Re: [PATCH] drm/i915: ensure HDMI port is disabled inside set_infoframes

Re: [PATCH] drm/i915: ensure HDMI port is disabled inside set_infoframes

[Paulo Zanoni]

Hi

2012/6/12 Daniel Vetter <daniel.vetter@ffwll.ch>:
> This function is supposed to be used at mode set time, so prevent
> against future mistakes by adding a WARN().
>
> Based on a patch by Paulo Zanoni, with the check extracted into a
> little assert_hdmi_port_disabled helper added to make things self
> documenting and move the assert stuff out of line.

I was going to write the patch, but thanks for doing it :)

Comments below:

>
> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
> Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  drivers/gpu/drm/i915/intel_hdmi.c |   23 +++++++++++++++++++++++
>  1 files changed, 23 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
> index 69637db..aac9a5d 100644
> --- a/drivers/gpu/drm/i915/intel_hdmi.c
> +++ b/drivers/gpu/drm/i915/intel_hdmi.c
> @@ -37,6 +37,19 @@
>  #include "i915_drm.h"
>  #include "i915_drv.h"
>
> +static void
> +assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi)
> +{
> +       struct drm_device *dev = intel_hdmi->base.base.dev;
> +       struct drm_i915_private *dev_priv = dev->dev_private;
> +       uint32_t enabled_bits;
> +
> +       enabled_bits = IS_HASWELL(dev) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;

My crystal ball tells me a check for IS_HASWELL_OR_NEWER would prevent
more changes to this code in the future.

> +
> +       WARN(I915_READ(intel_hdmi->sdvox_reg) & enabled_bits,
> +            "HDMI port enabled, expectin disabled\n");

expectinG?

> +}
> +
>  struct intel_hdmi *enc_to_intel_hdmi(struct drm_encoder *encoder)
>  {
>        return container_of(encoder, struct intel_hdmi, base.base);
> @@ -334,6 +347,8 @@ static void g4x_set_infoframes(struct drm_encoder *encoder,
>        u32 val = I915_READ(reg);
>        u32 port;
>
> +       assert_hdmi_port_disabled(intel_hdmi);
> +
>        /* If the registers were not initialized yet, they might be zeroes,
>         * which means we're selecting the AVI DIP and we're setting its
>         * frequency to once. This seems to really confuse the HW and make
> @@ -395,6 +410,8 @@ static void ibx_set_infoframes(struct drm_encoder *encoder,
>        u32 val = I915_READ(reg);
>        u32 port;
>
> +       assert_hdmi_port_disabled(intel_hdmi);
> +
>        /* See the big comment in g4x_set_infoframes() */
>        val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
>
> @@ -451,6 +468,8 @@ static void cpt_set_infoframes(struct drm_encoder *encoder,
>        u32 reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
>        u32 val = I915_READ(reg);
>
> +       assert_hdmi_port_disabled(intel_hdmi);
> +
>        /* See the big comment in g4x_set_infoframes() */
>        val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
>
> @@ -484,6 +503,8 @@ static void vlv_set_infoframes(struct drm_encoder *encoder,
>        u32 reg = VLV_TVIDEO_DIP_CTL(intel_crtc->pipe);
>        u32 val = I915_READ(reg);
>
> +       assert_hdmi_port_disabled(intel_hdmi);
> +
>        /* See the big comment in g4x_set_infoframes() */
>        val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
>
> @@ -516,6 +537,8 @@ static void hsw_set_infoframes(struct drm_encoder *encoder,
>        u32 reg = HSW_TVIDEO_DIP_CTL(intel_crtc->pipe);
>        u32 val = I915_READ(reg);
>
> +       assert_hdmi_port_disabled(intel_hdmi);
> +
>        if (!intel_hdmi->has_hdmi_sink) {
>                I915_WRITE(reg, 0);
>                POSTING_READ(reg);
> --
> 1.7.7.6
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx



-- 
Paulo Zanoni

[PATCH] drm/i915: ensure HDMI port is disabled inside set_infoframes

[Daniel Vetter]

This function is supposed to be used at mode set time, so prevent
against future mistakes by adding a WARN().

Based on a patch by Paulo Zanoni, with the check extracted into a
little assert_hdmi_port_disabled helper added to make things self
documenting and move the assert stuff out of line.

Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/i915/intel_hdmi.c |   23 +++++++++++++++++++++++
 1 files changed, 23 insertions(+), 0 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 69637db..aac9a5d 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -37,6 +37,19 @@
 #include "i915_drm.h"
 #include "i915_drv.h"
 
+static void
+assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi)
+{
+	struct drm_device *dev = intel_hdmi->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t enabled_bits;
+
+	enabled_bits = IS_HASWELL(dev) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;
+
+	WARN(I915_READ(intel_hdmi->sdvox_reg) & enabled_bits,
+	     "HDMI port enabled, expectin disabled\n");
+}
+
 struct intel_hdmi *enc_to_intel_hdmi(struct drm_encoder *encoder)
 {
 	return container_of(encoder, struct intel_hdmi, base.base);
@@ -334,6 +347,8 @@ static void g4x_set_infoframes(struct drm_encoder *encoder,
 	u32 val = I915_READ(reg);
 	u32 port;
 
+	assert_hdmi_port_disabled(intel_hdmi);
+
 	/* If the registers were not initialized yet, they might be zeroes,
 	 * which means we're selecting the AVI DIP and we're setting its
 	 * frequency to once. This seems to really confuse the HW and make
@@ -395,6 +410,8 @@ static void ibx_set_infoframes(struct drm_encoder *encoder,
 	u32 val = I915_READ(reg);
 	u32 port;
 
+	assert_hdmi_port_disabled(intel_hdmi);
+
 	/* See the big comment in g4x_set_infoframes() */
 	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
 
@@ -451,6 +468,8 @@ static void cpt_set_infoframes(struct drm_encoder *encoder,
 	u32 reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
 
+	assert_hdmi_port_disabled(intel_hdmi);
+
 	/* See the big comment in g4x_set_infoframes() */
 	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
 
@@ -484,6 +503,8 @@ static void vlv_set_infoframes(struct drm_encoder *encoder,
 	u32 reg = VLV_TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
 
+	assert_hdmi_port_disabled(intel_hdmi);
+
 	/* See the big comment in g4x_set_infoframes() */
 	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
 
@@ -516,6 +537,8 @@ static void hsw_set_infoframes(struct drm_encoder *encoder,
 	u32 reg = HSW_TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
 
+	assert_hdmi_port_disabled(intel_hdmi);
+
 	if (!intel_hdmi->has_hdmi_sink) {
 		I915_WRITE(reg, 0);
 		POSTING_READ(reg);
-- 
1.7.7.6

Re: [PATCH] drm/i915: ensure HDMI port is disabled inside set_infoframes

[Eugeni Dodonov]

On 06/12/2012 11:36 AM, Daniel Vetter wrote:
> This function is supposed to be used at mode set time, so prevent
> against future mistakes by adding a WARN().
> 
> Based on a patch by Paulo Zanoni, with the check extracted into a
> little assert_hdmi_port_disabled helper added to make things self
> documenting and move the assert stuff out of line.
> 
> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
> Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Apart from a typo in 'expectin' below:

Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>

> +	WARN(I915_READ(intel_hdmi->sdvox_reg) & enabled_bits,
> +	     "HDMI port enabled, expectin disabled\n");
                                 ^^
Eugeni

Re: [PATCH] drm/i915: ensure HDMI port is disabled inside set_infoframes

[Eugeni Dodonov]

On 06/12/2012 11:35 AM, Paulo Zanoni wrote:
> My crystal ball tells me a check for IS_HASWELL_OR_NEWER would prevent
> more changes to this code in the future.

I think that a .has_ddi (or similar) flag would be better than checking
for post-Haswell variants, as not necessarily all of them could work in
the same way. But we are still far far away from having to face this
case case anyway :).

Eugeni

Re: [PATCH] drm/i915: ensure HDMI port is disabled inside set_infoframes

[Daniel Vetter]

On Tue, Jun 12, 2012 at 12:43:42PM -0300, Eugeni Dodonov wrote:
> On 06/12/2012 11:36 AM, Daniel Vetter wrote:
> > This function is supposed to be used at mode set time, so prevent
> > against future mistakes by adding a WARN().
> > 
> > Based on a patch by Paulo Zanoni, with the check extracted into a
> > little assert_hdmi_port_disabled helper added to make things self
> > documenting and move the assert stuff out of line.
> > 
> > Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
> > Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> 
> Apart from a typo in 'expectin' below:
> 
> Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>

Queued for -next with the 'g' thrown in, thanks for the review.
-Daniel
-- 
Daniel Vetter
Mail: daniel@ffwll.ch
Mobile: +41 (0)79 365 57 48