=== Jack87|Away is now known as Jack87
shadeslayerlilstevie: I was trying to run sudo nvflash --sync with my transformer and I get : rcm version 0X4 Command send failed (usb write failed)06:26
shadeslayerlilstevie: could you give a hint as to what I might be doing wrong?06:26
twbThat error could just mean you forgot your SBK or so06:26
* shadeslayer looks06:27
shadeslayertwb: I also could not find good documentation about nvflash, any ideas where I can find that?06:28
shadeslayerI'm having a bit of a hard time figuring this out :)06:28
twbhttp://cyber.com.au/~twb/doc/tf101.txt now contains a slightly updated version06:29
twbshadeslayer: there isn't any, nvflash is a piece of shit internal nvidia bodge thing06:29
shadeslayerAlright, hopefully that link contains info on what actually goes on during booting06:30
twbThose links are my notes06:31
twbThat why they have my name in them06:31
shadeslayeroh ... awesome, I didn't notice that ...06:32
shadeslayerit's just been a long day ...06:32
twbno problem06:32
shadeslayerhmm ..06:35
shadeslayertwb: so I now use : --sbk 0x1682CCD8 0x8A1A43EA 0xA532EEB6 0xECFE1D9806:36
shadeslayer sudo ./nvflash --sbk 0x1682CCD8 0x8A1A43EA 0xA532EEB6 0xECFE1D98 --sync06:36
shadeslayerand I get the same thing06:36
shadeslayerand iirc the tablet installed a firmware update when it arrived06:36
twbDid you bounce it back into APX mode again?06:36
shadeslayerno, it just shutdown06:36
shadeslayerNvflash started06:37
shadeslayerrcm version 0X406:37
shadeslayerCommand send failed (usb write failed)06:37
twbTell you what, read that whole .txt file then come back if you still have problems06:37
shadeslayeruh ok06:37
shadeslayertwb: question, before you did all of this, did you update your tablet firmware?06:52
shadeslayerand can a firmware update change the sbk?06:52
twbIf you mean android, then no, I don't give a shit about android.  But the processes I describe should work for android as well.06:53
twbThe firmware does not decide the SBK, the SBK is baked into the hardware while in the factory06:53
shadeslayerright, I don't want android as well ...06:54
* shadeslayer goes back to reading the text file06:54
twbThat's fine then.  There was a jackass in here earlier that wanted help with android06:54
shadeslayertwb: another question, does it matter if I have sbk version 2?07:04
twbshadeslayer: what is an "sbk version 2" ?07:05
twbshadeslayer: do you mean a 3G transformer?07:05
shadeslayernope, It's a wifi version, but when i run sbkcheck i get : [~/tablet]$ sudo ./sbkcheck07:06
shadeslayerFound APX mode device07:06
shadeslayerChip UID: 0x428908841c0e11707:06
shadeslayerDetected SBKv207:06
twbNever heard of sbkcheck07:06
twbGoddamn munchkins and their pre-compiled ia32 binaries :-/07:07
twbThey don't even ship the sources07:08
shadeslayeryeah :(07:09
twbSo I have no idea what that program does and I'm not going to decompile it or run it to find out07:09
twblilstevie is here, he probably knows07:09
shadeslayertwb: I actually used that program because one of his posts mentioned it07:09
twbI am too old to deal well with stupid web fora that reinvent NNTP, poorly.07:10
shadeslayeralright, I'm off to sleep, will try this tomorrow07:13
shadeslayerthanks twb for the help :)07:14
=== Jack87 is now known as Jack87|Away
lilstevieshadeslayer: raymans app is far better than mine at testing for it08:01
lilstevieshadeslayer: basically sbkdetect v2 which is raymans version, uploads a command that is encrypted with the new sbk08:01
=== doko__ is now known as doko
=== lag` is now known as lag
misfitx7I just downloaded and imaged my panda board with the new 11.10 server release and I am being prompted for a login/pass. Does anyone know what the default account is? I was never prompted to create one as the install instructions suggested I would be.13:13
misfitx7I can't seem to find this documented anywhere.13:16
infinitymisfitx7: There is no default user/pass, it asks you to set one up during the install.13:16
infinitymisfitx7: If it didn't, then something when wrong, and you might want to try again.13:17
misfitx7infinity: Thanks. I've loaded it twice thinking that may be the case but both times I got the same result. I must be missing something, I'll give it another shot I suppose.13:19
infinityWell, what happens when you boot it the first time?13:19
infinityIt should expand the filesystem, then reboot into an installer.13:20
infinityIf that's not happening, the first bit may be failing.  Your card could be bad, or too small.13:20
misfitx7I watch the filesystem expand on the serial terminal, it reboots when it's done then I get a login prompt on my monitor.13:21
misfitx7I can try another card. The one I'm using now worked for 11.04. It's 8Gb13:21
infinityWait, on your monitor?  Are you still connected via serial as well?13:22
infinityThe installer should be popping up on the serial console.13:22
misfitx7I am still connected via serial. But it switches over to the monitor after the first reboot.13:24
misfitx7After the kernel loads and services are started13:24
misfitx7I'm flashing a larger card to see where that gets me.13:27
misfitx7I get the sys config with the a different/larger card.13:38
=== ericm|ubuntu is now known as ericm-Zzz
prpplagueGrueMaster / ogra_ turns out my sd card reader died yesterday15:32
GrueMasterOn your panda?15:32
prpplagueGrueMaster: no, on my host pc. it wasn't writing all the data to the sd card15:33
GrueMasterInteresting.  Could be just a cache not flushing properly.  I usually run a sync after flashing an image just to be sure.15:34
GrueMasterIs your PC Linux or Windows based?15:35
prpplague~lart GrueMaster15:36
prpplagueGrueMaster: silly GrueMaster15:37
prpplagueGrueMaster: ubuntu 10.1015:37
prpplagueGrueMaster: no, it was a complete failure of the sd card reader, looks like the 3.3v ldo was failing15:38
GrueMasterAh.  I had to ask.  Some people are using Windows on their PC but Linux on their dev boards.  We even have a tool for imaging SD cards for this situation.15:38
prpplagueGrueMaster: hehe indeed15:38
prpplagueGrueMaster: i've been MS free since 199715:38
GrueMasterHeh.  We should have Windows-Clean coins for our 10 year anniversaries.  :P15:39
austeregrimBut shouldn't mac users have something similar? lol15:47
prpplagueausteregrim: that;s kind of like someone stop smoking tobacco and switch to meth, sure stopping tobacco is good, but....15:48
GrueMasterNo, they are more like connoisseurs.  Twice the intake, but only 1/4 the guilt.15:49
=== gallth is now known as tgall_foo
MrCurioustested out pandaboard + ubuntu 11.10 with root fs on a USB thumb drive.  Got in excess of 100 FPS at 320x240.  GREAT work by the team on solving the usb speed bug!16:51
MrCuriousnot to mention that it was WAY more snappy than root on SD, and noticably more snappy than the last stable release (with root on usb)16:51
GrueMasterJust to clarify, the kernels for Maverick and Natty have been updated with this fix as well.16:53
MrCuriousyes, i am comparing pre-fix to post-fix16:53
NekoGrueMaster, I have an idea.. would it be possible to not have to create a swap partition in the main installer, or in any event, not ship a swap file, but as soon as ubiquity pops up or perhaps (since swap isn't needed for ubiquity, memory usage is well under 150MB) on first desktop login like ecryptfs or so used to, say "hey, I notice you have a small amount of memory and no swap, do you want me to fix this up for you?" so that it doesn't have t18:05
Nekoo be shipped on the SD card image?18:05
Nekoit'd help on VMs on x86 etc. too if you didn't have to create swap partitions inside the VM especially with memory ballooning and an assumption that the host has more than enough swap to handle overcommitting18:06
GrueMasterMaybe for P.  We can't change 11.10.18:06
Nekosure, P is good.. its been something I've been annoyed by since Karmic though :D18:07
infinityWe did it this way intentionally because dding a large file on an SD card on the target hardware takes forever.18:07
infinityIt's actually easier to do it at install time, just more annoying to the user.18:07
infinityNo one's stopping you from deleting the file if you didn't want it. :P18:07
Nekooh, for sure ... but it'd save some SD card space that basically doesn't get used until the user picks Firefox and opens 8 tabs..18:08
infinityDunno about that.  GrueMaster was showing me ubiquity OOMing on 512M systems.18:09
Nekoin the meantime by the time P comes out, zram will be nice and stable right? in the event of a missing swap partition or so maybe compressed in-memory swap would be a better idea just to have some swap around if it's truly, truly necessary18:09
infinityI dunno.  I'm not all that picky, to be honest.18:10
infinityI still think that users who are running systems on SD for anything other than quick test purposes are Doing It Wrong.18:10
NekoI think part of the problem is even on "512MB" systems on MX5 they're soaking a bunch for framebuffer/2D reservation18:11
Nekoso you really only have 400MB18:11
infinityBeagle in his case, and I don't recall how much it has available.18:11
GrueMasterOne thing to look at is zramswap.18:11
Neko256MB on my C4 and it's infuriating18:11
NekoI definitely agree something needs to be there just in case..18:12
GrueMasterSeems to work ok on AC100.18:12
infinityIt's not actually enabled on AC100.18:12
infinitySo, yes, works great.18:12
GrueMasterWell, the SD desktop images are really just for show & tell anyways.  It isn't hard to move the image to a usb drive using a separate system.18:13
infinityOr a USB hard drive, or whatever.  Yes.18:13
infinityBut at that point, you have enough space that losing some to swap is irrelevant.18:13
GrueMasterAnd since we don't have any real production systems beyond the AC100...18:13
infinityAnd as distasteful as swap on the SD card is (and man, I really think it is), I'd rather have something that sort of works for the 2 hours someone will play with it before they realise they need faster storage.18:14
NekoGrueMaster, zcache may well help too as an experiment.. that way you get the benefits of SD card data being put in the cache and when the system really needs it, it compresses them as they are evicted, which would help one hell of a lot in keeping real memory available18:16
Nekoor a USB key. or any slow-ass storage...18:17
Nekoactually you know this is something Linaro should be investing a ton of effort into :D18:18
lilstevieshadeslayer, you about23:25
lilstevieok, I just wanted to say to you, just because you have an SBKv2 device don't give up23:26
lilsteviework on your image creation23:26
shadeslayerlilstevie: oh ... ok23:26
shadeslayerI didn't give up :)23:26
lilstevieyou can still run the image in qemu or something23:26
shadeslayerI'm just busy with a sprint in MV23:26
lilsteviewe are working on a solution for the SBK23:27
shadeslayerok, I'd like to help, lemme know how i can :)23:27
lilsteviewell the problem is the miniloader23:27
lilstevieit is part of how nvflash work23:27
shadeslayeruh ok, I've never used nvflash, so I'll have to look that term up23:28
lilsteviethe bootrom shuts down communication with that 0x4 error if the messages are incorrectly encrypted23:28
lilsteviewe have communication23:28
lilstevieuntil we get to miniloader23:28
shadeslayerah ok23:28
lilstevieminiloader is a small loader (funny that), that inits ram and usb,23:29
lilstevieso that the bootloader can be sent to the device23:29
lilstevieon these SBKv2 devices the current miniloader crashes23:29
lilstevieor aborts23:29
lilstevieor something, we don't know exactly what is happening, no debugging interface is a pain23:30
lilstevieso we have been writing our own, but it is a very slow process23:30
shadeslayeralright, kind of understand it23:30
shadeslayerthe bootloader before the bootloader :P23:30
lilstevielook at it as a bootstrap23:31
lilsteviein APX the device hasn't initialized ram23:31
lilstevieor the dual A9's23:31
lilstevieall there is, little ARM7TDMI core23:32

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!