<Xogium>
so what happened is, I actually embedded xz compressed data into the kernel image, that is, already compressed
<Xogium>
but then, it turns out I also accidentally asked the kernel to do the compression
<Xogium>
double xz compression, needless to say linux was pissed
<Xogium>
I did not know it would compress it again, I thought this setting was to let it know what format the data was in
<Xogium>
also good news for us, John managed to get the audio working correctly on the odyssey. It turns out our previous pinout was wrong, and we could very well have damaged hte hardware had we tried
peterm6881 has joined #Speedsaver
peterm6881 has quit [Remote host closed the connection]
<sc6502>
Good afternoon Xogium :)
peterm6881 has joined #Speedsaver
<Xogium>
sc6502: hey :)
<Xogium>
how are you doing ?
<sc6502>
Not too bad. I've written the OLED echo program for use with RAUC
<Xogium>
oh yeah ?
<peterm6881>
hey guys
<Xogium>
the kernel stuff drove me mad. Again
<peterm6881>
ooo
<peterm6881>
thats quite a co-incidence sc6502, i logged on hoping to have a chat about whats required to get rauc up and running
<peterm6881>
now that the software side is done
<peterm6881>
brb
<Xogium>
well, rauc itself I could setup
<Xogium>
I've done it plenty of times already
<Xogium>
it's just the program that sc6502 made I'm not sure how to deal with and how to integrate into rauc
<peterm6881>
hi Xogium
<peterm6881>
as I said before, the plan is for us to get rauc working, and then for sc6502 to come in and do his bit to display user feedback
<peterm6881>
what do we need to do for basic rauc functionality
<Xogium>
I need to set it all up
<sc6502>
It's basically a command line utility you pass a few arguments to, along with a string, and it does it's stuff.
<peterm6881>
is there anything I need to do at my end?
<Xogium>
not at the start
<sc6502>
Have you ever seen an error like this peterm6881 ?
<Xogium>
sc6502: yeah that much I gathered, however no clue how to trigger it with rauc. I mean I usually just run rauc install and it does its thing
<peterm6881>
server is booting up
<peterm6881>
what error?
<sc6502>
I'm trying to copy/paste it now ;)
<peterm6881>
lol put a space in front of a forward slash
<sc6502>
Meh, I can copy into my VNC, not out of it :(
<sc6502>
Anyway
<sc6502>
Error parsing config file '/usr/share/navit/navit.xml': Error on line xxxx char yyy: Odd character (then some garbage)
<Xogium>
so, I think tomorrow I'm gonna start workin on a new branch for the repo called speedsaver_odyssey
<Xogium>
huh weird
<Xogium>
did you touch the xml ?
<Xogium>
I assume no
<sc6502>
Weird thing is, the sha1sum of the file on the device is identical to the sha1sum of the same file in the release dir on the host.
<Xogium>
hm so it didn't get corrupted
<sc6502>
And looking at the line:char on the host, it's all fine.
<Xogium>
sounds like navit going nut ?
<peterm6881>
pass. I havent gone near the dreaded navit.xml config file for a very long time.
<sc6502>
Also, printing out said line with awk on the speedsaver, it's fine too.
<Xogium>
is it the first time it does that ?
<sc6502>
Every time now :(
<peterm6881>
Xogium any joy with your bl issue?
<Xogium>
peterm6881: yes, as it turns out this was not a bootloader issue
<sc6502>
I just rebooted, and it's stuck in a failed to start loop thanks to systemctl
<Xogium>
hmm
<peterm6881>
maybe Xogium could send you a fresh navit.xml file, or grab one from a fresh image
<Xogium>
badly flashed micro sd ? Not quite sure
<peterm6881>
ooooo exciting times Xogium , what was it?
<sc6502>
Well it was working just fine earlier today.
<Xogium>
specifically here, it was not a bootloader issue. The kernel itself was at fault, more or less.
<Xogium>
did you brutally unplug it ? Don't forget this thing's not read-only
<Xogium>
so what happened is, I actually embedded xz compressed data into the kernel image, that is, already compressed
<Xogium>
but then, it turns out I also accidentally asked the kernel to do the compression
<Xogium>
double xz compression, needless to say linux was pissed
<Xogium>
I did not know it would compress it again, I thought this setting was to let it know what format the data was in
<Xogium>
also good news for us, John managed to get the audio working correctly on the odyssey. It turns out our previous pinout was wrong, and we could very well have damaged hte hardware had we tried
<peterm6881>
its great you got to the bottom of it, I was worried that was gonna be another 3 week problem!
<Xogium>
well, we had to use jtag to see that linux was panicking
<peterm6881>
im confused about the pinout, surely you were using headphones?
<peterm6881>
good knowledge Xogium
<Xogium>
I mean pinout as in how the device was connected, in the device tree
<peterm6881>
ohh
<Xogium>
the pins were not the right ones
<peterm6881>
brilliant
<peterm6881>
sc6502, I emailed you a fresh navit.xml file from my Speedsaver build directory
<peterm6881>
maybe you can compare them
<Xogium>
the file was not corrupted
<Xogium>
its checksum still match with the one in the build tree
<peterm6881>
it summed ok, and it looked ok, but its broke
<Xogium>
I doubt it is the file
<peterm6881>
yes but does that not only count the bytes
<Xogium>
it checks if every bit matches
<Xogium>
if even one bit was altered it would reflect in an altered checksum
<Xogium>
as far as I know, anyway
<peterm6881>
interesting, ive never looked into what it checks for
<Xogium>
but there's a far more nasty kind of corruption that can happen<
<Xogium>
they call that one bit rot
<Xogium>
and it's litterally what it sounds like, as far as I know. If you were to look physically at a device, you would see the bits getting flipped around and decaying
<peterm6881>
A checksum is a value that represents the number of bits in a transmission message
<Xogium>
that sort of corruption can't be fixed
<peterm6881>
but your hunch may be correct, since ive had personal experience of temperamental sd cards
<peterm6881>
although I THINK the one in sc6502's first prototype is a brand new Kingston one. sc6502 can you check its a Kingston 16GB
<sc6502>
Well it seems I have a "virus" :) I put the old card (which was working) in the new box, now it's busted too. I put both cards in the old speedsaver device, guess what, both are busted.
<peterm6881>
what the very heck
<peterm6881>
sometimes I think I must just lead a charmed life....although I didnt win on that scratchcard earlier
<Xogium>
the most impressive case of bit rot I saw was one file being asked for by systemd at boot
<peterm6881>
sc6502, what are you using to burn your cards?
<Xogium>
it wanted ld-lijux-arexf.qo.3
<Xogium>
come again ?
<Xogium>
that was supposed to be ld-linux-armhf.so.3
<peterm6881>
there you go, Im confident that would have passed checksum
<sc6502>
New card is a Kingston 16GB, old one is some unbranded generic 4GB
<sc6502>
I'll reflash them both tomorrow, but I'm done for today.
<peterm6881>
beware of others, but the Kingston is new and legit
<sc6502>
I'll push the oled_echo stuff tomorrow as well.
<Xogium>
actually bit rot is too low level to be fixed. It can't even be detected by tools
<peterm6881>
sc6502, are you just dd'ing?
<sc6502>
Both awk and sha1sum said the file was fine, and identical to what's on the dev machine.
<sc6502>
yes peterm6881 , just dd
<peterm6881>
thats what I use, that and dcfldd
<Xogium>
it could also be navit that's gone crazy
<peterm6881>
if you are burning one card at a time, just using dd, the process has been what I would describe as reliable.
<Xogium>
the file itself could be okay but navit read it bad, or something
<peterm6881>
its not a £1 card reader off Ebay is it.... no offence
<sc6502>
No, it's got a name on the front of it - Integral - and from Amazon
<peterm6881>
;)
<peterm6881>
just me then...
<Xogium>
also regarding kingston, despite what they'd have you believe, I know for a fact they are one of the worst brands on the market. In fact a linux engineer reversed engineered a couple of them and proved that every single one of those kingston would die before you ever filled them to capacity
<peterm6881>
crumbs
<Xogium>
the firmware is full of crap
<sc6502>
Until tomorrow - have fun :)
<peterm6881>
sc6502, I dont want you to think the process is inherently flaky. You've been unlucky and something got corrupted. I burn 6 cards every lime there is something to pull on the Speedsaver git
<peterm6881>
time
<Xogium>
seeya
<peterm6881>
and I always test each one
<peterm6881>
night sc6502 ;)
<peterm6881>
what were you saying about an Odyssey branch Xogium ?
<peterm6881>
cant wait
<Xogium>
well I'll port the stuff we have to the odyssey board
<peterm6881>
glad you got it sorted
<peterm6881>
cant wait to see the new branch
<Xogium>
well this was definitely exhausting
<peterm6881>
double compression, id never have guessed
<peterm6881>
amazing detective work, and great experience
<Xogium>
neither would I
<Xogium>
and I needed John's help
<Xogium>
to acces jtag on his board for me
<Speedsaver`>
Speedsaver/navit: peterbmckinley commented on issue #7 "Feature request: Blank OLED display after 10 minutes of little or no movement (5MPH / 8KM/H or less) " (https://git.io/JcpbH): Screen blanking became a screensaver in practice. Confirmed working...