01:28 <bkeys> https://u-boot.readthedocs.io/en/latest/board/sifive/unmatched.html

01:28 <bkeys> Hi all, I am trying to build u-boot for my unmatched board

01:28 <bkeys> Can someone tell me what is meant by <U-boot dir> in the second step?

01:31 <bkeys> I am guessing it is https://source.denx.de/u-boot/u-boot (assuming the unmatched supports mainline u-boot)

01:42 <m4t> yes, that is what is meant

01:48 <bkeys> m4t is the uboot supposed to be inside of the opensbi directory, or just in the same parent directory?

01:48 <bkeys> Or does it not matter?

01:49 <m4t> i think it should be outside, but i am not familiar with opensbi or how it relates to u-boot

01:50 <m4t> https://github.com/riscv/opensbi/blob/master/docs/platform/sifive_fu540.md

01:50 <m4t> makes it seem that u-boot is used as a payload for opensbi

01:50 <m4t> make PLATFORM=generic FW_PAYLOAD_PATH=<u-boot_build_dir>/u-boot-dtb.bin

01:54 <bkeys> m4t that documentation appears to be for the unleashed, not the unmatched

01:54 <m4t> i imagine the boot flow is similar

01:54 <bkeys> Since the unmatched uses FU740

01:54 <bkeys> Probably

01:54 <bkeys> I am relatively new to messing with u-boot at this level so I am feeling cautious

01:57 <m4t> yeah, so i think the sifive_fu540.md instructions are doing it differently

01:58 <m4t> i think you 1.) build opensbi 2.) export OPENSBI=/full/path/to/opensbi/fw_dynamic.bin 3.) clone / cd into u-boot 4.) sifive_unmatched_defconfig

01:58 <m4t> it appears that u-boot's Makefile picks up the OPENSBI env variable and does its thing to make a bootable image

01:59 <bkeys> Ah yes that is the connection

01:59 <m4t> https://github.com/u-boot/u-boot/blob/master/Makefile#L1316

02:00 <m4t> as long as you use an absolute rather than relative path for OPENSBI= it shouldn't matter the directory structure

02:01 <bkeys> Yeah I set it to an absolute path, then the error persisted so I am doing a clean build now

02:05 <bkeys> m4t did a clean build and error persisted, not sure if I should take the warning serious or not

02:05 <bkeys> Image 'main-section' is missing external blobs and is non-functional: opensbi

02:05 <bkeys> some images are invalid

02:06 <bkeys> Sounds serious to me

02:12 <m4t> if the file is where you say it is, and the environment is set correctly, not sure, sorry

02:13 <m4t> u-boot Makefile mentions 'make BINMAN_VERBOSE=3', perhaps you could pass that when building u-boot to get more info

02:13 <bkeys> Yeah for some reason echo was blank

02:14 <bkeys> Even though I set it, I set it again

02:14 <m4t> so OPENSBI wasn't set properly?

02:14 <bkeys> Seems so

02:15 <bkeys> I am gonna go through each step to make sure nothing is nuking my variable

02:15 <bkeys> Thank you for the help thus far.

02:15 <m4t> you can also pass it like 'make OPENSBI=/some/path/fw.bin'

02:15 <m4t> no ambiguity there

02:17 <bkeys> Running make without explicit variable, if that does not work then I will explicitly set it

02:17 <bkeys> Nope, error went away

02:17 <bkeys> Onto the next step

02:17 <bkeys> Yep, I have all of the files it mentions

02:18 <bkeys> I will work on this tomorrow

02:18 <bkeys> ttyl, thank you m4t

06:14 <LeSpocky> hello

06:16 <LeSpocky> there's currently no driver in u-boot for reading and using nvmem cells from dts, right?

06:31 <ceene> Hi! Any idea on what to do with a Xilinx Zynqmp based system which launches u-boot correctly using FSBL from Xilinx, but refuses to run uboot SPL? Latest uboot version should support loading of PMU firmware, I've contemplated it all but I can't get any serial output from SPL

13:44 <dgilmore> what could cause

13:44 <dgilmore> https://www.irccloud.com/pastebin/dnjgtmO3/

14:47 <marex> ceene: talk to monstr about all things Xilinx

15:42 <mrnuke> is the gitblab on source.denx.de also used as a bugtracker?

22:06 <pavelow> ceene: are you passing pm_cfg_obj, psu_init and pmufw.bin to the spl build? I think the flags should be: ZYNQMP_SPL_PM_CFG_OBJ_FILE, XILINX_PS_INIT_FILE and PMUFW_INIT_FILE