10:47 <robmur01_> nomis: without knowing what exactly yocto is producing and from what sources it's hard to say, but seeing no output whatsoever suggests that the initial loader (miniloader/idbloader.img depending on which u-boot flow) isn't getting flashed, since that normally says *something* even if it fails to progress to the main bootloader stage

10:50 <robmur01_> for boxes that stick closely to the reference design with LPDDR3 and RK805 PMIC, upstream u-boot for the EVB config should be close enough to at least boot from eMMC or SD (certainly that's the case on my Beelink) - for LPDDR4 or the cheapo boards that use PWM regulators instead of a PMIC I'm not so sure

11:01 <robmur01_> also, mixing upstream and downstream images on the same eMMC/SD card can lead to weirdness where you flash all the new images to the right places yet it still somehow boots the old downstream crap from somewhere - zeroing out the entire first 8MB of the device cleans that up.

11:09 <nomis> robmur01_: I am a little bit embarrassed to admit it, but I've confused rk3328 and rk3288. I think I've built a wrong bootloader.

11:10 <nomis> (right now trying to massage the u-boot for the rk3288 to build. It does not yet)

11:14 <nomis> robmur01_: yeah, that the "Loader" mode of the bootloader silently does not erase the start of the emmc on the "ef" command confused the heck out of me.

11:27 <robmur01_> oh yes, 3328 code on 3288 will be very dead as it's the wrong ISA :)

11:27 <nomis> well, it was 3288 code on a 3328, but I guess the same is valid there :)

11:30 <robmur01_> heh, well FWIW the other day I brought my 3288 box back to life with current upstream u-boot, and can confirm that that works too :)

11:32 <nomis> I am right now a bit in a pickle, since my yocto-tree is still based on the "thud" release which still has this "u-boot-rockchip_20171218.bb" recipe which fails to build for me.

11:33 <nomis> I wonder if I should just bite the bullet and switch to dunfell.

18:00 <JPEW> nomis: The rockchip support has improved significantly since thud

18:00 <JPEW> especially for rk3288 and rk3399

18:12 <nomis> JPEW: should I just use master?

18:27 <JPEW> either master or dunfell should be fine

18:28 <Cruft> Does the RK3399 take well to undervolting the big cores?

18:29 <Cruft> I noticed in tablet usage coreboot settings some places bumped the little core voltage and lowered big, but i wanted to know if both lows are acceptable and people are just being conservative