[Letux-kernel] DSI panels with omapdrm on 4.20-rc

H. Nikolaus Schaller hns at goldelico.com
Tue Dec 18 19:09:18 CET 2018


> Am 18.12.2018 um 15:24 schrieb Tomi Valkeinen <tomi.valkeinen at ti.com>:
> 
> On 18/12/18 13:21, H. Nikolaus Schaller wrote:
> 
>> Now the driver becomes enabled, sends DCS commands to the panel and
>> receives register values.
> 
> I have a DSI video mode panel that attaches to OMAP5 uevm (I think you
> sent it to me long time back), and I made some tests with that.
> 
> Now, oddly, I couldn't get even an old branch working with that, and I
> thought it should have worked fine. I get GFX underflows. But if I use
> kmstest and just show a plane of 400x400, it works. So I think the DSI
> side is ok, and DISPC side is mostly ok, but for some odd reason, DSS is
> not able to fetch pixels fast enough.
> 
> So I tried to get to the same stage (400x400) working with latest -rc7.
> 
>> But panel still stays (almost) black and shows no image.
> 
> And I got this same issue.
> 
>> Somtimes there are errors:
>> 
>> [   20.852710] DSI: omapdss DSI error: VC(0) busy when trying to
>> configure it!
>> [   20.852714] DSI: omapdss DSI error: VC(1) busy when trying to
>> configure it!
>> [   20.852718] DSI: omapdss DSI error: VC(2) busy when trying to
>> configure it!
>> [   20.852721] DSI: omapdss DSI error: VC(3) busy when trying to
>> configure it!
> 
> I did not get these...
> 
>> On other occasions we can see a single vertical line displayed which
>> appears to
>> show some X11 or framebuffer content. At least it changes if we launch
>> e.g. xterm
>> and changes back if we stop it.
> 
> And I saw a similar single line too.

Ok, interesting!

> Comparing it to the (mostly) working kernel, I found out that the bus
> flags are again broken. I attached a quick hack patch that seems to fix
> them, and with that I get a working 400x400 plane, but underflows with
> full-screen plane.
> 
> I couldn't quite figure out how to fix the bus-flags correctly, though.

Hm. For me this patch does not make a difference.

But I think we are close to a solution.

BR,
Nikolaus



More information about the Letux-kernel mailing list