<wpwrak>
xiangfu: hmm, my memory is hazy on that one, but i think you're right. you would just erase the flash parition and then yaffs should figure out the rest.
<wpwrak>
xiangfu: i know for sure that jffs2 works like this
<xiangfu>
wpwrak: is there a way to write file to yaffs2 partition without boot to Linux?
<xiangfu>
wpwrak: just like the 'usbboot' flash ubi file to nand flash?
<xiangfu>
s/?/.
<kristianpaul>
xiangfu: hi
<kristianpaul>
xiangfu: i dunno, but can urjtag help you on that?
<kristianpaul>
xiangfu: wellcome back?
<xiangfu>
kristianpaul: :)
<kristianpaul>
xiangfu: tha yaffs2 question is related to mm1?
<xiangfu>
kristianpaul: yes.
<xiangfu>
kristianpaul: I just think we should ship with a lot of patches. and some wallpapers :)
<kristianpaul>
indeed
<xiangfu>
kristianpaul:Â Â I am not sure if the fjmem can take care /dev/flash5.
<kristianpaul>
i wonder
<xiangfu>
fjmem.bit
<kristianpaul>
had you tried just dump yaffs2 partition and write it again?
<kristianpaul>
just curios and i wonder if it will work..
<xiangfu>
kristianpaul: I try dd in rtems. it's not like linux version dd.
<kristianpaul>
oh no
<kristianpaul>
xiangfu: fiy
<kristianpaul>
20938:10:57 < mwalle> lekernel: do you have a mkyaffsimage tool?
<kristianpaul>
20941:10:59 < lekernel> there's no mkyaffsimage that works with in-band tags (used with nor flash)
<kristianpaul>
20943:11:00 < lekernel> the response from the yaffs developer was "put the files on a test device, then dump its flash to get the image" ...
<kristianpaul>
loves his logs
<kristianpaul>
xiangfu: so you see the point about dump data now ;-)
<wolfspraul>
so that means one needs an m1 to build an image that is installable on another m1
<wolfspraul>
true?
<kristianpaul>
yeah
<wolfspraul>
or we don't build an image at all, and let a running rtems on the m1 write to nor?
<wolfspraul>
rather than trying to get support for in-band tags into mkyaffsimage (upstream or not)
<kristianpaul>
rtems is different from yaffs data partition
<wolfspraul>
maybe easier to forget about mkyaffsimage and use the yaffs code in rtems for the same purpose?
<kristianpaul>
well you can also use rtems for that
<wolfspraul>
yes sure, but I am thinking how we get the files to the m1
<wolfspraul>
if via mkyaffsimage, sounds like it will be difficult :-)
<kristianpaul>
yes :/
<wolfspraul>
no buildhost, unless you attach a physical m1 to the buildhost (just kidding)
<wolfspraul>
but we can go via rtems
<kristianpaul>
but
<kristianpaul>
rtems need pull data from somwhere...
<wolfspraul>
sure
<kristianpaul>
ethernet !
<kristianpaul>
buggy ;)
<kristianpaul>
i think fix memory card bugs will help
<wolfspraul>
well. we fix them one by one.
<kristianpaul>
sure
<wolfspraul>
memory card is fiddly (for installation)
<kristianpaul>
just wich one first? :-)
<wolfspraul>
of course memory card bugs also need to be fixed
<wolfspraul>
in the end installation/flashing/updating should work in many different ways
<wolfspraul>
but I think xiangfu is just thinking about the first way now, the one that works right away
<kristianpaul>
dump!
<wolfspraul>
I guess it involves an m1 to be used to create a nor-compatible image :-)
<wolfspraul>
yep
<wolfspraul>
xiangfu: there will be some manual steps in your process :-)
<xiangfu>
yes.
<wolfspraul>
and we know some more bug areas now: mkyaffsimage cannot create nor images. rtems cannot pull data via ethernet (buggy) or memory card (buggy)
<xiangfu>
kristianpaul: can you point me some document how to dump the /dev/flash5 in rtems?
<xiangfu>
kristianpaul: what is the bug of pull data from memory card?
<xiangfu>
kristianpaul: now I put all those patches files in memory card. so far never meet a bug.
<xiangfu>
512MB sd card.
<kristianpaul>
xiangfu: you mean you memory card is working??
<kristianpaul>
damm, mine not..
<kristianpaul>
how do you format it?
<xiangfu>
kristianpaul: work just fine. 512MB sd card
<kristianpaul>
can you try with 1Gb?
<xiangfu>
kristianpaul: in linux using mkfs.vfat.
<kristianpaul>
same here
<kristianpaul>
is odd, i think i tried also 32Mb and 512MB using the adapter wolfspraul gave me and no luck..
<kristianpaul>
damm
<xiangfu>
kristianpaul: yes. but it's need some time. tonight maybe. I don't have 1GB card now.
<wolfspraul>
what is your problem exactly?
<kristianpaul>
not even get recogniced..
<xiangfu>
kristianpaul: I will try all the sd card I have. 8GB and 16GB etc..
<kristianpaul>
xiangfu: (urjtag dump)Â Â readmem
<kristianpaul>
xiangfu: ah rtems..
<kristianpaul>
xiangfu: hmm, dunno
<kristianpaul>
xiangfu: just write to memory?
<kristianpaul>
xiangfu: can you share a  mm1 boot log with your 512mb card inserted?
<xiangfu>
kristianpaul: sure.
<xiangfu>
kristianpaul: wait one moment.
<kristianpaul>
xiangfu: that fat partition is like /dev/sdc1 not just /dev/sdc right?
<xiangfu>
kristianpaul: yes.
<xiangfu>
kristianpaul: have you run 'fdisk' before 'mkfs.vfat'
<kristianpaul>
any special recomendation with fdisk?
<xiangfu>
I just delete all partition in memory card. then create one new partition. that's all.
<kristianpaul>
yeah, i hope so, i read the pdf once and i'm aware flterm need to be used
<aw>
press middle button after typing "flterm ........."
<kristianpaul>
you mean "power-on-lik"Â Â button?
<aw>
then serial condole have outputs then immediately type 'ESC" key
<aw>
press SW2
<kristianpaul>
serialboot from bios, yes
<aw>
then type 'serialboot' under bios, yes
<kristianpaul>
aw: figure 1 should look like in pdf?..
<kristianpaul>
cause the right/bottom figure is not the same.
<kristianpaul>
jaa!!!!
<kristianpaul>
*** Running tests in category: Memory card
<kristianpaul>
Card initialization:Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â [Â Â FAILEDÂ Â ]
<kristianpaul>
Block read:Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â [Â Â FAILEDÂ Â ]
<kristianpaul>
:'(
<kristianpaul>
okay i have _a_ problem here
<kristianpaul>
aw: can you point me the TP fot the memory card?
<wpwrak>
xiangfu: (write to yaff2) hmm, if you just have a partition image, the same mechanisms should work. doesn't really matter what kind of file system you have inside.
<wpwrak>
ah, i see the issue has already ben discussed
<kyak>
xiangfu: hi
<kyak>
xiangfu: do you have an idea why there is xfce related stuff in config.full_system?
<kyak>
specifically, libxfconf and libxfce4util
<kyak>
which pull some other xfce related stuff
<kyak>
some of which, in turn, fails to compile
<kyak>
but most importantly, this doesn't work on Ben
<kyak>
xiangfu: in fact, i would disable xfce (and efl) feeds for good
<kyak>
they all require X support
<xiangfu>
kyak: don't know. yes. that is what I thinkg. just disable xfce.
<kyak>
ok.. perhaps we could leave the desktop feed
<kyak>
but xfce, efl and phoen are just useless and space consuming
<xiangfu>
kyak: I just send out some patches email to openwrt. and CC to you :)
<xiangfu>
kyak: all small fixed.
<xiangfu>
for trunk.
<kyak>
cool, i can see :)
<kyak>
xiangfu: i hope it won't be a problem for them to accept these and following patches, because all of them are in xburst/
<xiangfu>
kyak: yes.
<tuxbrain>
xiangfu: I have succesfully set up a qt-creator build enviroment (finally) but I have to copy the following libs copy(link) the folloing libs  libdl.so.0 librt.so.0 and ld-uClibc.so.0 from
<tuxbrain>
To make ld find them, I wanna make a tutorial to on how to setup qt-creator to develops apps for NN but this is a ugly hack, do you know how can be fixed in the origin?
<tuxbrain>
I'm using the SDK
<xiangfu>
tuxbrain: oh. the SDK have on issue.
<xiangfu>
some script will using the buildhost path. so you should make a symlink "/openwrt-xburst.full_system" to your SDK folder.
<xiangfu>
<xiangfu> tuxbrain: oh. the SDK have on issue.
<xiangfu>
<xiangfu> some script will using the buildhost path. so you should make a symlink "/openwrt-xburst.full_system" to your SDK folder.
<xiangfu>
<tuxbrain> ?
<xiangfu>
sorry.
<xiangfu>
(whey you select it in xchat. by default it copy them)
<xiangfu>
the symlinks should be at : "/home/xiangfu/openwrt-xburst.full_system"
<xiangfu>
since I using this folder in buildhost for compile SDK.
<tuxbrain>
ok, but I'm not using the SDK building scripts now, I'm using qt-creator directly
<tuxbrain>
but I take note of this when I try to build a proper package
<tuxbrain>
xianfu:when did you plan to make a new SDK release?
<tuxbrain>
xianfu-> xiangfu
<kyak>
when i adapted some qt apps for Ben screen (like qstardict) or created new (like Tile) i used qt-creator only to test it locally. Then i copied the source files to build environment and used the regular openwrt Makefile to build it for Ben
<tuxbrain>
kyak yes but I'm also setting up an automatic upload&run on NN script to launch from qt-creator so with one click you see the app pop up in NN :)
<kyak>
tuxbrain: this sounds cool :)
<tuxbrain>
put on this some electric guitar soundtrack and another cool video to upload :P
<kyak>
indeed~!
<tuxbrain>
but I want to make the related tutorial as clean as posible, and that ugly cp libs arrangement is annoying me
<xiangfu>
tuxbrain: those libs should be at : staging_dir/toolchain-mipsel_gcc-4.3.3+cs_uClibc-0.9.30.1/lib/ of openwrt.
<xiangfu>
tuxbrain, kyak: I never using qt-creator before. can we configure qt-creator search this folder?
<kyak>
i believe this path can be added to .pro file
<kyak>
i don't know if it's ugly :)
<tuxbrain>
can be added to the qmake.conf?
<tuxbrain>
how to put two dirs on a -rpath-link
<tuxbrain>
?
<tuxbrain>
I have tried an I fail
<tuxbrain>
this is my LD conf now, how I modified it?
<wolfspraul>
kristianpaul: I think most browsers have a root certificate of the china whatever network organization as a 'trusted' certifier.
<GNUtoo>
indeed that's a big concern
<GNUtoo>
and it's nothing new
<GNUtoo>
that's in my top concerns of unfixed security problemns
<GNUtoo>
other include:
<GNUtoo>
ldd on GNU/Linux
<wolfspraul>
that borders on insanity, as I can guarantee you 100% that the Chinese governtment posesses the practical capability, and uses it in practical reality, to eavesdrop on sessions with certificates generated on the fly
<GNUtoo>
and gnome's .desktop stuff
<wolfspraul>
if you think that is not happening, you are more than naive
<GNUtoo>
indeed that's happening
<GNUtoo>
ldd is a script that can be abused to get arbitrary code execution with a crafted binary, nothing was done
<kyak>
GNUtoo: bash script can be written to execute the arbitrary code, too. Or did I misunderstand you?
<GNUtoo>
you mnissunderstood
<GNUtoo>
someone gives you a crafted binary
<GNUtoo>
let's call it binary
<GNUtoo>
you do
<GNUtoo>
ldd ./binary
<GNUtoo>
and it execute some code in binary
<kyak>
ah ok
<kyak>
not good at all
<GNUtoo>
the .desktop are also problematic
<kyak>
what's up with .desktop?
<GNUtoo>
basically let's say someone get user access to your computer
<GNUtoo>
not root
<GNUtoo>
but user
<kyak>
GNUtoo: is there a bug report/activity for ldd issue?
<GNUtoo>
then he replace some menu
<GNUtoo>
you can imagine the issue
<GNUtoo>
because he could replace administrative stuff within the gnome menus
<GNUtoo>
instead of calling the normal program
<GNUtoo>
it calls a different one
<GNUtoo>
you run it
<GNUtoo>
you see no difference
<GNUtoo>
you put your password
<GNUtoo>
and boom he has your password and can become root
<kyak>
hm you are right.. But this is the only way for the user to overwrite systems menu himself?
<kyak>
his ~/.local menu has more priority than system-wide
<kristianpaul>
gnome have many security issues, i discover the network manager cant detect changes between two access-point with same name (dunno if there is a bug for this)
<kyak>
you shouldn't keep it to yourself :)
<kristianpaul>
:p
<GNUtoo>
I know something even worse
<GNUtoo>
and that could do huge damage
<GNUtoo>
the most dangerous command ever
<GNUtoo>
is this channel logged?
<kristianpaul>
ye
<kristianpaul>
s
<kyak>
yes it are
<GNUtoo>
ok better not telling
<GNUtoo>
basically you tell that command to a newbie ubuntu user
<GNUtoo>
he will run it
<GNUtoo>
for rm -rf /* he won't
<GNUtoo>
but for this one he will
<GNUtoo>
even experienced people didn't know what the command did
<GNUtoo>
*really did
<kyak>
there are numerous ways to obfuscate that rm -rf / :)
<GNUtoo>
that one is worse
<GNUtoo>
if you really want to know PM me
<kyak>
can you pm it?
<wpwrak>
rm -rf /Â Â ... why ? just tried it and got no error
<GNUtoo>
yes
<kristianpaul>
okay ,lets kick out here qi-bot ;)
<wpwrak>
there are countless commands you can use to cause nasty damage. it's often sufficient to add some syntax error to config file to cripple a system for a newbie
<GNUtoo>
I told to kyak and kristianpaul
<kyak>
now we have the Sacred Knowledge ;)
<kristianpaul>
:p
<GNUtoo>
Scary too
<GNUtoo>
they can tell that the command is very dangerous....
<kyak>
anyone reading the log should now go and hide
<kristianpaul>
kyak: hahah ;)
<GNUtoo>
note that I found the command myself
<GNUtoo>
I wasn't told by someone
<kristianpaul>
You should file a bug or told doms distros to get a workaround i think
<kristianpaul>
at least to prevent
<GNUtoo>
hmmm
<kristianpaul>
:-)
<GNUtoo>
how could that be prevented?
<GNUtoo>
removing the command is not an option since it's needed
<kristianpaul>
a warning about what it should really do
<kristianpaul>
so user can regret before is late
<kristianpaul>
anyway :-)
<GNUtoo>
it was not used in the wild
<GNUtoo>
and I only told 4 people until now
<kristianpaul>
ok
<GNUtoo>
a university friend that is trusted, Cotulla kristianpaul  and kyak
<GNUtoo>
no one of theses people really knew what the command did
<lars_>
what does the command do?
<GNUtoo>
it's an equivalent of rm -rf /* under another name
<GNUtoo>
and it's name has no indices on the fact that it rm -rf /*
<wpwrak>
date >/dev/root? :)
<GNUtoo>
worse
<wpwrak>
or, better: yes >/dev/root
<kyak>
i guess it's time we form the masonic lodge :)
<kyak>
only 4 people!
<GNUtoo>
lol
<kristianpaul>
xD
<GNUtoo>
indeed it start to become out of control
<GNUtoo>
maybe I shouldn't have mentioned it
<GNUtoo>
because everyone wants to know now
<kristianpaul>
yeap, but you DID ! :)
<kristianpaul>
run way !
<kristianpaul>
:)
<GNUtoo>
lol
<GNUtoo>
sorry
<wpwrak>
now .. where's that bloody mkimage ... yet another great thing about getting rid if u-boot will be the elimination of that beast
<wpwrak>
tuxbrain: any news about the data sheets ?
<wpwrak>
lars_: what's the best way for working on the wpan extensions for the nanonote kernel ? make a branch on git://projects.qi-hardware.com/qi-kernel.git ? or keep it somewhere else ?
<wpwrak>
lars_: i haven't worked with a public qit branch yet, so i don't know how high the risk of unintended consequences is :)
<kristianpaul>
wpwrak: you can clone it and make a wpwrak-branch later lars_ can safelly merge it..
<lars_>
wpwrak: since I can always reupload my local copy of the git tree in case you seriously messed something up, using a branch in qi-kernel.git should be fine
<wpwrak>
excellent. let's keep all things in one place then. branch name ben-wpan ?
<kristianpaul>
just give a small nock to his scope and the screen go on-line again
<lars_>
wpwrak: sounds good
<wpwrak>
kristianpaul: it's nice to see what you're doing, isn't it ? :)
<kristianpaul>
wpwrak: :)
<kristianpaul>
i'm debuggin right now the memory card
<kristianpaul>
xiangfu's 512Mb card works
<kristianpaul>
mine not :/
<tuxbrain>
is gmenu2x killable? I try but it comes to live again and again
<xMff>
you have to nuke it from /etc/inittab
<kristianpaul>
yeap :-)
<wpwrak>
tuxbrain: no data sheets or at least a part number from the smt fab ?
<wpwrak>
tuxbrain: and how did the visit to the master of torture go ? :)
<tuxbrain>
wpwrak: nothing dude :(, we must wait to the quote to see which refs they propose
<tuxbrain>
wpwrak: (MoT) nah he is a good artisan and I was listen good metal meanwhile , I was near to get asleep a lot of times :)
<wpwrak>
tuxbrain: (torturer) so he gave you the good drugs :)
<tuxbrain>
here in spain is not usual the dentist use drugs :) maybe some local anestesia, but no this time, just fix some little caries :)
<wpwrak>
tuxbrain: (smt) grmbl. their response times suck a bit. well, maybe they didn't actually have the connector yet but just think they can find one that's 100% smt, and are searching now. that would explain why they didn't send the spec along right away.
<tuxbrain>
two weeks ago he extract a teeth, and there yes he uses a loooot of local anestesia
<wpwrak>
tuxbrain: (dentist without drugs) that's what you get for searching planes coming from south america :)
<wpwrak>
tuxbrain: (i suppose the little incident with the medical plane from argentina caused some waves over there as well ?)
<tuxbrain>
wpwrak: (time suck) totally agree
<tuxbrain>
wpwrak: medical plane? what was that? any link?
<tuxbrain>
hohohoho, well whatever this type of drug will be not very conveniet to apply to dentist patient... you better need one that relax , and this one is totally the opposite :)
<wpwrak>
tuxbrain: yeah, seems that the ship with the opium made it through fine :)
<tuxbrain>
sure that meanwhile police was keeping the eye to that plane, a cargo ship was happilly discharging in the other side of the map let's say Galicia ... police happy due he catch some bad guys, narcos happy due his plan goes fine
<tuxbrain>
people at ibiza happy due the party will not stop
<kristianpaul>
:-|
<tuxbrain>
mothers of the dead boy by overdoze crying...
<tuxbrain>
and on and on and on
<tuxbrain>
let's change topic I'm starting to distract from coding with sad thoughs
<kristianpaul>
so qt is taking fast developing on your side?
<Jay7>
hehe.. come here, my father is dentist :)
<wpwrak>
tuxbrain: hmm, let's see if this is a happier topics: did you sell any of those ubb-arduino cable kits ?
<wpwrak>
Jay7: i suppose the anesthetic of choice would be vokda in this case ? should also help with keeping things properly disinfected :)
<Jay7>
wpwrak: hammer :)
<Jay7>
better than vodka even ;)
<tuxbrain>
not a killer sell but some about 20Â Â in individual units and a pair of 10x bags , but we gain a lot of visibility, I will try to keep things hot that's why I'm developing Nanokoder, but I thing I will return to the original idea of doing a more specific arduino oriented thing than a general editor
<kristianpaul>
I dont get it "the change has to be accounted for in software"
<kristianpaul>
bah, lets just inject some clock and see what happens..
<wpwrak>
Jay7: (hammer) good old russian pragmatism ;-)
<Jay7>
wpwrak: yep! :)
<wpwrak>
tuxbrain: 30 units sounds like a pretty good start for something like this
<wpwrak>
tuxbrain: and how's UBB doing ? still selling after the first wave ?
<tuxbrain>
kristianpaul: (qt is taking fast devel) well fast is not the word, is my first contact with Qt world, I have spended a lot of time setting up things and understanding, thanks to the Nanomap creator that helps me in the setting up I achieve to have a very good developent enviroment setting up, also his Nanomap code is real clear and understandable to have and idea how things works,also some parts are totally reusable like the fileselector, or the batt
<tuxbrain>
ery and time layer, so I think next moves will be more productive in development.
<wpwrak>
rjeffries_: (ITead) ah yes, minimum via hole size. that's what separates the boys from the men ;-)
<wpwrak>
rjeffries_: a 12 mil drill isn't too bad, though. e.g., i use a 13.5 mil drill. i give my vias a 30 mil copper diameter, so even 15 mil holes would be acceptable. (rule of thumb: via hole:via copper = 1:2 (for FR4 PCB material))
<kristianpaul>
wolfspraul: Are you aware of people related with digital signal processing at the qi list?
<kristianpaul>
i was thinking drop some questions there about that topic.. :-)
<tuxbrain>
courious if I launch the qt app trhough ssh as command it' doesn't clean exit, but lauched thoug a console from NN or trhough ssh yes it does....
<kristianpaul>
he
<tuxbrain>
if lauched as ssh commant it let beutifull grey screen
<tuxbrain>
and in fact it's still showing on ps
<wolfspraul>
kristianpaul: no you are the pioneer. There may be someone, or maybe not. Just email and see whether anybody replies.