* pabs
suggests getting the old wiki saved to archive.org by the ArchiveTeam folks before the switchover happens
<libv>
pabs: we have several full backups
<libv>
i pull a daily one here for local storage myself
<libv>
karlp: good catch, thanks
<libv>
pabs: the data is one thing, and the images/ dir, that's all backed up
<libv>
that's the most important bits
<libv>
the issues we might have is mediawiki extensions
<libv>
the ones we were using on the old version i had to go and find newer versions matching the latest mediawiki, and sometimes had to find replacement packages
<libv>
and it's those things that are, thanks to good mediawiki infrastructure (db upgrade is nice), easy to miss, or not catch all corner cases
<libv>
like these pngs
<libv>
even though i somehow feel like they are not really necessary here
<karlp>
I've got no comment on page design :)
<libv>
:)
<libv>
yeah, this needs to be fixed
<libv>
the png issue
<libv>
not the page
<libv>
because if i fix the contents, it will vanish anyway
<libv>
hrm, could this be because we renamed images/ uploads/...
grming has joined #linux-sunxi
JohnDoe_71Rus has quit []
bauen1 has joined #linux-sunxi
<mnemoc>
libv: MBOX_icon_information.png ended in overlay/images instead of overlay/uploads
JohnDoe_71Rus has joined #linux-sunxi
cnxsoft has quit []
evgeny_boger has quit [Quit: evgeny_boger]
evgeny_boger has joined #linux-sunxi
apritzel has quit [Remote host closed the connection]
<libv>
mnemoc: i put it there to see if that would fix it
<libv>
which it didn't
<libv>
and then my 6y old sidetracked me again
<libv>
mnemoc: google gives me nothing on that issue
<libv>
it's in templates only
Net147_ has joined #linux-sunxi
Net147 has quit [Read error: Connection reset by peer]
<libv>
heh, works ok on my build machine
NekoMay has quit [Ping timeout: 480 seconds]
<mnemoc>
meh :-/
apritzel has joined #linux-sunxi
<libv>
no rename of images to uploads there
grming has quit [Remote host closed the connection]
grming has joined #linux-sunxi
NekoMay has joined #linux-sunxi
<mnemoc>
maybe it would be simpler to revert the uploads/ change and rely on the overlay
pmp-p has joined #linux-sunxi
vagrantc has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
apritzel has quit []
ftg has joined #linux-sunxi
<upnix>
After writing u-boot-sunxi-with-spl.bin to the correct location on an SD card, should I notice the boot process of my device acting any differently when the SD card is inserted? I see on the "Bootable SD card" wiki page that there are several more steps to get a rootfs loaded, but I'd like to make sure I got the bootloader part right before I move on.
<upnix>
This is a T95H-like device, and U-Boot was compiled using the X96 Mate defconf file on U-Boot v2023.01-rc2.
<upnix>
The bin file was written with: dd if=u-boot-sunxi-with-spl.bin of=/dev/sdd bs=1024 seek=8
<DuClare>
Surely u-boot prints its version at some point during bootup?
<upnix>
It does. It's the original version. I should say - there's not change in boot behaviour whether the SD card is loaded or not. It's identical according to a diff
<DuClare>
The first thing I'd do is figure out the boot priority for your device
<DuClare>
(Or just try to neuter whatever other thing it is booting from)
<upnix>
It's booting from the eMMC.. And the way I'd neuter it would be to short something.. I suppose. But there may be no way coming back from that, and I don't even know if that would be sufficient
<DuClare>
Is there a fel mode button?
<upnix>
There is
<upnix>
I'm reading now about booting over USB using FEL. Is that what you're thinking?
<DuClare>
Possibly. Also wondering whether it might affect boot order
<DuClare>
Your chip may also have boot select lines
<DuClare>
H616 does, according to its datasheet
<DuClare>
Incredible, it has 5 bits to select from all of 6 possible boot devices
<DuClare>
And one of them is "no support"
<DuClare>
:D
apritzel has joined #linux-sunxi
JohnDoe_71Rus has quit []
<apritzel>
upnix: that's odd, it should boot from the SD card
<apritzel>
if you check the "3.4.2.1.1. GPIO Pin Select" table in the H616 manual, you will see that SMHC0 is always the first, regardless of the boot select pins
<apritzel>
upnix: can you make sure that it reboots correctly?
<apritzel>
sometimes the UART leaks some current, that prevents proper booting
<apritzel>
also possibly try a different SD card
<jernej>
IIRC secure boot also changes boot order, but that isn't enabled on vast majority of TV boxes
<DuClare>
There is efuse based boot select too
<upnix>
apritzel: Secure boot might be enabled? On the console I see: NOTICE: secure os exist \ secure enable bit: 1 \ secure storage read hdcpkey fail \ secure storage read hdcpkey fail with:-1
<apritzel>
upnix: it's possible, though the word "secure" doesn't mean much when it comes to Allwinner ;-)
<apritzel>
it's very rare to see secure-boot enabled Allwinner boxes, the only one I know is the Remix Mini PC
<apritzel>
upnix: to verify that, you can enable TOC0 secure boot support in U-Boot
<upnix>
Am I wrong to expect *any* sort of indication from the stock UBoot that it's reading the SD card? That's what worries me a bit.
<apritzel>
make menuconfig, type '/' for the search dialog, type 'TOC0', then press '1' to go to the menu and enable TOC0
<apritzel>
U-Boot isn't involved at all in reading the SD card, that's all done by the boot ROM (BROM)
<upnix>
I'll try to TOC0 after I try a different micro SD card
<upnix>
Ah, I see. So I'm not seeing that going on at all
<apritzel>
the BROM always tries the SD card first
<apritzel>
it if doesn't like that, or if it doesn't find a valid bootable image, it tries the other boot sources
<apritzel>
looking for TOC0 (secure boot) instead of eGON (non-secure) would explain it
dikiy has joined #linux-sunxi
<apritzel>
but again it's rare, and by experience there is most likely something else going on
<apritzel>
remove power, disconnect the UART cables, re-connect the UART, then re-apply the power
<upnix>
Hey! Compiling with TOC0 actually got me something new:
<upnix>
U-Boot SPL 2023.01-rc2 (Dec 01 2022 - 13:43:28 -0700) \ DRAM: 4096 MiB \ Failed to set core voltage! Can't set CPU frequency \ SP
<upnix>
Is this double bad? Secure boot is enabled, and the x96 Mate config isn't setting the CPU frequency?
<apritzel>
upnix: that's very interesting, you have found the second secure boot device ;-)
<apritzel>
but as you can see the banner, the "secure" part is already solved
<apritzel>
that's what I meant above with Allwinner's loose interpretation of secure
<apritzel>
upnix: do you have the box still open, or did you take pictures?
<apritzel>
so this message tells me that this box doesn't have a PMIC (which is quite common for the cheaper boxes)
<apritzel>
ah yes, that's what I see all over Aliexpress ;-)
<upnix>
Do you want a picture of the board itself?
<apritzel>
upnix: yes, those pictures help tremendously, but it would be even better to have a Wiki page
<apritzel>
so that you can upload your pictures there, and that we have all information in one place
<apritzel>
so did you have Trusted Firmware compiled and copied or linked as bl31.bin?
<upnix>
Yeah, I cloned the arm-trusted-firmware repository, compiled with "make CROSS_COMPILE=aarch64-linux-gnu- PLAT=sun50i_h616 DEBUG=1 bl31", then compiled U-Boot with "make CROSS_COMPILE=aarch64-linux-gnu- BL31=../arm-trusted-firmware/build/sun50i_h616/debug/bl31.bin"
<apritzel>
and for U-Boot: please copy the x96_mate_defconfig file, remove every line with I2C in the copy, and add "CONFIG_SPL_IMAGE_TYPE_SUNXI_TOC0=y"
<apritzel>
very good, that's exactly right
<upnix>
Ok, I was going to try compiling again from the master branch of U-Boot, but I'll go back to v2023.01-rc2
<apritzel>
also add "# SUNXI_NO_PMIC is not set"
<apritzel>
wrong
<apritzel>
also add "SUNXI_NO_PMIC=y"
<apritzel>
the H616 defaults to a PMIC, because 2 out of 2 supported boards have one ;-)
<apritzel>
upnix: do you have a download link to any vendor firmware? any version would do
<upnix>
I do have an image that I was able to use to unbrick using PhoenixSuit
<apritzel>
or can you post the serial output of the vendor firmware, to check the DRAM type? If it's LPDDR3 or DDR3
<apritzel>
yeah, that one
<upnix>
I'm thinking it has DDR3 - DRAM Type =3 (3:DDR3,4:DDR4,7:LPDDR3,8:LPDDR4)
<apritzel>
that's good, because that's already supported by U-Boot
<apritzel>
did the mainline U-Boot produce more output?
<upnix>
The mainline U-Boot only produced those 4 lines
<apritzel>
mmh, interesting, maybe disabling the PMIC helps, though it looks non-fatal in the code - except for the CPU still running at a very low frequency
bauen1 has joined #linux-sunxi
<upnix>
Power on now gets me two fewer lines: U-Boot SPL 2023.01-rc2 (Dec 01 2022 - 14:20:54 -0700) \ DRAM: 8192 MiB
<upnix>
And my RAM doubled?
<apritzel>
ah, that would explain something
<apritzel>
so the DRAM setup isn't right
<apritzel>
it's close enough to pass the setup, but anything trying to run from it fails (like TF-A or the U-Boot stack from that point on)
<upnix>
In the menuconfig I saw there were some options for increased output - is that worth setting?
<apritzel>
I don't think menuconfig helps here. Adding #define DEBUG at the very beginning of source files does, but in this case I doubt it
dikiy has quit [Ping timeout: 480 seconds]
<apritzel>
upnix: if you look at the orangepi_zero2_defconfig, there are more DRAM parameters you could try
<upnix>
I'll give them a shot
<apritzel>
or you could also lower the DRAM frequency
<apritzel>
try CONFIG_DRAM_CLK=624
<apritzel>
we default to 720 MHz, which is somewhat optimistic, and requires that all other parameters are spot on
evgeny_boger has quit [Ping timeout: 480 seconds]
<upnix>
U-Boot compiled with CONFIG_DRAM_CLK=624 shows the right RAM amount, but it hangs on "SPL:"
<upnix>
So I'm going to carry on with seeing what the orangepi_zero2_defconfig has that I might be able to use
<apritzel>
try just: "CONFIG_DRAM_SUN50I_H616_READ_CALIBRATION=y" and "CONFIG_DRAM_SUN50I_H616_UNKNOWN_FEATURE=y"