drivers/gpu/drm/tidss/tidss_crtc.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
TIDSS uses crtc_* fields to propagate its registers and set the
clock rates. So set the CRTC modesetting timing parameters with
the adjusted mode when needed, to set correct values.
Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Jayesh Choudhary <j-choudhary@ti.com>
---
Hello All,
After the DSI fixes[0], TIDSS is using crtc_* timings while programming
hardware[1]. But while testing on TI's J784S4-EVM platform, I noticed
that crtc_timings are not propagated properly.
The display pipeline there looks like:
TIDSS -> CDNS-DSI -> SN65DSI86 bridge -> DisplayPort
Consider the case of 1920x1080 resolution where the EDID mode has clock
of 148500kHz. After adjustment, the clock changes to 148800kHz. While
this change is reflected in mode->clock, its not propagated to
mode->crtc_clock.
[0] provides the **essential** fixes to get DSI working and its
patches are Reviewed and Tested.
The series improves the condition of DSI. I have observed that
800x600 and 1280x1024 modes are working now after [0].
This patch helps to enables other modes. So taking this up as a
delta patch so as to avoid respining v5 of [0].
I hope this approach is okay!
[0]: https://lore.kernel.org/all/20250618-cdns-dsi-impro-v4-0-862c841dbe02@ideasonboard.com/
[1]: https://patchwork.kernel.org/project/dri-devel/patch/20250618-cdns-dsi-impro-v4-3-862c841dbe02@ideasonboard.com/
drivers/gpu/drm/tidss/tidss_crtc.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/tidss/tidss_crtc.c b/drivers/gpu/drm/tidss/tidss_crtc.c
index 17efd77ce7f2..da89fd01c337 100644
--- a/drivers/gpu/drm/tidss/tidss_crtc.c
+++ b/drivers/gpu/drm/tidss/tidss_crtc.c
@@ -91,7 +91,7 @@ static int tidss_crtc_atomic_check(struct drm_crtc *crtc,
struct dispc_device *dispc = tidss->dispc;
struct tidss_crtc *tcrtc = to_tidss_crtc(crtc);
u32 hw_videoport = tcrtc->hw_videoport;
- const struct drm_display_mode *mode;
+ struct drm_display_mode *mode;
enum drm_mode_status ok;
dev_dbg(ddev->dev, "%s\n", __func__);
@@ -108,6 +108,9 @@ static int tidss_crtc_atomic_check(struct drm_crtc *crtc,
return -EINVAL;
}
+ if (drm_atomic_crtc_needs_modeset(crtc_state))
+ drm_mode_set_crtcinfo(mode, 0);
+
return dispc_vp_bus_check(dispc, hw_videoport, crtc_state);
}
--
2.34.1
Hi, On 24/06/2025 11:04, Jayesh Choudhary wrote: > TIDSS uses crtc_* fields to propagate its registers and set the > clock rates. So set the CRTC modesetting timing parameters with > the adjusted mode when needed, to set correct values. > > Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> > Signed-off-by: Jayesh Choudhary <j-choudhary@ti.com> > --- > > Hello All, > > After the DSI fixes[0], TIDSS is using crtc_* timings while programming > hardware[1]. But while testing on TI's J784S4-EVM platform, I noticed > that crtc_timings are not propagated properly. > > The display pipeline there looks like: > TIDSS -> CDNS-DSI -> SN65DSI86 bridge -> DisplayPort > > Consider the case of 1920x1080 resolution where the EDID mode has clock > of 148500kHz. After adjustment, the clock changes to 148800kHz. While > this change is reflected in mode->clock, its not propagated to > mode->crtc_clock. > > [0] provides the **essential** fixes to get DSI working and its > patches are Reviewed and Tested. > The series improves the condition of DSI. I have observed that > 800x600 and 1280x1024 modes are working now after [0]. > > This patch helps to enables other modes. So taking this up as a > delta patch so as to avoid respining v5 of [0]. > I hope this approach is okay! Yes, I think this makes sense. I'll pick this up after my series has been merged. Tomi > > [0]: https://lore.kernel.org/all/20250618-cdns-dsi-impro-v4-0-862c841dbe02@ideasonboard.com/ > [1]: https://patchwork.kernel.org/project/dri-devel/patch/20250618-cdns-dsi-impro-v4-3-862c841dbe02@ideasonboard.com/ > > drivers/gpu/drm/tidss/tidss_crtc.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/tidss/tidss_crtc.c b/drivers/gpu/drm/tidss/tidss_crtc.c > index 17efd77ce7f2..da89fd01c337 100644 > --- a/drivers/gpu/drm/tidss/tidss_crtc.c > +++ b/drivers/gpu/drm/tidss/tidss_crtc.c > @@ -91,7 +91,7 @@ static int tidss_crtc_atomic_check(struct drm_crtc *crtc, > struct dispc_device *dispc = tidss->dispc; > struct tidss_crtc *tcrtc = to_tidss_crtc(crtc); > u32 hw_videoport = tcrtc->hw_videoport; > - const struct drm_display_mode *mode; > + struct drm_display_mode *mode; > enum drm_mode_status ok; > > dev_dbg(ddev->dev, "%s\n", __func__); > @@ -108,6 +108,9 @@ static int tidss_crtc_atomic_check(struct drm_crtc *crtc, > return -EINVAL; > } > > + if (drm_atomic_crtc_needs_modeset(crtc_state)) > + drm_mode_set_crtcinfo(mode, 0); > + > return dispc_vp_bus_check(dispc, hw_videoport, crtc_state); > } >
Hi, On 24/06/2025 11:04, Jayesh Choudhary wrote: > TIDSS uses crtc_* fields to propagate its registers and set the > clock rates. So set the CRTC modesetting timing parameters with > the adjusted mode when needed, to set correct values. > > Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> > Signed-off-by: Jayesh Choudhary <j-choudhary@ti.com> > --- > > Hello All, > > After the DSI fixes[0], TIDSS is using crtc_* timings while programming > hardware[1]. But while testing on TI's J784S4-EVM platform, I noticed > that crtc_timings are not propagated properly. > > The display pipeline there looks like: > TIDSS -> CDNS-DSI -> SN65DSI86 bridge -> DisplayPort > > Consider the case of 1920x1080 resolution where the EDID mode has clock > of 148500kHz. After adjustment, the clock changes to 148800kHz. While > this change is reflected in mode->clock, its not propagated to > mode->crtc_clock. Hmm, so CDNS-DSI changes the adjusted_mode->clock, but in the end tidss doesn't actually use the adjusted clock at all? I'm pretty sure I tested that... I need to try it (and this) again. Tomi > > [0] provides the **essential** fixes to get DSI working and its > patches are Reviewed and Tested. > The series improves the condition of DSI. I have observed that > 800x600 and 1280x1024 modes are working now after [0]. > > This patch helps to enables other modes. So taking this up as a > delta patch so as to avoid respining v5 of [0]. > I hope this approach is okay! > > [0]: https://lore.kernel.org/all/20250618-cdns-dsi-impro-v4-0-862c841dbe02@ideasonboard.com/ > [1]: https://patchwork.kernel.org/project/dri-devel/patch/20250618-cdns-dsi-impro-v4-3-862c841dbe02@ideasonboard.com/ > > drivers/gpu/drm/tidss/tidss_crtc.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/tidss/tidss_crtc.c b/drivers/gpu/drm/tidss/tidss_crtc.c > index 17efd77ce7f2..da89fd01c337 100644 > --- a/drivers/gpu/drm/tidss/tidss_crtc.c > +++ b/drivers/gpu/drm/tidss/tidss_crtc.c > @@ -91,7 +91,7 @@ static int tidss_crtc_atomic_check(struct drm_crtc *crtc, > struct dispc_device *dispc = tidss->dispc; > struct tidss_crtc *tcrtc = to_tidss_crtc(crtc); > u32 hw_videoport = tcrtc->hw_videoport; > - const struct drm_display_mode *mode; > + struct drm_display_mode *mode; > enum drm_mode_status ok; > > dev_dbg(ddev->dev, "%s\n", __func__); > @@ -108,6 +108,9 @@ static int tidss_crtc_atomic_check(struct drm_crtc *crtc, > return -EINVAL; > } > > + if (drm_atomic_crtc_needs_modeset(crtc_state)) > + drm_mode_set_crtcinfo(mode, 0); > + > return dispc_vp_bus_check(dispc, hw_videoport, crtc_state); > } >
Hi Tomi, On Tue Jun 24, 2025 at 1:47 PM CEST, Tomi Valkeinen wrote: > On 24/06/2025 11:04, Jayesh Choudhary wrote: > > TIDSS uses crtc_* fields to propagate its registers and set the > > clock rates. So set the CRTC modesetting timing parameters with > > the adjusted mode when needed, to set correct values. > > > > Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> > > Signed-off-by: Jayesh Choudhary <j-choudhary@ti.com> > > --- > > > > Hello All, > > > > After the DSI fixes[0], TIDSS is using crtc_* timings while programming > > hardware[1]. But while testing on TI's J784S4-EVM platform, I noticed > > that crtc_timings are not propagated properly. > > > > The display pipeline there looks like: > > TIDSS -> CDNS-DSI -> SN65DSI86 bridge -> DisplayPort > > > > Consider the case of 1920x1080 resolution where the EDID mode has clock > > of 148500kHz. After adjustment, the clock changes to 148800kHz. While > > this change is reflected in mode->clock, its not propagated to > > mode->crtc_clock. > > Hmm, so CDNS-DSI changes the adjusted_mode->clock, but in the end tidss > doesn't actually use the adjusted clock at all? I'm pretty sure I tested > that... I need to try it (and this) again. FWIW, without this patch, DSI isn't working on my board (DSI -> DSI85 -> eDP). At least without the (now dropped) patch "drm/tidss: Adjust the pclk based on the HW capabilities" [1]. That is, it was working with v3 of your DSI patch series, but not with v4. I'll need this patch together with v4 to get DSI working. Maybe that helps, -michael [1] https://lore.kernel.org/all/20250402-cdns-dsi-impro-v2-3-4a093eaa5e27@ideasonboard.com/
© 2016 - 2025 Red Hat, Inc.