc_state. You also don't have to make everything you need
> from crtc_state a local variable.
Sure, will modify this.
> > +
> > + if (samples_room < 3) {
> > + val &= ~NUMBER_SAMPLES_PER_LINE_MASK(pipe);
> > + val |= NUMBER_SAMPLES_PER_LINE(samples_room, pipe);
> > + } else {
> > + val &= ~NUMBER_SAMPLES_PER_LINE_MASK(pipe);
> > + val |= NUMBER_SAMPLES_PER_LINE(0x0, pipe);
> > + }
> > + }
> > +
> > + intel_de_write(dev_priv, AUD_CONFIG_BE, val); }
> > +
> > static void hsw_audio_codec_enable(struct intel_encoder *encoder,
> > const struct intel_crtc_state *crtc_state,
> > const struct drm_connector_state *conn_state)
> @@ -529,6
> > +635,10 @@ static void hsw_audio_codec_enable(struct intel_encoder
> > *encoder,
> >
> > mutex_lock(&dev_priv->av_mutex);
> >
> > + /* Enable Audio WA for 4k DSC usecases */
> > + if (encoder->type == INTEL_OUTPUT_DP)
> > + enable_audio_dsc_wa(encoder, crtc_state);
> > +
> > /* Enable audio presence detect, invalidate ELD */
> > tmp = intel_de_read(dev_priv, HSW_AUD_PIN_ELD_CP_VLD);
> > tmp |= AUDIO_OUTPUT_ENABLE(cpu_transcoder);
> > diff --git a/drivers/gpu/drm/i915/i915_reg.h
> > b/drivers/gpu/drm/i915/i915_reg.h index 8cebb7a86b8c..ff293e9d5a63
> > 100644
> > --- a/drivers/gpu/drm/i915/i915_reg.h
> > +++ b/drivers/gpu/drm/i915/i915_reg.h
> > @@ -9395,6 +9395,22 @@ enum {
> > #define AUD_PIN_BUF_CTL _MMIO(0x48414)
> > #define AUD_PIN_BUF_ENABLE REG_BIT(31)
> >
> > +/* Display Audio Config Reg */
> > +#define AUD_CONFIG_BE _MMIO(0x65ef0)
> > +#define HBLANK_EARLY_ENABLE_ICL(pipe) (0x1 << (20 - (pipe)))
> > +#define HBLANK_EARLY_ENABLE_TGL(pipe) (0x1 << (24 + (pipe)))
> > +#define HBLANK_START_COUNT_MASK(pipe) (0x7 << (3 + ((pipe) * 6)))
> > +#define HBLANK_START_COUNT(val, pipe) (((val) & 0x7) << (3 +
> ((pipe)) * 6))
>
> I'd swap pipe and val params around. Same below.
>
> You might find REG_FIELD_PREP etc. useful.
Ok Sure.
> > +#define NUMBER_SAMPLES_PER_LINE_MASK(pipe) (0x3 << ((pipe) * 6))
> > +#define NUMBER_SAMPLES_PER_LINE(val, pipe) (((val) & 0x3) << ((pipe) *
> 6))
> > +
> > +#define HBLANK_START_COUNT_8 0x0
> > +#define HBLANK_START_COUNT_16 0x1
> > +#define HBLANK_START_COUNT_32 0x2
> > +#define HBLANK_START_COUNT_64 0x3
> > +#define HBLANK_START_COUNT_96 0x4
> > +#define HBLANK_START_COUNT_128 0x5
>
> They're just numbers, hex doesn't provide anything useful here.
Will update.
Thanks Jani for review and useful pointers. Will send out next version addressing all comments.
Regards,
Uma Shankar
> > +
> > /*
> > * HSW - ICL power wells
> > *
>
> --
> Jani Nikula, Intel Open Source Graphics Center
More information about the Intel-gfx
mailing list