how to make S-Video out working?

Hi!

I'm trying to understand how can I get S-Video output working,
because I don't have any DVI-D capable monitor to use.
I can see the colored band on u-boot boot, and I'm running Koen's
Angstrom demo (ones with e17), with uImage from google code (uImage_OTG).

I've done echo "tv" > /sys/class/display_control/omap_disp_control/graphics
as suggested in hints page.
On my TV I can see something confused in a white background, but I can't see
what it is (the login form?).
Maybe the "problem" is related to resolution and refresh rate or something else...

xrandr --verbose returns:
Screen 0: minimum 1024 x 768, current 1024 x 768, maximum 1024 x 768
default connected 1024x768+0+0 (0x42) normal (normal left inverted right x axis y axis) 0mm x 0mm
         Identifier: 0x41
         Timestamp: 1947596985
         Subpixel: unknown
         Clones:
         CRTC: 0
         CRTCs: 0
   1024x768 (0x42) 0.0MHz
         h: width 1024 start 0 end 0 total 1024 skew 0 clock 0.0KHz
         v: height 768 start 0 end 0 total 768 clock 0.0Hz

I'm a newby, and I need some suggestions on how to procede...

Thank you very much!

G.

apk said the following on 09/03/2008 12:55 PM:

Hi!

I'm trying to understand how can I get S-Video output working,
because I don't have any DVI-D capable monitor to use.
I can see the colored band on u-boot boot, and I'm running Koen's
Angstrom demo (ones with e17), with uImage from google code (uImage_OTG).
  

Here is a patch which just puts the stuff in line with the latest git
kernel changes. sending the same as attachment to prevent code wrap -
The original patch is from Hunyue . Thanks for keeping the s-video folks
happy :wink:

NOTE: things have not worked for me with Angstrom -> might need some
help on that... :frowning: probably x and y res? or some sort of udev fixes? I
am not sure...

Regards,
Nishanth Menon

0001-modifications-for-s-video.patch (355 KB)

setenv bootargs 'console=ttyS2,115200n8 noinitrd root=/dev/mmcblk0p2
rootfstype=ext3 rw rootdelay=1 nohz=off'

What's with the nohz=off? I keep seeing TI folks doing that, but no-
one has been able to good reason for it. There's also a correlation
between people saying "stuff does not work" and using 'nohz=off' :slight_smile:

regards,

Koen

Koen,

Koen Kooi said the following on 09/08/2008 03:58 AM:

setenv bootargs 'console=ttyS2,115200n8 noinitrd root=/dev/mmcblk0p2
rootfstype=ext3 rw rootdelay=1 nohz=off'

What's with the nohz=off? I keep seeing TI folks doing that, but no-
one has been able to good reason for it. There's also a correlation
between people saying "stuff does not work" and using 'nohz=off' :slight_smile:
  

I think I blame myself copying the bootargs from
http://code.google.com/p/beagleboard/wiki/BootingBeagleBoard
section:
"Bootargs for MMC File System"

Regards,
Nishanth Menon

When you use MUSB OTG controller MODE0 DMA, number interrupts generated are huge, say 1 per 512 byte packet and if you enable dynamic tick along with this, it will add some more latency in interrupt path and significantly reduce throughputs.

Temporary work around is to disable dyntick by specifying nohz=off. Ideally we need to use MODE1 DMA wherever possible to reduce the number of interrupts.

Hi Menon,

Thanks for the update. :slight_smile:

I tested your patch with latest l-
o(de1121fdb899f762b9e717f44eaf3fae7c00cd3e) and some changes are
necessary. I'm sending a link[1] that contain a new version with small
changes.

[1] - http://franciscoalecrim.com/down/beagle/s-video/0001-modifications-for-s-video.patch

I'm checking/studying the best way to summit for l-o.

Hi Menon,

Thanks for the update. :slight_smile:

I tested your patch with latest l-
o(de1121fdb899f762b9e717f44eaf3fae7c00cd3e) and some changes are
necessary. I'm sending a link[1] that contain a new version with small
changes.

[1] -http://franciscoalecrim.com/down/beagle/s-video/0001-modifications-fo

I'm checking/studying the best way to summit for l-o.

l-o will never accept this patch, it would be better to work on the
DSS rewrite with nokia & ti instead of polishing this driver.

regards,

Koen

Alec,Koen,

Hi Menon,

Thanks for the update. :slight_smile:

I tested your patch with latest l-
o(de1121fdb899f762b9e717f44eaf3fae7c00cd3e) and some changes are
necessary. I'm sending a link[1] that contain a new version with small
changes.

[1] -http://franciscoalecrim.com/down/beagle/s-video/0001-modifications-fo

I'm checking/studying the best way to summit for l-o.

l-o will never accept this patch, it would be better to work on the
DSS rewrite with nokia & ti instead of polishing this driver.

I agree with Koen. The code which is was "hacked up" is in complete
contrast to Tomi's patch here:
'[PREVIEW] New display subsystem for OMAP2/3' - MARC - this patch set
is more in line with what is possibly acceptable to l-o list.

I think it makes sense to get involved in that discussion and help it
go into mainstream.

Regards,
Nishanth Menon