Dear Laurent, Dear Kieran, On Di, Apr 26, 2022 at 10:37:07 +0200, Eugeniu Rosca wrote: > A barely noticeable (especially if hardly reproducible) display flicker > may not be the biggest concern in the development environment. However, > an automotive OEM will not only notice it, but will also be haunted by > its phenomenon/nature till it is understood in the greatest detail and > ultimately eradicated, to avoid impairing user experience. > > Troubleshooting the above without the right tools becomes a nightmare. > > Since VSPD underruns may indeed cause [1] display flicker, we believe > that having a minimal/lightweight support for detecting and logging > such events would be extremely beneficial. Obviously, this only applies > to VSP2 modules having an interface to DU (i.e. not mem2mem). > > This implementation is heavily inspired by Koji Matsuoka's work [2-3], > but has been refactored to hopefully become production/mainline-friendly > (the original feature is intended for the development environment only). > > [1] https://lore.kernel.org/linux-renesas-soc/20220421161259.GA2660@lxhi-065 > [2] https://github.com/renesas-rcar/linux-bsp/commit/3469001c3098 > ("v4l: vsp1: Add underrun hung-up workaround") > [3] https://github.com/renesas-rcar/linux-bsp/commit/12ea79975a10 > ("v4l: vsp1: Add underrun debug messege option") Would you please provide brief high-level guidance/steering on this patch proposal, to define further direction? Many thanks in advance. Any criticism and/or simply the gut feeling would be very welcome. Best regards, Eugeniu
