Thanks Stefano,
On 09/12/2013 02:47 AM, Stefano Babic wrote:
Hi Fabio,
CC to Eric and Troy.
On 11/09/2013 23:14, Fabio Estevam wrote:
diff --git a/board/freescale/mx6sabresd/mx6sabresd.c b/board/freescale/mx6sabresd/mx6sabresd.c index c832bd9..0f91fe2 100644 --- a/board/freescale/mx6sabresd/mx6sabresd.c +++ b/board/freescale/mx6sabresd/mx6sabresd.c @@ -313,7 +313,7 @@ int board_video_skip(void) if (!panel) { for (i = 0; i < ARRAY_SIZE(displays); i++) { struct display_info_t const *dev = displays+i;
if (dev->detect(dev)) {
if (dev->detect && dev->detect(dev)) {The same should happen on the Nitrogen board. Should this fix extended to the other boards using hdmi ?
This isn't needed yet in the stock (main-line) code base, since we haven't added any panels without detection.
We do have a form this patch and a lot of panels in our local tree on Github, but wanted to avoid unnecessary noise on the list because we've integrated a dozen or so panels and the existing structure really doesn't scale.
We had a separate discussion regarding treating the displays as data (environment), but have stalled somewhat on that front.
The current device-tree code for i.MX6 uses mode strings instead of the detailed timing data that's really needed for a proper solution.
Regards,
Eric