Media graph flow for an hybrid device as discussed at the media workshop
During the discussions at the Media Workshop, we came with some dot files that would describe a hybrid PC-consumer TV stick with radio, analog video, analog TV and digital TV on it.
I consolidated all the dot files we've worked there, and added the connectors for RF, S-Video and Composite.
The dot file and the corresponding picture is at:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.dot
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.png
As my plan is to start working on some real driver to produce such graph, please validate if the entities, interfaces, data links and interface links are correct, and if the namespace nomenclature is ok, or if I miss something.
Thanks!
Mauro
Hi Mauro,
Quoted 7 lines · Mauro Carvalho Chehab
During the discussions at the Media Workshop, we came with some dot files that would describe a hybrid PC-consumer TV stick with radio, analog video, analog TV and digital TV on it.
I consolidated all the dot files we've worked there, and added the connectors for RF, S-Video and Composite.
Quoted 7 lines
The dot file and the corresponding picture is at:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.dot
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.pngAs my plan is to start working on some real driver to produce such graph, please validate if the entities, interfaces, data links and interface links are correct, and if the namespace nomenclature is ok, or if I miss something.
This looks OK to me, except for one small detail: I wouldn't use the name "Source entities" for connectors. Instead use "Connector entities" since such entities correspond to actual real connectors on a backplane. A proper source entity would be a sensor or test pattern generator. Which actually can occur with the em28xx since it's used in webcams as well.
And a really, really small detail: in the legend the 'interface link' is an arrow, but it should be a line, since there is no direction. The graph itself is fine.
As you mentioned on irc, the v4l-subdevX nodes won't be created for this device since all the configuration happens via the standard interfaces.
But if they were to be created, then they would appear where they are in this example.
Regards,
Hans
Em Mon, 10 Aug 2015 14:43:50 +0200
Hans Verkuil <[email protected]> escreveu:
Quoted 1 line
Hi Mauro,
Thanks for the review!
Quoted 13 lines · Mauro Carvalho Chehab
Quoted 11 lines
During the discussions at the Media Workshop, we came with some dot files that would describe a hybrid PC-consumer TV stick with radio, analog video, analog TV and digital TV on it.
I consolidated all the dot files we've worked there, and added the connectors for RF, S-Video and Composite.
The dot file and the corresponding picture is at:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.dot
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.png
Quoted 7 lines
Quoted 3 lines
As my plan is to start working on some real driver to produce such graph, please validate if the entities, interfaces, data links and interface links are correct, and if the namespace nomenclature is ok, or if I miss something.
This looks OK to me, except for one small detail: I wouldn't use the name "Source entities" for connectors. Instead use "Connector entities" since such entities correspond to actual real connectors on a backplane.
Yeah. Well, they're actually "Source connector entities" ;) But I see your point. All connectors should be marked with a different type at the media_graph_obj.
Quoted 3 lines
A proper
source entity would be a sensor or test pattern generator. Which actually can occur with the em28xx since it's used in webcams as well.
Ah, true. I'll add that in the graph and use a different color to distinguish between "source" and "connector" entities.
Quoted 4 lines
And a really, really small detail: in the legend the 'interface link' is an arrow, but it should be a line, since there is no direction. The graph itself is fine.
Well, I didn't find a way to put a line there. The legend is produced by an html code. I would need to have a "line" character, or to add an image.
Perhaps I should look deeper to find a bold horizontal line at the UTF-8 charset. — and ― are too thin. Do you know any char that would look better there?
Quoted 5 lines
As you mentioned on irc, the v4l-subdevX nodes won't be created for this device since all the configuration happens via the standard interfaces.
But if they were to be created, then they would appear where they are in this example.
Thanks!
Mauro
Quoted 9 lines
Regards,
Hans
_______________________________________________
media-workshop mailing list
[email protected]
http://www.linuxtv.org/cgi-bin/mailman/listinfo/media-workshop
Em Mon, 10 Aug 2015 10:05:24 -0300
Mauro Carvalho Chehab <[email protected]> escreveu:
Quoted 44 lines
Em Mon, 10 Aug 2015 14:43:50 +0200
Hans Verkuil <[email protected]> escreveu:Quoted 1 line
Hi Mauro,
Thanks for the review!
Quoted 13 lines · Mauro Carvalho Chehab
Quoted 11 lines
During the discussions at the Media Workshop, we came with some dot files that would describe a hybrid PC-consumer TV stick with radio, analog video, analog TV and digital TV on it.
I consolidated all the dot files we've worked there, and added the connectors for RF, S-Video and Composite.
The dot file and the corresponding picture is at:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.dot
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.pngQuoted 7 lines
Quoted 3 lines
As my plan is to start working on some real driver to produce such graph, please validate if the entities, interfaces, data links and interface links are correct, and if the namespace nomenclature is ok, or if I miss something.
This looks OK to me, except for one small detail: I wouldn't use the name "Source entities" for connectors. Instead use "Connector entities" since such entities correspond to actual real connectors on a backplane.
Yeah. Well, they're actually "Source connector entities" ;) But I see your point. All connectors should be marked with a different type at the
media_graph_obj.Quoted 3 lines
A proper
source entity would be a sensor or test pattern generator. Which actually can occur with the em28xx since it's used in webcams as well.Ah, true. I'll add that in the graph and use a different color to distinguish between "source" and "connector" entities.
Quoted 4 lines
And a really, really small detail: in the legend the 'interface link' is an arrow, but it should be a line, since there is no direction. The graph itself is fine.
Quoted 6 lines
Well, I didn't find a way to put a line there. The legend is produced by an html code. I would need to have a "line" character, or to add an image.
Perhaps I should look deeper to find a bold horizontal line at the UTF-8 charset. — and ― are too thin. Do you know any char that would look better there?
Found one character ;)
I also added a webcam sensor and fixed the legend. See below:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v3.png
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v3.dot
Quoted 19 lines
Quoted 6 lines
As you mentioned on irc, the v4l-subdevX nodes won't be created for this device since all the configuration happens via the standard interfaces.
But if they were to be created, then they would appear where they are in this example.
Thanks!
MauroQuoted 9 lines
Regards,
Hans
_______________________________________________
media-workshop mailing list
[email protected]
http://www.linuxtv.org/cgi-bin/mailman/listinfo/media-workshop
Quoted 54 lines · Mauro Carvalho Chehab
Em Mon, 10 Aug 2015 10:05:24 -0300
Mauro Carvalho Chehab <[email protected]> escreveu:Quoted 44 lines
Em Mon, 10 Aug 2015 14:43:50 +0200
Hans Verkuil <[email protected]> escreveu:Quoted 1 line
Hi Mauro,
Thanks for the review!
Quoted 13 lines · Mauro Carvalho Chehab
Quoted 11 lines
During the discussions at the Media Workshop, we came with some dot files that would describe a hybrid PC-consumer TV stick with radio, analog video, analog TV and digital TV on it.
I consolidated all the dot files we've worked there, and added the connectors for RF, S-Video and Composite.
The dot file and the corresponding picture is at:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.dot
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v2.pngQuoted 7 lines
Quoted 3 lines
As my plan is to start working on some real driver to produce such graph, please validate if the entities, interfaces, data links and interface links are correct, and if the namespace nomenclature is ok, or if I miss something.
This looks OK to me, except for one small detail: I wouldn't use the name "Source entities" for connectors. Instead use "Connector entities" since such entities correspond to actual real connectors on a backplane.
Yeah. Well, they're actually "Source connector entities" ;) But I see your point. All connectors should be marked with a different type at the
media_graph_obj.Quoted 3 lines
A proper
source entity would be a sensor or test pattern generator. Which actually can occur with the em28xx since it's used in webcams as well.Ah, true. I'll add that in the graph and use a different color to distinguish between "source" and "connector" entities.
Quoted 4 lines
And a really, really small detail: in the legend the 'interface link' is an arrow, but it should be a line, since there is no direction. The graph itself is fine.
Quoted 6 lines
Well, I didn't find a way to put a line there. The legend is produced by an html code. I would need to have a "line" character, or to add an image.
Perhaps I should look deeper to find a bold horizontal line at the UTF-8 charset. — and ― are too thin. Do you know any char that would look better there?
Quoted 6 lines
Found one character ;)
I also added a webcam sensor and fixed the legend. See below:
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v3.png
http://linuxtv.org/downloads/presentations/mc_ws_2015/dvb-pipeline-v3.dot
Looks good. But if you have a sensor, then there should also be a v4l-subdev2 interface for the sensor entity, and it is also controlled by video0, so that interface-to-entity link is missing.
And the saa7115 output pads need to be renumbered to 4-6 (there are two pads '3' at the moment and the mixer is linked to the input pad 3).
Regards,
Hans
Quoted 25 lines
Quoted 20 lines
Quoted 7 lines
As you mentioned on irc, the v4l-subdevX nodes won't be created for this device since all the configuration happens via the standard interfaces.
But if they were to be created, then they would appear where they are in this example.
Thanks!
MauroQuoted 9 lines
Regards,
Hans
_______________________________________________
media-workshop mailing list
[email protected]
http://www.linuxtv.org/cgi-bin/mailman/listinfo/media-workshop--
To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html