[Intel-gfx] Need your advice: Add a new communication inteface between HD-Audio and Gfx drivers for hotplug notification/ELD update

Lin, Mengdong mengdong.lin at intel.com
Wed Feb 19 01:08:39 PST 2014


> -----Original Message-----
> From: Ville Syrjälä [mailto:ville.syrjala at linux.intel.com]
> Sent: Tuesday, February 18, 2014 10:23 PM
> To: Lin, Mengdong
> Cc: Daniel Vetter; Takashi Iwai; alsa-devel at alsa-project.org; Barnes, Jesse;
> Zanoni, Paulo R; dri-devel; intel-gfx at lists.freedesktop.org
> Subject: Re: [Intel-gfx] Need your advice: Add a new communication
> inteface between HD-Audio and Gfx drivers for hotplug notification/ELD
> update
> 
> On Tue, Feb 18, 2014 at 01:58:22PM +0000, Lin, Mengdong wrote:
> > Sorry to pick up this thread after a long time.
> >
> > > -----Original Message-----
> > > From: daniel.vetter at ffwll.ch [mailto:daniel.vetter at ffwll.ch] On
> > > Behalf Of Daniel Vetter
> > > Sent: Thursday, January 23, 2014 4:27 PM
> > > To: Takashi Iwai
> >
> > > On Thu, Jan 23, 2014 at 8:57 AM, Takashi Iwai <tiwai at suse.de> wrote:
> > > >> Thanks for clarification!
> > > >> Maybe we can add output info (eg. display port number) to the eld
> > > >> entries
> > > under /proc/asound/cardx. Is it okay?
> > > >
> > > > It's possible, but the proc file is just a help.  It can't be the API.
> > > > For accessing the information, we'll need some new API, or let
> > > > inform via sysfs of the new device.
> > >
> > > Links in sysfs sound like the best approach. drm already has nodes
> > > for each connector, so on the gfx side there's a natural endpoint
> already.
> > > sysfs links also avoids any naming issues from the start, e.g. the
> > > above DP connector id might lead to clashes with multiple cards.
> >
> > Hi Daniel,
> >
> > Is there a 1:1 mapping between these connector nodes and ports of Gfx
> display engine?
> > Eg. For Haswell Ultrabook, under
> > /sys/devices/pci0000:00/0000:00:02.0/drm/card0/
> > There are four connector nodes,
> > card0-DP-1		-> DDI port B
> > card0-eDP-1/		-> DDI port A
> > card0-HDMI-A-1/	-> DDI Port C
> > card0-HDMI-A-2/	-> Which DDI port ?  Haswell-ULT does not
> support port D, and I think port E is for VGA.
> 
> There's no fixed mapping with the port and the connector name. The
> number in the connector name is basically just a running number per
> connector type. However I do believe we do register the connectors in the
> order of the ports more or less always, so you can
> *sometimes* deduce the port name from the connector.
> 
> I suppose in this example HDMI-A-1 is port B, HDMI-A-2 is port C, and
> DP-1 can be either port B or port C. DP++ is the reason why we have
> overlapping DP and HDMI connectors for the same port.

Thanks for clarification, Ville!

Does Haswell support DP++ on port B/C/D?
And will the name of these connector node change after system boot, e.g. after S3/S4 cycles or hot-plug?


As long as the connector name is constant, it can help to find out the screen status no matter the port works in HDMI or DP mode.

There is 1:1 mapping between audio output pins and DDI ports.
And the new ALSA PCM 'screen' control for a pin can reflect the connector name in right mode. E.g. for pin/port B, it can return HDMI-A-1 or DP-1 depending on what kind of monitor is connected.

Thanks
Mengdong

> >
> > Hi Takashi,
> >
> > To make user space figure out which audio output is connected to which
> screen (connector), maybe we can define a new ALSA control for each
> HDMI/DP PCM device:
> > e.g. numid=x,iface=PCM,name='Screen',device=3
> > Reading the control will return the name of the DRM connector nodes
> like ' card0-DP-1'. The audio driver can get the connector name from the
> gfx driver.
> >
> > For DP1.2 Multi-stream transport, it's not supported by i915 and HD-A
> driver now. But probably there will be sub-nodes for the DP connector
> node in the future and an index in their name can be used distinguish
> monitors connected to the same DP port, like card0-DP-1.1, card0-DP-1.2,
> card0-DP-1.3 ... These names can be used by the above ALSA PCM 'Screen'
> control, so we can still know which audio output is to which monitor.
> >
> > Thanks
> > Mengdong
> >
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> --
> Ville Syrjälä
> Intel OTC


More information about the dri-devel mailing list