[Intel-gfx] [PATCH 2/2] drm/i915: Use the fixed pixel clock for eDP in intel_dp_set_m_n()

Takashi Iwai tiwai at suse.de
Mon Mar 18 10:29:16 CET 2013


At Sun, 17 Mar 2013 23:12:03 +0100,
Daniel Vetter wrote:
> 
> On Tue, Mar 12, 2013 at 04:32:28PM +0100, Takashi Iwai wrote:
> > The eDP output on HP Z1 is still broken when X is started even after
> > fixing the infinite link-train loop.  The regression was introduced in
> > 3.6 kernel for cleaning up the mode clock handling code in intel_dp.c.
> > 
> > In the past, the clock of the reference mode was modified in
> > intel_dp_mode_fixup() in the case of eDP fixed clock, and this clock was
> > used for calculating in intel_dp_set_m_n().  This override was removed,
> > thus the wrong mode clock is used for the calculation, resulting in a
> > psychedelic smoking output in the end.
> > 
> > This patch corrects the clock to be used in the place.
> > 
> > Cc: <stable at vger.kernel.org>
> > Signed-off-by: Takashi Iwai <tiwai at suse.de>
> 
> I truly hate this mess of dotclock vs portclock vs. whatever. Can you pls
> apply a little bikeshed and use the existing intel_edp_target_clock like
> in ironlake_set_m_n? And if you have the regressing commit a little
> citation to assign the blame (it's probably me) would be good.

OK, the revised patch is below.


thanks,

Takashi

---
From: Takashi Iwai <tiwai at suse.de>
Subject: [PATCH v2] drm/i915: Use the fixed pixel clock for eDP in intel_dp_set_m_n()

The eDP output on HP Z1 is still broken when X is started even after
fixing the infinite link-train loop.  The regression was introduced in
3.6 kernel for cleaning up the mode clock handling code in intel_dp.c
by the commit [71244653: drm/i915: adjusted_mode->clock in the dp
mode_fix].

In the past, the clock of the reference mode was modified in
intel_dp_mode_fixup() in the case of eDP fixed clock, and this clock was
used for calculating in intel_dp_set_m_n().  This override was removed,
thus the wrong mode clock is used for the calculation, resulting in a
psychedelic smoking output in the end.

This patch corrects the clock to be used in the place.

v1->v2: Use intel_edp_target_clock() for checking eDP fixed clock
instead of open code as in ironlake_set_m_n().

Cc: <stable at vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai at suse.de>
---
 drivers/gpu/drm/i915/intel_dp.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 6f728e5..2606811 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -820,6 +820,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	struct intel_link_m_n m_n;
 	int pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	int target_clock;
 
 	/*
 	 * Find the lane count in the intel_encoder private
@@ -835,13 +836,22 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 		}
 	}
 
+	target_clock = mode->clock;
+	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
+		if (intel_encoder->type == INTEL_OUTPUT_EDP) {
+			target_clock = intel_edp_target_clock(intel_encoder,
+							      mode);
+			break;
+		}
+	}
+
 	/*
 	 * Compute the GMCH and Link ratios. The '3' here is
 	 * the number of bytes_per_pixel post-LUT, which we always
 	 * set up for 8-bits of R/G/B, or 3 bytes total.
 	 */
 	intel_link_compute_m_n(intel_crtc->bpp, lane_count,
-			       mode->clock, adjusted_mode->clock, &m_n);
+			       target_clock, adjusted_mode->clock, &m_n);
 
 	if (IS_HASWELL(dev)) {
 		I915_WRITE(PIPE_DATA_M1(cpu_transcoder),
-- 
1.8.2




More information about the Intel-gfx mailing list