Hi, * Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> [190501 19:03]: > On 2019-05-01 10:44:31 [-0700], Tony Lindgren wrote: > > Hmm so I tried without "earlycon" in command line thinking it might be > > happening with just "earlyprintk" but still no luck. > > > > BTW, in general you might want to update your kernel command line > > options to: > > > > debug earlyprintk earlycon > > debug. Let me look if I manage to hide that `debug' from systemd… Oh that.. I've been quite happy with openrc now for years :) > > Looking at the oops 0xfa1cc000, so 0x481cc000 I guess which is d_can0? > > That node around it I guess. OK I found two issues. It seems that d_can also needs osc clock on am335x. And there's no revision register for d_can.. We're now reading the CTL register unnecessarily. Below is what I hope fixes the boot issue for you, care to boot test? If this helps I'll send out proper patches for for both issues. Regards, Tony 8< ---------------------- diff --git a/arch/arm/boot/dts/am33xx-l4.dtsi b/arch/arm/boot/dts/am33xx-l4.dtsi --- a/arch/arm/boot/dts/am33xx-l4.dtsi +++ b/arch/arm/boot/dts/am33xx-l4.dtsi @@ -1762,8 +1762,9 @@ reg = <0xcc000 0x4>; reg-names = "rev"; /* Domains (P, C): per_pwrdm, l4ls_clkdm */ - clocks = <&l4ls_clkctrl AM3_L4LS_D_CAN0_CLKCTRL 0>; - clock-names = "fck"; + clocks = <&l4ls_clkctrl AM3_L4LS_D_CAN0_CLKCTRL 0>, + <&dcan0_fck>; + clock-names = "fck", "osc"; #address-cells = <1>; #size-cells = <1>; ranges = <0x0 0xcc000 0x2000>; @@ -1785,8 +1786,9 @@ reg = <0xd0000 0x4>; reg-names = "rev"; /* Domains (P, C): per_pwrdm, l4ls_clkdm */ - clocks = <&l4ls_clkctrl AM3_L4LS_D_CAN1_CLKCTRL 0>; - clock-names = "fck"; + clocks = <&l4ls_clkctrl AM3_L4LS_D_CAN1_CLKCTRL 0>, + <&dcan1_fck>; + clock-names = "fck", "osc"; #address-cells = <1>; #size-cells = <1>; ranges = <0x0 0xd0000 0x2000>;
