me again, 2016-06-10 13:20 GMT+02:00 Andreas Fenkart <afenkart@xxxxxxxxx>: > >> The omap_hsmmc.c driver supports configuring two interrupts, and >> two named pinctrl states. See PINCTRL_STATE_IDLE in the driver. >> >> You need to define the dat1 GPIO mux mode as the PINCTRL_STATE_IDLE >> in the dts file, and add that as the second interrupt using the >> interrupts-extended binding. > > Actually it should work without this. The throughput would be lame > though, since it would fallback to polling mode. > > http://www.mjmwired.net/kernel/Documentation/devicetree/bindings/mmc/ti-omap-hsmmc.txt#61 just realized that am33xx.dtsi, declares the mmc ports to be compatible with "ti,omap4-hsmmc" mmc1: mmc@48060000 { compatible = "ti,omap4-hsmmc"; That way the omap_hsmmc driver will never emit a warning for missing pinctrl states, and NOT fallback to polling. That would explain the problems that Michael observed. This probably should be changed to: --- a/arch/arm/boot/dts/am33xx.dtsi +++ b/arch/arm/boot/dts/am33xx.dtsi @@ -357,7 +357,7 @@ }; mmc1: mmc@48060000 { - compatible = "ti,omap4-hsmmc"; + compatible = "ti,am33xx-hsmmc"; ti,hwmods = "mmc1"; ti,dual-volt; ti,needs-special-reset; Guess this makes sense for all platforms based on am33xx and points people into the right direction quicker. /Andreas -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html