| === asac_ is now known as asac | ||
| sveinse | We've been running our prototypes from maverick using rootstock which does not install a kernel nor a initrd file. Why do Ubuntu need a initrd script and why does it apparently seem to work without one? | 07:44 |
|---|---|---|
| ppisati | morning | 09:00 |
| sveinse | We've been running our prototypes from maverick using rootstock which does not install a kernel nor a initrd file. Why do Ubuntu need a initrd script and why does it apparently seem to work without one? | 10:40 |
| sveinse | I.e. all necessary features are compiled into the kernel, so I don't need modules. Apart from that, what is initrd doing? | 10:42 |
| ogra_ | read up about initramfs-tools, its deeply integrated with the packaging system (packages can put bits and pieces into the initrd) ... | 10:43 |
| ogra_ | some packages rely on functions that are only possible while / isnt mounted yet (fsck, lvm, dmraid, entrypted filesystems etc) | 10:44 |
| ogra_ | *encrypted | 10:44 |
| ogra_ | while ubuntu will work fine without initrd, if one of your users installs something like the encryption stuff, encrypts / and reboots without initrd, he trashed his install | 10:45 |
| ogra_ | if you can take that risk ... live without initrd :) | 10:46 |
| sveinse | one point though (and only one from your list): fsck. Rather important. | 10:47 |
| sveinse | Thanks | 10:47 |
| sveinse | in this natty release frenzy, how is armel doing? | 10:49 |
| ogra_ | quite well, though there was a heavy issue with libreoffice | 11:00 |
| === sbambrough-lunch is now known as sbambrough | ||
| === JamieBen1ett is now known as JamieBennett | ||
| sempil | how | 12:23 |
| ericb2 | hi | 12:42 |
| ericb2 | I did an apt-get dist-upgrade on my OMAP3 (Beagleboard xM) and X.org is frozen ... :/ | 12:42 |
| ericb2 | more precisely, the mouse cursor is frozen | 12:42 |
| ericb2 | is there a known reason and in this case, what is the guilty archive, what has to be done ? .. and so on | 12:43 |
| ericb2 | thanks in advance | 12:43 |
| ericb2 | looks like ssh works ... better than nothing :/ | 12:58 |
| === zyga is now known as zyga-afk | ||
| rsalveti | ericb2: are you using it with a sd card? | 13:16 |
| rsalveti | janimo: have any idea why some packages from libreoffice didn't contain armel at it's architecture? | 13:31 |
| janimo | rsalveti, no, I only saw your bugreport today | 13:32 |
| janimo | AFAIK LibO used to build and work fine on arm during natty | 13:32 |
| janimo | since about 3.3.1 at least when it started building fully | 13:32 |
| rsalveti | janimo: hm, ok, do you remember how many hours it usually takes to finish the build at our builders? | 13:32 |
| janimo | over 1 day | 13:32 |
| rsalveti | 10 hours until now, and still building, hopefully it will work just fine | 13:33 |
| janimo | did some armhf patches get in via debian? | 13:33 |
| rsalveti | janimo: yes | 13:33 |
| ogra_ | rsalveti, so you are rebuilding locally atm ? | 13:45 |
| rsalveti | ogra_: both locally and at the PPA | 13:45 |
| ogra_ | ah, k | 13:46 |
| ogra_ | so we could pocket-copy it from the PPA ? | 13:46 |
| ogra_ | (once the binaries are there) | 13:46 |
| rsalveti | ogra_: that's the idea | 13:47 |
| ogra_ | ah, cool | 13:47 |
| ogra_ | bah, i cant commit to desktop-webmail upstream | 13:48 |
| ogra_ | JamieBennett owns it | 13:48 |
| * JamieBennett hides | 13:48 | |
| JamieBennett | ogra_: you can have it now ;) | 13:48 |
| ogra_ | heh, assign it to canonical-arm then | 13:49 |
| JamieBennett | ogra_: infact -webmail is asac's | 13:49 |
| ogra_ | though probably a general core-dev approval would be better | 13:49 |
| ogra_ | asac, ^^^ can you hand that over to core-dev ? | 13:50 |
| ericb2 | rsalveti: yes, I'm using it using an sd card | 13:50 |
| rsalveti | ericb2: so it could be that the heavy i/o from dpkg is making your beagle useless | 13:50 |
| ericb2 | rsalveti: :-/ | 13:51 |
| ericb2 | rsalveti: what has to be done ? | 13:51 |
| ogra_ | fix dpkg to require less resources ;) | 13:51 |
| ogra_ | or more likely debconf than dpkg | 13:51 |
| ogra_ | or switch to use an USB disk | 13:51 |
| ericb2 | ogra_: what does the bug exactly ? (what are the symptoms ? ) | 13:51 |
| rsalveti | even a faster sd card can help already | 13:52 |
| ogra_ | yeah | 13:52 |
| === ian_brasil___ is now known as ian_brasil | ||
| ericb2 | rsalveti: I can connect to the beagle using ssh | 13:53 |
| JamieBennett | mine is -office | 13:53 |
| ericb2 | rsalveti: at least I did once. e.g. I restarted gdm, but unsuccessfully | 13:53 |
| ericb2 | janimo: FYI, OOoLight works well on the beagleboard too | 13:54 |
| janimo | ericb2, great to hear that :) | 13:54 |
| janimo | fast enough? | 13:54 |
| ericb2 | janimo: first launch is 10s | 13:54 |
| ogra_ | form SD ? | 13:54 |
| ogra_ | thats fast ! | 13:54 |
| ericb2 | janimo: the proc is 800MHz and I got 512 MB of RAM | 13:54 |
| ericb2 | ogra_: from sd, yes | 13:54 |
| ericb2 | compiled from sd too | 13:55 |
| asac | ogra_: what do you need it for? | 13:55 |
| ericb2 | and starting second launch around 3 s | 13:55 |
| asac | ogra_: committing? i dont even see an attempt to propose or merging ;) | 13:55 |
| ogra_ | asac, we ship it by default ... bug 769235 has a fix | 13:55 |
| ubot2 | Launchpad bug 769235 in desktop-webmail "Icons & app names identical for desktop-email & desktop-email-config" [Low,Confirmed] https://launchpad.net/bugs/769235 | 13:55 |
| asac | ogra_: junk is probably not the best place to push such patches | 13:56 |
| ogra_ | asac, well, i'm lazy with LP ... | 13:56 |
| ogra_ | asac, well, bzr push lp:desktop-webmail didnt work and i need to move on ... +junk is still better than losing it from my disk ;) | 13:56 |
| asac | ogra_: can you push that to lp:~ogra/desktop-webmail/lp769235 and just hit propose for merge? | 13:56 |
| asac | ogra_: you dont know how launchpad works, right? | 13:57 |
| ericb2 | rsalveti: do you have the dpkg bug report at hand please ? | 13:57 |
| asac | i think bzr could be improved to have a command: bzr push-merge-proposal-for lp:desktop-webmail ;) | 13:58 |
| ogra_ | asac, slow, right ? | 13:58 |
| asac | ogra_: no ... but you can just push to the proper project and dont need to use junk | 13:58 |
| asac | ogra_: like i said: lp:~ogra/desktop-webmail/lp769235 would be appropriate | 13:58 |
| ogra_ | pushed that way now | 13:58 |
| asac | cool | 13:58 |
| ogra_ | opening the page will take me 20min though | 13:58 |
| ogra_ | so i will do the proposal later | 13:58 |
| StevenK | asac: bzr lp-propose | 13:58 |
| asac | ogra_: i can propose. thanks | 13:59 |
| === zyga-afk is now known as zyga | ||
| pgraner | GrueMaster, ping | 14:42 |
| rsalveti | ogra_: the slideshow seems to be disabled at the latest netbook daily image, is this expected? | 14:46 |
| rsalveti | got from 20110426.4 | 14:47 |
| rsalveti | ogra: GrueMaster: from 20110426.4: slideshow disabled at the installed, software center gave me a "not found" while trying to install the omap 4 extras because it seems that it didn't run apt-get update and I'm getting a lot of fifo underflow :-( | 15:08 |
| rsalveti | *installer | 15:08 |
| ogra | missing apt-get update is a fee we have to pay for changing the sources.list from jasper | 15:09 |
| ogra | (no network, so no update run) | 15:10 |
| rsalveti | ogra: and still a lot of "asoc" error messages at the kernel log, shouldn't this be fixed already? | 15:17 |
| rsalveti | or are we waiting the fix to hit the updates? | 15:17 |
| ogra | no, its fixed, but tobin seems to have found that one reboot is needed for it to properly take effect | 15:17 |
| rsalveti | hm, interesting, will try | 15:18 |
| ogra | i guess we need to create the state file from jasper | 15:18 |
| ogra | or some such ... | 15:18 |
| ogra | havent looked deeper yet | 15:18 |
| rsalveti | ogra: showing just pandahdmi output and still reporting the error messages | 15:34 |
| rsalveti | and no packages to update, using latest stuff | 15:34 |
| ogra | yes, the fixes are in the packages sine a week | 15:34 |
| ogra | *since | 15:34 |
| ogra | it shoudl just work | 15:35 |
| rsalveti | ogra: what was the fix? | 15:35 |
| ogra | adding the patches and a proper udev rule to select the right ucm profiles | 15:35 |
| GrueMaster | pgraner: Pong | 15:38 |
| rsalveti | ogra: udev rule is there, but still broken | 15:38 |
| GrueMaster | ogra: Audio does not come up after a single reboot. It appears that alsaucm is never run. | 15:38 |
| ogra | GrueMaster, well, thats strange sinc ethe udev rule runs it | 15:42 |
| ogra | and that just matches against the device name | 15:42 |
| * ogra has a netbook image ready, will test myself | 15:42 | |
| GrueMaster | Does it? Because I am not seeing audio until I run alsaucm manually then reboot. | 15:43 |
| ogra | well, there *should* be a udev rule :) | 15:43 |
| GrueMaster | There is, but I don't think it gets executed. | 15:44 |
| rsalveti | udev rule is there, could be that the device name is not right | 15:44 |
| ogra | ask diwic if it actually ended up in the upload where the changelog claims it did :) | 15:44 |
| rsalveti | or something similar | 15:44 |
| === fairuz is now known as fairuz-away | ||
| ogra | for anyone intrested, i think we can have proper images for the ac100 now with the .37 kernel ... you can actually update kernel and initrd from the running device | 15:48 |
| ericb2 | looks like I got the network working. Can I downgrade dpkg or something to make it work again ? | 15:48 |
| ogra | i think i'll roll a community image for it in oneiric | 15:48 |
| * ericb2 still wondering why x.org does not work after a big apt-get dist-upgrade on Beagleboard (xM) | 15:49 | |
| ogra | err | 15:52 |
| ogra | where is the slideshow ? | 15:52 |
| ogra | (and more important, why isnt there a bug for it missing) | 15:52 |
| rsalveti | ogra: I said you earlier, the slideshow seems to be gone | 15:52 |
| rsalveti | I asked you if this is expected or not | 15:53 |
| GrueMaster | ogra: I was trying to debug it and got sidetracked yesterday. It is only affecting armel. | 15:53 |
| rsalveti | that's why I didn't fill a bug yet | 15:53 |
| rsalveti | ogra: http://paste.ubuntu.com/599329/ | 15:53 |
| GrueMaster | I tested it on x86 in a VM. | 15:53 |
| ogra | hmm | 15:53 |
| ogra | it surely speeds up the install | 15:53 |
| ogra | ARGH ! | 15:54 |
| ogra | rsalveti, thanks, that says everything | 15:54 |
| ogra | oh my | 15:54 |
| pgraner | GrueMaster, FYI I tested the OMAP4 images and no issues according to the test on the ISO tracker | 15:55 |
| ogra | GrueMaster, rsalveti: sudo ln -s /usr/bin/alsaucm /lib/udev/alsaucm | 15:55 |
| GrueMaster | pgraner: How exactly did you test them? There are several issues. | 15:55 |
| rsalveti | ogra: yep, testing that | 15:56 |
| ogra | that should fix it | 15:56 |
| ericb2 | what is the last dpkg version correctly working ? | 15:56 |
| ogra | i dont really know if i can use absolute paths in udev rules | 15:56 |
| pgraner | GrueMaster, like I said following the tests on the iso tracker | 15:56 |
| ogra | but i guess that would be the right fix | 15:56 |
| ogra | i wouldnt have expected the helper to look in a hardcoded path | 15:56 |
| rsalveti | 85-usbmuxd.rules:ACTION=="remove", SUBSYSTEM=="usb", ENV{PRODUCT}=="5ac/12[9a][0-9a-f]/*", ENV{INTERFACE}=="255/*", RUN+="/usr/sbin/usbmuxd -x" | 15:57 |
| rsalveti | ogra: seems we can | 15:57 |
| pgraner | GrueMaster, http://testcases.qa.ubuntu.com/Install/ARM/PreinstalledImage | 15:57 |
| rsalveti | let me test with the full path | 15:57 |
| ogra | rsalveti, ah, didnt think of just grepping :P | 15:57 |
| pgraner | GrueMaster, if there are others let me know I have two Panda's here to test with | 15:57 |
| rsalveti | ogra: :-) | 15:58 |
| GrueMaster | pgraner: Bugs I had listed yesterday: bug 758486, bug 746023, bug 769235. | 15:58 |
| ubot2 | Launchpad bug 758486 in linux-ti-omap4 "omapdss DISPC error on Panda platform" [Medium,New] https://launchpad.net/bugs/758486 | 15:58 |
| ubot2 | Launchpad bug 746023 in alsa-utils "No sound on omap4" [High,Fix released] https://launchpad.net/bugs/746023 | 15:58 |
| ubot2 | Launchpad bug 769235 in desktop-webmail "Icons & app names identical for desktop-email & desktop-email-config" [Low,Confirmed] https://launchpad.net/bugs/769235 | 15:58 |
| ogra | 769235 has a fix already and will be an SRU | 15:59 |
| GrueMaster | Also, there is an installer bug yet to be filed, oem-config doesn't run the slideshow during installation. | 15:59 |
| ogra | which is cosmetic ... | 15:59 |
| pgraner | GrueMaster, I'm not seeing test cases for any of those issues, I only see the one I posted to you or am I missing something | 15:59 |
| GrueMaster | pgraner: Also, the image seed was changed last Thursday, so there are a lot of apps that haven't been tested. | 16:00 |
| GrueMaster | pgraner: I list all known bugs during release testing. | 16:00 |
| GrueMaster | Not just what the testcase says to look for. | 16:00 |
| pgraner | GrueMaster, do you have additional test cases? | 16:01 |
| GrueMaster | And since we had the seed change, I have been going over everything. | 16:01 |
| GrueMaster | No, I don't. | 16:01 |
| GrueMaster | I just run programs and see what works & what doesn't. | 16:01 |
| GrueMaster | The main testcase you refer to is specific to armel installation on omap3/omap4 platforms. The rest is just like testing the desktop apps. | 16:03 |
| ogra | pgraner, "use the image and all apps" is our main testcase | 16:03 |
| pgraner | ogra, GrueMaster, thanks | 16:05 |
| rsalveti | ogra: http://paste.ubuntu.com/599337/ worked when I triggered udev add events by hand | 16:06 |
| rsalveti | let me try rebooting | 16:06 |
| ogra | ah, cool | 16:06 |
| * GrueMaster goes to get massively caffeinated prior to starting testing. | 16:06 | |
| ogra | hmm, the slideshow is still isted for the netbook-live task | 16:07 |
| ogra | but it doesnt end up on the image | 16:07 |
| * ogra scratches head | 16:07 | |
| GrueMaster | It is in the manifest. | 16:10 |
| ogra | not in mine | 16:13 |
| GrueMaster | which package are you looking at? | 16:13 |
| ogra | i search for slideshow | 16:14 |
| GrueMaster | ubiquity-slideshow-ubuntu 40 | 16:14 |
| ogra | i dont have it in the current manifest | 16:15 |
| GrueMaster | Thats what fgrep ubiquity 20110426/natty-preinstalled-netbook-armel+omap4.manifest reports. | 16:15 |
| ogra | not sure what you look at | 16:15 |
| ogra | GAR ! | 16:15 |
| * ogra curses chromium | 16:15 | |
| ogra | just FYI ... the slideshow is definitely not installed on headless | 16:15 |
| rsalveti | ogra: ok, worked fine, will prepare the debdiff | 16:15 |
| * ogra grumbles | 16:15 | |
| GrueMaster | why would we have a slideshow in headless? :P | 16:16 |
| * GrueMaster offers coffee infusion to ogra | 16:16 | |
| ogra | GrueMaster, ask chromium why it gave me the cdimage page i just *didnt* want | 16:16 |
| ogra | it should guess the right one from my thoughts !!! | 16:16 |
| GrueMaster | Actually, I should ask why you don't just pull a local copy with wget when you zsync the image. | 16:17 |
| ogra | i rarely look at the manifest unless something is wrong | 16:17 |
| GrueMaster | I always look at the manifest because it is rare that something isn't wrong. | 16:18 |
| ogra | heh | 16:19 |
| GrueMaster | Besides, doing a diff between builds lets me know what changed and what to focus testing on. | 16:20 |
| GrueMaster | (especially when someone changes the image seed close to release). | 16:20 |
| GrueMaster | :P | 16:20 |
| ericb2 | is there somebody able to help me on the channel : got Ubuntu on BeagleBoard, and after I did a big apt-get dist-upgrade, I got x.org started, but no keyboard, nor mouse .... thanks in advance for any help or hint | 16:22 |
| ogra | rsalveti, hmm, just using the udev rule doesnt help for me | 16:23 |
| ericb2 | could be gdm broken, or one lib gdm depends. But can be something completel different | 16:23 |
| ogra | rsalveti, it only works after i ran alsaucm manually | 16:23 |
| rsalveti | ogra: hm, I just triggered udev by hand and then rebooted | 16:23 |
| rsalveti | now it's fine | 16:23 |
| ericb2 | ssh works, the network too and I can install, update | 16:23 |
| GrueMaster | ericb2: Try logging into a console and rerunning "sudo apt-get -f install" to make sure all packages are installed and updated. | 16:23 |
| ericb2 | GrueMaster: I did one minute ago | 16:24 |
| ogra | rsalveti, well, it might be that it needs the init once | 16:24 |
| ericb2 | GrueMaster: maybe some missing modules (or not loaded ) | 16:24 |
| ericb2 | something around usb | 16:25 |
| ogra | i definitely had no sound (and did about 5 reboots to make sure) without running the command once | 16:25 |
| rsalveti | ogra: ok, will fire up a new image with just this fix included to see how it goes | 16:25 |
| ogra | sigh, and GDM takes about a century on a class2 card | 16:26 |
| ericb2 | GrueMaster: the exact state is : my login name (eric) selected, and under the mouse cursor, I can read "Login as eric" | 16:26 |
| ericb2 | GrueMaster: but the cursor is stalled, like the keyboard and mouse | 16:26 |
| GrueMaster | ericb2: Do you see any messages in dmesg regarding keyboard & mouse? | 16:27 |
| ogra | rsalveti, i didnt run the Record verb and even after several reboots i have no input device (which is what record creates) | 16:28 |
| rsalveti | ogra: hm, weird | 16:29 |
| ogra | so your manual trigger has done the initialization i guess and the state file brings it back to the state it had on shutdown | 16:29 |
| ogra | i bet even removing the rule will still get you working sound after reboot | 16:30 |
| * ogra tries that | 16:30 | |
| rsalveti | ogra: the manual trigger should be the same trigger that happens while booting the device | 16:31 |
| ericb2 | GrueMaster: [ 4.711303] input: PS/2+USB Mouse as /devices/platform/ehci-omap.0/usb1/1-2/1-2.3/1-2.3:1.1/input/input2 | 16:31 |
| ericb2 | GrueMaster: looks ok at least for the mouse | 16:31 |
| ogra | rsalveti, well, lets see, i removed the rule and am just rebooting | 16:32 |
| ericb2 | GrueMaster: imho, this rather could be an x.org issue with keyboard + mouse or something | 16:32 |
| ogra | rsalveti, yup, sound is there even without the rules file, it is restored from the state file | 16:34 |
| ericb2 | GrueMaster: and /var/log/syslog says : http://paste.lisp.org/display/121603 | 16:34 |
| ogra | (and record is still missing as well) | 16:35 |
| ericb2 | GrueMaster: interesting : click either the mouse buttons or the keyboard do nothing (meant : nothing is detected) | 16:36 |
| rsalveti | ogra: janimo: why do we have mpurate=900 at the headless image for omap 3? | 16:37 |
| rsalveti | that's not compatible with old beagles, like C4 | 16:37 |
| ogra | rsalveti, does it do any harm ? | 16:37 |
| rsalveti | I can't even boot at my C4 | 16:37 |
| ogra | argh | 16:37 |
| GrueMaster | ericb2: Why do you have oem-config installed? | 16:37 |
| ericb2 | GrueMaster: let me check | 16:37 |
| ogra | rsalveti, i added that a while ago to get the XM to full speed | 16:38 |
| GrueMaster | rsalveti: My beagle C4 is working fine on headless. | 16:38 |
| ogra | i thought kernels not capable of using 900 would just ignore it | 16:38 |
| rsalveti | ogra: did you add this just at the headless image? | 16:38 |
| ogra | GrueMaster, check your cmdline | 16:38 |
| ogra | rsalveti, i dont think so | 16:38 |
| rsalveti | hm, let me check it again, but after removing it I'm able to boot | 16:39 |
| ogra | should be on all omap3 ones | 16:39 |
| * ogra checks the code | 16:39 | |
| ericb2 | GrueMaster: looks like it isn"t. Shall I install it ? | 16:39 |
| GrueMaster | It is in your pastebin, that is why I ask. | 16:39 |
| ogra | yeah, its in all omap3 images | 16:40 |
| rsalveti | ok, so it could be a different but, let me check it properly | 16:40 |
| rsalveti | *bug | 16:40 |
| GrueMaster | ogra: rsalveti http://paste.ubuntu.com/599358/ | 16:40 |
| ericb2 | GrueMaster: don't ask me why :) I'll try to install it, and I'll keep you informed about the progress. Thanks anyway :) | 16:41 |
| rsalveti | GrueMaster: is this a C4? | 16:41 |
| GrueMaster | ericb2: YOu don't need it. | 16:41 |
| ericb2 | GrueMaster: arf .. | 16:41 |
| GrueMaster | rsalveti: yes. | 16:41 |
| ericb2 | GrueMaster: so what shall I do ? uninstall it ? | 16:41 |
| ogra | GrueMaster, thats a C4 ? | 16:41 |
| GrueMaster | ogra: yes. | 16:41 |
| ogra | wow | 16:41 |
| ogra | i didnt know they could run that fast :) | 16:42 |
| rsalveti | hm, the only difference is that I'm also using it with the zippy expansion board | 16:42 |
| GrueMaster | It is the only beagleboard I have. | 16:42 |
| rsalveti | it probably can, but not that stable I'd say | 16:42 |
| ogra | oh, i thought you got an XM replacement at some point | 16:42 |
| GrueMaster | That is an XM. | 16:42 |
| ogra | you said C4 | 16:42 |
| GrueMaster | This is an original beagleboard. | 16:42 |
| ogra | C4 != XM | 16:42 |
| GrueMaster | I know that. | 16:42 |
| * ogra is confused now | 16:43 | |
| GrueMaster | I have both. But I don't have any other rev of beagleboard. | 16:43 |
| GrueMaster | BeagleXM is a different product. | 16:43 |
| ericb2 | GrueMaster: I got an xM if this can help you | 16:43 |
| ogra | you have a C4 and an XM but you only have one beagleboard ?!? | 16:43 |
| * ogra is even more confused now | 16:43 | |
| GrueMaster | ericb2: Sadly, I am a bit usy with last minute release work. | 16:43 |
| ericb2 | GrueMaster: I understand. thanks anyway :) | 16:44 |
| GrueMaster | ogra: The beagleboards include up to Rev C4. The beagleXM is an entirely different product. | 16:44 |
| ogra | GrueMaster, XM and Cx or Bx are the same product ... | 16:44 |
| GrueMaster | Similar to Panda. | 16:44 |
| ogra | different revision, but same product | 16:45 |
| GrueMaster | No, XM has a different rev CPU. | 16:45 |
| phh | beagleXM similar to panda ? | 16:45 |
| phh | wut ? | 16:45 |
| GrueMaster | Check your specs. | 16:45 |
| ogra | still, the product is called beagleboard | 16:45 |
| GrueMaster | sigh. | 16:45 |
| GrueMaster | No, it is called beagleXM. | 16:45 |
| ogra | ask TI why they prevented us from finding the difference in cpuinfo :) | 16:45 |
| phh | BeagleBoard-xM delivers extra ARM ® Cortex TM -A8 MHz now at 1 GHz and extra memory with 512MB of low-power DDR RAM, | 16:45 |
| phh | I stopped at the first line. | 16:45 |
| phh | it's not like panda. | 16:45 |
| GrueMaster | Ok, there is too much noise in the channel. | 16:46 |
| * GrueMaster goes back to testing. | 16:46 | |
| ogra | GrueMaster, so does your C4 behave stable then at 900MHz ? | 16:47 |
| GrueMaster | Seems to, at least for headless. | 16:47 |
| * ogra wouldnt have expected it to even scale to 900) | 16:47 | |
| ppisati | do anyone know where i can find a gdb for arm? i.e. arm-linux-gnueabi-gdb | 16:47 |
| GrueMaster | It was running all night untill I reimaged it this morning. | 16:47 |
| ppisati | or do i have to roll my own? | 16:47 |
| ogra | apt-get install gdb ? | 16:48 |
| ppisati | i need ar arm one | 16:48 |
| ericb2 | GrueMaster: apt-get autoremove + disconnect/reconnect all usb stuff fixed the issue | 16:48 |
| ogra | (which should actually be installed) | 16:48 |
| phh | ppisati: apt-get install gdb should install an arm one :p | 16:48 |
| phh | oh well you need an arm system. | 16:48 |
| ogra | ppisati, that gets you an arm one on the right platform ;) | 16:48 |
| ppisati | :) | 16:48 |
| ppisati | ok | 16:48 |
| phh | everyone here got an arm as main system no ? :p | 16:48 |
| ppisati | i'm chasing the kexec stuff | 16:48 |
| ppisati | it hangs after relocation | 16:48 |
| ppisati | so i need jtag now | 16:48 |
| ppisati | i managed to setup openocd on panda | 16:49 |
| ppisati | so fa so good | 16:49 |
| ppisati | but now i need a proper gdb | 16:49 |
| ogra | phh, one at least ... some of us are exaggerating though like GrueMaster ... he has many :P | 16:49 |
| ogra | one arm for every hand !!! | 16:49 |
| ppisati | one arm for my reign! | 16:49 |
| phh | one ac100 for every hand here yeah | 16:49 |
| ppisati | ok, i tri the codesourcery toolchain | 16:49 |
| ogra | phh, oh, you have a second one ? | 16:49 |
| * ogra is envious | 16:49 | |
| phh | ogra: how the fuck do you test on your main system ? :p | 16:50 |
| ogra | by using a different one while testing | 16:50 |
| ogra | thats why i do it so rarely :) | 16:50 |
| ogra | over easter i switched completely to 2.6.37 though | 16:50 |
| phh | i was fed up of doing tests only once a month, so I got another ac100 :p | 16:51 |
| ogra | that makes everything easier | 16:51 |
| phh | it's still the cheap french ones though | 16:51 |
| phh | no 3G no BT | 16:51 |
| GrueMaster | We can't even get them in the States. | 16:51 |
| ogra | well, the shop i got mine from still wants 299€ for it ... and i havent gotten around ordering one from the cheaper shops online | 16:51 |
| phh | hihihi | 16:51 |
| ogra | but i plan to ... and a mPCI SD | 16:52 |
| ogra | *SSD | 16:52 |
| phh | GrueMaster: you can't have everything | 16:52 |
| phh | ogra: I only plan to change my panel | 16:52 |
| phh | with planning being something really long term :D | 16:52 |
| GrueMaster | Sure we can, as long as it runs Windows 7. | 16:52 |
| ogra | that possibly too | 16:52 |
| ogra | first of all i want a usable disk | 16:52 |
| phh | ogra: and your '...' is soldering some more ram ? | 16:53 |
| phh | (just saying.) | 16:53 |
| ogra | HD screen is next on my list :) | 16:53 |
| phh | well, the panel has two goals: | 16:53 |
| phh | get an HD screen | 16:53 |
| ogra | more ram would be awesome but i'm not good at SMD soldering | 16:53 |
| phh | get a new screen | 16:53 |
| phh | not that I killed my current panel | 16:53 |
| phh | but err, yes. | 16:53 |
| ogra | heh | 16:53 |
| ericb2 | For the record : disconnect/reconnect all USB devices , is mandatory when you log out / login | 16:54 |
| ogra | well, $90 isnt bad for a replacement that also does HD | 16:54 |
| phh | yeah | 16:54 |
| ericb2 | now it works ... | 16:54 |
| GrueMaster | I have 8G of DDR2 sitting here on my desk waiting for testing to finish so I can upgrade my desktop from the current 3G. | 16:54 |
| ogra | but thats lame ... doesnt involve soldering at all | 16:54 |
| phh | :) | 16:55 |
| ogra | its more exciting if you can break it :) | 16:55 |
| GrueMaster | ericb2: Sounds like something else is wrong. The natty images work fine here on beagle/XM/Panda w/o hotplugging the USB. | 16:55 |
| GrueMaster | Whoa, this is interesting. mpurate=900 is on both headless & netbook images. My beagleboard (C4) is showing 893 bogomips, and my XM is showing 782. | 16:58 |
| GrueMaster | Oh, and cpuinfo does show a difference between the two. C4: CPU Variant 0x1 XM: CPU Variant 0x3 | 16:59 |
| ppisati | code sourcery toolchain had one, and it works | 17:02 |
| phh | ppisati: good to know | 17:02 |
| ogra | ppisati, i'm pretty sure we have a packaged one in the cross toolchain, hrw would know where though | 17:05 |
| ogra | GrueMaster, yeah, but they explicitly have not changed the Hardware entry between the revisions | 17:06 |
| ogra | btw, all apps that were added with the recent meta change have been working here | 17:06 |
| ogra | i gave all of them some small function tests | 17:07 |
| GrueMaster | Yes, I tested them here over the weekend. | 17:07 |
| ogra | great | 17:07 |
| * ogra would like to know if mpurate can actually cause any instabilities ... honestly i would prefer to go with 900MHz on the XM (instead of 600) and rather say we stop supporting C4 officially without image modification | 17:08 | |
| GrueMaster | I would say leave headless w/o mpurate and mark the netbook for XM only. | 17:09 |
| ogra | GrueMaster, i think bogomips are more than just cpu speed, to get the actual MHz value look in dmesg | 17:09 |
| ogra | ugh, adding a flavour specific hack additionally? | 17:09 |
| ogra | i dont think that flies two days before release ... its either all omap3 image on or all of them off | 17:10 |
| GrueMaster | Well, you could always add a hack in jasper to detect the cpu variant. | 17:10 |
| GrueMaster | I was thinking of Oneiric. | 17:10 |
| ogra | sure, and i can grow grass on my knee if i dont move for long enough :) | 17:11 |
| GrueMaster | For now, we release note it. | 17:11 |
| ogra | yeah | 17:11 |
| ogra | well, i sense a respin anyway | 17:11 |
| GrueMaster | why? | 17:11 |
| ogra | so we still have a chance to drop it ... but its an all or nothing thing in natty | 17:11 |
| ogra | GrueMaster, if we find an easy fix for slideshow and sound, wouldnt you want them in ? | 17:11 |
| GrueMaster | slideshow is not a showstopper. Audio can easily be fixed with an update. | 17:12 |
| * ogra doubts audio is fixable easily outside of jasper | 17:12 | |
| GrueMaster | I'd rather not respin unless it is critical or the entire mess is respun (all arches). | 17:12 |
| GrueMaster | Can't the udev rule be updated with the correct path to alsaucm? | 17:13 |
| ogra | sure, but it changes nothing | 17:13 |
| rsalveti | ogra: GrueMaster: up, can't boot: http://paste.ubuntu.com/599371/ | 17:13 |
| ogra | at least here | 17:13 |
| rsalveti | finished the headless installation without the mpurate argument, and then added it again by hand | 17:14 |
| rsalveti | at next reboot the kernel get stuck | 17:14 |
| ogra | weird, i dont even see it setting the mpurate | 17:14 |
| ogra | [ 0.000000] Clocking rate (Crystal/Core/MPU): 26.0/332/500 MHz | 17:15 |
| ogra | you should actually have another line where it tells you that it re-sets to a different value | 17:15 |
| rsalveti | let me try with different values | 17:15 |
| rsalveti | that's all I have | 17:15 |
| ogra | yeah, i belive you :) | 17:16 |
| ogra | i just know it looks quite different on my XM if it actually changes the value during boot | 17:16 |
| GrueMaster | Out of curiosity, why is the uboot delay on omap3 10 seconds, and only 3 seconds on panda? | 17:17 |
| ogra | GrueMaster, blame jcrigby :) | 17:17 |
| rsalveti | GrueMaster: there's a patch to fix this upstream already | 17:18 |
| ogra | its a compile time default | 17:18 |
| rsalveti | yeah | 17:18 |
| rsalveti | ogra: mpurate=720 boots fine | 17:18 |
| ogra | grmbl | 17:18 |
| rsalveti | ogra: this is what should be the next line: | 17:18 |
| rsalveti | [ 0.527679] Switched to new clocking rate (Crystal/Core/MPU): 26.0/332/720 MHz | 17:18 |
| ogra | yeah | 17:18 |
| rsalveti | so it seems it got stuck while trying to set the clock | 17:18 |
| ogra | thats what i have on XM (with different values) | 17:18 |
| ogra | damned | 17:19 |
| rsalveti | OMAP3530-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 720 mHz | 17:19 |
| rsalveti | from u-boot | 17:19 |
| rsalveti | beagle c4 should be stable with 720 | 17:19 |
| ogra | NCommander, janimo, rsalveti, GrueMaster, persia ... votes ... do we respin or not ? | 17:19 |
| rsalveti | ogra: I think we have a kernel patch that changes that at the kernel side, let me try to look for it first | 17:20 |
| ogra | doesnt help much if you cant install at all | 17:20 |
| GrueMaster | I vote no unless it is required. Make the changes so they are there if we do need to respin. | 17:20 |
| rsalveti | yeah, we would need a respin or at least a big note at the release notes | 17:20 |
| * ogra waits for more votes ... thats +1 vs -1 atm | 17:22 | |
| GrueMaster | List the fixes that would be picked up in a respin. | 17:23 |
| ogra | GrueMaster, well, we dont have any fix for sound nor for slideshow atm | 17:23 |
| ogra | so mpurate would be the only one if we respun right now | 17:24 |
| GrueMaster | If you don't have a fix for them, respin currently does nothing. | 17:24 |
| ogra | it makes old beagle work | 17:24 |
| ogra | and new beagle slower | 17:24 |
| GrueMaster | I haven';t seen a problem on my old beagle yet. | 17:24 |
| GrueMaster | I only run it headless though. | 17:24 |
| rsalveti | that probably depends on the hardware | 17:24 |
| rsalveti | it's an overclock | 17:25 |
| GrueMaster | right. | 17:25 |
| ogra | right | 17:25 |
| ogra | for C series its an overclock | 17:25 |
| GrueMaster | Can be release noted. | 17:25 |
| rsalveti | I can't get mine to boot even with 800 | 17:25 |
| ogra | XM is actually supposed to run up to 1GHz | 17:25 |
| rsalveti | let me check if our u-boot is setting the mpurate value | 17:25 |
| ogra | my XM runs at 600MHz if you dont set it on cmdline | 17:26 |
| GrueMaster | ubuntu@beagle:~$ dmesg|fgrep Crystal | 17:26 |
| GrueMaster | [ 0.000000] Clocking rate (Crystal/Core/MPU): 26.0/332/500 MHz | 17:26 |
| GrueMaster | [ 0.357208] Switched to new clocking rate (Crystal/Core/MPU): 26.0/332/900 MHz | 17:26 |
| ogra | so i would say u-boot doesnt | 17:26 |
| GrueMaster | From my beagle C4 | 17:26 |
| rsalveti | u-boot just sets the mpurate var | 17:26 |
| ogra | yeah | 17:26 |
| rsalveti | you still need something like mpurate=$mpurate | 17:26 |
| rsalveti | something like that | 17:26 |
| ogra | but no value | 17:26 |
| rsalveti | yup, seems to be setting it up | 17:27 |
| ogra | davidm, thoughts ? | 17:27 |
| rsalveti | let me see if I can get that with our kernel | 17:27 |
| rsalveti | if we can get it to get the right value from u-boot, then I believe we should make a respin | 17:28 |
| rsalveti | because it'll always be the right clock | 17:28 |
| rsalveti | ogra: http://paste.ubuntu.com/599374/ | 17:28 |
| rsalveti | 600 for beagle a/b | 17:28 |
| rsalveti | 720 for c4 | 17:28 |
| rsalveti | and 1000 for xM | 17:28 |
| ogra | 1000 doesnt work | 17:28 |
| rsalveti | why not? | 17:29 |
| ogra | 900 is the highest i could get (i tested the option before adding it ) | 17:29 |
| rsalveti | hm | 17:29 |
| ogra | (indeed i didnt test on Cx) | 17:29 |
| phh | the bootloader scripts i've seen states that there is a bug in kernel that prevents it from going to 1G | 17:29 |
| ogra | i think there is some linux-omap patch that enhances it to 1000 | 17:29 |
| ogra | which we dont have in our mainline kernel | 17:30 |
| ogra | some koen magic :) | 17:30 |
| rsalveti | argh, we can include it but then it'll probably not boot with 1000 | 17:30 |
| rsalveti | ogra: yup, just need to use mpurate=${mpurate} | 17:33 |
| rsalveti | and you'll get the value from u-boot | 17:33 |
| rsalveti | with that I'm getting 720 at my C4 | 17:33 |
| ogra | hmm, could somone try with XM? | 17:33 |
| rsalveti | sure, a sec | 17:34 |
| ogra | i clearly saw probs with 1000 here | 17:34 |
| davidm | ogra, rsalveti is the issue some boards currently can't boot but with the single change mpurate=${mpurate} all boards will boot? | 17:34 |
| ogra | davidm, not sure yet, as we stand atm we can only boot XM reliably and Cx with luck | 17:35 |
| rsalveti | with current setup we may have problems with old beagles, like beagle a/b and c | 17:35 |
| rsalveti | but it should work better for xM | 17:35 |
| rsalveti | using mpurate=${mpurate} will fix for beagle a/b/c but will also set 1000 for xM | 17:35 |
| rsalveti | and it seems our kernel doesn't support it | 17:35 |
| rsalveti | testing right now | 17:35 |
| ogra | davidm, dropping the aparm completely will drop XMs to 600MHz but make Cx and Bx work | 17:35 |
| ogra | davidm, setting mpurate=${mpurate} will set a value that didnt work on XM during my tsets | 17:36 |
| phh | you have no control on bootloader's config file ? | 17:36 |
| ogra | phh, we have a release on thu ... | 17:36 |
| rsalveti | ogra: I prefer doing a respin without any mpurate, putting on the release notes how to use the mpurate from u-boot while we try to fix the kernel bug at the xM | 17:37 |
| ogra | the archive is frozen ... uploads and respins of images cost a lot | 17:37 |
| GrueMaster | Just rebooted with mpurate=${mpurate} on my XM. Logging in now. | 17:37 |
| phh | ogra: ah | 17:37 |
| phh | ogra: the way i've seen is changing boot script, if beagleboardxm then mpurate=600 | 17:37 |
| ogra | GrueMaster, check dmesg :) | 17:37 |
| GrueMaster | Failed to set it. | 17:37 |
| ogra | phh, not that easy, but yeah. thats the proper fix | 17:37 |
| ogra | GrueMaster, yeah., thats what i had too | 17:38 |
| phh | no the proper fix is to fix your kernel ;) | 17:38 |
| rsalveti | ogra: yup, failed to set it | 17:38 |
| rsalveti | [ 0.000000] Clocking rate (Crystal/Core/MPU): 26.0/332/600 MHz | 17:38 |
| rsalveti | but that's fine! | 17:38 |
| rsalveti | once the kernel is fixed it'll work | 17:38 |
| rsalveti | and the kernel fix can be an SRU | 17:38 |
| ogra | no, thats to slow to use netbook | 17:38 |
| GrueMaster | Along with the DVI EDID code. | 17:38 |
| rsalveti | well, we don't have other choices | 17:38 |
| ogra | we do | 17:38 |
| ogra | go with GrueMaster's suggestion and dont do a respin | 17:39 |
| GrueMaster | The other choice is to document. | 17:39 |
| ogra | right | 17:39 |
| rsalveti | but that will break on older beagles | 17:39 |
| GrueMaster | It is easy enough to edit the boot.script. | 17:39 |
| rsalveti | and even after the kernel fix is included, it'll still run with 900 and not 1000 | 17:39 |
| rsalveti | the user would then need to edit again by hand | 17:39 |
| phh | GrueMaster: easy enough but ubuntu can't do it itself :D | 17:39 |
| ogra | yes | 17:39 |
| rsalveti | I don't like this solution, but not my call | 17:40 |
| GrueMaster | phh: We have a wiki documenting how to change it. | 17:40 |
| davidm | OK let me get a complete picture,this is beagle only | 17:40 |
| ogra | phh, thats rather a safety thing ... we could but we usually dont touch boot stuff after installation | 17:40 |
| rsalveti | davidm: yes | 17:40 |
| rsalveti | davidm: with mpurate=900 (current setup) we may break old beagles, like beagle a/b/c | 17:40 |
| davidm | if we do nothing at this point early beagles won't boot but current beagle anb beagle xm will work? | 17:40 |
| rsalveti | but have xM working with 900mhz | 17:40 |
| ogra | davidm, current = XM, but yes | 17:41 |
| davidm | How hard is iti to mount the image and add something to make it work from release notes? | 17:41 |
| GrueMaster | davidm: beagle C4 is hit or miss. XM works. | 17:41 |
| rsalveti | with mpurate=${mpurate} we'll have the best mhz available for beagle a/b/c, but xM will behave with 600mhz | 17:41 |
| rsalveti | because it fails to set 1000 | 17:41 |
| ogra | davidm, as GrueMaster said, not hard | 17:41 |
| phh | GrueMaster: if it boots on C4 you might damage it no, | 17:41 |
| phh | ? | 17:41 |
| ogra | davidm, but our screwup is permanent as rsalveti says | 17:41 |
| rsalveti | GrueMaster: and probably always miss with B or A | 17:41 |
| rsalveti | as it can do at most 600 | 17:42 |
| rsalveti | 900 is too high | 17:42 |
| davidm | What is the process for the community? Can it be made the default sothey don't have to edit u-boot each boot? | 17:42 |
| rsalveti | let u-boot to decide | 17:42 |
| ogra | davidm, while mpurate=${mpurate} could fix the issue through a kernel and bootloader update | 17:42 |
| rsalveti | that's why mpurate=${mpurate} | 17:42 |
| davidm | Whoops, we might damage boards with the current setting? That is different | 17:42 |
| rsalveti | the mpurate is a variable that's set by u-boot | 17:42 |
| rsalveti | davidm: yeah | 17:42 |
| rsalveti | it's overclock | 17:42 |
| ogra | davidm, we dont damage them but we will shorten their lifecycle | 17:42 |
| phh | there is a dynamic voltage thing ? | 17:43 |
| rsalveti | I don't mind having xM running with 600mhz until the first kernel upgrade | 17:43 |
| davidm | can this be fixed with an SRU post release? | 17:43 |
| ogra | davidm, no, because we dont re-write boot.scr | 17:44 |
| GrueMaster | davidm: If we respin with mpurate=${mpurate}, then XM can go full speed with a kernel SRU. | 17:44 |
| rsalveti | that's why I believe we should release with mpurate=${mpurate}, even with xM running only at 600mhz | 17:44 |
| ogra | mpurate=${mpurate} could be fixed by a bootloader and kernel update | 17:44 |
| davidm | to me shortening the life of a board is damage | 17:44 |
| rsalveti | the kernel fix can go with SRU | 17:44 |
| ogra | davidm, we'Re the only distro running the XM at 600MHz afaik | 17:44 |
| rsalveti | then at the first kernel update the user will have his xM running with 1000 | 17:44 |
| rsalveti | even better than 900 | 17:44 |
| rsalveti | well, it's a bug | 17:45 |
| ogra | yes | 17:45 |
| rsalveti | and we'll be the only distro that the image doesn't even boot on older beagles :-( | 17:45 |
| rsalveti | even now that we have the headless option | 17:45 |
| davidm | so fewer people are effected if we switch to mpurate=${mpurate}, and there is no chance of damage to any boards? | 17:45 |
| rsalveti | more and more people will try installing ubuntu on older beagles | 17:45 |
| rsalveti | davidm: yes | 17:46 |
| GrueMaster | On one side, we release note (similar to BeagleXM DVI issue with maverick requiring user intervention before booting), on the other, we respin and SRU. | 17:46 |
| davidm | OK I'm going to talk to pgraner and skaet | 17:46 |
| * ogra just pinged cjwatson | 17:46 | |
| ogra | to see if we can respin only omap3 | 17:46 |
| ogra | which i think we cant | 17:47 |
| ogra | which in turn means re-testing *all* images | 17:47 |
| GrueMaster | which is why I vote no, but that's just me. | 17:47 |
| ogra | we can respin just omap3 | 17:48 |
| rsalveti | just omap3 seems fine | 17:48 |
| rsalveti | I think this is our best option | 17:48 |
| davidm | Why is this just being found now? | 17:48 |
| GrueMaster | Because all of my systems work fine as is. | 17:48 |
| ogra | same here | 17:49 |
| rsalveti | because I decided to try an older beagle here, to do another work, not related with testing, and boom | 17:49 |
| ogra | and we discussed the setting a while ago with the conclusion that it wouldnt do harm | 17:49 |
| rsalveti | didn't boot | 17:49 |
| ogra | i had it on my TODO list from end of maverick still | 17:49 |
| GrueMaster | ogra: Will this require a jasper fix? | 17:50 |
| ogra | GrueMaster, no, jasper doesnt touch the cmdline (only preseeding options and root= are parsed) | 17:50 |
| GrueMaster | ok | 17:50 |
| ogra | jasper could carry a board specific fix (looking for cpu revision) but apparently we dont need that with ricardos solution | 17:51 |
| GrueMaster | I'm going to test here with a fresh image to make sure it gets passed through anyways. | 17:51 |
| ogra | the sound thing really annoys me :( | 17:52 |
| ogra | having put in weeks of work over two releases and still not having it working proper is depressing :( | 17:52 |
| rsalveti | yeah, the sound issue doesn't help | 17:53 |
| GrueMaster | Well, we still don't have sound on omap3 either, and this fix won't even help it even if I were able to get working ucm scripts for omap3. | 17:56 |
| ogra_ | who cares about omap3 (as long as we dont break the HW) | 17:57 |
| rsalveti | sound is kind of a minor bug, annoying but not critical | 17:59 |
| davidm | how long does it take to resping OMAP 3 | 17:59 |
| GrueMaster | All I am saying is that omap3 audio and omap4 audio are pretty much at the same state at this point. kernel-side they both work. | 17:59 |
| * NCommander is alive now and reading backscroll | 17:59 | |
| ogra_ | davidm, 90min for netbook, about 30 for headless | 17:59 |
| GrueMaster | retesting is minimal as the changes are small. | 18:00 |
| ogra_ | indeed, its just a boot option | 18:00 |
| ogra_ | if it boots everywhere we're fine | 18:00 |
| rsalveti | yup | 18:00 |
| phh | ah release is in two days | 18:00 |
| phh | no need to hurry | 18:00 |
| ogra_ | thats what i said above, yeah :) | 18:00 |
| rsalveti | will have a look for the xM kernel fix | 18:00 |
| GrueMaster | The headache is the respin/post/download process. | 18:00 |
| ogra_ | rsalveti, ask koen, i think he wrote it | 18:01 |
| rsalveti | should probably be at l-o already, just need to find the patch | 18:01 |
| GrueMaster | The question is if we do a respin, do we try to get other fixes in as well? The ubiquity slideshow issue would affect all platforms (even though they already work). | 18:02 |
| davidm | OK call made | 18:03 |
| GrueMaster | Possibly the same with audio. | 18:03 |
| davidm | please fix package upload we are going to resping | 18:03 |
| ogra_ | no package | 18:03 |
| ogra_ | :) | 18:03 |
| ogra_ | its a one line buildsystem change | 18:03 |
| davidm | rsalveti, ogra OK make change and let colin and skaet know its done | 18:03 |
| ogra_ | on it already | 18:04 |
| rsalveti | davidm: cool | 18:04 |
| skaet | davidm, which is the version of u-boot we are keeping an eye on? | 18:05 |
| davidm | skaet, it's not even a package change it's a build system change | 18:05 |
| davidm | no archive change needed | 18:05 |
| davidm | one line change to adding mpurate=${mpurate} | 18:06 |
| cjwatson | where's that? | 18:06 |
| cjwatson | (sorry, lacking scrollback) | 18:06 |
| rsalveti | ogra: should know where to change | 18:07 |
| cjwatson | ogra should but I want to know for planning purposes | 18:07 |
| davidm | ogra, rsalveti please explain where change is going to cjwatson | 18:07 |
| cjwatson | is this in post-boot-natty-armel+omap? | 18:07 |
| davidm | GrueMaster, no changes except this boot fix change NONE | 18:08 |
| GrueMaster | davidm: Understood. Testing fix here to ensure it doesn't break something else. | 18:08 |
| cjwatson | ok, so it's just in debian-cd on antimony - so ogra can commit that, we can deploy it on antimony, and then we don't even need to rebuild the livefs | 18:10 |
| cjwatson | so it's actually a really quick rebuild | 18:10 |
| cjwatson | headless or netbook or both? | 18:10 |
| rsalveti | both | 18:10 |
| cjwatson | which subarch? | 18:10 |
| ogra_ | cjwatson, oh, you are here | 18:11 |
| * ogra_ ignored the channel | 18:11 | |
| apw | davidm, i think i heard there would be a kernel change needed after for performance on some cards ... whats the bug number, and do we already have the patch | 18:11 |
| ogra_ | cjwatson, fis is already in debian-cd in the tools/boot/natty/post-boot-armel+omap script | 18:11 |
| ogra_ | *fix | 18:11 |
| rsalveti | apw: I'm looking for the patch right now | 18:12 |
| rsalveti | will also report the bug soon | 18:12 |
| davidm | cjwatson, thanks | 18:12 |
| davidm | apw, rsalveti has point and will get info into bug and to you on this one | 18:13 |
| apw | rsalveti, can you sub me to the bug, and drop the number in privmsg | 18:16 |
| rsalveti | apw: sure | 18:16 |
| apw | rsalveti, cool, so i can make sure any change is in the earliest sru | 18:17 |
| cjwatson | I'm just waiting for existing builds to complete | 18:20 |
| cjwatson | shouldn't be too long | 18:20 |
| rsalveti | apw: ogra: davidm: got the patch, will try it and open the bug when I'm sure it works | 18:25 |
| ogra_ | yeah, no hurry | 18:26 |
| ogra_ | updating the bootloaders is complex anyway | 18:26 |
| ogra_ | well, not complex, but requires manual cmdline action | 18:26 |
| rsalveti | ok, out for lunch while the kernel is building | 18:27 |
| cjwatson | headless/netbook omap3 rebuilding | 18:29 |
| ogra_ | thx | 18:30 |
| GrueMaster | ogra: Not sure that it matters, but if we have mpurate=${mpurate} in the initial boot.scr, jasper will only use the u-boot translated value for subsequent boots (i.e. mpurate=1000 on my XM after oem-config runs). | 18:57 |
| ogra_ | GrueMaster, hmm, right, that needs a note (and an oneiric bug i guess) | 18:58 |
| GrueMaster | This should be release noted in case someone wants to use the same SD for beagle C4 (or earlier) & XM. | 18:58 |
| ogra_ | yep | 18:59 |
| ogra_ | and jasper needs special casing for that var | 18:59 |
| GrueMaster | Another issue for Oneiric: wifi in omap4 headless. Not sure what to file a bug against, but we need wireless-tools and possibly some other package for this to work ootb. | 19:01 |
| cjwatson | ubuntu headless omap3 posted | 19:02 |
| GrueMaster | cjwatson: Thanks. | 19:02 |
| ogra_ | awesome ! | 19:02 |
| GrueMaster | ogra_: Will the omap3 image build change also affect kubuntu-[desktop|mobile]? If so, they also need to be respun. | 19:58 |
| ogra_ | bah, sigh, indeed | 19:59 |
| GrueMaster | omap3 headless retested on beagle. XM is finishing install. Looks good. | 20:16 |
| ogra_ | great | 20:17 |
| GrueMaster | I updated bug 746023 as the udev rule is still failing. I'm looking into how to fix it completely now. | 20:23 |
| ubot2 | Launchpad bug 746023 in alsa-utils "No sound on omap4" [High,In progress] https://launchpad.net/bugs/746023 | 20:23 |
| ogra_ | i fear we need to call alsaucm once from the alsa-utils postinst | 20:24 |
| GrueMaster | Shouldn't need to. | 20:25 |
| ericb2 | do you need testers ? | 20:25 |
| ogra_ | with UCM you need to init the default verbs you want once | 20:25 |
| GrueMaster | ericb2: do you have a pandaboard? | 20:25 |
| ericb2 | GrueMaster: no, sorry. only an xM | 20:25 |
| ericb2 | GrueMaster: I think I'll buy one soon, but I think I'll have to wait some weeks anyway | 20:26 |
| GrueMaster | ericb2: If you can come up with a ucm config for XM, great. We can add it to natty-updates. | 20:26 |
| ericb2 | GrueMaster: sorry, what means ucm ? | 20:26 |
| ericb2 | GrueMaster: I'm not native speaker, and sometimes I need more explanations, sorru | 20:27 |
| GrueMaster | ogra_: calling alsaucm the way we do in the current udev rule looks right, but needed a path. Now we just need to figure out how to fire it. | 20:27 |
| ericb2 | sorry | 20:27 |
| GrueMaster | ericb2: ucm == Use Case Manager, part of the recent alsa release. | 20:27 |
| ericb2 | GrueMaster: ah, ok. no problem for me to make some tests | 20:27 |
| ogra_ | GrueMaster, well, when i tested that was a few kernels ago, not sure the device name udev sees is still the same | 20:28 |
| ogra_ | though rsalveti's lost looked like the rule fired just fine | 20:29 |
| GrueMaster | ogra_: run "sudo udevadm test /class/sound/card0" and it will show you. | 20:29 |
| ogra_ | s/lost/logs/ | 20:29 |
| ericb2 | GrueMaster: what is the process ? | 20:29 |
| rsalveti | ogra: GrueMaster: it worked when I fired the udev events again | 20:29 |
| GrueMaster | Well, I updated the alsa-ucm.rules to have the proper path prior to starting a fresh image and it still failed. | 20:30 |
| rsalveti | and I could also see it at the logs | 20:30 |
| rsalveti | just didn't check yet if that's also happening during boot | 20:30 |
| ogra_ | rsalveti, yeah, it didnt fire automatically here though | 20:30 |
| rsalveti | need to find a way to debug udev | 20:30 |
| GrueMaster | I'll try some more things here. | 20:30 |
| rsalveti | see why it's not calling the rule | 20:30 |
| GrueMaster | udevadm has a lot of debug capabilities. | 20:30 |
| rsalveti | I know, I just want to trace the events that udev gets while booting | 20:31 |
| ogra_ | rsalveti, hmm, try to remove the conditions at the top and bottom and only leave the ID lines in place | 20:31 |
| rsalveti | when I trigger them by hand it works | 20:31 |
| ogra_ | might be they prevent execution somehow | 20:31 |
| GrueMaster | udevadm info --attribute-walk --name=/dev/snd/by-path/platform-soc-audio.0 will tell you. | 20:32 |
| GrueMaster | at any rate, I need lunch. biab. | 20:35 |
| rsalveti | ogra: put "exec udevd --daemon --debug >> /var/run/udev.log 2>&1" at your /etc/init/udev.conf | 20:49 |
| rsalveti | it should give you all events | 20:49 |
| rsalveti | and here it's still running the alsaucm just fine | 20:49 |
| ogra_ | how do you get back to a virgin state ? or do you test it with a fresh image every time ? | 20:51 |
| rsalveti | ogra: I'm currently just checking if the udev rule is being called | 20:52 |
| rsalveti | and it should be called on everyboot | 20:52 |
| ogra_ | hmm | 20:52 |
| ogra_ | yes | 20:52 |
| rsalveti | as it doesn't depend on the alsaucm state | 20:52 |
| ogra_ | every time an SDP4430 device shows up in the sound stack it should be called | 20:52 |
| ogra_ | but that doesnt happen for tobin or me | 20:53 |
| rsalveti | ogra: ok, at this boot I'm getting only the hdmi output | 20:54 |
| rsalveti | even after calling alsaucm by hand | 20:54 |
| ogra_ | killall pulseaudio | 20:54 |
| ogra_ | and re-open the prefs | 20:54 |
| ogra_ | pulse doesnt update properly apparently | 20:54 |
| rsalveti | yup, it's there now | 20:55 |
| rsalveti | let me reboot | 20:55 |
| rsalveti | maybe a racing condition? | 20:56 |
| ogra_ | hmm | 20:56 |
| ogra_ | i think the SDP4430 is builtin, the trigger should happen very early | 20:56 |
| rsalveti | ogra_: now it's working all the time =\ | 21:07 |
| rsalveti | it was working before, stopped working and now it got back to work after I called alsaucm by hand | 21:08 |
| ogra_ | you run from SD, right ? | 21:08 |
| rsalveti | and restarted pulseaudio | 21:08 |
| rsalveti | yes | 21:08 |
| ogra_ | not from USB disk :) | 21:08 |
| rsalveti | I have one with usb disk and another with sd :-) | 21:08 |
| ogra_ | you only need the alsaucm call once | 21:08 |
| ogra_ | we dont need it on every boot | 21:09 |
| rsalveti | what happens after you call it? | 21:09 |
| rsalveti | can we restore the state to where it was before calling it? | 21:09 |
| ogra_ | its is the same f*cking problem we had last release | 21:09 |
| rsalveti | let me strace it | 21:09 |
| ogra_ | i dont know, i guess removing the upstart job that stores state and removing the state file might help | 21:09 |
| ogra_ | if you called alsaucm and alsa stores the state it will restore that state so you get working sound all the time | 21:10 |
| ogra_ | i have that here | 21:10 |
| rsalveti | yeah, I tried removing the alsa state file but still worked | 21:10 |
| ogra_ | the alsaucm call should actually not come from a udev rule | 21:11 |
| rsalveti | even after a cold reboot | 21:11 |
| ogra_ | well, did you remove the upstart job too ? | 21:11 |
| rsalveti | yeah, makes sense | 21:11 |
| rsalveti | not yet, thought that would be part of the package | 21:11 |
| ogra_ | or only the state file | 21:11 |
| rsalveti | and not something generated while calling the command | 21:11 |
| rsalveti | let me check | 21:11 |
| ogra_ | sure it is, but it will store your state in any case | 21:11 |
| rsalveti | ok, alsa-store is not going to be called anymore | 21:13 |
| rsalveti | and removed /var/lib/alsa/asound.state | 21:13 |
| ogra_ | k | 21:13 |
| ogra_ | so reboot :) | 21:13 |
| rsalveti | alsaucm is just loading the configuration from /usr | 21:14 |
| rsalveti | not writing anything | 21:14 |
| rsalveti | so this state file should be all | 21:14 |
| ogra_ | yep | 21:14 |
| rsalveti | ok, rebooted, with the working udev rule, without alsa.state file | 21:16 |
| rsalveti | and sound is working | 21:16 |
| ogra_ | now remove the udev rule and reboot :) | 21:16 |
| ogra_ | to see if its really gone | 21:16 |
| rsalveti | yup, doing that right now | 21:17 |
| ogra_ | the prob with the udev rule is that it will always set to hifi on boot, regardless what use case you selected | 21:17 |
| rsalveti | got it | 21:18 |
| rsalveti | ogra: ok, finally, [ 28.309753] asoc: no valid backend routes for PCM: SDP4430 Media | 21:19 |
| ogra_ | yeah | 21:19 |
| rsalveti | so the udev rules was actually working just fine | 21:19 |
| ogra_ | weird that it didnt for GrueMaster and me | 21:19 |
| rsalveti | why the hell it didn't work for you neither GrueMaster once you fixed the udev rule? | 21:19 |
| rsalveti | yeah | 21:19 |
| ogra_ | hmm | 21:19 |
| rsalveti | even after removing the state file? | 21:19 |
| ogra_ | probably a different race :) | 21:19 |
| rsalveti | ok, trying again, without the rule and saving the broken state file | 21:20 |
| ogra_ | i.e. the state file with the non working state gets loaded *after* the udev rule fires | 21:20 |
| ogra_ | that would revert the udev changes | 21:20 |
| rsalveti | yup, that seems to be possible | 21:20 |
| ogra_ | the only real fix is to call the commands from the rules brefore the first stae file is created | 21:22 |
| ogra_ | for oneric i would solve that in jasper | 21:22 |
| ogra_ | sadly thats not possible for natty | 21:23 |
| ogra_ | not sure how we can actually solve it reliably for natty ... as i said, same prob as in maverick | 21:23 |
| rsalveti | yeah, we should put the correct files and remove old state file | 21:23 |
| rsalveti | otherwise this is not going to work | 21:23 |
| rsalveti | if we remove the state file it should always work | 21:24 |
| rsalveti | let me know fix the udev rule with the broken state file to see if I can get the race | 21:24 |
| ogra_ | right, we would need to remove the state file from the postinst, call alsaucm ... and for safety, create the new statefile too | 21:24 |
| rsalveti | *now | 21:24 |
| rsalveti | yeah | 21:25 |
| rsalveti | true, if we don't create a new one once the user reboots the old state will go to the file | 21:25 |
| ogra_ | another super ugly hack in alsa postinst script :( | 21:25 |
| rsalveti | weeeee :-) | 21:25 |
| ogra_ | this is depressing | 21:25 |
| rsalveti | yeah =\ | 21:26 |
| rsalveti | ogra: and we should also fix the alsaucm path, that's not yet fixed | 21:26 |
| ogra_ | the path ? | 21:27 |
| ogra_ | if we remove the rule ? | 21:27 |
| ogra_ | the SRU should remove it and just call the above postinst stuff | 21:27 |
| rsalveti | yeah, right | 21:30 |
| rsalveti | once the state file is there it should be fine | 21:30 |
| rsalveti | the only problem is that the user end up removing the state file it'll never be right again | 21:31 |
| rsalveti | *if the user | 21:31 |
| rsalveti | as you're setting up the state file at postinst | 21:31 |
| rsalveti | so I believe we should still keep the udev rule | 21:31 |
| rsalveti | as it's called before restoring the alsa-state, it should be fine | 21:31 |
| rsalveti | unless the racing condition is really happening | 21:32 |
| rsalveti | then it can be annoying for the user | 21:32 |
| ogra_ | right | 21:32 |
| rsalveti | ok, confirmed that with the proper udev rule and with the wrong state file it fails | 21:32 |
| ogra_ | \o/ | 21:32 |
| rsalveti | let me reboot a couple of times to see if I can see the racing condition | 21:33 |
| rsalveti | failed again | 21:34 |
| rsalveti | one more time | 21:34 |
| rsalveti | failed again | 21:35 |
| rsalveti | last time | 21:35 |
| rsalveti | ogra_: ok, 4 times and it always failed | 21:37 |
| rsalveti | let me now remove the state file | 21:37 |
| ogra_ | sounds good | 21:38 |
| rsalveti | ogra_: yup, worked like a charm | 21:40 |
| ogra_ | awesome | 21:41 |
| rsalveti | ogra_: let me know once you have the debdiff, I can help you testing | 21:41 |
| ogra_ | k, wont do that tonight anymore though (nearly 11pm here, i need some rest) | 21:42 |
| rsalveti | ogra_: haha, true :-) | 21:42 |
| rsalveti | let me comment this at the bug report | 21:42 |
| ogra_ | thx | 21:42 |
| rsalveti | ok, this should finally solve the sound issue | 21:51 |
| ogra_ | heh ... dejavu | 21:51 |
| rsalveti | :P | 21:51 |
| ogra_ | :) | 21:52 |
| GrueMaster | back | 22:26 |
| ericb2 | GrueMaster: booting | 22:58 |
| ericb2 | GrueMaster: "checking filesystem before resizing ..." | 22:58 |
| GrueMaster | yep | 23:04 |
| ericb2 | GrueMaster: what is the default root password ? | 23:04 |
| GrueMaster | ??? | 23:04 |
| GrueMaster | It should boot into oem-config. | 23:04 |
| GrueMaster | There is no password. | 23:05 |
| ericb2 | GrueMaster: indeed. I just did CTRL +ALT + Fn :) | 23:05 |
| GrueMaster | Ah. Don't do that. :P | 23:05 |
| ericb2 | GrueMaster: got the install in french. Nice :) | 23:06 |
| ericb2 | GrueMaster: there is something strange with the windows size: too little, or the fonts too big ..choose one | 23:08 |
| ericb2 | GrueMaster: I meant the "System configuration window" after, the TZ and the user name | 23:09 |
| GrueMaster | What type of monitor do you have? The default resolution is 1280x720 in the image. | 23:09 |
| rsalveti | argh, pyside doesn't fail when building at my panda | 23:10 |
| ericb2 | GrueMaster: maybe 1400 x ... not sure, but for sure > 1280 x720 | 23:10 |
| GrueMaster | It looks wonky on a widescreen that is capable of higher resolutions. | 23:10 |
| ericb2 | GrueMaster: this monitor is able to display higher resolutions, indeed | 23:10 |
| ericb2 | GrueMaster: got mpurate=1000, but 600MHz in the dmesg | 23:12 |
| GrueMaster | Yes, that will be fixed with the next kernel update after release. | 23:12 |
| GrueMaster | The kernel can't handle mpurate>900 at the moment. | 23:13 |
| ericb2 | clock : dpll1_ck: unable to set MPU rate to 1000: -22 | 23:13 |
| ericb2 | GrueMaster: ok | 23:13 |
| ericb2 | GrueMaster: so recreate a boot.something + use mpurate=900 should basicaly work ? | 23:14 |
| GrueMaster | rsalveti is checking on a patch now, should be ready for the first kernel update. | 23:14 |
| GrueMaster | Yes, you can edit /boot/boot.script then rerun flash-kernel. | 23:14 |
| ericb2 | GrueMaster: great ... thanks rsalveti :) | 23:14 |
| rsalveti | ericb2: mpurate=900 should work for you, I'm just testing the patch that should make it work with 1000 | 23:14 |
| ericb2 | GrueMaster: what test case do you need now ? | 23:15 |
| GrueMaster | rsalveti: Once you have a kernel .deb, care to share? That way we can get more testers on it. | 23:15 |
| rcn-ee_at_work | for the xM, 800 was the recommend max for the default voltage, so 900 may or may not work.. ;) | 23:15 |
| rsalveti | GrueMaster: sure | 23:15 |
| rsalveti | rcn-ee_at_work: yup | 23:15 |
| rsalveti | rcn-ee_at_work: is your kernel still using 800 as default? | 23:15 |
| GrueMaster | ericb2: Start running some apps and see what you can break. :P | 23:16 |
| ericb2 | I tested 800 with sakoman distro, works fine | 23:16 |
| rcn-ee_at_work | yeap it is.. i was betting on dvfs entering 2.6.39, but it didn't... | 23:16 |
| rsalveti | yeah... | 23:16 |
| rcn-ee_at_work | sounds like it might not for 2.6.40 either now with the arm shakeup.. ;) | 23:17 |
| ericb2 | GrueMaster: I'll test OOo4Kids for OMAP3 | 23:20 |
| ericb2 | OOoLight in fact | 23:20 |
| GrueMaster | Haven't heard of that. Not really on my radar this week. | 23:21 |
| rsalveti | ok, found the cause of the pyside ftbfs | 23:27 |
| GrueMaster | nice. | 23:32 |
| GrueMaster | rsalveti: Did you get a bug number for the omap mpurate issue? | 23:34 |
| rsalveti | GrueMaster: nope, not yet | 23:34 |
| rsalveti | GrueMaster: you can create it if you want, but I'm planning to do this in a few hours | 23:34 |
| rsalveti | just need to get dinner | 23:34 |
| GrueMaster | Yea, I'll file in the next few minutes. | 23:35 |
| GrueMaster | Need it for iso tracker. | 23:35 |
| rsalveti | GrueMaster: oh, ok | 23:35 |
| GrueMaster | Don't worry about it. You can add the patch once it is filed. | 23:36 |
| GrueMaster | grrr. Still seeing random oem-config crashes midway through. Nothing in the logs. | 23:37 |
| GrueMaster | And it doesn't happen often. | 23:37 |
| rsalveti | sigh | 23:38 |
| rsalveti | ok, dinner time | 23:38 |
| * rsalveti brb | 23:38 | |
| GrueMaster | rsalveti: Bug #771537 | 23:53 |
| ubot2 | Launchpad bug 771537 in linux "mpurate=1000 fails on beagleXM" [Medium,In progress] https://launchpad.net/bugs/771537 | 23:53 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!