Hi,
Am Mittwoch, 24. Juli 2019, 14:33:06 CEST schrieb Linus Walleij:
On Tue, Jul 23, 2019 at 5:19 PM Fabian Vogt fabian@ritter-vogt.de wrote:
I gave the series a try (virtual CX and TP so far, will do on a real CX later): OOPS with a nullptr deref during probe. This diff which just moves some lines around fixes that and the LCD appears to work properly.
OK I folded this into my patch, thanks!
Once I verified the 24bit depth and clock speed config on HW as well, I can give you my Tested-by, or would you prefer that I resubmit your series with the fix below?
It's fine if you test it just with your patch as-is, I didn't change anything else.
I would be amazed if it "just works" now.
Indeed, you won't be. On a real CX the LCD displays shows content without any other changes to the set, but has a ~3Hz pulsating effect scrolling from the top to the bottom and the gray text changes color slightly.
Without the patchset applied everything looks perfectly.
I tried setting vrefresh to 20, fb_bpp to 16 and forcing an inverted panel clock, but the pulsing didn't change.
Using the emulated CX I compared the contents of the registers and found that only the IPC bit (which I tried to set, so that's likely not it) and the interrupt registers have a different content.
Any idea?
Thanks, Fabian
Yours, Linus Walleij