tl_lim has quit [Read error: Connection reset by peer]
florian_kc has joined #linux-sunxi
AneoX has joined #linux-sunxi
fevv8[m] has quit [Quit: Idle for 30+ days]
suprothunderbolt has quit [Ping timeout: 244 seconds]
[right] has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
dev1990 has quit [Remote host closed the connection]
dev1990 has joined #linux-sunxi
matthias_bgg has quit [Ping timeout: 256 seconds]
matthias_bgg has joined #linux-sunxi
florian_kc is now known as florian
\\Mr_C\\ has joined #linux-sunxi
matthias_bgg has quit [Ping timeout: 256 seconds]
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-sunxi
matthias_bgg has joined #linux-sunxi
matthias_bgg has quit [Ping timeout: 256 seconds]
<Jin^eLD>
doh.. I still cant let it go :P so I booted with an old image, I have access to /dev/nand which is not an mtd device, so can't use nandwrite and so on; if I just dd it into /dev/nand i.e. like this: dd if=u-boot-sunxi-with-spl.bin of=/dev/sdX bs=1024 seek=8 I get a lot of garbage when booting before it jumps into fel
<Jin^eLD>
so I am kind of almost there but still missing something
<KotCzarny>
legacy uboot uses different 'encoding' than legacy linux
<KotCzarny>
so uboot part is written to nand in different bits
<Jin^eLD>
so what would be the way to write it from legaccy linux?
<KotCzarny>
when i was hacking my a13 based ereader, i cycled between uboot and linux
<Jin^eLD>
well.. no legacy linux does not see /dev/nand
<KotCzarny>
i copy nand to ram
<Jin^eLD>
and legacy linux does, but I understand - flashes incorrectly
<KotCzarny>
then from ram to sdcard
<KotCzarny>
and viceversa
<Jin^eLD>
do you still have your tool?
<KotCzarny>
no tool, manual work
<Jin^eLD>
oh
<KotCzarny>
but i assume livesuit and/or wiki has a way to write uboot to nand to be bootable
<Jin^eLD>
well thats the thing - eveything points to livesuit but it did not work for me
<Jin^eLD>
probably because its ancient and outdated? not ssure
<Jin^eLD>
livesuit is not opensoure right? i could not find the code anywhere
<KotCzarny>
afaik livesuit is just implementation of fex protocol on the pc side
<Jin^eLD>
I would assume so, but that nand part of it is missing in the command line tools
<Jin^eLD>
but could you elaborate on your RAM method a bit?
<KotCzarny>
boot legacy uboot, nand read to memory, reboot to mainline linux, write from memory to file
<KotCzarny>
then when i was restoring, i did in reverse
<KotCzarny>
i had to reboot because legacy uboot was compiled without ext3/4 support
<KotCzarny>
and without sdcard support which was way more important
<Jin^eLD>
ok wait... so lets assume I have a nand.mbr backup which I want to write to nand
<Jin^eLD>
and I can boot from sdcard or fel/usb
<Jin^eLD>
I boot into legacy u-boot
<KotCzarny>
most likely i could've gotten away by recompiling legacy uboot, but i wasnt able to do it (some files were missing)
<Jin^eLD>
ah, then load the file into RAM and write from there to nand?
<KotCzarny>
although hmm, i wonder if i could've gotten away by simply using legacy kernel
<Jin^eLD>
is "Linux-3.4.79-sun7i" a legacy kernel?
<KotCzarny>
yup
<Jin^eLD>
well, dd'ing u-boot from it did not help
<KotCzarny>
was it the uboot-with-spl or just uboot?
<Jin^eLD>
good question.. the only thing I managed to get booted from teh old stuff was "cubian", because it had a normal dumped image which you could dd onto the sdcard
<KotCzarny>
although i wonder if spl has to have nand support enabled to read the uboot
<Jin^eLD>
and on cubian they apparently had a nandmbr.backup
<Jin^eLD>
I tried flashing it, did not help
<Jin^eLD>
I need to find a "proven" and working legacy u-boot binary
<Jin^eLD>
and try to flash that
<Jin^eLD>
such a mess
<KotCzarny>
i wonder if nand can be made bootable with mainline uboot/kernel
<Jin^eLD>
I tried enabling nand boot media in mainline u-boot but I did not really see anything... there are some older dts files floating around which were supposedly workng with mainline a few years back
<KotCzarny>
well, you can try some 4.x kernel those patches were made for
<Jin^eLD>
I am not that much of a lowlevel hardware guy so I could only trial and error ancient patches.
<Jin^eLD>
ah.. I got set off by "MLC NANDs are not well supported right now. You can enable MLC NAND support, but remember that you risk data loss." and started my quest in the legacy section :P
<Jin^eLD>
assuming I could get original cubietruck binaries and be done with the flashing in no time...
<Jin^eLD>
what a naive thought that was
<Jin^eLD>
:)
<KotCzarny>
one nice thing about allwinner is it's almost unbrickable
<KotCzarny>
even if you destroy the nand, it will happily boot from fel/sdcard
<Ashleee>
the boot order is safe indeed
<mru>
that's true of most socs these days
<KotCzarny>
mru: qualcomm?
<Ashleee>
I have rk3399 on its way, I think I need to bridge two pins to get it to prefer SD card over eMMC
<mru>
no idea
<mru>
KotCzarny: is anyone allowed within a 10-mile radius of the datasheets?
<Jin^eLD>
Ashleee: yes, it can be told to prefer internal emmc or sdcard by a jumper setting.. at least thats how it is on an rk3288/asus tinker board s
<Jin^eLD>
mru: I managed to kill an stm32mp1 eval board by accidentally flashing wront tf-a firmware
<Jin^eLD>
so my unbrickable expectations got a bit shattered there :>
<Ashleee>
people have killed AVRs by somehow overwriting its device ID and thus not being recognized by programmers anymore
<Ashleee>
(ie something you'd expect to be one time fuse burn)
<KotCzarny>
theoretically you can burn fusebits on aw too
<mru>
I was thinking of normal application processors like TI and nxp
<Jin^eLD>
I think thats what happened on my ST board, I flashed an "ssp" version of tf-a by mistake which apparently burned some otp fuses without asking or checking - the board did not support ssp, which was pretty much the last error message I saw from it before it stopped reacting to anything :>
[right] has quit [Quit: shhhhhhhhhhhhhhhh.]
<mru>
well, if you start burning efuses, you can obviously get into an unrecoverable mess
<mru>
don't some chips have a master efuse that if burned blocks any further changes?
gaston1980 has joined #linux-sunxi
matthias_bgg has joined #linux-sunxi
suprothunderbolt has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
vagrantc has joined #linux-sunxi
andy25225 has quit [Ping timeout: 256 seconds]
andy25225 has joined #linux-sunxi
<Jin^eLD>
mru: thing is that I was neither planning to burn anything nor did I know that this fimrware would decide to do it itself...
florian has quit [Quit: Leaving]
AneoX has quit [Ping timeout: 260 seconds]
AneoX has joined #linux-sunxi
netlynx has quit [Quit: Ex-Chat]
lurchi_ is now known as lurchi__
<Jin^eLD>
./spl/sunxi-spl-with-ecc.bin :> now I'm curious how this will work out
<mru>
I have managed to make it work in the past
<Jin^eLD>
I am not sure if I got the flash parameters, i.e. page size etc right
cnxsoft has quit [Remote host closed the connection]
cnxsoft has joined #linux-sunxi
jstein has joined #linux-sunxi
<Jin^eLD>
mru: did you have to enable NAND support in SPL too? the readme does not say anything about that
<Jin^eLD>
I got to the stage where I am in u-boot now after loading those special ecc spls
<Jin^eLD>
but nand erase.chip says "no devices available"
<Jin^eLD>
so I must have messed up somewhere along the way
<KotCzarny>
if you have uboot loaded, then you are past spl already
<mru>
since it's spl that loads u-boot, it has to support whatever medium contains it
<mru>
it's a long time since I messed with this
<mru>
I convinced everybody to use emmc instead
<Jin^eLD>
well spl has been loaded via fel
<Jin^eLD>
hmm
<KotCzarny>
is there 'nand start' or something?
<Jin^eLD>
I dont see anything like it..
<KotCzarny>
or nand detect
<Jin^eLD>
looking at some sources I may need to check what my CONFIG_SYS_MAX_NAND_DEVICE says
<Jin^eLD>
altough thats a setting from an old uboot
<Jin^eLD>
not sure if the new one has it
<Jin^eLD>
hmm
<Jin^eLD>
that should have been set
<Ashleee>
oh hexdump isn't here now :) wanted to say thanks for the emmc patch, it seems to work when forced to SDR mode instead of DDR :P. Ah well, the 64GB card will go into another SBC then :D
<Ashleee>
the memory stuff will be done on sunday possibly
<Ashleee>
maybe tomorrow..
JohnDoe_71Rus has quit [Quit: KVIrc KVIrc Aria 5.0.1, revision: 5.0.1+git-7433-0df9f22f2, build type: debug, sources date: 20160102, built on: 2019-12-08 19:19:20 UTC 5.0.1+git-7433-0df9f22f2 http://www.kvirc.net/]