Hi,
we are currently porting our system from L4T 32.7.2 to 35.1.0 and experienced some issues with the GMSL cameras we use. After a lot of debugging we pinpointed the problem now to the newly introduced deskewing mechanism if the data rate is >1.5Gbps.
To understand the situation a bit better:
Who waits for the deskewing? The RT-CPU?
In case we want to implement deskewing correct we are required to get our camera firmware patched that it supports it, right?
Would there be a way to deactivate the wait-for-deskewing for legacy cameras (where we aren’t able to change the firmware)? (thinking of patching the kernel module or the RCE/SCE firmware)
Also i checked for the device tree flags for deskewing, but it appears that they aren’t read anywhere (and serve only documentation purposes?)
kernel/nvidia/drivers/media/platform/tegra/camera/sensor_common.c: "deskew_initial_enable", &temp_str);
kernel/nvidia/drivers/media/platform/tegra/camera/sensor_common.c: signal->deskew_initial_enable =
kernel/nvidia/drivers/media/platform/tegra/camera/sensor_common.c: signal->deskew_initial_enable = 0;
kernel/nvidia/drivers/media/platform/tegra/camera/csi/csi.c: deskew_enable = sig_props->deskew_initial_enable;
kernel/nvidia/include/media/tegra-v4l2-camera.h: __u32 deskew_initial_enable;
Best Regards
Thomas