/srv/irclogs.ubuntu.com/2010/09/06/#ubuntu-arm.txt

prpplague^2ho ho hum01:45
prpplague^2new laptop is working pretty well01:45
rsavoyemine too04:26
rsavoyeI added a 16GB SD card to a a Sharp NetWalker, so very portable ARM build machine04:27
DanaGhmm, can you run ubuntu on it?  Lucid and Maverick, specifically.04:32
DanaGah, Freescale i.MX515.04:33
DanaGA bit expensive for the specs, though.04:33
rsavoyeI'm running a maverick chroot04:36
rsavoyeyeah, pricy compared to the other imx51 machines, but smaller, and has a touchscreen and keyboard04:36
rsavoyeI couldn't find anything else that was actually available. got a road trip ion a few days...04:37
rsavoyeat least it'04:37
rsavoyes twice as fast as a beagleboard :-)04:37
ojnTalk about setting the bar low. :-)04:42
nimaerror: implicit declaration of function '__cpu_to_be16'04:45
=== ApOgEE__ is now known as ApOgEE
=== plars` is now known as plars
=== hrw|gone is now known as hrw
hrwmorning08:56
Baybalevening09:39
ndecogra: hi. is it possible for me to know which version of package was pulled in a daily image?10:24
ogra_cmpcndec, not currently, i'm working on a fix for deploying the .manifest files for preinstalled images10:29
ndecogra_cmpc: thx. so which kernel version was used in the 09/06 image?10:30
ogra_cmpcndec, the latest from sebjan's tree10:30
ogra_cmpcndec, but still the ES1 x-loader so i think they wont boot ... new x-loader was just uploaded and i'll do a 0906.1 image as soon as it shows up in the archive10:31
ndecogra_cmpc:  thx. I think this isn't the latest but this one: https://launchpad.net/ubuntu/+source/linux-ti-omap4/2.6.35-903.8, right? bryan has made a new pull request over the week end for 8-layer board.10:31
ndecogra_cmpc: ok. i see. so the current image will work on 6 layer ES2.0. on 8-layer the USB host isn't working (as per my email)10:32
ogra_cmpcyes, its the one we got during the beta freeze10:32
ogra_cmpcndec, ok, i'll make sure we'll get the patch in asap10:33
=== amitk is now known as amitk-afk
=== ApOgEE__ is now known as ApOgEE
ogracooloney, are you on top of the above ^^^^ ?10:52
cooloneyogra: i just got my new ES2.0 board booting with the kernel 903.9 based on sebjan's latest tree10:59
ograbuilds for 20100906.1 are running, we should have a working ES2.0 (6 layer) image in about 2h10:59
ogracooloney, 6 or 8 layer ?10:59
cooloneyogra: will test the usb host soon10:59
ogra(6 layer is painted black, 8 is green)10:59
cooloneyogra: it's 8 layer, green one10:59
ograok11:00
cooloneyogra and ndec, usb host works on my es2.0 8 layer board11:03
ograwith the current kernel package from the archive ?11:03
ndeccooloney: ^^^^11:06
cooloneyndec: http://people.canonical.com/~roc/kernel/omap4/11:08
cooloneyi just prepared a branch for our ti-omap4 updates11:08
ndeccooloney: so this is with all patches from sebjan, not current archive, right?11:08
cooloneyit is based on sebjan's latest branch11:08
cooloneyndec: yeah, exactly11:09
cooloneyit is still waiting for review and pull11:09
ograah, k11:14
=== amitk-afk is now known as amitk
persiandec, If you've *booted* a daily image, `dpkg -l ${packagename}` will tell you which version you have (although the .manifest files will let one discover that without booting)11:41
ograi just added a fix for the manifest files11:42
ogranext builds should have it11:42
persiaAre you running one manually, or tomorrow's daily?11:43
ogramanually atm11:43
ografor ES2.0 ...11:44
ograwe had a gcc breakdown over the weekend, so i held back the x-loader fixes11:44
ograand only uploaded them this morning11:44
persiaSaw that.  Much thanks to ScottK for helping sort that.11:44
ogra_grmpf11:49
=== ogra_ is now known as ogra
* ogra mumbles something about disconnects11:50
persiaSomeone needs to write a good multi-host IRC client.  I've heard of irssi scripts that try, but there's no good reason we oughtn't be able to have two proxies on two hosts, and have our clients connect to the proxies, and have it all get sorted.11:50
ograwell, if my line goes down a proxy behind the gw doesnt help much11:51
ogradown is down11:51
persiaThe idea would be to put one proxy behind e.g. your line, and one proxy behind e.g. my line (and let me have a proxy behind yours), and we'd not lose backscroll.11:52
persiawhich makes disconnects less bad, assuming the client has infinite backscroll features (like e.g. quassel)11:52
ograor xchat11:52
persiaxchat doesn't have infinite backscroll, does it?11:53
* persia hasn't found a way to have more than about a week before xchat crashes on startup11:53
ograi have logs since 2004 for the ubuntu channels on my disk11:53
ograseveral gig11:53
persiaBut can you reach them with PgUp in your client?11:54
ograno and i wouldnt want to :)11:54
ograi have 5000 line scrollback set or some such, for more i look at the local logs11:54
persiaMakes sense.  Having infinite backscroll basically just integrates the experience of looking at the logs, as one does it in one's client.11:55
persiaAnyway, well-off-topic for this channel :)11:55
ograheh, yeah11:55
ograpersia, did you see the new pulse bug above ?11:55
ograoh, it wasnt on IRC ... only in mail11:56
ograbug 63136211:56
ubot2Launchpad bug 631362 in pulseaudio (Ubuntu Maverick) (and 1 other project) "Include several configuration files (affects: 1) (heat: 14)" [Medium,Confirmed] https://launchpad.net/bugs/63136211:56
persiaI don't like it architecturally (we ought sort things so that we don't need to do that), but that's long-term, really.  I have no objections for maverick, assuming that it either backports sanely, or there's a new upstream pull for Maverick anyway.11:58
ograyeah, i'm waiting for TheMuso to reply11:58
ograi pinged him in -devel ... assuming he's back from vac.11:58
persiaHe is, but you pinged him late local time.11:59
ograyep11:59
ograpersia, oh, btw, did you see my ping on the weekend ?11:59
ograi have access to toshiba ac100's for 380€ here12:00
ogra(tegra2 based netbook)12:00
persiaI did.  It's larger and heavier than I like, and has less battery power than other devices I can source locally.12:00
persiaErr, less battery life.12:00
persiaOh, and it has the same annoying issue of needing special kernels :(12:01
ograyes12:01
ograits very very light though12:01
persiaThat said, if you're planning to make a clean kernel available (or know someone who is), I'll take one as a buildd.12:01
ograi took a look on saturday12:01
persiaIt's heavier than a netwalker :p12:01
ograbut has a full sized kbd and reasonable display12:01
ograsure, even my n900 is heavier than a netwalker i guess :)12:02
persiaI think n900 is about 30g lighter than netwalker.  netwalker and ac100 have the same resolution.12:02
ogra(given the amount of metal in the case)12:02
ograi did say resolution above on purpose (since i knew you would come up with that) ;)12:02
ogra*didnt12:03
persiaheh, you said "reasonable display": I prefer DPI to size :)12:03
persiaAnyway, food.12:03
ograenjoy12:03
rsalvetimorning12:46
ograheya12:46
rsalvetihm, cooloney is not here anymore12:50
ograany issues ?12:51
rsalvetiogra: not much, I sent an email with my current issues already, just wanted to hear more testing feedback from current sebjan's tree12:53
ograah12:53
ograwell, the USB issues with 8 layer boards seem to be in the works12:53
rsalvetiit's the last bit to have es2 6 and 8 layer support12:53
rsalveticool12:54
ograright, i guess there is a pull request for it (that was discussed above)12:54
rsalvetinow we just have the weird bug we face when building the kernel12:54
rsalvetiwhen using 1gb12:54
ograrsalveti, on a sidenote the 20100906.1 image should have all your x-loader changes12:54
* ogra is just zsyncing12:54
rsalvetiogra: nice, will do the same, thanks12:54
rsalvetiogra: then we can start using 1gb already, but need to change the kernel cmd line at jasper12:55
sebjanrsalveti: I did some more testing: I cannot reproduce the issue if I disable highmem support in the kernel config (so around 760M are acces-able without highmem)12:55
ograi would have uploaded earlier, but the gcc desaster somehow kept me from it (wouldnt have built until this morning anyway due to it)12:55
ograrsalveti, jasper was uploaded too12:56
rsalvetiogra: oh, cool12:56
rsalvetisure, np12:56
ograi still need to modify persia's patch but decided to make the change first12:56
rsalvetisebjan: interesting12:56
rsalvetisebjan: did you test without smp?12:56
sebjanrsalveti: I can easily reproduce problems by using memtester12:56
persiaogra, Anything you didn't like about my patch, or just integration with the moving trunk?12:56
rsalvetisebjan: nice, easier12:57
ograpersia, the changes of the defaults in omap4 need to be adjusted12:57
sebjanrsalveti: with memtester, I can reproduce the issue with 'nosmp'12:57
ograpersia, your patch is fine by the looks of it12:57
sebjanrsalveti: sudo memtester -p 0xb0000000 12012:57
persiaAh, good.  Yeah, any changes need to get integrated into that :)12:57
sebjanrsalveti: when it cashes, it's quite fast12:58
rsalvetisebjan: hm, interesting12:58
ndecpersia: thanks for your comment... i was looking for package version before booting the image ;-)13:06
persiaThat *should* be available soon, based on the traffic in the relevant channel.13:07
ogramy patch sdaly had a thinko ... already doing a rebuild13:08
ograndec, et voila ... http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100906.2/maverick-preinstalled-netbook-armel+omap.manifest13:12
ograerr13:12
ograhttp://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100906.2/maverick-preinstalled-netbook-armel+omap4.manifest indeed :)13:12
rsalvetiawesome13:13
ograGrueMaster, http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100906.2/maverick-preinstalled-netbook-armel+omap4.manifest with love and kisses :)13:13
ogra(he is waiting the longest for it i think :) )13:13
* ogra takes a break13:15
ndecogra: thanks for the image, and thanks for the manifest !13:32
ndecogra: i downloaded beta image, how do I use zsync now? never used it...13:33
ogra_cmpcyou need the .gz file from last download around13:34
ndecogra: so I can't do this from the beta .gz...13:34
ogra_cmpcthen in the same dir you just call zsync http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100906.2/maverick-preinstalled-netbook-armel+omap4.img.gz.zsync13:35
ogra_cmpcit pulls down the .zsync file, compares the changes and then downloads the differences and applies them to the existing .img.gz file13:36
=== Riotta is now known as Riot777
ogra_cmpchmm, seems i dont have *any* USB support on the ES213:50
ogra_cmpcneither on the 6 nor on the 8 layer13:51
ndecogra: guess what... it worked ;-) I had to use '-i' option for the beta file since it has a difference name, but this is much better!13:54
ogra_cmpcyeah, its quite fast :)13:54
ograhrm, i though the USB issue only affects 8 layer boards14:02
rsalvetiogra: you can try the uImage from http://people.canonical.com/~rsalveti/maverick/kernel/es2/linux-image-2.6.35-903-omap4_2.6.35-903.8rsalveti1_armel.deb, that includes the latest not-yet released patches14:05
rsalvetiI created to test the new patches we're getting14:05
rsalvetibut I also thought that it'd affect just the 8 layer board14:06
rsalvetionce the image download finish I'll try on my 6 layer14:07
ograhmm, i should probably also drop the hardcoded usb0 stuff from jasper now14:21
ograhmm, it also looks like the display driver is completely broken now14:24
rsalvetiogra: yep, should work ok without it14:24
ograno output during resize :/14:25
ogradang ... it all worked so well before14:25
rsalvetiogra: this is what I'm getting http://www.flickr.com/photos/rsalveti/4932601965/, but beside that it seems to be working fine14:25
ogralucky you14:25
rsalvetiogra: hm, you probably testing the new hdmi patches for the first time14:26
rsalvetiogra: are you testing with current deb file or from mine?14:26
ograi dont even get a signal14:26
ograstill with the current image14:26
rsalvetihm, ok, you could try with mine because we also have some more new patches14:26
ograi freshly dd'ed it again, booting the 6 layer ... no display no nothing14:26
rsalvetiand one that fixed a possible regression with many monitors14:27
ograi *saw* the oem-config screen already on that image14:27
* ogra doesnt get it14:27
rsalvetihm14:28
ogranow thats weird14:32
ograseems it didnt even boot at all14:32
rsalvetiouch14:32
ograat the fifth try with the same SD i now get output on the screen and it is *resizing* !14:32
ograi dont get why it didnt resize the 4 times before yet14:34
ograi guess i should start from scratch and check with serial console set upü14:34
persiaGruemaster reported a couple cases where it didn't resize on some boots but did on others, with same SD and same image.14:35
persiaMight be something transient, or a race, etc.14:36
rsalvetiogra: and check your sd card :-)14:36
* rsalveti never saw it14:36
ograpersia, it doent boot14:37
ogra*doesnt14:37
ograso it doesnt get to the resize step14:37
ograi assume the kernel locks up before14:37
* ogra blames lag 14:38
persiaI don't have enough information to say if that is the same behaviour or not.14:38
rsalvetiogra: add earlyprintk=ttyO2 and serial console14:38
rsalvetithen you'll see what's wrong14:38
rsalvetisome instability is expected with the 2.6.35 hehe :-)14:39
ograwell, thats very odd instability14:39
rsalvetiwhat we know from now is that the highmem support is making our kernel instable atm14:40
rsalvetibut it should work most of the time, we only got issues when stressing the system14:40
* ogra is brave and tries the ES114:42
rsalvetihaha, at least your usb will be broken, if you get the kernel to boot14:43
ograwell, if it would boot :)14:44
ograsame hang it seems14:44
rsalvetiogra: any message at your console?14:44
ogranope, at least not without fiddling with the cdmline (which i didnt do yet)14:44
ograStarting kernel ...14:45
ograUncompressing Linux... done, booting the kernel.14:45
ograrsalveti, so dropping the mem args seems to make it boot on all boards15:01
rsalvetiogra: interesting, so the memory issue we saw is affecting more than we wanted15:04
rsalvetiso it seems that the best for now would be to disable highmem and use 760M, as sebjan pointed before15:05
* ogra takes back the 'all' ... doesnt boot the ES115:05
ograbut who cares :P15:05
ograwell, currently i dotn have any mem arg at all15:05
ograthats likely wrong too15:06
rsalvetiyup15:06
ograbut i get through to oem-config at least15:06
ogralet me add the stuff back and fiddle more with the cmdline15:06
rsalvetiok15:06
rsalvetiargh, just finished the download, very slow15:07
rsalvetiwill test it now15:07
ograHA !15:10
rsalvetiogra: and?15:10
* ogra blames lag ... i said so from the beginning ! i get the same errors he had last week !15:10
ogra(indeed its not his fault *g*)15:10
rsalvetiogra: what errors?15:11
ograhttp://paste.ubuntu.com/489261/15:11
ogragood old ASoC15:11
rsalvetiWARNING: at /build/buildd/linux-ti-omap4-2.6.34/sound/soc/soc-jack.c:67 snd_soc_jack_report+0x30/0x100()15:12
* rsalveti asks why 2.6.34...15:12
rsalvetiprobably just old builddir15:12
ograoh, intresting15:12
ograthat shouldnt be ...15:12
rsalvetiI2C write to TSP61305 failed15:13
rsalvetiogra: which board are you using?15:13
ograthe black one15:14
ogra6 layer if i'm not wrong15:14
rsalvetiyep15:15
rsalvetiogra: and for the led to work we still need a patch at u-boot15:22
=== Riot777 is now known as Riotta
rsalvetiI sent it already for sakoman, and as soon he get it applied at his tree I'll ping jcrigby to update linaro's package15:22
rsalvetithen we'll have it working15:23
ograk15:23
rsalvetiok, time to try the image with my es2 615:23
rsalvetifirst boot is still with mem=463M, so not mem related15:28
rsalvetiouch, booom :-)15:29
rsalvetihm....15:29
rsalvetiLinux version 2.6.34-903-omap4 (buildd@hawthorn) (gcc version 4.4.5 20100728 (prerelease) (Ubuntu/Linaro 4.4.4-8ubuntu1) ) #7-Ubuntu SMP PREEMPT Thu Aug 5 16:12:58 UTC 201015:29
rsalvetiogra: ^15:30
rsalvetilinux-image-2.6.34-903-omap4 2.6.34-903.715:30
rsalvetilinux-image-omap4 2.6.35.903.415:30
rsalvetino meta upload?15:31
ograawwwww !!!!!15:32
ograbah !15:32
ograrsalveti, you rock !15:32
rsalveticool :-)15:37
rsalvetiogra: we need to ping cooloney to always request a meta-upload when sending a tree update15:38
rsalvetiit's the second time we face similar issues15:38
=== playya__ is now known as playya
ndecogra: this is really bad luck ;-)15:52
ogra_cmpcheh, yeah15:52
ogra_cmpci could have looked at the manifest file :P15:53
rsalvetilunch!16:02
=== bjf is now known as _bjf
ndecogra: if the initial install fails for some reasons, is it possible to restart the installer from scratch without reflashing the SD card?16:16
ograndec, fails ? in what way ?16:21
ndecogra: i don't know... just an open question ;-)16:21
ndecogra: if I don't plug the hdmi before booting, what happens?16:22
ograwell, if the resizing fails you are kind of screwed16:22
ograif the resizing works you will always get to eom-config until you finished it properly16:22
ograthe critical part is the re-partitioning and resizing16:23
ogra*oem-config16:23
ndecogra: can we get a console prompt on the console by default?16:25
ndecogra: I am not asking for a root console, just a prompt16:25
ograwell, i'm still planning to add teh code that brings up a getty on serial, but that requires that you fiddle with boot.scr and add a console= line16:26
* ogra wont add that by default16:27
ograjasper will see a bunch of changes before final freeze, thats one of the ones on my list16:27
ndecogra, rsalveti, robclark: i just tested with a new display (Full HD, ACER AT 2355) and it worked out of the box. xrandr tells me 1920x1080p. I just had to set the 'HDMI scan info' to underscan in the TV menu.16:35
ograndec, i suppose not with the kernel on the image though, right ?16:36
ndecogra: right, I should have said this... i am using a different MLO, uboot and uImage. the ones that you will get soon too16:37
ograi think we wont get any newer MLO anymore16:37
ograours should be the latest16:38
ogra(uploaded this morning)16:38
ndecit depends the latest of whom ;-)16:38
ograrsalveti pulled teh recent changes from sakoman i think ... i uploaded that today16:39
ogra  * new upstream release16:39
ogra   - adds Omap 4 ES2 6 layer and 8 layer support (LP: #624652)16:39
ogra   - uses branch omap4_panda_L24.9 from http://gitorious.org/pandaboard/x-loader16:39
ogra   - es2 compatible only16:39
ogra  * adding 02-panda-fix-ddr-timings.patch and 03-panda-x-loader-emif-1gb-support.patch16:39
ogra    to have 1gb support16:39
ografor u-boot we wait for some additional fixes and for linaro to pull them into the package16:40
ndecogra: ok. I logged in with the desktop, not UNE, but apps still open full window. is that expected or is it a bug?16:42
* ogra thinks the bug is rather that you can select a desktop session :P 16:42
ograwell, file one, i'll take a look16:42
ograsmells like metacity is somehow overridden16:43
ograits just silly that we have to ship a full desktop just to make GDM happy16:43
ograsuch a waste of space ...16:44
* ogra really looks forward to http://bobthegnome.blogspot.com/2010/07/lightdm.html16:45
ograthen we only need to ship what the session needs (the panel and some applets)16:45
ndecogra: i need to open the bug against which package? metacity?16:45
ndecor gdm?16:45
ogranah16:46
ogratry gnome-session but make clear its arm related and only happens if you have efl too16:46
* ogra goes afk now16:53
=== hrw is now known as hrw|gone
=== cwillu_at_work is now known as cwlu
=== cwlu is now known as cwillu_at_work
playyait seems, that we might get EFL 1.0 in 11.04 :)23:29

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