On 13.05.2023 06:25, Adam Ford wrote: > On Fri, May 12, 2023 at 4:02 PM Marek Szyprowski > <m.szyprowski@xxxxxxxxxxx> wrote: >> On 12.05.2023 22:00, Adam Ford wrote: >>> On Fri, May 12, 2023 at 2:37 PM Lucas Stach <l.stach@xxxxxxxxxxxxxx> wrote: >>>> Am Samstag, dem 06.05.2023 um 14:24 -0500 schrieb Adam Ford: >>>>> The DPHY timings are currently hard coded. Since the input >>>>> clock can be variable, the phy timings need to be variable >>>>> too. Add an additional variable to the driver data to enable >>>>> this feature to prevent breaking boards that don't support it. >>>>> >>>>> The phy_mipi_dphy_get_default_config function configures the >>>>> DPHY timings in pico-seconds, and a small macro converts those >>>>> timings into clock cycles based on the pixel clock rate. >>>>> >>>> This week I finally had some time to take a deeper look at this series >>>> and test it on some of my systems. >>> Thanks for testing this! >>>> This patch causes issues when the burst clock rate is fixed by >>>> supplying the DT entry. Instead of describing the issue below, I'm >>>> attaching the patch that makes things work on my system. >>> Oops, sorry about that. >>> >>>> I would appreciate if you could test this one on your side. Feel free >>>> to squash it into yours if you find it working properly. >>> I reviewed your patch, and it looks like it makes a lot of sense. >>> If it works, I'll squash them together and add your name to the sign-off. > That worked really well, I'll add it to my WIP directory since Marek S > said he'd test the other proposal of dropping the dynamic phy flag and > corresponding check in favor of pushing everyone to the same code. > >>>> Also I would almost bet that dynamic_dphy is working on the Exynos >>>> boards with that fix added. So if anyone with access to those boards >>>> would like to give it a shot, we may be able to get rid of the >>>> hardcoded PHY parameters altogether, which would be a nice cleanup. >>> I wondered the same thing, but I didn't want to create more work for >>> Marek S and since there was so much churn getting the original driver >>> ported, I thought it would be the safest thing to try to give the >>> imx8m m/n/p the features without breaking the Exynos. >>> >>> Marek S - Do you want me to post this file without the extra checks to >>> see if it still works with Exynos? >> Feel free to send me patches to test or just point to your >> work-in-progress git repo. > Thanks for testing this, Marek S. My work-in-progress branch is: > > https://protect2.fireeye.com/v1/url?k=2eeb1ed9-4e098384-2eea9596-000babd9f1ba-9ad5c339e5ea6e4d&q=1&e=652be603-d622-4d0e-95d3-639656ab1af1&u=https%3A%2F%2Fgithub.com%2Faford173%2Flinux%2Ftree%2Fdsim-updates-wip > > Depending on what you find will determine how I modify the next > revision of the code I push, so I very much appreciate your feedback. > Hopefully the suggestion from Lucas will work for your applications > and we can reduce some of the code complexity. The above mentioned 'dsim-updates-wip' branch works fine on all my Exynos based boards. Best regards -- Marek Szyprowski, PhD Samsung R&D Institute Poland