Weston 10.0 New Shader leads to performance drop

Hi all, Already discussed with Pekka Paalanen it turns out new Shader management has ended up in performance drop in out side (STM32MP1 MPU from STMicroelectronics) The pre_mult code added, pre_curve and color mapping changes leads to extra GPU cycles that can't be optimized by the compiler: RCP, MUL and worse JMP

The reason pre-multiplied alpha is first undone is that otherwise it is not possible to apply anything non-identity in color_pipeline(). view-alpha is not needing that, color_pre_curve() and color_mapping() are. For identity pipeline, the pre-mult is not required at all.

I believe some additional "if" prior to run color_pipeline is likely needed here to prevent extra GPU cycles (even more un-wanted)

Thanks in advance Regards