[alsa-devel] [RFC PATCH] ALSA: hda - Ignore default pin config on Haswell HDMI codecs
According to Daniel Vetter, these pin configs should be ignored, because the graphics driver/userspace could route audio to these pins even if they are disabled by BIOS.
Signed-off-by: David Henningsson david.henningsson@canonical.com --- sound/pci/hda/patch_hdmi.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
It would be good to have this acked by someone at Intel before committing.
diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index e12f7a0..8c1bede 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1307,12 +1307,14 @@ static int hdmi_add_pin(struct hda_codec *codec, hda_nid_t pin_nid) if (!(caps & (AC_PINCAP_HDMI | AC_PINCAP_DP))) return 0;
- config = snd_hda_codec_get_pincfg(codec, pin_nid); - if (get_defcfg_connect(config) == AC_JACK_PORT_NONE) - return 0; - if (codec->vendor_id == 0x80862807) intel_haswell_fixup_connect_list(codec, pin_nid); + else { + /* On Haswell, default pin config should be ignored */ + config = snd_hda_codec_get_pincfg(codec, pin_nid); + if (get_defcfg_connect(config) == AC_JACK_PORT_NONE) + return 0; + }
pin_idx = spec->num_pins; per_pin = snd_array_new(&spec->pins);
Hi David,
Do you mean that a pin's default configuration may tell "No physical connection" by mistake?
Thanks Mengdong
-----Original Message----- From: David Henningsson [mailto:david.henningsson@canonical.com] Sent: Tuesday, June 04, 2013 4:03 PM To: alsa-devel@alsa-project.org Cc: tiwai@suse.de; Wang, Xingchao; daniel@ffwll.ch; Lin, Mengdong; Girdwood, Liam R; Li, Jocelyn; David Henningsson Subject: [RFC PATCH] ALSA: hda - Ignore default pin config on Haswell HDMI codecs
According to Daniel Vetter, these pin configs should be ignored, because the graphics driver/userspace could route audio to these pins even if they are disabled by BIOS.
Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_hdmi.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
It would be good to have this acked by someone at Intel before committing.
diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index e12f7a0..8c1bede 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1307,12 +1307,14 @@ static int hdmi_add_pin(struct hda_codec *codec, hda_nid_t pin_nid) if (!(caps & (AC_PINCAP_HDMI | AC_PINCAP_DP))) return 0;
- config = snd_hda_codec_get_pincfg(codec, pin_nid);
- if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)
return 0;- if (codec->vendor_id == 0x80862807) intel_haswell_fixup_connect_list(codec, pin_nid);
else {
/* On Haswell, default pin config should be ignored */config = snd_hda_codec_get_pincfg(codec, pin_nid);if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)return 0;}
pin_idx = spec->num_pins; per_pin = snd_array_new(&spec->pins);
-- 1.7.9.5
On 06/04/2013 12:10 PM, Lin, Mengdong wrote:
Hi David,
Do you mean that a pin's default configuration may tell "No physical connection" by mistake?
A pin's default configuration can be set to "No physical connection". I've seen this on one machine that we're trying to enable.
Whether that is a mistake or not, that's the million dollar question here:
If it is not a mistake, then the graphics driver/stack should pay attention to this value, and not route audio to that pin.
If it is a mistake, then the bspec [1] should mention that programming of this value should not be allowed (so we can tell the OEM to correct their BIOS).
I'm still trying to get a clear answer from someone at Intel, and it's becoming a bit frustrating not to get that.
Thanks Mengdong
-----Original Message----- From: David Henningsson [mailto:david.henningsson@canonical.com] Sent: Tuesday, June 04, 2013 4:03 PM To: alsa-devel@alsa-project.org Cc: tiwai@suse.de; Wang, Xingchao; daniel@ffwll.ch; Lin, Mengdong; Girdwood, Liam R; Li, Jocelyn; David Henningsson Subject: [RFC PATCH] ALSA: hda - Ignore default pin config on Haswell HDMI codecs
According to Daniel Vetter, these pin configs should be ignored, because the graphics driver/userspace could route audio to these pins even if they are disabled by BIOS.
Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_hdmi.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
It would be good to have this acked by someone at Intel before committing.
diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index e12f7a0..8c1bede 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1307,12 +1307,14 @@ static int hdmi_add_pin(struct hda_codec *codec, hda_nid_t pin_nid) if (!(caps & (AC_PINCAP_HDMI | AC_PINCAP_DP))) return 0;
- config = snd_hda_codec_get_pincfg(codec, pin_nid);
- if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)
return 0;- if (codec->vendor_id == 0x80862807) intel_haswell_fixup_connect_list(codec, pin_nid);
else {
/* On Haswell, default pin config should be ignored */config = snd_hda_codec_get_pincfg(codec, pin_nid);if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)return 0;}
pin_idx = spec->num_pins; per_pin = snd_array_new(&spec->pins);
-- 1.7.9.5
Hi David,
-----Original Message----- From: David Henningsson [mailto:david.henningsson@canonical.com] Sent: Tuesday, June 04, 2013 6:20 PM To: Lin, Mengdong Cc: alsa-devel@alsa-project.org; tiwai@suse.de; Wang, Xingchao; daniel@ffwll.ch; Girdwood, Liam R; Li, Jocelyn; James M Leddy Subject: Re: [RFC PATCH] ALSA: hda - Ignore default pin config on Haswell HDMI codecs
On 06/04/2013 12:10 PM, Lin, Mengdong wrote:
Hi David,
Do you mean that a pin's default configuration may tell "No physical
connection" by mistake?
A pin's default configuration can be set to "No physical connection". I've seen this on one machine that we're trying to enable.
Whether that is a mistake or not, that's the million dollar question here:
If it is not a mistake, then the graphics driver/stack should pay attention to this value, and not route audio to that pin.
Well by default, all the pins choose converter 0 as data source. So when playing audio on the pin without external physical device connected, it would throw data to converter 0. Meanwhile converter 0 maybe data source for other pin in use with physical device connected, that's why you hear sound.
I'm working on the patch to fix this issue. in my test, you cannot hear sound on an pin with nothing connected.
If it is a mistake, then the bspec [1] should mention that programming of this value should not be allowed (so we can tell the OEM to correct their BIOS).
I'm still trying to get a clear answer from someone at Intel, and it's becoming a bit frustrating not to get that.
Thanks --xingchao
Thanks Mengdong
-----Original Message----- From: David Henningsson [mailto:david.henningsson@canonical.com] Sent: Tuesday, June 04, 2013 4:03 PM To: alsa-devel@alsa-project.org Cc: tiwai@suse.de; Wang, Xingchao; daniel@ffwll.ch; Lin, Mengdong;
Girdwood,
Liam R; Li, Jocelyn; David Henningsson Subject: [RFC PATCH] ALSA: hda - Ignore default pin config on Haswell HDMI codecs
According to Daniel Vetter, these pin configs should be ignored, because the graphics driver/userspace could route audio to these pins even if they are disabled by BIOS.
Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_hdmi.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
It would be good to have this acked by someone at Intel before committing.
diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c
index
e12f7a0..8c1bede 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1307,12 +1307,14 @@ static int hdmi_add_pin(struct hda_codec
*codec,
hda_nid_t pin_nid) if (!(caps & (AC_PINCAP_HDMI | AC_PINCAP_DP))) return 0;
- config = snd_hda_codec_get_pincfg(codec, pin_nid);
- if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)
return 0;- if (codec->vendor_id == 0x80862807) intel_haswell_fixup_connect_list(codec, pin_nid);
else {
/* On Haswell, default pin config should be ignored */config = snd_hda_codec_get_pincfg(codec, pin_nid);if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)return 0;}
pin_idx = spec->num_pins; per_pin = snd_array_new(&spec->pins);
-- 1.7.9.5
-- David Henningsson, Canonical Ltd. https://launchpad.net/~diwic
[1] Or some other document. I don't have access to the bspec, so I can't check this for myself.
Hi David,
Did you test the patch and what exactly the issue it could fix? We do have issue when play audio on the pin without physical device connected.i.e. you can hear sound on that pin. I'm working on some patches to fix similar issues, also trying to fix issue like dual monitor audio routing. So please let me know what's the scenario of this patch trying to fix.
thanks --xingchao
-----Original Message----- From: David Henningsson [mailto:david.henningsson@canonical.com] Sent: Tuesday, June 04, 2013 4:03 PM To: alsa-devel@alsa-project.org Cc: tiwai@suse.de; Wang, Xingchao; daniel@ffwll.ch; Lin, Mengdong; Girdwood, Liam R; Li, Jocelyn; David Henningsson Subject: [RFC PATCH] ALSA: hda - Ignore default pin config on Haswell HDMI codecs
According to Daniel Vetter, these pin configs should be ignored, because the graphics driver/userspace could route audio to these pins even if they are disabled by BIOS.
Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_hdmi.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
It would be good to have this acked by someone at Intel before committing.
diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index e12f7a0..8c1bede 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1307,12 +1307,14 @@ static int hdmi_add_pin(struct hda_codec *codec, hda_nid_t pin_nid) if (!(caps & (AC_PINCAP_HDMI | AC_PINCAP_DP))) return 0;
- config = snd_hda_codec_get_pincfg(codec, pin_nid);
- if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)
return 0;- if (codec->vendor_id == 0x80862807) intel_haswell_fixup_connect_list(codec, pin_nid);
else {
/* On Haswell, default pin config should be ignored */config = snd_hda_codec_get_pincfg(codec, pin_nid);if (get_defcfg_connect(config) == AC_JACK_PORT_NONE)return 0;}
pin_idx = spec->num_pins; per_pin = snd_array_new(&spec->pins);
-- 1.7.9.5
participants (3)
-
David Henningsson -
Lin, Mengdong -
Wang, Xingchao