Hello Mike, Stephen & All, I am working with device drivers for some ROHM components. Recently I clk did a driver for bd71837/bd71847 PMICs. The PMIC is connected using i2c and contains a simple clk gate. If the connecting bus could be mmap'd I would have used the clk-gate provided by common clk framework. But as it is i2c I ended up making the driver and enable/disable functions myself.
I am now working with next PMIC - which once again has similar clk gate in it. I could write new driver again - but it feels a bit stupid as it would be mostly just a copy-paste of bd71837/bd71847 - driver. I first thought of doing regmap based 'clk-gate, clk-div and clk-mux'. But after some googling I found https://patchwork.kernel.org/patch/5955831/ https://lore.kernel.org/lkml/[email protected]/ and reading comments makes me to assume there would be more... Current way seems to be that we have vendor/platform specific sub folders which all seem to more or less implement own regmap-based basic clocks. (Well, I randonly skimmed through the qcom and meson subfolders and aty least they did - I guess same holds with others). I sure could create 'rohm' subfolder and build up the 'basic clocks using regmap' there - and put ROHM component related drivers in it. But that does feel really stupid. Maybe the series worked by Matthias (last link in the mail) could be reworked - or maybe we could drop one of the platform specific regmap implementatios from subfolder to generic portion? I think that for example moving the drivers/clk/qcom/clk-regmap.c to drivers/clk/clk-regmap.c so that it could be used without enabling qualcomm clk configs would be a good start. Seeing the number of chips connected with i2c/spi/... increasing - and seeing the comments from community when Matthias sent his patch series - I'd say the basic clocks supporting regmap would be very welcome... But maybe I am missing something? I would like to start working with the clk driver for the new ROHM PMIC - but I would like to take sane approach. Any good adviece on how to proceed? Best Regards Matti Vaittinen -- Matti Vaittinen ROHM Semiconductors ~~~ "I don't think so," said Rene Descartes. Just then, he vanished ~~~
