Hi @cdsousa,
It sounds like there are some bugs in our CMU transform.
You can refer to the Xavier TRM here → chapter 7.5 display controller.
Below is the pipeline of display controller. Before sending data to raster, it will have de-gamma (pre-comp), and re-gamma (post-comp) process. It looks like some data loss happens during this.
The detail is in the TRM section 7.5.2.3. I need to consult this with internal team.
Hi @WayneWWW, thanks.
Is there any news from the internal team?
Also, is this something that can be fixable in the kernel/driver code?
Hi cdsousa,
Sorry for late reply. Is it possible to use the method you found out in previous comment to workaround this issue and proceed your development first. We still need some time internally for this issue.
Also want to clarify, when you set the force_user_degamma. What is the result of below node? Is it still 0?
/sys/kernel/debug/tegradc.common/tegra_win.0/degamma/degamma
Before setting the force_user_degamma to 1, degamma is 1.
After setting the force_user_degamma to 1, degamma is 0.
Thanks for the update!
