drm/i915/audio: extend get_saved_enc() to support more scenarios
Subsystems (2)
PATCH 1/2 drm/i915/audio: extend get_saved_enc() to support more scenarios
From: Libin Yang <[email protected]>
When bootup, audio driver may not know it is MST or not. The audio driver will poll all the port & pipe combinations in either MST or Non-MST mode. get_saved_enc() should handle this situation.
--- drivers/gpu/drm/i915/intel_audio.c | 32 ++++++++++++++++++++++++++++---- 1 file changed, 28 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/intel_audio.c +28 -4
@@ -737,25 +737,49 @@ static int i915_audio_component_get_cdclk_freq(struct device *kdev)return dev_priv->cdclk_freq; } +/* +* get the intel_encoder according to the parameter port and pipe +* intel_encoder is saved by the index of pipe +* MST & (pipe >= 0): return the av_enc_map[pipe], +* when port is matched +* MST & (pipe < 0): this is invalid +* Non-MST & (pipe >= 0): only pipe = 0 (the first device entry) +* will get the right intel_encoder with port matched +* Non-MST & (pipe < 0): get the right intel_encoder with port matched +*/ static struct intel_encoder *get_saved_enc (struct drm_i915_private *dev_priv,int port,int pipe) { +struct intel_encoder *encoder;if (WARN_ON (pipe >=I915_MAX_PIPES ))return NULL ;/* MST */ -if (pipe >=0 ) -return dev_priv->av_enc_map[pipe]; +if (pipe >=0 ) { + encoder = dev_priv->av_enc_map[pipe]; +/* +* when bootup, audio driver may not know it is +* MST or not. So it will poll all the port & pipe +* combinations +*/ +if (encoder !=NULL && encoder->port == port && + encoder->type ==INTEL_OUTPUT_DP_MST ) +return encoder; + }/* Non-MST */ -for_each_pipe (dev_priv, pipe) { -struct intel_encoder *encoder; +if (pipe >0 ) +return NULL ; +for_each_pipe (dev_priv, pipe) { encoder = dev_priv->av_enc_map[pipe];if (encoder ==NULL )continue ; +if (encoder->type ==INTEL_OUTPUT_DP_MST ) +continue ; +if (port == encoder->port)return encoder; }
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 2/2 drm/i915/audio: Extend audio sync rate support for DP MST
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_audio.c +2 -1
@@ -807,7 +807,8 @@ static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {DRM_DEBUG_KMS ("Not valid for port %c\n" ,port_name (port)); err = -ENODEV ; goto unlock;
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Jani Nikula
maintainerRe: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MST
Quoted 15 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/i915/intel_audio.c +0
-0
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 811
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +2
-1@@ -807,7 +807,8 @@ static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {
I think the better option is to make absolutely sure we never store other kinds of encoders in dev_priv->av_enc_map[pipe] to begin with. I think that's true already, but please add
if (WARN_ON(intel_encoder->type != INTEL_OUTPUT_HDMI &&
intel_encoder->type != INTEL_OUTPUT_DP &&
intel_encoder->type != INTEL_OUTPUT_DP_MST))
return;near the beginning of intel_audio_codec_enable(), and remove the type checks here. This reduces the confusion about different kinds of checks after calling get_saved_enc().
BR,
Jani.
Quoted 3 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0DRM_DEBUG_KMS ("Not valid for port %c\n" ,port_name (port)); err = -ENODEV ; goto unlock;
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MST
Quoted 27 lines · Jani Nikula
Quoted 15 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/i915/intel_audio.c +0
-0drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 811
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +2
-1@@ -807,7 +807,8 @@ static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {I think the better option is to make absolutely sure we never store other kinds of encoders in dev_priv->av_enc_map[pipe] to begin with. I think that's true already, but please add
Quoted 8 lines
if (WARN_ON(intel_encoder->type != INTEL_OUTPUT_HDMI && intel_encoder->type != INTEL_OUTPUT_DP && intel_encoder->type != INTEL_OUTPUT_DP_MST)) return;near the beginning of
intel_audio_codec_enable(), and remove the type checks here. This reduces the confusion about different kinds of checks after callingget_saved_enc().
This while encoder->type thing is pretty broken actually. Currently we're semi-randomly flipping DDI encoders between UNKNOWN,HDMI and DP. I want to eliminate that so that their type will always be just "DDI". For iguring out what kind of signal we are driving out you should be looking at crtc_state->output_types.
Signature · 6 lines
Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MST
Quoted 51 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Wednesday, November 30, 2016 12:50 AM
To: Jani Nikula <[email protected]>
Cc: Yang, Libin <[email protected]>; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Subject: Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MSTQuoted 11 lines · Jani Nikula
Quoted 10 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c static int
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -807,7 +807,8 @@ static inti915_audio_component_sync_audio_rate(struct device *kdev, int port,
Quoted 11 lines
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 811
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +2
-1intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {I think the better option is to make absolutely sure we never store other kinds of encoders in dev_priv->av_enc_map[pipe] to begin with. I think that's true already, but please add
Quoted 8 lines
if (WARN_ON(intel_encoder->type != INTEL_OUTPUT_HDMI && intel_encoder->type != INTEL_OUTPUT_DP && intel_encoder->type != INTEL_OUTPUT_DP_MST)) return;near the beginning of
intel_audio_codec_enable(), and remove the type checks here. This reduces the confusion about different kinds of checks after callingget_saved_enc().
Quoted 5 lines
This while
encoder->typething is pretty broken actually. Currently we're semi- randomly flipping DDI encoders between UNKNOWN,HDMI and DP.
I want to eliminate that so that their type will always be just "DDI". For iguring out what kind of signal we are driving out you should be looking atcrtc_state->output_types.
Do you mean the code should be like:
@@ -807,10 +807,7 @@ static int i915_audio_component_sync_audio_rate(struct device *kdev, int port,
/* 1. get the pipe */
intel_encoder = get_saved_enc(dev_priv, port, pipe);
- if (!intel_encoder || !intel_encoder->base.crtc ||
- (intel_encoder->type != INTEL_OUTPUT_HDMI &&
- intel_encoder->type != INTEL_OUTPUT_DP &&
- intel_encoder->type != INTEL_OUTPUT_DP_MST)) {
+ if (!intel_encoder || !intel_encoder->base.crtc) {
DRM_DEBUG_KMS("Not valid for port %c\n", port_name(port));
err = -ENODEV;
goto unlock;
@@ -819,7 +816,12 @@ static int i915_audio_component_sync_audio_rate(struct device *kdev, int port,
/* pipe passed from the audio driver will be -1 for Non-MST case */
crtc = to_intel_crtc(intel_encoder->base.crtc);
crtc_state = crtc->config;
+ if (WARN_ON((crtc_state->output_types &
+ ((1 << INTEL_OUTPUT_HDMI) |
+ (1 << INTEL_OUTPUT_DP) |
+ (1 << INTEL_OUTPUT_DP_MST))) == 0))
+ return 0;
+
pipe = crtc->pipe;
I did the test, and it seems intel_encoder->type can show the correct type on my platform. It is similar with crtc_state->output_types.
Regards,
Libin
Quoted 4 lines
--
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MST
In reply to Jani NikulaQuoted 40 lines
-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Wednesday, November 30, 2016 12:34 AM
To: Yang, Libin <[email protected]>; [email protected]; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Cc: Yang, Libin <[email protected]>
Subject: Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MSTdrivers/gpu/drm/i915/intel_audio.c static int
Quoted 15 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -807,7 +807,8 @@ static inti915_audio_component_sync_audio_rate(struct device *kdev, int port,
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 811
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +2
-1intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {I think the better option is to make absolutely sure we never store other kinds of encoders in dev_priv->av_enc_map[pipe] to begin with. I think that's true already, but please add
Quoted 8 lines
if (WARN_ON(intel_encoder->type != INTEL_OUTPUT_HDMI && intel_encoder->type != INTEL_OUTPUT_DP && intel_encoder->type != INTEL_OUTPUT_DP_MST)) return;near the beginning of
intel_audio_codec_enable(), and remove the type checks here. This reduces the confusion about different kinds of checks after callingget_saved_enc().
We can't put the check at the beginning of the function.
We still need the check in the same place as we need get intel_encoder at first.
Regards,
Libin
Quoted 5 lines
BR,
Jani.
Quoted 7 lines
drivers/gpu/drm/i915/intel_audio.c static int · line 812
Quoted 1 line
drivers/gpu/drm/i915/intel_audio.c +0
-0DRM_DEBUG_KMS("Not valid for port %c\n",
port_name(port));drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 814
Quoted 2 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0err = -ENODEV ; goto unlock;--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Jani Nikula
maintainerRe: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MST
Quoted 44 lines · Libin
Quoted 44 lines
-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Wednesday, November 30, 2016 12:34 AM
To: Yang, Libin <[email protected]>; [email protected]; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Cc: Yang, Libin <[email protected]>
Subject: Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MSTdrivers/gpu/drm/i915/intel_audio.c static int
Quoted 15 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -807,7 +807,8 @@ static inti915_audio_component_sync_audio_rate(struct device *kdev, int port,
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 811
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +2
-1intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {I think the better option is to make absolutely sure we never store other kinds of encoders in dev_priv->av_enc_map[pipe] to begin with. I think that's true already, but please add
if (WARN_ON(intel_encoder->type != INTEL_OUTPUT_HDMI && intel_encoder->type != INTEL_OUTPUT_DP && intel_encoder->type != INTEL_OUTPUT_DP_MST)) return;
Quoted 6 lines
Quoted 3 lines
near the beginning of
intel_audio_codec_enable(), and remove the type checks here. This reduces the confusion about different kinds of checks after callingget_saved_enc().We can't put the check at the beginning of the function.
We still need the check in the same place as we need getintel_encoderat first.
We do not need to duplicate the checks here if we ensure only good stuff is ever put into av_enc_map.
Quoted 16 lines
Regards,
LibinQuoted 12 lines
BR,
Jani.drivers/gpu/drm/i915/intel_audio.c static int · line 812
Quoted 1 line
drivers/gpu/drm/i915/intel_audio.c +0
-0DRM_DEBUG_KMS("Not valid for port %c\n",
port_name(port));drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 814
Quoted 2 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0err = -ENODEV ; goto unlock;--
Jani Nikula, Intel Open Source Technology Center
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MST
Quoted 61 lines
-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Wednesday, November 30, 2016 11:52 PM
To: Yang, Libin <[email protected]>; [email protected]; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Subject: RE: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MSTQuoted 49 lines · Libin
Quoted 48 lines
-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Wednesday, November 30, 2016 12:34 AM
To: Yang, Libin <[email protected]>; [email protected]; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Cc: Yang, Libin <[email protected]>
Subject: Re: [PATCH 2/2] drm/i915/audio: Extend audio sync rate support for DP MSTdrivers/gpu/drm/i915/intel_audio.c static int
Quoted 15 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
This patch extends the support of audio sample rate sync up to DP MST.
Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -807,7 +807,8 @@ static inti915_audio_component_sync_audio_rate(struct device *kdev, int port,
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 811
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +2
-1intel_encoder =get_saved_enc (dev_priv, port, pipe);if (!intel_encoder || !intel_encoder->base.crtc || (intel_encoder->type !=INTEL_OUTPUT_HDMI && - intel_encoder->type !=INTEL_OUTPUT_DP )) { + intel_encoder->type !=INTEL_OUTPUT_DP && + intel_encoder->type !=INTEL_OUTPUT_DP_MST )) {I think the better option is to make absolutely sure we never store other kinds of encoders in dev_priv->av_enc_map[pipe] to begin with. I think that's true already, but please add
if (WARN_ON(intel_encoder->type != INTEL_OUTPUT_HDMI && intel_encoder->type != INTEL_OUTPUT_DP && intel_encoder->type != INTEL_OUTPUT_DP_MST)) return;near the beginning of
intel_audio_codec_enable(), and remove the type checks here. This reduces the confusion about different kinds of checks after callingget_saved_enc().
Quoted 6 lines
Quoted 2 lines
We can't put the check at the beginning of the function.
We still need the check in the same place as we need getintel_encoderatfirst.
We do not need to duplicate the checks here if we ensure only good stuff is ever put into
av_enc_map.
Yes, I agree. I will remove the type judgement.
Regards,
Libin
Quoted 12 lines
Quoted 12 lines
Regards,
LibinQuoted 7 lines
BR,
Jani.drivers/gpu/drm/i915/intel_audio.c static int · line 812
Quoted 1 line
drivers/gpu/drm/i915/intel_audio.c +0
-0DRM_DEBUG_KMS("Not valid for port %c\n",
port_name(port));
Quoted 8 lines
Quoted 5 lines
Quoted 5 lines
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, · line 814
Quoted 2 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0err = -ENODEV ; goto unlock;--
Jani Nikula, Intel Open Source Technology Center--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✓ Fi.CI.BAT: success for series starting with [1/2] drm/i915/audio: extend get_saved_enc() to support more scenarios
In reply to [email protected]== Series Details ==
== Series Details ==
Series: series starting with [1/2] drm/i915/audio: extend get_saved_enc() to support more scenarios
URL : https://patchwork.freedesktop.org/series/15321/
State : success
== Summary ==
Series 15321v1 Series without cover letter
https://patchwork.freedesktop.org/api/1.0/series/15321/revisions/1/mbox/
fi-bdw-5557u total:244 pass:229 dwarn:0 dfail:0 fail:0 skip:15 fi-bsw-n3050 total:244 pass:204 dwarn:0 dfail:0 fail:0 skip:40 fi-bxt-t5700 total:244 pass:216 dwarn:0 dfail:0 fail:0 skip:28 fi-byt-j1900 total:244 pass:216 dwarn:0 dfail:0 fail:0 skip:28 fi-byt-n2820 total:244 pass:212 dwarn:0 dfail:0 fail:0 skip:32 fi-hsw-4770 total:244 pass:224 dwarn:0 dfail:0 fail:0 skip:20 fi-hsw-4770r total:244 pass:224 dwarn:0 dfail:0 fail:0 skip:20 fi-ilk-650 total:244 pass:191 dwarn:0 dfail:0 fail:0 skip:53 fi-ivb-3520m total:244 pass:222 dwarn:0 dfail:0 fail:0 skip:22 fi-ivb-3770 total:244 pass:222 dwarn:0 dfail:0 fail:0 skip:22 fi-kbl-7200u total:244 pass:222 dwarn:0 dfail:0 fail:0 skip:22 fi-skl-6260u total:244 pass:230 dwarn:0 dfail:0 fail:0 skip:14 fi-skl-6700hq total:244 pass:223 dwarn:0 dfail:0 fail:0 skip:21 fi-snb-2520m total:244 pass:212 dwarn:0 dfail:0 fail:0 skip:32 fi-snb-2600 total:244 pass:211 dwarn:0 dfail:0 fail:0 skip:33
2f21978cfd8984c79e4cbd77ce63d9f73fe226ef drm-intel-nightly: 2016y-11m-14d-21h-23m-10s UTC integration manifest de72750 drm/i915/audio: Extend audio sync rate support for DP MST
06279ec drm/i915/audio: extend get_saved_enc() to support more scenarios
== Logs ==
For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_2995/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Jani Nikula
maintainerQuoted 18 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
When bootup, audio driver may not know it is MST or not. The audio driver will poll all the port & pipe combinations in either MST or Non-MST mode.
get_saved_enc()should handle this situation.Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 32 ++++++++++++++++++++++++++++---- 1 file changed, 28 insertions(+), 4 deletions(-)drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -737,25 +737,49 @@ static int i915_audio_component_get_cdclk_freq(struct device *kdev)return dev_priv->cdclk_freq; }
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_get_cdclk_freq(struct device *kdev) · line 748
Quoted 10 lines
drivers/gpu/drm/i915/intel_audio.c +9
-0+/* +* get the intel_encoder according to the parameter port and pipe +* intel_encoder is saved by the index of pipe +* MST & (pipe >= 0): return the av_enc_map[pipe], +* when port is matched +* MST & (pipe < 0): this is invalid +* Non-MST & (pipe >= 0): only pipe = 0 (the first device entry) +* will get the right intel_encoder with port matched +* Non-MST & (pipe < 0): get the right intel_encoder with port matched
This essentially describes the code in English. I can read the code. I need to know more about the higher level of what this does and *why*.
I also look at the call sites of get_saved_enc and wonder why they have different checks for the return values. Also patch 2/2 only modifies one of them, not both. Why?
So maybe I'm just too clueless about DP (MST) audio in general, and I wouldn't even get offended if you told me that. But I don't think the commit message and the comments in this patch properly help me either.
BR,
Jani.
Quoted 41 lines
drivers/gpu/drm/i915/intel_audio.c +19
-4+*/ static struct intel_encoder *get_saved_enc (struct drm_i915_private *dev_priv,int port,int pipe) { +struct intel_encoder *encoder;if (WARN_ON (pipe >=I915_MAX_PIPES ))return NULL ;/* MST */ -if (pipe >=0 ) -return dev_priv->av_enc_map[pipe]; +if (pipe >=0 ) { + encoder = dev_priv->av_enc_map[pipe]; +/* +* when bootup, audio driver may not know it is +* MST or not. So it will poll all the port & pipe +* combinations +*/ +if (encoder !=NULL && encoder->port == port && + encoder->type ==INTEL_OUTPUT_DP_MST ) +return encoder; + }/* Non-MST */ -for_each_pipe (dev_priv, pipe) { -struct intel_encoder *encoder; +if (pipe >0 ) +return NULL ; +for_each_pipe (dev_priv, pipe) { encoder = dev_priv->av_enc_map[pipe];if (encoder ==NULL )continue ; +if (encoder->type ==INTEL_OUTPUT_DP_MST ) +continue ; +if (port == encoder->port)return encoder; }
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 35 lines
-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Friday, November 25, 2016 10:46 PM
To: Yang, Libin <[email protected]>; [email protected]; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Cc: Yang, Libin <[email protected]>; Libin Yang <[email protected]>
Subject: Re: [PATCH 1/2] drm/i915/audio: extendget_saved_enc()to support more scenariosdrivers/gpu/drm/i915/intel_audio.c static int
Quoted 18 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
When bootup, audio driver may not know it is MST or not. The audio driver will poll all the port & pipe combinations in either MST or Non-MST mode.
get_saved_enc()should handle this situation.Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 32 ++++++++++++++++++++++++++++---- 1 file changed, 28 insertions(+), 4 deletions(-)drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -737,25 +737,49 @@ static int
i915_audio_component_get_cdclk_freq(struct device *kdev)Quoted 3 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0return dev_priv->cdclk_freq; }
Quoted 13 lines
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_get_cdclk_freq(struct device *kdev)
Quoted 10 lines
drivers/gpu/drm/i915/intel_audio.c +10
-0+/* +* get the intel_encoder according to the parameter port and pipe +* intel_encoder is saved by the index of pipe +* MST & (pipe >= 0): return the av_enc_map[pipe], +* when port is matched +* MST & (pipe < 0): this is invalid +* Non-MST & (pipe >= 0): only pipe = 0 (the first device entry) +* will get the right intel_encoder with port matched +* Non-MST & (pipe < 0): get the right intel_encoder with port +matched This essentially describes the code in English. I can read the code. I need to know more about the higher level of what this does and *why*.
In initialization, audio driver will call functions get_eld() and etc. But at that time, audio driver may not know whether it is DP MST or not. In the original function get_saved_enc(), if it is DP MST, it will set the pipe to the correct value, otherwise, it will be set -1.
Although audio driver can get the knowledge whether it is in DP MST mode or not by reading the codec register. But it will drop performance each time it call the get_eld and other functions. As gfx driver can easily know whether it is in DP MST mode or not. I would like to extend the get_saved_enc() function to handle the situation that audio driver still send the device id info even it is in DP mode. In this situation, get_saved_enc() will return the correct intel_encoder instead of panic.
Quoted 4 lines
I also look at the call sites of
get_saved_encand wonder why they have different checks for the return values. Also patch 2/2 only modifies one of them, not both. Why?
The 2 check should be similar? In sync_audio_rate(), it calls get_saved_enc() and checked base.crtc, I think it is because it will use base.crtc later.
And it will check the intel_encoder->type because this function only need handle HDMI/DP/DP MST situation to setup the audio config.
Regards,
Libin
Quoted 45 lines
So maybe I'm just too clueless about DP (MST) audio in general, and I wouldn't even get offended if you told me that. But I don't think the commit message and the comments in this patch properly help me either.
BR,
Jani.drivers/gpu/drm/i915/intel_audio.c static int · line 751
Quoted 2 lines
drivers/gpu/drm/i915/intel_audio.c +1
-0+*/ static struct intel_encoder *get_saved_enc(struct drm_i915_private
*dev_priv,Quoted 33 lines
drivers/gpu/drm/i915/intel_audio.c +15
-4int port, int pipe) { +struct intel_encoder *encoder;if (WARN_ON (pipe >=I915_MAX_PIPES ))return NULL ;/* MST */ -if (pipe >=0 ) -return dev_priv->av_enc_map[pipe]; +if (pipe >=0 ) { + encoder = dev_priv->av_enc_map[pipe]; +/* +* when bootup, audio driver may not know it is +* MST or not. So it will poll all the port & pipe +* combinations +*/ +if (encoder !=NULL && encoder->port == port && + encoder->type ==INTEL_OUTPUT_DP_MST ) +return encoder; + }/* Non-MST */ -for_each_pipe (dev_priv, pipe) { -struct intel_encoder *encoder; +if (pipe >0 ) +return NULL ; +for_each_pipe (dev_priv, pipe) { encoder = dev_priv->av_enc_map[pipe];if (encoder ==NULL )continue ;
Quoted 9 lines
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_get_cdclk_freq(struct device *kdev) · line 785
Quoted 6 lines
drivers/gpu/drm/i915/intel_audio.c +3
-0+if (encoder->type ==INTEL_OUTPUT_DP_MST ) +continue ; +if (port == encoder->port)return encoder; }--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Jani Nikula
maintainerQuoted 53 lines · Libin
Quoted 34 lines
-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Friday, November 25, 2016 10:46 PM
To: Yang, Libin <[email protected]>; [email protected]; [email protected]; Vetter, Daniel <[email protected]>; Pandiyan, Dhinakaran <[email protected]>; Kp, Jeeja <[email protected]>; [email protected]
Cc: Yang, Libin <[email protected]>; Libin Yang <[email protected]>
Subject: Re: [PATCH 1/2] drm/i915/audio: extendget_saved_enc()to support more scenariosdrivers/gpu/drm/i915/intel_audio.c static int
Quoted 18 lines · On Tue, 15 Nov 2016, [email protected] wrote:
From: Libin Yang <[email protected]>
When bootup, audio driver may not know it is MST or not. The audio driver will poll all the port & pipe combinations in either MST or Non-MST mode.
get_saved_enc()should handle this situation.Signed-off-by Libin Yang <[email protected]>--- drivers/gpu/drm/i915/intel_audio.c | 32 ++++++++++++++++++++++++++++---- 1 file changed, 28 insertions(+), 4 deletions(-)drivers/gpu/drm/i915/intel_audio.c +0
-0@@ -737,25 +737,49 @@ static int
i915_audio_component_get_cdclk_freq(struct device *kdev)Quoted 3 lines
drivers/gpu/drm/i915/intel_audio.c +0
-0return dev_priv->cdclk_freq; }Quoted 13 lines
drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_get_cdclk_freq(struct device *kdev)
Quoted 10 lines
drivers/gpu/drm/i915/intel_audio.c +10
-0+/* +* get the intel_encoder according to the parameter port and pipe +* intel_encoder is saved by the index of pipe +* MST & (pipe >= 0): return the av_enc_map[pipe], +* when port is matched +* MST & (pipe < 0): this is invalid +* Non-MST & (pipe >= 0): only pipe = 0 (the first device entry) +* will get the right intel_encoder with port matched +* Non-MST & (pipe < 0): get the right intel_encoder with port +matched This essentially describes the code in English. I can read the code. I need to know more about the higher level of what this does and *why*.
In initialization, audio driver will call functions
get_eld()and etc. But at that time, audio driver may not know whether it is DP MST or not. In the original functionget_saved_enc(), if it is DP MST, it will set the pipe to the correct value, otherwise, it will be set -1.
Quoted 7 lines
Although audio driver can get the knowledge whether it is in DP MST mode or not by reading the codec register. But it will drop performance each time it call the
get_eldand other functions. As gfx driver can easily know whether it is in DP MST mode or not. I would like to extend theget_saved_enc()function to handle the situation that audio driver still send the device id info even it is in DP mode. In this situation,get_saved_enc()will return the correctintel_encoderinstead of panic.
Please describe that in the commit message and/or comments in the patch.
Quoted 10 lines
Quoted 4 lines
I also look at the call sites of
get_saved_encand wonder why they have different checks for the return values. Also patch 2/2 only modifies one of them, not both. Why?The 2 check should be similar? In
sync_audio_rate(), it callsget_saved_enc()and checked base.crtc, I think it is because it will use base.crtc later.And it will check the
intel_encoder->typebecause this function only need handle HDMI/DP/DP MST situation to setup the audio config.
I think addressing my review comment on patch 2 will help this.
BR,
Jani.
Quoted 58 lines
Regards,
LibinQuoted 54 lines
So maybe I'm just too clueless about DP (MST) audio in general, and I wouldn't even get offended if you told me that. But I don't think the commit message and the comments in this patch properly help me either.
BR,
Jani.drivers/gpu/drm/i915/intel_audio.c static int · line 751
Quoted 2 lines
drivers/gpu/drm/i915/intel_audio.c +1
-0+*/ static struct intel_encoder *get_saved_enc(struct drm_i915_private
*dev_priv,Quoted 29 lines
drivers/gpu/drm/i915/intel_audio.c +15
-4int port, int pipe) { +struct intel_encoder *encoder;if (WARN_ON (pipe >=I915_MAX_PIPES ))return NULL ;/* MST */ -if (pipe >=0 ) -return dev_priv->av_enc_map[pipe]; +if (pipe >=0 ) { + encoder = dev_priv->av_enc_map[pipe]; +/* +* when bootup, audio driver may not know it is +* MST or not. So it will poll all the port & pipe +* combinations +*/ +if (encoder !=NULL && encoder->port == port && + encoder->type ==INTEL_OUTPUT_DP_MST ) +return encoder; + }/* Non-MST */ -for_each_pipe (dev_priv, pipe) { -struct intel_encoder *encoder; +if (pipe >0 ) +return NULL ; +for_each_pipe (dev_priv, pipe) {drivers/gpu/drm/i915/intel_audio.c static int i915_audio_component_get_cdclk_freq(struct device *kdev) · line 785
Quoted 10 lines
drivers/gpu/drm/i915/intel_audio.c +3
-0encoder = dev_priv->av_enc_map[pipe];if (encoder ==NULL )continue ; +if (encoder->type ==INTEL_OUTPUT_DP_MST ) +continue ; +if (port == encoder->port)return encoder; }--
Jani Nikula, Intel Open Source Technology Center
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx