<robmur01>
tuxd3v: yup, just write the images as that page says - the low-level boot doesn't care about any kind of partition table, only that the right things are present at the right absolute addresses
LargePrime has quit [Ping timeout: 250 seconds]
<tuxd3v>
robmur01, many thanks!
<tuxd3v>
in the situation above I still need boot.img writen at 0x8000 ?
<robmur01>
with upstream U-Boot you can have it fetch the kernel from wherever you like, more or less
<robmur01>
on mine I use a separate partition for /boot, but that's mostly out of habit ;)
LargePrime has joined #linux-rockchip
<tuxd3v>
but boot.bin contains any kinf of u-boot related stuff?
<tuxd3v>
boot.img
<tuxd3v>
sorry
<tuxd3v>
:)
<robmur01>
nope, that's the kernel/dtb for the downstream (Android) boot flow
<tuxd3v>
ok
<tuxd3v>
many thanks
<tuxd3v>
documentation is not clear about that
<tuxd3v>
does you use a boot.src script file there for uboot?
<tuxd3v>
to load
<tuxd3v>
or can we use one
<tuxd3v>
to pass kernel parameters and such
<tuxd3v>
instead of extlinux
<robmur01>
that's really down to however you prefer to drive u-boot
<robmur01>
personally I use distro boot because I can't be bothered to faff around compiling scripts to update parameters
<robmur01>
but you're free to use a bootscript if you want to do more complex stuff at boot-time
<tuxd3v>
yea its easy with extlinux
<tuxd3v>
but If i already have a boot.scr script I would like to use use it
<tuxd3v>
:)
vicencb has quit [Quit: Leaving.]
wens has quit [Ping timeout: 265 seconds]
begin has quit [Quit: ]
wens has joined #linux-rockchip
<wens>
hmm, nanopi m4v2 doesn't seem very stable
<wens>
getting what appear to be memory errors
nashpa has quit [Ping timeout: 268 seconds]
nashpa has joined #linux-rockchip
<robmur01>
wens: what sort of errors?
<robmur01>
if it's external aborts and you're using the rkbin "trust" blob, beware the 'missing TEE reservation' issue
drrty has joined #linux-rockchip
<wens>
mem aborts
<wens>
first I was using mainline U-boot with some tweaks to use lpddr4 instead of ddr3, but that always has errors during the boot process
<wens>
now I'm using the bootloader for the Rock Pi 4 from Armbian, which seems to work better
<wens>
this seems to use rkbin trust blob?
vagrantc has joined #linux-rockchip
wens has quit [Ping timeout: 268 seconds]
wens has joined #linux-rockchip
<robmur01>
OK, in that case you might need to hack up a memory reservation from 0x8400000-0x9600000
<robmur01>
that's what the downstream u-boot carves out, and which I've been using as a workaround
<wens>
does mainline u-boot / atf use that as well?
<robmur01>
by default mainline ATF won't include a TEE, which also avoids the problem
<robmur01>
(but currently it also can't reboot properly; you win some, you lose some...)
<wens>
I had more issues with mainline :/
<robmur01>
I've also found that mainline u-boot seems to go a bit wonky if there's no spl-boot-order property
<robmur01>
with that and the ATF power domain patch which is floating around I did finally get upstream u-boot/ATF working nicely on nanopc-t4 the other night
field^Zzz4 has quit [Ping timeout: 268 seconds]
cristian_c has joined #linux-rockchip
<anarsoul>
robmur01: reboot issue has been fixed
<robmur01>
anarsoul: I saw your patch was merged, has the other one landed too?
<anarsoul>
I'm not sure
<anarsoul>
you should ask Kever
<anarsoul>
or whoever was the author for the patch from armbian
<anarsoul>
(it's a bit annoying that armbian folks don't care to submit patches upstream)
<robmur01>
but yeah, with the armbian patch manually applied I can confirm it's now working on my board too :)
<anarsoul>
great
<anarsoul>
wanna submit it? :)
<robmur01>
aww, but I like having precisely one patch in ATF :P
<anarsoul>
but it's so fragile :)
<robmur01>
I dunno, it's coming up on 5 years now and I've not given in yet...
stikonas has joined #linux-rockchip
begin has joined #linux-rockchip
vicencb has joined #linux-rockchip
<tuxd3v>
doesn't mainline idbloader.img + u-boot.itb, already contain optee and all that stuff?
<begin>
you need to build optee independently and provide the resultant ta.bin to uboot
<begin>
s/ta/tee/
<tuxd3v>
I tough that uboot.itb already had that..
<begin>
it will have it if tee-pager.bin/tee.bin was copied to the path that fit_spl_optee.its specifies, before compilation
<tuxd3v>
its a mess the rockchip scheme..
<tuxd3v>
this troubles should have been fixed in makefile..
<begin>
why not fix it then? :)
<tuxd3v>
the problem is that the mess is so big that we don't know were to start..
<tuxd3v>
anny one knows some good reference how to build uboot for rk3399?
afaerber has quit [Quit: Leaving]
afaerber has joined #linux-rockchip
stikonas has quit [Remote host closed the connection]
stikonas has joined #linux-rockchip
ldevulder_ has joined #linux-rockchip
ldevulder has quit [Ping timeout: 252 seconds]
vagrantc has quit [Quit: leaving]
<mmind00>
robmur01: with the current review speed you can easily get 2 more years even if you submit the patch now ;-)
<mmind00>
tuxd3v: put bl31.elf (from building atf) into the build directory, put tee.elf from building optee into the build-directory ... make u-boot.itb
<mmind00>
begin: fit_spl_optee-foo is solely for socs like the rk3229 that only take optee without atf