On Monday 11 March 2013 05:58 PM, Tomi Valkeinen wrote:
> On 2013-03-05 16:17, Archit Taneja wrote:
>> The support outputs struct for overlay managers is incorrect for OMAP4. Make
>> these changes:
>>
>> - DPI isn't supported via the LCD1 overlay manager, remove DPI as a supported
>> output.
>> - the TV manager can suppport DPI, but the omapdss driver doesn't support
>> that
>> yet, we require some muxing at the DSS level, and we also need to
>> configure
>> the hdmi pll in the DPI driver so that the TV manager has a pixel clock.
>> We
>> don't support that yet.
>>
>> Signed-off-by: Archit Taneja <archit at ti.com>
>> ---
>> drivers/video/omap2/dss/dss_features.c | 6 ++----
>> 1 file changed, 2 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/video/omap2/dss/dss_features.c
>> b/drivers/video/omap2/dss/dss_features.c
>> index d7d66ef..7f791ae 100644
>> --- a/drivers/video/omap2/dss/dss_features.c
>> +++ b/drivers/video/omap2/dss/dss_features.c
>> @@ -202,12 +202,10 @@ static const enum omap_dss_output_id
>> omap3630_dss_supported_outputs[] = {
>>
>> static const enum omap_dss_output_id omap4_dss_supported_outputs[] = {
>> /* OMAP_DSS_CHANNEL_LCD */
>> - OMAP_DSS_OUTPUT_DPI | OMAP_DSS_OUTPUT_DBI |
>> - OMAP_DSS_OUTPUT_DSI1,
>> + OMAP_DSS_OUTPUT_DBI | OMAP_DSS_OUTPUT_DSI1,
>>
>> /* OMAP_DSS_CHANNEL_DIGIT */
>> - OMAP_DSS_OUTPUT_VENC | OMAP_DSS_OUTPUT_HDMI |
>> - OMAP_DSS_OUTPUT_DPI,
>> + OMAP_DSS_OUTPUT_VENC | OMAP_DSS_OUTPUT_HDMI,
>>
>> /* OMAP_DSS_CHANNEL_LCD2 */
>> OMAP_DSS_OUTPUT_DPI | OMAP_DSS_OUTPUT_DBI |
>>
>
> Thanks, I'll apply this to omapdss fixes branch.
Hi, just one point here, this patch is a prerequisite for the patch 2/4
in this series. So we need to make sure that the 2/4 patch is not
without this one in a kernel.
Tomi,
About patch '2/4', could you have a look at it too? It basically tries
to do a dynamic assignment of channels to outputs. I worked on this
before you posted the misc series with recommended_channel for outputs.
This patch tries to figure out managers with supported_outputs. It isn't
the most optimal way, as it can't back track and chose a better manager,
but it still seems to do a reasonable job.
We could also use the recommended channel way for omapdrm, I can't
figure out what's the better approach at the moment.
Archit
