/srv/irclogs.ubuntu.com/2010/01/15/#ubuntu-arm.txt

=== asac_ is now known as asac
persiaSo, I think I figured out how to make plymouth work on armel, at least for ATI or nVidia cards.02:43
persiaBut someone needs to write KMS drivers for more typical video adaptors if it's really going to work for most folk.02:44
rcn-eehey persia what do you need? omap's have dss2 which went mainline in 2.6.33 which might work...02:54
persiarcn-ee: From what I'm told, I need a KMS driver, and then someone to write something likehttp://cgit.freedesktop.org/plymouth/commit/?id=f2048af97dcc862dbbb587a0cc2546ddbdbd2b0c that works with the KMS driver.02:55
persiaAnd preferably something that can build and (kinda) work on all architectures, so we don't have to special-case it.02:56
persiaMind you, some cases won't get much testing (like I'm building libdrm-intel for armel right now)02:56
persiaNote that the results of this will probably either require some customisation or more maturity of Ubuntu armel, just because with the current way arm kernels boot, only a couple dev boards get supported, which means very little chance of working drivers for fancy graphics chips.02:57
rcn-eevery true....  which reminds me, it would be nice if meta packages had 'arch' support, no need for ati/intel/nvidia stuff for armel...02:58
persiaBut getting the code written and into linux and plymouth makes it easy to get working once the hardware is there.02:58
persiaActually, metapackages *do* have per-arch support.02:58
persiaBut I disagree with your assertion.  I own an ARM device (sadly ARMv5) that has an ATI graphics accellerator.02:58
persiaAnd I know there exist ones with nVidia chips.02:59
rcn-eeyeah, that's right the msm7200 has one..  (we almost started stocking that one..) and nvidia's tegra...02:59
persiaDunno if the current radeon and nouveau drivers work for those, but we may as well publish packages and maybe someone will find a bug and fix it :)03:01
oyotatHello. Does anyone know if ubuntu-arm would run on a Gumstix Xscale? like a Gumstix Basix?03:01
persiaoyotat: I can say that you'll probably need a custom kernel, at least.  Let me check the gumstix site.03:02
rcn-eeactually plymouth's framebuffer mode should work, i need to test that tomorrow.. (dss2 sets the resolution at boot)03:02
persiarcn-ee: I don't know if I will get the necessary bits into the archive by tomorrow, but in short: build libdrm-intel for armel, then build plymouth using that libdrm-intel1-dev.03:03
oyotatpersia, thanks. one last question, anyone happen to know what's the difference between a marvell armada xscale like the pxa1xx versus the older intel xscale like pxa270?03:03
persiaYou end up with a currently-useless intel driver for plymouth, but it lets you test.03:03
persiaoyotat: You should be able to run Jaunty on a verdex-based system with a custom kernel, and up through lucid on an overo-based system with a custom kernel.03:05
persiaDunno which is "Basix".03:05
rcn-eeoyotat, pxa270 i think are only armv4 based...03:06
persiaI'm fairly sure they are ARMv503:06
persiaBecause some of the pxa26x series were ARMv503:06
persiaBut I could be mistaken.03:06
rcn-eeyeah they are armv5, + intel instructions... they were just weird...03:07
persiaWell, one can ignore the instruction extensions to make something work :)  My collection of Zauri have been happily doing their thing for some time (although I find that, despite the better form-factor, I have replaced all my use cases with the Netwalker)03:08
persiaErr, maybe I should have written "despite the inferior shape"03:09
persia(Zaurus case >> Netwalker case but Netwalker chips >> Zaurus chips)03:09
Sarvattif anyone wants to try a newer pixman with more neon optimizations out -- http://sarvatt.com/downloads/armel/lucid/04:33
=== bizkut-miau is now known as bizkut-redhat
=== suihkulo1ki is now known as suihkulokki
=== jamie is now known as JamieBennett
ogramumble12:56
ograso i cant load boot.scr for no apparent reason :/12:56
xrealHi there. Is it even possible to run Linux on a PNA with ARM-compatible CPU?13:09
persiaxreal: What's a PNA?13:10
ograpicking nose appliance ?13:11
xrealpersia: Personal Navigation Assistant13:11
persiaxreal: Well, there's probably three factors you'd need to investigate.13:12
xrealpersia: Okay, which ones?13:12
persia1) Which instruction set is supported by the processor: if it's new enough, you may be able to run Ubuntu13:12
xrealpersia: http://pdadb.net/index.php?m=cpu&id=a9g3&c=centrality_atlas_iii13:13
persia2) Is there a way to boot an arbitrary kernel and run against an arbitrary filesystem : if the device is hackable, you may be able to run Ubuntu13:13
xrealInstructions:  ARMv5TEJ13:13
persiaFor that instruction set, you could only run Ubuntu 9.04.  Newer versions will not run.13:13
xrealpersia: I've talked to the manufactor and he can offer me a "clean" device, without any software.13:13
persia3) Can you build (or find) a kernel that supports Ubuntu 9.04 userspace that runs on the device.13:14
xrealpersia: I think 4) will be: is the other hardware support, like graphics, touchscreen etc.13:14
persia4) Are there sufficient drivers for the available input/output devices to make it worthwhile?13:14
persiaYep :)13:14
persiaBut the first three are the key ones.13:14
xrealI think, I will stay on WM5 and will code software for Windows Mobile.13:15
persiaAnd 4) is kinda just personal style.  Some people would like to run Ubuntu on their toaster, even if the entire interface consists of a lever one pushes down, which pops up when the toast is done.13:15
xrealI wanted to use Linux on the device and code linux software ... but that's too hard to figure out.13:15
ograthat really depends on your toaster !13:16
persiaogra: Well, sure.  If you have a cool toaster, there might be a point.13:16
ograi do !13:16
xrealI'm not that deep into linux...13:16
* ogra got a cool toaster for chriatmas 13:16
ogra*christmas13:16
persiaOn the other hand, I've found that the smarter an appliance I have, the less likely I can use it successfully.  The processor on my rice cooker died, and now I can make rice reliably.13:17
persiaWhich one?13:17
ograhttp://fs1.kauflux.de/slot/358/artimg/large/13226207_401557.jpg13:17
persiablinkenlights!13:19
ograhehe13:19
persiaSo, did you install Ubuntu?13:20
persiaOr does it come pre-installed?13:20
ogranope, sadly not ... still working on an installation toast :)13:20
ian_brasilthat sir is an impressive toaster13:20
ograheh13:20
persiaWhat due the dark spots beside the slots do?13:21
ograthey hover out the breadroll device13:21
ograit slides out if you touch them13:22
persiaSo you can slow-roast stuff over the toaster?13:22
ograright13:22
persiaWow!13:22
ian_brasilthat rocks13:22
ograit also has a freezer program to thaw frozen toast13:22
persiaDoes it auto-toast once the bread is defrosted?13:23
ograyep13:23
* persia should really put on an op hat and declare this off topic, but deep fascination wars with this instict13:23
ogralol13:23
persiaThat's just a really nice toaster.13:23
ograit is, was my only item on the whishlist this year :)13:24
koczishi all, anyone interested in ARM/Cortex talk? - have a few questions... rather regarding hardware, but... , anyone?13:35
persia!ask | koczis13:36
ubot4koczis: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-)13:36
koczisself-introducing and the subject may be handy - especially when group subject isn't really in sync...13:39
koczisI have questions about interrupt handling and interaction between Cortex Core and peripherals in STM3213:40
koczisthe 1st is: do I have to clear pending flags early or lately13:40
koczis?13:40
persiaHrm.  You might have to wait for some time to pass to get an answer to something that specific.13:44
persiaMost of what we've been doing is taking code that compiled for ARMv4 and recompiling for newer versions, based on some toolchain changes.13:44
persiaAnd I don't think I've seen anyone talk about using STM chips, although that doesn't mean nobody does.13:45
persiaDepending on whether that was one of your more specific questions, or one of your less specific questions, it may be worth asking others, and hoping someone gets back to you about the first.13:45
persiaBut if that was one of your more general questions, you may find faster/better answers somewhere else (although I'd be happy to be proved wrong about that)13:46
koczisthat was less specific  - cause in fact: 1. I didn't find a word about it in STM32 doc, 2. I have own experience with early/late clear behavior on other arch, 3. I tested it and it looks the late clearing has flaws that make it outcomes unpredictable13:49
koczis...so I use only early clearing now13:50
koczisso the more specific question about same thing - how do you do it?13:50
asaclool:  * Bug:488267: ffmpeg: "should be built with -marm for lucid on armel"14:46
asaccant we just upload that?14:46
asacor is there more?14:46
asacbug 48826714:46
ubot4Launchpad bug 488267 in ffmpeg "ffmpeg should be built with -marm for lucid on armel" [Low,Fix committed] https://launchpad.net/bugs/48826714:46
asacbug 45665914:46
ubot4Launchpad bug 456659 in linux-fsl-imx51 "suspend/resume failure on imx51" [High,Confirmed] https://launchpad.net/bugs/45665914:46
asacplars: isnt that fixed in latest kernel?14:47
asaccooloney: ?14:47
asacis that still "fix available" ... or should we move that back to "needs investigation" ?14:47
asacdmart: hi14:48
dmartHi there14:48
plarsasac: still fix available, I've been told that they have a fix for it, but I have not heard that it has gone in yet14:48
asacdmart: we have  * Bug:458537: linux-fsl-imx51: hibernate does not work14:49
cooloneyasac: i prepared some kernel deb to do some test,14:49
asacsomeone said, that this bug doesnt make sense for arm14:49
asacis that a hardware restriction? or was that just none-sense14:49
asaccooloney: can you get plars that deb for verification? or post the url on the bug?14:49
* asac sets the bug back to in progress14:49
dmart(Bug #458537)14:50
ubot4Launchpad bug 458537 in linux-fsl-imx51 "hibernate does not work" [High,Triaged] https://launchpad.net/bugs/45853714:50
cooloneyasac: ok, cool, will update the status of the bug14:51
dmartWhy doesn't the bog make sense for arm?14:52
dmartbug14:52
loolasac: 15:55 < lool> asac: The changes are ready in git since Wed; I've tried to  confirm twice with sirestart how he want to handle some other  intrusive changes which are in git (symbol versioning); once  that's clear I can prepare the upload from git14:57
plarscooloney: got it, will test shortly15:11
cooloneyplars: thanks a lot. frankly, i am not sure whether this will fix the issue.15:24
cooloneyplars: but i will dig into it after i return to Shanghai since my board is there15:24
plarscooloney: installing now, should be able to let you know shortly15:24
cooloneyplars: great, so what is the symptom on your side of that suspend/resume issue15:25
plarscooloney: does this kernel also have things to possibly help with the suspend/resume?15:25
cooloneyplars: yeah, might be, because i merged some patches from FSL latest BSP15:26
plarscooloney: on imx51, it seems to suspend (although I have some doubts as to whether it suspends completely), but will not resume15:26
cooloneyplars: those patches are relatedt to suspend/resume15:26
plarsif I connect to serial, I can see a message like "<PWR> key pressed" when I push the power button, but nothing happens15:27
plarscooloney: oh, you posted it on the hibernate bug15:27
cooloneyplars: right, i posted both of them, since i applied 30+ patches from FSL latest BSP15:29
plarscooloney: will do15:29
cooloneyplars: need your help to test, because I am traveling now. heh15:29
cooloneyplars: thx a lot15:29
plarscooloney: It's what I do :)  Feel free to point me to stuff like that anytime15:30
cooloneyplars: appreciate, oh, do you also have BB3 or just BB2?15:32
plarscooloney: I'm testing on bb3 at the moment15:32
plarscooloney: I have a 2.5, but it is in the box at the moment15:33
cooloneyplars: cool, thanks a lot15:33
plarscooloney: suspend/resume appears to be unaffected by the new kernel, still broken15:33
cooloneyplars: ok, got you,15:34
cooloneyplars: i plan to ping fsl guys for help know15:34
cooloneyplars: if we wanna to resume the suspended board, we press keyboard or the PWR key15:38
asacbug 50577215:41
ubot4Launchpad bug 505772 in linux-mvl-dove "system freezes sometimes after X is up for a while" [Critical,Confirmed] https://launchpad.net/bugs/50577215:41
plarscooloney: neither keyboard, nor power key seem to work.  Also, hibernate still seems to not be supported (not listed in /sys/power/state)15:41
cooloneyplars: ok, thanks, will ping fsl guys soon15:42
plarsasac: that one could be related to 504880, definitely still broken15:42
plarsbug #50488015:42
ubot4Launchpad bug 504880 in linux-mvl-dove "[dove] possibility of thumb2 instructions invalidly being handled" [High,Confirmed] https://launchpad.net/bugs/50488015:42
plarscooloney: thanks, I'll update the bugs15:42
cooloneyasac: yeah, i think ericm might know that freezing bug, right?15:43
asacbug 50488015:45
ubot4Launchpad bug 504880 in linux-mvl-dove "[dove] possibility of thumb2 instructions invalidly being handled" [High,Confirmed] https://launchpad.net/bugs/50488015:45
asaccooloney: yes. sorry. just posting here so i can copy that text to release team meeting report15:45
asaccooloney: plars: ogra: JamieBennett: https://wiki.ubuntu.com/MobileTeam/ReleaseStatus/Lucid15:47
asacplease check if you want something to get added to summary15:47
asacand if i missed anything on the RC bug front15:48
ograasac, uboot NIC issues ?15:48
asacogra: bug?15:48
plarsbug #50788715:49
ubot4Launchpad bug 507887 in uboot-imx "NIC not properly initialized in uboot code" [High,New] https://launchpad.net/bugs/50788715:49
ograhttps://bugs.launchpad.net/ubuntu/+source/uboot-imx/+bug/50788715:49
asacgive me id .. .ensure its targetted for lucid and milestoned15:49
ubot4Launchpad bug 507887 in uboot-imx "NIC not properly initialized in uboot code" [High,New]15:49
plarsha! I win :)15:49
ograheh :)15:49
asac"better, strong faster ..." ;)15:49
asacplars: ogra: please remember to target such things for lucid ...15:49
asacdone15:49
ograoh, sorry, i milestoned it for A3 though15:50
asacyes, both is required15:50
asacalso all that are targetted need to be milestoned (i failed in the past)15:50
asacand have an assignee15:51
* asac checks bug list again to see if anything is missing15:51
plarsasac: 462798 needs a target also, but is medium and I don't know when/if NCommander plans to work on.  NCommander is a3 reasonable for this?15:52
NCommanderbug #46279815:52
ubot4Launchpad bug 462798 in ubiquity "selecting 'new partition table' confuses the partitioning" [Medium,Triaged] https://launchpad.net/bugs/46279815:52
plarssorry, asac, NCommander bug #46279815:52
NCommanderplars, medium, alpha 3, I'm going to work with persia on this during the sprint if we can fix the dove issues15:53
plarsgreat!15:53
plarswill target15:53
persiaIs that when we're doing it?  I've been wondering :)15:53
asacogra: i will put that uboot net bug on your plate for now ...15:54
asaci will also try to help at least to verify15:55
asacbut i need an assignee15:55
asac;)15:55
ograhrm, k15:55
asacogra: feel free to reassign to someone else from the team ;)15:56
ograheh15:56
asacogra: bug 50676115:57
ubot4Launchpad bug 506761 in uboot-imx "lucid uboot hangs on fatload uImage on fsl TO2 TO2.5 and TO3" [Undecided,Confirmed] https://launchpad.net/bugs/50676115:57
asacisnt RC anymore15:57
asacor should we even close?15:57
ograno, there is still something wonky15:58
ograext2load works flawless15:58
ograwhile fatload doesnt15:58
asacright. but its not RC15:59
asacas we managed to work around15:59
asacor is the script loading bug also gone on ext215:59
asac?15:59
plarstoo bad we don't have something for the "fixed this week" category.  dmucs ftbfs was fixed iirc, but not exactly rc15:59
asacright15:59
asacwell15:59
asaclool: i dont understand what i means if you say that ffmpeg is committed to git16:00
asacwhen does that come down to us?16:00
asaccan we just upload it to ubuntu ?16:00
ograasac, the script *only* loads on ext216:00
asac;)16:00
asacogra: yeah. ok16:00
ograi cant make it load on fat at all16:00
ograBBG U-Boot > fatload mmc 0:2 0x90800000 boot.scr16:00
ograreading boot.scr16:00
ogra... hangs forever ...16:00
asacyeah16:01
ograBBG U-Boot > ext2load mmc 0:2 0x90800000 boot.scr16:01
ograLoading file "boot.scr" from mmc device 0:2 (xxa2)16:01
ogra407 bytes read16:01
ograBBG U-Boot >16:01
asaci will check fat code16:01
ograworks flawless16:01
asacwe need to add debugging16:01
asacu-boot really needs better debugging facilities imo16:01
asaceven the DEBUG flags are all broken16:01
asacogra: i can reproduce it, so i can try to work on that if we cant go for ext216:01
ograwell, i wouldnt be surprised if thats due to some vendor patches16:01
ograwe cant16:01
ograext2 needs root16:01
ograwe dont have root on the build machine16:02
asacsure16:02
asaci will check that16:02
asacthought e2progs work ... or does that need root too?16:02
ograits to slow16:02
ograNCommander tested it and said its unusable16:03
asacerr e2tools16:03
asacok. noted16:03
NCommanderogra, well, I never looked at optimizing it to reduce the number of e2tools calls16:03
ograNCommander, but you said it takes hours to copy the files16:04
ograeven if you optimize that will be to slow16:04
NCommanderogra, no, it was about ~10-15 minutes per image16:04
asacsounds not that much16:04
ogravs 1-2 min :)16:04
ograthats a lot16:05
asacright. but its by far not a dramatic problem16:05
asacor how many image runs are we doing?16:05
asacone-per-image ... right?16:05
asacaka 2 ;)16:05
asacrather 4 (with alternates)16:05
ogradont forget a live build already takes 90min16:05
ograduring alpha and release time thats already to much16:06
ogra(which is why we're still waiting for a second livefs builder to not have to deal with 3h buildtime)16:06
asacstill. its acceptable imo16:06
asacat least as a backup16:07
asaci will try to fix fat and talk to uboot folks that did the fsl patches16:07
asacits really annoying how fat behaves ;)16:07
ograyou mean to fsl folks that did the uboot patches :)16:07
asacyes16:07
asaci will escalate that if i dont find the issue quickly16:08
ograyeah16:09
asacogra: so you already tried vfat?16:10
asacthat felt like a good pointer16:10
ograyes, but its already enabled in some common.h file16:10
ogradidnt change a thing, just spilled redefine wanrings16:10
persiaThe other massive benefit to FAT is that if we get it working, and get USB support working, we can do bootable USB flash solutions without requiring users to reformat their flash.16:11
persia(or FAT-on-SD, with the same benefits)16:12
persiaWhich means no needing to use dd (or wrappers).16:12
cooloneyasac: just got email from fsl guys, they do not support hibernate, are we going to do that?16:12
asaccooloney: if they dont support it, we wont do it ... will remove it from RC bug list after meeting16:13
cooloneyasac: got it, i think you are in the email loop i sent, let me update the status of the bug hibernate16:14
asacthanks16:15
asaccooloney: do you know if its a hardware limitation thing? or just because there is no code yet? (or is that in the mail i havent read yet)?16:15
cooloneyasac: i just checked code quickly, i think there is no code for hibernate.16:18
asacok. but doesnt that need hardware support too?16:19
asacogra: can you link the uboot bug to the spec?16:21
asacthx16:21
asacslangasek will otherwise poke i guess16:21
ograasac, both ?16:22
ogra(NIC and FAT)16:22
asacboth16:22
cooloneyasac: and from my experience, it seems no arm soc supports hibernate16:23
ogradonr16:23
ograe16:23
ografun16:23
ogranow i can load boot.scr but uimage hangs16:24
asaccooloney: right16:26
asacthanks16:26
asacogra: make the partition even bigger ;)16:26
asacor bad the file with zeros16:26
ogranah, thats cant be it :P16:26
asacso its at least a full block :)16:26
asacwho knows16:26
asaci saw code in uboot that said: "always full in full blocks"16:27
asacpull16:27
ograboot.scr is 315 byte big and loads fine16:27
ogradoesnt load uinitrd either16:30
loolasac: You pinged me on this bug16:33
loolasac: 15:46 < asac> lool:  * Bug:488267: ffmpeg: "should be built with -marm for  lucid on armel"16:33
loolasac: I'm saying that I fixed this and it's pending upload; it's commited in the packaging git repo16:34
asacsure16:34
asaci understood that. just wasnt sure what that means16:34
asacbut nevermind16:34
loolI dont understand where the ambiguity is16:34
asace.g. is that git repo for debian16:34
asacwill we wait until it migrates to testing16:34
asacetc.16:35
loolasac: It's a git repo for debian and ubuntu16:45
loolThe changes aren't uploaded to Debian16:45
loolIt's like regular Vcs-Git stuff; not sure what to say16:46
asacsure. git just felt like debian16:48
asacthats why i double check16:48
ogradebian isnt *that* bad, come on16:49
loolWhat's in Vcs-Git for ffmpeg and -extra is correct16:50
loolNo way to encode the branch infor,mation though16:51
asaci dont mind debian, just the time delay if we go through debian first17:38
asacthats why i wanted to clarify it17:38
asacso all fine ;)17:39
=== bizkut-miau is now known as bizkut-redhat
=== bizkut-miau is now known as bizkut-redhat
=== bizkut-miau is now known as bizkut-redhat
=== keesj_ is now known as keesj
loolsuihkulokki: You might be interested in kees patch for the ld-linux segfaults22:07
loolhttp://bugs.launchpad.net/bugs/45217522:08
ubot4Launchpad bug 452175 in linux "Random segfaults when using ld.so explicitly to start a program" [Medium,In progress]22:08

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