ganbold has quit [Read error: Connection reset by peer]
vagrantc has joined #linux-rockchip
ldevulder has quit [Remote host closed the connection]
repk has quit [Quit: WeeChat 2.7.1]
ldevulder has joined #linux-rockchip
repk has joined #linux-rockchip
<vagrantc>
oh no ... on a firefly-rk3288 installed u-boot-tpl/u-boot-spl on eMMC and the u-boot.bin on microSD ... and it's trying to load from eMMC ... but there's nothing there and hangs and doesn't appear to fallback to microSD
<vagrantc>
i think the old way around this is shorting the eMMC? ... or somehow writing to the eMMC ?
<vagrantc>
i should set up udev rules making the emmc non-writeable on these systems...
<vagrantc>
testing bootloaders way too often
<vagrantc>
or rather, the SPL fails to load from eMMC and tries microSD ... but fails for the same reason
<inode>
some kind of boot media is still present if you're seeing bootloader output with the short
<vagrantc>
yes, the eMMC
<vagrantc>
which seems to fail badly enough SPL can't load u-boot from eMMC or microSD, but not badly enough that TPL can't be loaded
<vagrantc>
or i haven't gotten the short to work properly until TPL/SPL has loaded...
<vagrantc>
e.g. not quite shorted while loading TPL/SPL, and then shorted ... but that seems unlikely
<inode>
are you shorting it prior to power on and holding it in place for a few seconds after?
<vagrantc>
yes
<vagrantc>
but can't say my hand doesn't slip at some point
<vagrantc>
e.g. it's a tenuous short at best
<inode>
fair enough :)
<inode>
it might be worth soldering a jumper across temporarily to be sure
<vagrantc>
got it!
<vagrantc>
really wish these boards had a jump for this
<vagrantc>
i fried the eMMC on a firefly-rk3399 trying to fix a similar issue
<inode>
how did you do that?
<vagrantc>
slipped at some point and shorted the wrong things
<vagrantc>
really wish these boards had a jumper for this!
<vagrantc>
now ... do i put u-boot.img or u-boot.bin at 16384 ?
<inode>
make your own? solder a switch or a pin header each to clk and gnd :)
<vagrantc>
i've often thought of asking a freind to do that... but yeah
<vagrantc>
my only solder experience is decades old and with jewelry
<inode>
i had an unresponsive emmc by similar circumstances, but it turned out to be unintentional bridging between emmc clk and gnd (adjacent pads) after shorting them out - almost as though tin whiskers had formed
<inode>
scraping between the pads was enough to solve the issue
<vagrantc>
i think i've occasionally seen the eMMC show up on that machine ... but it's spotty ... so yeah, maybe some stray metal whiskers causing a intermitant short
<vagrantc>
ok, now i'm set to be able to upload u-boot to debian
* vagrantc
does one more cold powercycle to check
<vagrantc>
inode: thanks for the pointers and being a sounding board :)
<inode>
np, i know the just-bricked feeling all too well :)
<vagrantc>
huh. seems like both u-boot.bin and u-boot.img work ... u-boot.img has some headers that help with data corruption, so may as well go with that
elektirnis has quit [Remote host closed the connection]