07:35 <naoki> I am testing .dts for mainline Linux/U-Boot on a Radxa E54C

07:35 <naoki> But so far gmac is not working :(

07:39 <naoki> I got the error "EQOS_DMA_MODE_SWR stuck" in U-Boot and "stmmac_hw_setup: DMA engine initialization failed" in Linux.

07:41 <naoki> I think the clock is not running when the gmac is reset, but I don't know where the problem is in the dts.

11:09 <robmur01> ISTR getting weird DMA errors from stmmac on RK3328 when the I/O domain voltage was wrong, can't remember if it was a similar symptom for having the clock input from the phy misconfigured

12:09 <wens> getting a bunch of "unhandled SMC 0x82000003" ATF debug messages on my rk3399

12:20 <robmur01> looks like ROCKCHIP_SIP_SUSPEND_MODE from the power domain driver, where the design seems more "can firmware tolerate this call?" rather than "does firmware actually want it?"

12:24 <wens> doesn't seem upstreamed in atf

12:33 <robmur01> I think it's only actually needed for RK3576 anyway

12:40 <robmur01> even if there isn't a proper discovery mechanism, I guess it could at least check the return value and suppress further calls when it's not supported