<rejon>
Maximum number of bytes in multi-byte program: 1024
<rejon>
Number of Erase Block Regions within device: 1
<rejon>
Erase Block Region Information:
<rejon>
Region 0:
<rejon>
Erase Block Size: 131072 B (128 KiB)
<wolfspraul>
the jtag executable itself, plus liburjtag.so.0
<rejon>
Number of Erase Blocks: 256
<rejon>
Manufacturer: Intel
<rejon>
Chip: 28F256J3A
<rejon>
program:
<rejon>
block 0 unlocked
<wolfspraul>
yes yes, don't post it all here
<rejon>
erasing block 0: 0
<rejon>
addr: 0x0001F000
<rejon>
block 1 unlocked
<rejon>
erasing block 1: 0
<rejon>
addr: 0x0003F000
<rejon>
block 2 unlocked
<rejon>
erasing block 2: 0
<rejon>
addr: 0x0005F000
<rejon>
block 3 unlocked
<rejon>
erasing block 3: 0
<rejon>
addr: 0x00078DD2
<rejon>
verify skipped
<rejon>
Manufacturer: Intel
<rejon>
Chip: 28F256J3A
<rejon>
program:
<rejon>
got to there and froze
<rejon>
hahaha
<rejon>
i love information
<wolfspraul>
install the 2 binaries I sent you
<rejon>
doing
<rejon>
ok, i will reset my m1 and remove power
<wolfspraul>
you may need to setup a liburjtag.so.0 symlink
<wolfspraul>
you'll figure it out
<wolfspraul>
then run /usr/local/bin/jtag --help, if that works it's setup
<wolfspraul>
in reflash_m1.sh, there is only 1 line where it invokes the jtag binary, lines 88
<wolfspraul>
line 88
<wolfspraul>
change that to /usr/local/bin/jtag
<rejon>
done
<wolfspraul>
I'm still downloading, then I can tell you how your stdout is supposed to continue after you see the freezing (not sure this helps you much, but still)
<wolfspraul>
so try running reflash_m1.sh with my jtag binaries now
<rejon>
trying
<wolfspraul>
make sure it picks up the new liburjtag (check with ldd)
<rejon>
i'm running 64bit btw
<wolfspraul>
same here
<rejon>
k
<wolfspraul>
check with ldd /usr/local/bin/jtag which liburjtag it takes...
<wolfspraul>
this is how it continues for me (flashing now):
<wolfspraul>
addr: 0x00078DD2
<wolfspraul>
verify skipped
<wolfspraul>
Manufacturer: Intel
<wolfspraul>
Chip: 28F256J3A
<wolfspraul>
program:
<wolfspraul>
block 5 unlocked
<wolfspraul>
erasing block 5: 0
<wolfspraul>
addr: 0x000BF000
<wolfspraul>
block 6 unlocked
<wolfspraul>
erasing block 6: 0
<wolfspraul>
addr: 0x000DF000
<wolfspraul>
so the first line that is missing for you is "block 5 unlocked"
<rejon>
UrJTAG 0.10 #1902
<rejon>
ok
<wolfspraul>
you are only a few percent into the flashing
<wolfspraul>
in total the flashing should take about a minute, and work its way up to about block 100 (and erase some more up to 255)
<wolfspraul>
there never is any freeze of more than 1 second or so, it should always move further and output stuff
<wolfspraul>
basically just like at the beginning for you, but then continue all the way up to about 100
<wolfspraul>
ok mine is reflashed now, I'm rebooting :-)
<rejon>
right, it worked before like that
<wolfspraul>
unfortunately since you are on fedora and a new kernel I guess, the problem could also come from any of the other libs or kernel
<wolfspraul>
what do you get now? (with my jtag binaries)
<rejon>
UrJTAG 0.10 #1902
<rejon>
ok trying now
<wolfspraul>
don't paste everything again
<wolfspraul>
does your m1 still boot?
<wolfspraul>
there are other ways to reflash (ftp), but I don't know them well/at all
<wolfspraul>
that's assuming that you have at least something running on your m1, like rescue image or whatever (I don't know those things well)
<wolfspraul>
jtag should normally always work, it just pumps data directly into the nor flash chip
<rejon>
ok
<wolfspraul>
ah great
<wolfspraul>
my m1 booted but into a totally broken update screen
<wolfspraul>
in compliance with the milkymist culture, I can say it's a ridiculously fucked up piece of crap I just booted into
<wolfspraul>
I can move the mouse, but I cannot click any of the fuckingly stupid buttons
<wolfspraul>
maybe I wait a little :-)
<rejon>
hmmm
<rejon>
should i go back an image?
<wolfspraul>
:-)
<wolfspraul>
I was joking
<wolfspraul>
how does your flashing go?
<rejon>
getting errors
<rejon>
Unable to open file '/usr/local/share/urjtag/MANUFACTURERS'
<rejon>
  Unknown manufacturer! (00001001001) (/usr/local/share/urjtag/MANUFACTURERS)
<wolfspraul>
the urjtag software should automatically retry and lower the clock, or there should be a way to tell it to do so
<wolfspraul>
all just should unfortunately
<wolfspraul>
rejon lekernel I was thinking about changing the URL on the product label and the box to twitter.com/milkymistvj
<wolfspraul>
good idea or bad?
<wolfspraul>
I want to get people talking
<wolfspraul>
twitter has a clean interface and we have enough ways to link from there to a shop or other material
<wolfspraul>
just thinking...
<lekernel>
rejon, issue the commands manually... flash each file in turn
<lekernel>
so when it fails you can retry at that file
<lekernel>
your jtag cable probably has a problem, it should work at 6MHz. all other cables do.
<lekernel>
wolfspraul, project URL is milkymist.org. and there is a twitter box right on it...
<wolfspraul>
lekernel: you mean the usb cable?
<wolfspraul>
he should try another usb cable?
<lekernel>
yeah, i think so
<wolfspraul>
about the url, my personal estimate is with every redirect you loose 80% of people
<wolfspraul>
so writing twitter.com/milkymistvj on the box, for example, may make someone write it down when seeing the box, or typing it into his android/iphone twitter client on the spot
<lekernel>
write both URLs if you want
<wolfspraul>
I'm not sure, just thinking
<wolfspraul>
lekernel: you sign off the final artwork (at least that's the plan, unless something even bypasses me :-))
<wolfspraul>
but I've seen dramatic losses over redirects, it's amazing
<wolfspraul>
but then when you think about your own behavior when running into new things, maybe it's not so amazing after all
<wolfspraul>
I'm just not willing to follow link after link to something new
<wolfspraul>
and a twitter link may really kick people to ask a question or note it down... (that's the theory)
<wolfspraul>
let's see what rejon thinks about it
<lekernel>
oh, I won't be picky about the artwork you know... my main concern atm is reducing delays.
<lekernel>
so as long as it's reasonably good, fine
<wolfspraul>
well, I mean which url is there etc.
<wolfspraul>
lekernel: how did okcon go? anybody interested in milkymist, or good leads?
<lekernel>
not really, well, I didn't have high hopes about it and only went because it was only a few blocks away
<lekernel>
and some very uninteresting ones as well, like the ohanda panel... omg... they embarked on several hour long discussions mixing the process of a process (yeah right), design, patents, with a good bureaucracy seasoning... WTF
<wolfspraul>
hmm, OK I see.
<lekernel>
ah, and I watched my first rms talk *g*
<wolfspraul>
I feel pretty good moving with m1 just to music and video lovers, so I will not try that much more in outreach to core foss guys, for now
<wolfspraul>
time is limited, I will try my luck with music and video folks first :-)
<lekernel>
yeah way to go
<wolfspraul>
so far reactions there are very positive, I need to find a way to sell on the spot
<wolfspraul>
maybe I always carry a boxed m1 with me :-)
<lekernel>
I always get good feedback from these people :)
<lekernel>
even within the busy production companies of the world's biggest bands
<lekernel>
that being said - Mako's talk was enlightening. I think we aren't bad either as a FLOSS project, it's just that FLOSS itself is bad in general.
<lekernel>
wolfspraul, carry a brochure instead. it's much easier to explain when you can show figures etc.
<xiangfu>
ok. I ask some days off, will full back to work on Sunday evening.
<wolfspraul>
has anybody heard from hadez again, with the freezing problem?
<lekernel>
no
<wolfspraul>
hmm, OK
<wolfspraul>
I need to come up with a good idea to really make people go crazy about our video synthesizer :-) so they finally jump in and enjoy and give feedback, if only to make it work better for themselves...
<wolfspraul>
seems hard to keep people attracted, but we will find the magic sauce eventually
<lekernel>
btw, since I fixed the video chip my M1 never crashed again in dozens of hours of operation
<wolfspraul>
fixed video chip?
<lekernel>
yeah, the ferrite bead problem
<wolfspraul>
wait which problem is this? not the L19 one, right? the L9/L10 we fixed at some point earlier?
<wolfspraul>
hmm, no
<wolfspraul>
lekernel: you mean shorting L19?
<wolfspraul>
I want to start a way to document the patches we have on m1, each patch with a little description, maybe a video of the patch in action with different types of music (don't have a vga grabber yet)
<wolfspraul>
where would you want to do this?
<wolfspraul>
in the milkymist wiki?
<wolfspraul>
same I want to start working on an illustrated patch programming guide, basically explaining the commands and showing what the effects are. is there some milkdrop stuff we can use as a starting point? and where would you do that? milkymist wiki?
<wolfspraul>
I'm just wondering where you would put such things, or whether there are starting points already...
<lekernel>
wolfspraul, L19 yes
<lekernel>
for milkdrop there are two "preset authoring guides" around
<lekernel>
both are largely relevant to flickernoise patches
<lekernel>
btw, you should see Mako's hard numbers about how little people contribute to free software :)
<lekernel>
in terms of contribs, we definitely are in the top 5% projects today
<wolfspraul>
yes sure I read the article
<wolfspraul>
my strategy for new contributors now is not via the gpl & 'free cpu' angle. instead, I want to make this a truly fantastic video synthesizer, whose usefulness as just that immediately comes across
<wolfspraul>
then we can go back to core hackers and see whether that motivates them to do more on this box, and out of this box
<wolfspraul>
so instead of the moral angle, let's try the fun angle :-)
<wpwrak>
argh. the mother of all hangovers is just shaping up
<wpwrak>
rejon_: how's life today ?
<wolfspraul>
wpwrak: he had a rough night
<wolfspraul>
nightmares about jtag clock frequencies
<wpwrak>
;-)
<wpwrak>
did he wake up happy, though ?
<wolfspraul>
probably still sleeping
<wpwrak>
still time left until his talk :)
<wpwrak>
is his m1 working now ?
<wpwrak>
ah, there are mails. checking ...
<wpwrak>
kewl. he solved it :)
<wpwrak>
i wouldn't have trusted myself today to be of much help ;-)
<wpwrak>
4 am. not bad :) that must be around the time i returned from drinking :)
<wolfspraul>
really he solved it? checking mail...
<wolfspraul>
wpwrak: where is the mail that he solved it? are you sure?
<wolfspraul>
last he was tryign to reduce the jtag speed to crazy slow, but it will still failing after some percentage
<wpwrak>
well, whatever "mm1 working" and "i have a solution"-means :) he didn't copy you
<wpwrak>
off to the bathroom now
<wolfspraul>
ah ok
<rejon_>
arg
<rejon_>
i thought solved, but woke up and the flashing froze
<rejon_>
morning wpwrak wolfspraul
<rejon_>
wpwrak, yes, would be great to try flashing from your system
<rejon_>
mine is freezing
<wolfspraul>
sebastien had the idea of trying another usb cable
<wolfspraul>
not a very likely root cause but why not, if you have another one try switching
<wolfspraul>
try flashing from werner's machine, yes
<rejon_>
i already swapped that
<wpwrak>
rejon_: argh
<rejon_>
yeah
<rejon_>
wpwrak, must be my system
<rejon_>
i got my setup here in this room if you want to pop over with your machine
<rejon_>
i was really trying to do this without your help ;)
<wpwrak>
rejon_: i'll be presentable in ~40'
<rejon_>
ok, cool
<rejon_>
sorry dude
<wpwrak>
rejon_: do you want to do the experiments here or at the conference ?
<rejon_>
here
<rejon_>
i have it setup, ready, projector
<wpwrak>
ok. room #?
<rejon_>
719
<wpwrak>
wonders if we'll even be able to build the toolchain on my gentoo
<wpwrak>
an old gentoo, i have to add
<rejon_>
ok
<rejon_>
i'm confident
<wpwrak>
meanwhile, i'm almost certain it won't work :)
<wpwrak>
but there ought to be many people with debian/ubuntu at the venue. so there's a plan B
<wolfspraul>
wpwrak: do you have libusb and libftdi?
<rejon_>
wpwrak, yes
<wpwrak>
libftdi.so.1.14.0
<rejon_>
wolfspraul, lekernel what do the 3 buttons do on mm1 again?
<rejon_>
from left to right?
<rejon_>
next, on, random
<wolfspraul>
next, on, update
<wolfspraul>
or update, on, next
<rejon_>
ok
<wolfspraul>
I'm out for about 30 min, xiangfu will also try to help if there are problems on werner's machine
<wolfspraul>
have to rush buy a few things
<wolfspraul>
good luck!
<rejon_>
ok thanks wolfspraul
<lekernel>
rejon_, flash the files one by one
<lekernel>
at maximum speed that doesn't outright fail
<rejon_>
lekernel, yes, was doing that
<lekernel>
and you can skip the rescue images, they are normally not needed
<lekernel>
in fact the script should never have touched them, just to avoid the situation you are in right now ...
<rejon_>
lekernel, i successfully flashed the rescue images
<lekernel>
oh, and flashing stuff just before demos is generally a bad idea :p
<rejon_>
would that help?
<lekernel>
ok, so just boot in rescue mode now
<rejon_>
lekernel, yeah, we started this yesterday, 24 hours before
<rejon_>
anyway
<lekernel>
if you have standby + *rescue flashed
<lekernel>
booting in rescue mode should work
<lekernel>
then use FTP (not flash) to send the patches
<lekernel>
generally you can always use FTP and the GUI to flash anyway... often simpler than using the cable
<lekernel>
(and it won't break the rescue partitions)
<lekernel>
rejon_, does rescue mode work?
<rejon_>
yes!
<lekernel>
ok, from now you can 1) upload some patches to the filesystem 2) flash the main partitions from the GUI
<lekernel>
2) is optional
<lekernel>
just set a login/password and IP address in system settings and this enables a FTP server
<rejon_>
ok
<rejon_>
shit my keyboard not working, wpwrak you have a mouse? i forgot my mouse, arg
<lekernel>
what version are you running?
<rejon_>
the rescue images are from 5-9-2011
<lekernel>
mh, ok. the latest images (not those) have keyboard shortcuts to emulate the mouse.
<rejon_>
k
<lekernel>
if you are lucky, the filesystem will not have been damaged by your failed flashing attempts and you'll still have the patch files ...
<rejon_>
can you run patches in rescue mode?
<rejon_>
(waiting to get a keyboard
<rejon_>
)
<lekernel>
yes
<rejon_>
ok cool
<rejon_>
what do i need to flash a microsd chip to?
<rejon_>
vfat?
<lekernel>
yes
<wpwrak>
has rodent
<wpwrak>
almost ready
<rejon_>
great
<rejon_>
lekernel, so i have sd card, i put patches onto it
<lekernel>
yeah and you hope the sd card will be recognized
<lekernel>
if you have only a mouse, you should be able to run patches
<lekernel>
keyboard is not needed for this
<rejon_>
ok, i will try this first
<rejon_>
(once wpwrak comes over)
<lekernel>
by the way, have you shorted L19 on your board?
<rejon_>
no we decided to not do, because was too dangerous ;)
<rejon_>
lekernel, so if i have to flash from sd, what files are needed on the sd card?
<lekernel>
use a ftp or the memory card to throw the images from latest xiangfu's snapshot then select the files in the gui
<rejon_>
ok, so i don't need to do: dd if=flickernoise.fbi of=boot.bin bs=1 skip=8
<lekernel>
no, you don't
<rejon_>
ok
<lekernel>
just use images as they are on the sd card or via ftp
<lekernel>
the gui will shout at you if you don't have the fbi header anyway
<lekernel>
it's much safer than jtag...
<rejon_>
ok
<rejon_>
how does the sd card go in?
<rejon_>
its tricky
<lekernel>
open the connector
<lekernel>
then _drop_ (not slide) the sd card in
<rejon_>
ok we got
<rejon_>
will rc3 ship with jtag?
<lekernel>
wolfspraul wants to, but I disagree
<rejon_>
from werner: congratulations to the clever design of JTAG vs. uSD sharing air space (-:C
<xiangfu>
:D
<wolfspraul>
rejon_: of course rc3 ships with jtag
<wolfspraul>
the reasons are:
<rejon_>
i don't want to hear right now
<wolfspraul>
1) just earlier lekernel pointed out that he is still in the 95th percentile of 1-man projects
<rejon_>
ok, we flashed it
<wolfspraul>
:-)
<rejon_>
all is well!
<rejon_>
which is camera input again?
<wolfspraul>
2) I do not have enough feedback from people saying they don't need it
<rejon_>
does it work on 2010-05-09
<wolfspraul>
that's because we have pretty much zero feedback from rc2, and the few people that use it mostly need jtag
<rejon_>
makes sense wolfspraul
<rejon_>
can't talk now
<rejon_>
which color for video in ?
<wolfspraul>
there is absolutely no basis to remove jtag from rc3, zero
<lekernel>
wolfspraul, flickernoise is, but the whole project isn't
<rejon_>
green ?
<wolfspraul>
it will be removed once we have enough data to justify removing it. of cours it costs me money...
<rejon_>
pink ?
<wolfspraul>
lekernel: I was joking
<wolfspraul>
jtag is a developer tool
<wolfspraul>
milkymist needs more developers
<wolfspraul>
I do not have enough data to justify removing it
<wolfspraul>
hopefully we can collect that data with rc3 and then make an educated decision for rc4
<wolfspraul>
right now I mostly have silence. and the few active people we have all need it, for now
<rejon_>
wolfspraul, shut up and tell us which is the video in :)
<lekernel>
green
<wolfspraul>
green one
<rejon_>
thanks. is it working in the latest build ?
<lekernel>
yes
<wolfspraul>
rejon_: the new jtag-serial are not sharing memcard airspace anymore
<rejon_>
this is 05-09
<lekernel>
the only things that do not work with the latest build are
<lekernel>
1) web update (but i'm on it)
<lekernel>
2) german translation
<rejon_>
note, wpwrak and i are rejon
<lekernel>
all the rest is supposed to work fine
<rejon_>
so should we flash to latest?
<rejon_>
from sd?
<wolfspraul>
don't flash to xiangfu's 07-01
<rejon_>
ok
<rejon_>
so will keep at 05-09
<wolfspraul>
maybe sebastien has something that is better than 05-09?
<rejon_>
not getting camera to work with
<wolfspraul>
why not?
<wolfspraul>
camera will work
<lekernel>
xiangfu 07-01 is supposed to have only those two features missing and no bugs, unless xiangfu made a mistake
<rejon_>
wolfspraul, you're a smart one. we should use you for breeding :)
<lekernel>
rejon_, if the camera doesn't work, reboot
<lekernel>
with the video signal present
<lekernel>
or better, short L19
<rejon_>
oh great, my little button glue came off
<rejon_>
broken button
<rejon_>
FAIL
<rejon_>
those buttons have to change
<rejon_>
realtime death
<wolfspraul>
totally
<wolfspraul>
but roh is working on a much better gluing solution, and that will be in rc3
<rejon_>
no glue
<rejon_>
gluebad
<rejon_>
wears out
<wolfspraul>
he will use a chemical glue
<wolfspraul>
that will not wear out :-)
<wolfspraul>
it will bond the pieces together for life (supposedly)
<wolfspraul>
but that doesn't help you now, sorry to hear about the broken button
<wolfspraul>
not much I can do, and not surprised
<wolfspraul>
it's extremely fiddly to get them back together, just use a tool to push the button :-)
<lekernel>
i seem particularly lucky here, my buttons were the first milkymist buttons to be assembled, i'm probably the one who uses them most, and they still hold...
<lekernel>
glued with cyanoacrylate
<wolfspraul>
lekernel: xiangfu's 07-01 boots my m1 into a big update window with several buttons (update from file, update from web, check versions, etc), but it's completely stuck. can move mouse but not click on anything.
<lekernel>
huh? wtf...
<lekernel>
i've never seen anything like that
<xiangfu>
wolfspraul: if you don't connect the network cable. it will exit in 180 seconds.
<lekernel>
so i'll try breaking mine as well....
<rejon_>
focusing on the camera not working
<rejon_>
ok packing up
<wolfspraul>
xiangfu: ahh, ok. I wasn't patient enough, just flashed back to 05-09
<lekernel>
wolfspraul, are you holding the auto update button or ...?
<wolfspraul>
why does it go to that window at all? shouldn't it boot into simple mode?
<wolfspraul>
no way, I'm just pressing the middle button
<wolfspraul>
rejon_: camera should work
<wolfspraul>
insert it before booting your m1, that will always work
<wolfspraul>
or have werner short L19, should be really easy actually
<wolfspraul>
good luck!!! You are making it to the finish line with your last men, I guess :-)
<lekernel>
wolfspraul, sounds like the auto update shortcut is triggered at boot, for some reason
<lekernel>
then you run into the "freeze without network connectivity" bug I've bothered xiangfu with
<rejon_>
yes, we got it
<rejon_>
presentation rockin
<rejon_>
all will be on video ifyou want to watch me fail out ;)
<wolfspraul>
all small details, I just wrote it so rejon_ doesn't fall into those 07-01 issues now
<rejon_>
bbiab
<wolfspraul>
url?
<wolfspraul>
ok I google
<lekernel>
wolfspraul, yeah but this is super pesky. i'm tired of endlessly fixing trivial crap like that.
<lekernel>
rejon_, gluck
<lekernel>
wolfspraul, have you found a streaming url?
<wolfspraul>
lekernel: reflashed 07-01, pretty sure no button is pressed, still got to the update screen
<lekernel>
works for me... as always
<lekernel>
shit!!!!!
<lekernel>
see why software needs to speed up as well? unexpected pesky problems
<lekernel>
the boards go through SMT on 11th (hopefully)
<lekernel>
then they will be ready for flashing
<wolfspraul>
you asked me to test this again, now what? I'm not surprised the bug is still there, that's why I went back to 05-09 :-) but this is a small thing, I'm not worried about it.
<wolfspraul>
xiangfu is currently traveling because of an emergency and doesn't have his m1 with him. he probably released the 07-01 image in a rush, without testing much, if at all.
<wolfspraul>
I think no need to worry about this tiny bug.
<wolfspraul>
I wait 180 seconds now :-)
<wolfspraul>
actually I already waited 180 seconds, I doubt that's enough...
<wolfspraul>
the last msg I get is "Failed to download bitstream.", then it's stuck
<lekernel>
the current web update code has tons of bugs, that's why I insisted it be fixed. but in your case this is a new and unexpected problem - the code shouldn't be run at all
<wolfspraul>
but it's very minor, it shouldn't distract you
<wolfspraul>
I will just go back to 05-09 now, and when xiangfu is back with his m1 he will get to the bottom of this small bug
<wolfspraul>
I just rebooted again, still the same (no buttons pressed for sure)
<wolfspraul>
also no 180 seconds delay, it's either longer or permanently stuck
<wolfspraul>
lekernel: so xiangfu's 07-01 image boots into simple mode on your m1?
<lekernel>
i haven't tested this particular image yet, but the one I have compiled theoretically from the same source does
<wolfspraul>
ah
<wolfspraul>
:-)
<wolfspraul>
well maybe you test the 07-01 image
<wolfspraul>
if you are interested in this bug
<wolfspraul>
but I suggest you to not do that, and let xiangfu get back to his m1 and in control first
<wolfspraul>
he sent this image out in a hurry, I think while traveling, without access to his m1. maybe even without any testing.
<wolfspraul>
this will be cleared up soon.
<wolfspraul>
it's a good reminder for me that once we have products on the market, we need a test plan and regression tests before pushing updates out
<wolfspraul>
so that was a welcome wakeup call :-)
<wolfspraul>
lekernel: ?
<lekernel>
yes?
<wolfspraul>
what's the solution?
<lekernel>
let xiangfu have a look?
<wolfspraul>
we wait for xiangfu to fix it? I go back to 05-09?
<wolfspraul>
ok good
<wolfspraul>
I agree
<GitHub18>
[flickernoise] sbourdeauducq pushed 1 new commit to master: http://bit.ly/kO2e7s
<GitHub18>
[flickernoise/master] Web update: version check working, verify HTTP response codes - Sebastien Bourdeauducq
<kristianpaul>
wolfspraul: roh , for my home made buttons those are are glued with silicone, no problem so as i reboot and poweron a lot