/srv/irclogs.ubuntu.com/2010/09/03/#ubuntu-arm.txt

devilhornsogra, are you around ?02:34
=== gcl_ is now known as gcl
=== hrw|gone is now known as hrw
hrwmorning08:17
=== kmargar is now known as markos_
Baybalevening08:34
lagogra09:05
lagograaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa09:05
* apw waits to seee if he is here09:06
lagapw: Are you after him too?09:06
apwthere is an upload pending beta being out the way, and his approval yes09:07
lagOGRAAAAAAAAAAAAAAAAAAAAAAAAA09:16
sebjancooloney: to summarize, I just pushed some patches with latest display fixes and panda patches aligned with the pandaboard gitorious tree.10:04
sebjancooloney: these patches are required for 8 layers boards support10:05
sebjancooloney: they also add blaze support (all ES2.0)10:05
cooloneysebjan: awesome, i just cloned it10:07
sebjancooloney: regarding bootloaders, the gitorious pandaboard tree (omap4_panda_L24.9 branch) is the most up to date, thouhg x-loader still misses some patches to support 1GB memory (from Rob x-loader tree)10:09
sebjanlag, rsalveti, ogra  ^10:11
cooloneysebjan: great, so after i got the board, i need to upgrade the bootloader firstly?10:11
lagcooloney: I have a bootloader you can use10:13
lagsebjan: I'm assuming the new bootloader is still lacking our patches?10:14
cooloneylag: great, man10:14
cooloneylag: my board is on the way, it will arrive next week, i guess10:15
lagcooloney: In fact, rsalveti already has them uploaded: http://people.canonical.com/~rsalveti/maverick/boot/es2/10:16
sebjanlag: I am not clear with the 'our patches'?10:16
lagYou need u-boot.bin-with-ubuntu-patch and MLO-1gb10:16
cooloneylag: great, bookmarked10:16
lagsebjan: To read boot.scr instead of bootargs10:16
sebjanlag: I have not seen boot.scr support in pandaboard u-boot tree.10:18
lagOkay, that answers my question10:18
sebjanlag: anyway, the official bootlaoder ogra is packaging is based on an upstream u-boot, right? So you shall check the patches added in the pandaboard tree are integrated in the bootloader packages?10:19
sebjanlag: fyi, I personnaly build a version of u-boot / x-loader for my usage now, which include all these various patches. I can send the binaries for testing if required?10:21
lagsebjan: I have binary bootloaders that work - that's all I'm interested in at the moment10:24
lagsebjan: Bootloaders is ogra's thing :)10:24
sebjanlag: ok, I'll check with him for patches alignment then10:25
lagsebjan: np :)10:28
cooloneysebjan: thanks a lot for heads up10:37
lagogra: I found the issue I was facing11:05
lagMy scripts weren't picking up the 2.6.34 kernel at all11:05
ogralet me guess from the backlog, its bootloader related :)11:05
ograoh11:05
ograheh11:05
ograk11:05
lagThe first boot did use the 2.6.35 kernel (hence the successful install)11:06
ograthe es2.0 kernel should have built by now11:06
lagThen on second boot it was using the installed kernel11:06
lagWell it's no good now!11:06
lagWhy didn't you tell me the kernel is replaced?11:06
ograi did11:07
lagThat was why the kernel failed on second boot - old kernel!11:07
ograseveral times11:07
lagWhere?11:07
ograif you call flash-kernel it takes everything from /boot11:07
ogradunno several times during the last weeks11:07
ograsince you play with that script11:07
lagYes, you told me that11:07
lagBut I don't call flash-kernel11:08
lagNot ever11:08
ograyou did manually yesterday11:08
lagI did not11:08
ografor /boot/boot.script11:08
lagNope11:08
lagNever used it11:08
lagI pasted you the lines of script that I use11:08
lagSo why is the kernel in /boot used?11:08
lagIs flash-kernel called during the install process?11:09
ograby oem-config, yes11:09
ograit even tells you about it11:09
ogra"running triggers for update-initramfs"11:09
ograupdate-initramfs calls flash-kernel11:10
ograflash-kernel takes the new initramfs, boot.script and kernel from /boot and puts them into the vfat11:11
ograto make sure that a) the initramfs has exactly what you have installed in the fs and b) as a sideeffect, make sure jasper is gone11:11
ogranote that our images focus on the enduser who will never even see there is a vfat11:13
sebjanogra: did you follow the bootloader updates required to support the ES2.0 boards, 1GB mem, ...? Do you need some help here?11:39
ograsebjan, nope, we just have to make sure linaro includes them in the upstream branch11:40
ograwell, for u-boot that is, for x-loader i think rsalveti has a branch prepared so far that still needs some touching11:40
ograi'll talk to jcrigby later today about the u-boot stuff11:41
ograand ricardo said he would look into finishing x-loader later today11:41
ograthe kernel is uploaded but sits in the NEW queue i think, i'll take care for getting it out there11:42
sebjanogra: ok, I'll ask rsalveti for checking his x-loader tree then. You will use this tree to update the x-loader package?11:42
ograso with some luck es2.0 support is ready on monday11:42
sebjancool!11:42
ograsebjan, right, thats the plan11:42
sebjanok, thx11:43
lagogra: The lastest daily build is still 2.6.3412:34
ogra_cmpcyes12:37
ogra_cmpckernel is still building and will then sit in the NEW queue12:37
ogra_cmpclag, ^^12:38
lagETA?12:38
ogra_cmpcEOD i think12:38
ogra_cmpcbut that wont help you, bootloaders are still ES1.012:38
lagI have my own bootloaders ;)12:39
ogra_cmpci hope we'll have the x-loader for 1GB support in today and u-boot upgraded over the weekend12:40
ogra_cmpcwith luck we'll have usable ES2 images on monday12:40
lagogra_cmpc: Don't jinx it12:53
ogra_cmpcheh12:53
ogra_cmpci'm trying :)12:53
ograasac, did you ever find out what your sholder pain actually was ? seems i have the same since two weeks (got really bad during this week)13:06
=== xfaf is now known as zul
asacogra: yes, it was overwork13:27
ograhmm13:27
asacogra: it went away when i stopped typing so much and just having phone calls13:27
ograheh13:28
asacogra: no i am taking notes all the time on phone, it seems to come back13:28
asacat least the last few days its started to be a bit of a problem again (far away from before)13:28
ograi should move to linaro ... you guys never work ... just chat on the phone :P13:28
asacheh13:28
asacnow i chat on the phone and type almost the same amount of stuff ;13:28
asac)13:28
asacso much email stuff flying around :/13:28
ograit goes away for me during the morning. but comes back twice as hard at night13:29
rsalvetisebjan: yep, just found we were missing some patches14:43
ogramopdenacker, its not complete yet, but so you see i didnt forget you ... https://wiki.ubuntu.com/ARM/OMAPImageBuilding14:44
rsalvetisebjan: will build one x-loader with full 1gb support for panda and test it here with my boards14:44
ogra(the script isnt complete)14:44
rsalvetiso we can then have a good package14:44
devilhornsogra, quick question ... the packages you need for efl .. is that going to be for lucid or maverick or ?14:45
ogradevilhorns, maverick and beyond14:46
ogradevilhorns, did you manage to get a working vm setup yet ?14:46
devilhornsogra, no :(14:46
devilhornsogra, I've been at it for days and days now too :(14:47
ogradid you use rootstock ?14:47
devilhornsyes14:47
ograshould just work14:47
ograin maverick ?14:47
devilhornsyea14:47
ograhmm14:47
ograworks for others ... weird14:47
devilhornsworks mostly in that the image builds sure14:47
devilhornsbut14:47
devilhornswhen running it in qemu, whenever I try to apt-get something, it segfaults14:48
ograwhat kernel do you use ?14:48
devilhornsusing the maverick one from ports14:48
devilhornsone second, I will get you a link14:48
devilhornskernel from here: http://ports.ubuntu.com/ubuntu-ports/dists/maverick/main/installer-armel/current/images/versatile/netboot/14:49
ograrsalveti, didnt you test with tehz maverick kernel in qemu back when you fixed rootstock ?14:49
ogradevilhorns, probably try the lucid one, though i think we use the same kernel during build (which uses apt-get a lot)14:50
rsalvetiogra: yup, worked fine when I tested it14:50
devilhornsogra, yes, believe I did also try the lucid one and it still segfaulted :(14:50
rsalvetibut qemu changed, and don't know if we got a new versatile kernel14:50
devilhornsbut I will give it another try14:50
rsalvetistill didn't test with the latest combination14:51
ograrsalveti, well, if rootstock finishes without errors with that kernel i'D expect it to still work14:51
rsalvetibut it should work, weird14:51
rsalvetiI can check later14:51
devilhornsok, I have rootstock running now, so we'll find out soon :)14:52
devilhornsshould I be using the rootstock from bazaar ?14:55
devilhornsright now, I am just using the normal rootstock and qemu stuff from the standard maverick repo14:55
rsalvetidevilhorns: you can try it, but maverick's version should work the same way14:55
devilhornsrsalveti, ok :)14:56
mopdenackerogra: many thanks!14:56
devilhornsogra, oh btw, talked to Lynda in H.R. and we have reached an agreement so I am certainly "in" for this :)14:57
ogradevilhorns, cool ! welcome aboard then :)14:58
devilhornsthanks :)15:01
devilhornsok rootstock is done15:01
devilhornsok image is built, gonna try qemu now15:04
devilhornsogra, rsalveti ok, using the lucid kernel with the qemu image and it just hangs ... gets passed most of the init (last thing I seen was Freeing kernel memory) but now looking @ a blank qemu window with a blinking cursor15:09
devilhornserr Freeing init memory15:10
devilhornsso that we are on the same page, here are the commands I am using:15:11
devilhornshttp://www.pastebin.ca/193179515:11
ericm|ubuntusaeed, ping15:13
ogradevilhorns, sorry, in a call atm15:13
devilhornsnp15:13
* ogra will looks afterwards15:14
devilhornslemme know when you are free :)15:14
ograhttps://wiki.ubuntu.com/ARM/OMAPImageBuilding15:17
ograndec, ^^^15:17
apwogra, where do we stand on the omap4 branch?15:36
ograapw, still going forward, but the main chunk shoudl be in with the currently building package15:36
apwogra, so is the latest up and in, i thought it got rejected for safety15:37
ograit was uploaded again yesterday and is building atm15:37
apwoh strange, he specifically asked me to talk to you today to see if you wanted it up ... oddness ... as long as you are happy i can cross it off15:38
ograhttps://launchpad.net/ubuntu/maverick/+source/linux-ti-omap4/2.6.35-903.8 ;)15:39
devilhornsgrrr silly qmeu, why do you keep segfaulting ? :(15:44
devilhornss/qemu15:45
rsalvetidevilhorns: yep :-( seems very buggy15:46
devilhornsrsalveti, indeed ... been fighting with it all week :(15:46
rsalvetidevilhorns: why are you trying qemu?15:46
rsalvetino omap boards?15:47
devilhornsrsalveti, no15:47
rsalvetihm, ok15:47
devilhornsand no arm hardware either :( david mentioned sending some hardware out to me, but I dunno what15:47
devilhorns(or when) :)15:47
rsalvetidevilhorns: where are you based?15:47
rsalveticountry15:47
devilhornsUSA15:47
rsalvetidevilhorns: hm, ok, so it should be fine and fast to get you a board15:48
devilhornsrsalveti, yea, don't think it will be a problem15:48
rsalvetia lot better then debugging qemu :-)15:48
devilhornsbut I wanted to get a jump on things, so I decided todo an image and try qemu15:48
devilhornsso I could at least run the current netbook-efl stuff and see where the problems were, etc etc15:48
devilhornsand get a jump on package building15:49
rsalvetiok, makes sense15:49
devilhornsif I try using the lucid kernel, the qemu image never fully boots. If I use the maverick kernel, it boots and I can login, but any attempts w/ apt-get just segfault :(15:50
devilhornsfrustrating15:50
devilhornsif I try to seed the image w/ the netbook stuff before hand, then it just stalls at trying to configure xul-runner15:51
rsalvetiargh, really frustrating :-)15:51
devilhornsindeed :) spent the whole week so far trying different permutations of the process just to get a usable image15:52
devilhornsbut there is one ray of hope....15:53
devilhornsSamsung is supposed to send me a development phone (arm based) for the efl work we are doing for them, so I can fiddle w/ that when it gets here, but who knows when that is15:54
rsalvetihm, nice15:55
devilhornsyea :) when they get off their bumms and decide to send it :)15:56
dcordesrsalveti: Congrats on the visa16:02
rsalvetidcordes: haha, thanks :-)16:02
dcordesAnybody seen the maverick on HD2 video ?16:03
devilhornsnot yet16:05
armin76rsalveti: i want a pandaboard too :(16:15
ograarmin76, you will soon be able ot order one16:16
ogra*to16:16
armin76ogra: you didn't order it :P16:16
ograarmin76, well, start working for a company that works with the vendor or work on some really significant software the vendor cares about and you can get one too :)16:17
armin76its easy to say when you get paid for it16:18
ograelse you have to wait until its out in the world ... thats how it works (and thats not vendor specific :) )16:18
GrueMasterarmin76: The work we do now on prerelease boards ensures that when you get it, you can focus on what you want to do with the system, instead of tediously working around major bugs that hang the system.16:27
hrwarmin76: I want pandaboard too but looking at amount of base work which ubuntu-arm people need to do to get working system I am happy with waiting for final, production device16:29
armin76GrueMaster: i don't mind doing that kind of work16:30
armin76hrw: i like challenges16:30
hrwarmin76: first was ES1.0 pandas which failed to have ehci working so ugly cable was soldered for faking it. then ES2.0-6layer was released just to be replaced with ES2.0-8layer 'few days' later.16:30
hrwand production ones will probably use ES2.1 which is yet another version to care about16:31
ograwell, all es2 should work with our images16:31
hrwand es1 is something which could be dropped probably16:31
ograit was dropped with todays kernel upload16:32
ograwe only build for es2 now16:32
hrwcool16:32
ograpeople using es1 should use beta and not upgrade :)16:32
ograwith all its bugs and warts16:33
hrwnow let TI get final versions manufactured and I will wait for Fedex guy16:33
devilhornsogra, well, looking like I am stuck until I get some arm hardware :(16:44
devilhornssad, but what can I do ? :) I did try tho16:45
ograi'll have a talk with david once he is back (he is sick atm)16:45
devilhornsogra, ahh ok :) well I hope he feels better16:45
devilhornsbeing sick sucks :(16:45
devilhornsogra, thanks :)16:45
ograit does16:45
=== hrw is now known as hrw|gone
prpplagueanyone recommend a cheap netbook that works really well with ubuntu?17:06
ogradell mini17:07
ograits the reference platform used for netbook stuff (or it used to be, not sure it still is)17:07
ograyou can buy them preinstalled17:08
prpplaguei was hope'n to pick something up this evening17:08
ograhmm, dunno then, but most netbooks should just work imho17:09
ograjust dont buy a model that only came out in sept. :)17:09
prpplaguehehe17:09
prpplagueogra: my main concern was support for the wifi and such17:10
prpplaguehttp://www.bestbuy.com/site/Dell+-+Inspiron+Mini+Netbook+/+Intel%26%23174%3B+Atom%26%23153%3B+Processor+/+10.1%22+Display+/+1GB+Memory+/+250GB+Hard+Drive+-+Obsidian+Black/9980598.p?id=1218204946938&skuId=998059817:11
prpplagueogra: that look about right?17:11
ograyup17:12
ograthough there should abe an ubuntu version too17:12
ograusually cheaper without the win fee17:12
ograoh, indeed probably not at best buy17:13
sebjanrsalveti, lag : I was able to natively build a kernel image package with the 2.6.35 kernel, by reducing to mem=460M17:14
lagsebjan: How do you mean?17:16
lagsebjan: We've been building 2.6.35 kernels for a while now17:17
ogralag, on your panda ?17:17
ograsebjan, thats our default atm (well, its 463M in the images)17:17
sebjanlag: natively it was difficult because of crashes. Here it went to the end17:18
sebjanthe good news is that it was much faster than with ES1.0 (less than 2.5h)17:18
rsalvetisebjan: hm, interesting17:19
rsalvetiso it could be a memory issue (on sw layer)17:20
rsalvetisebjan: I tried yesterday but wasn't able to build it17:20
rsalveticool we now have a workaround17:20
rsalveties2 should build much faster than es117:20
robclarkogra: btw, I think with next release that 463 becomes 460..17:21
rsalvetirobclark: I got the patch for jasper already :-)17:21
ograrobclark, k, will adjust, i think i saw you posting other cmdline options too here or in #pandaboard for the 1G17:21
rsalvetipending review17:21
rsalvetiogra: ^17:22
lagAh I see17:22
robclarkyeah.. bootargs are a bit different for 1g..17:22
ograrsalveti, yeah, i somehow need to merge that with persia's cmdline parsing change17:22
lagInteresting17:22
rsalvetibug 62926917:22
ograwill do that over the weekend17:22
robclarkbut sebjan and I still see some instabilities w/ 1g after some time...  I suspect there are some highmem vs smp issues, perhaps17:22
ubot2Launchpad bug 629269 in jasper-initramfs (Ubuntu) "Panda ES2.0: Correct mem arguments for 1gb (affects: 1) (heat: 8)" [Undecided,In progress] https://launchpad.net/bugs/62926917:22
rsalvetiogra: cool17:22
ografor now, i'm off ... (but will do the x-loader stuff later tonight as promised)17:24
rsalvetiogra: ok, will push the x-loader-omap4 update also, then you can review over the weekend or monday17:25
ograoki17:25
rsalveti/home/ubuntu/kernel/ubuntu-maverick/include/linux/page-flags.h:256: internal compiler error: Bus error17:25
rsalvetisebjan: the error I got when compiling the kernel17:26
rsalveti[ 3513.993927] Unhandled fault: imprecise external abort (0x1406) at 0x0006502817:26
* rsalveti never saw a similar error before17:26
robclarkrsalveti: similar to what I've seen17:26
robclarkbuilding GStreamer stuff17:27
rsalvetihm, weird17:27
robclarkrsalveti / sebjan:  one thought.. I suppose you could try booting with 'nosmp' in bootargs, and see if highmem issues go away..  that would be one way to see if they are smp related17:28
sebjanrsalveti: same error I see while building the kernel17:30
rsalvetirobclark: yep, would be good if we could isolate it17:30
robclarkmaybe over the weekend I'll leave it rebuilding some stuff in a loop overnight.. and see how it is in the morning17:31
Vic_Sage_The_QueI have a question what apps are available for ubuntu ARM17:42
Vic_Sage_The_Que??17:42
Vic_Sage_The_QueI'd be looking for stuff like GIMP, Scribus or openoffice17:42
Vic_Sage_The_QueI am asking because I was looking at getting the tegra powered notion ink adam17:43
rsalvetigimp and scribus should work fine, I believe17:45
rsalvetiopenoffice there is the java issue17:45
rsalvetino java at the moment17:45
rsalvetiyou'll probably have issues with mono-based applications17:45
=== pcacjr__ is now known as pcacjratwork
Vic_Sage_The_Quemono-based??17:51
Vic_Sage_The_Quewell openoffice isn't really a must have but I might need just a basic office suite17:53
Vic_Sage_The_QueI definitely need Karbon14, or some other vector based appz17:56
Vic_Sage_The_QueI am a designer incase you didn't know LOL17:58
Vic_Sage_The_Queanybody home?18:01
Vic_Sage_The_Quehello?18:15
=== jcrigby is now known as jcrigby_lunch
Vic_Sage_The_Quedoes xarax work on arm based ubuntu?18:21
=== james_w` is now known as james_w
tommdI'm having issues booting Mavrick Alpha on the touchbook - has anyone gotten that going?20:12
=== jcrigby_lunch is now known as jcrigby
DanaGES1.0?  is that ES as in GL ES, or Engineering Sample?  I had never considered it meaning the second.21:08
DanaGSay, is NDA still in place on specs and such?21:09
DanaGAnd release date?21:09
DanaGIt's also a bummer most of the connectivity is just USB 2.0... it'd be nice to have the stuff Marvell has, such as mPCIe and SATA and GbE.  Too bad Marvell lacks NEON.21:11
robclarkDanaG: ES == engineering sample..21:32
robclarkI'm not sure about board specs and that sort of thing for panda..  but I don't think it is a big secret..21:32
robclarkand OMAP4430 TRM is public now21:32
DanaGAh, what I'm most curious about is whether it's more like Beagle C4, or like Beagle XM.  Also curious when it'll be buy-able.21:33
robclarkI think closer to beagle XM21:34
DanaGCool.21:34
tommdWhat device?21:34
robclarkpanda21:35
tommdah21:35
robclark(well, either that or DanaG and I are talking about two different things ;-))21:35
DanaGYup, panda.21:36
DanaGI think that'd be a much cooler thing to have as a router, than a real router.21:36
DanaGs/real/consumer/21:36
DanaGWould just need an access point to go with it.21:36
dcordesDanaG: Use some usb wifi device which supports master mode to create your panda ap ?23:15
dcordesDanaG: ap = access point23:15
DanaGDo any USB wifi things even do Master?23:19
DanaGThe only USB wifi-stick I have on hand is a crappy rtl8187, that actually won't even work at ALL on ARM.23:19
DanaGIt tries to join SSIDs with garbage characters as names... seems like uninitialized memory.23:20
=== orbarron is now known as orbarron|OoO

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