**** BEGIN LOGGING AT Wed Apr 13 02:59:57 2011 Apr 13 04:51:48 lol Apr 13 05:19:19 yeah... the jaunty thing was surprising to me too Apr 13 08:17:53 they used 9.04 on original sheevaplug and nearly nothing changed probably in rootfs for new devices Apr 13 09:46:58 kgilmer: hi Apr 13 09:48:06 hi hrw. saw your tweet, sorry not enough people in the meeting :/ Apr 13 09:50:24 kgilmer: happens Apr 13 09:50:41 kgilmer: I saw Froyo on bug2 post - nice Apr 13 09:51:22 thx, yeah good to get a release out on time. even if no one is using it... Apr 13 09:56:16 hrw, dinner calls. ttyl. Apr 13 14:58:30 here comes the 1M question: what the heck is musb? Apr 13 15:31:52 ogra: ping. When you did the netbook image, did you select autologin? Apr 13 15:32:03 nope Apr 13 15:32:13 why ? Apr 13 15:32:18 do you see issues ? Apr 13 15:32:24 Could you retry it? I want to see if this is just me. Apr 13 15:32:29 hrm Apr 13 15:32:37 oem-config restarted on my system. Apr 13 15:36:23 I'm going to restart fresh and retest. Apr 13 15:36:35 Then I have to leave for the Dr. Apr 13 15:39:00 ogra: will enable the release ppa with our current sgx packages Apr 13 15:39:14 then we can test if it's working fine Apr 13 15:48:49 ogra: I am now seeing the color issues on headless oem-config. Colors are slightly different here, bug fugly none the less. Apr 13 15:49:11 Didn't see this yesterday (nor did I see oem-config, but hey). :P Apr 13 15:50:03 GrueMaster, hmm, my test might take a while, cron decided that it wants to run update-apt-xapian index right before the package removal step Apr 13 15:50:22 so it sitting there now "waiting for apt to finish" Apr 13 15:50:55 rsalveti, yeah, one of us needs to do that before release Apr 13 15:51:11 ogra_: should be done in a few minutes Apr 13 15:51:16 though the current image still points to maverick Apr 13 15:51:39 hm, yeah, saw that you just fixed these bugs Apr 13 15:51:41 fix is uploaded, but i didnt want a rebuild for that Apr 13 15:51:49 I'll test the ppa packages later today or tomorrow. After I get the release testing done. Apr 13 15:52:08 you will have to fiddle with the channel setup first Apr 13 15:52:14 np Apr 13 15:52:17 else it pulls maverick Apr 13 15:52:19 k Apr 13 15:52:48 if you dropped back into oem-config with autologin enabled, the package removal step cant have run Apr 13 15:53:02 that looks more like an installer crash Apr 13 15:53:17 (if package removal would have run, there wouldnt be oem-config binaries) Apr 13 15:53:18 Nothing was in the logs. Apr 13 15:53:26 True. Apr 13 15:54:03 Possible SD corruption. That's why I reflashed and retried. Apr 13 15:54:27 yup, i found a bad SD too today Apr 13 15:54:32 my first one actually Apr 13 15:54:51 i have five worn out USB sticks, but never managed to trash an SD Apr 13 15:55:17 * ogra_ goes upstairs to see how far the xapian crap is Apr 13 15:57:08 Looks like a normal glitch. It is in package removal phase now. Apr 13 15:58:39 yup, finished and logged in automatically here Apr 13 15:58:51 Also, I think I found a uboot issue. On my beagle (C4), I hold the user button when resetting, then it loads uboot from SD. But it fails to load the boot.scr from mmc unless I do "mmc init ; mmcinfo". Apr 13 15:59:03 If I don't do mmcinfo, fatload fails. Apr 13 15:59:18 sounds like a broken u-boot Apr 13 15:59:39 are you sure the one from SD is actually used in that case ? Apr 13 15:59:45 (compare the versions) Apr 13 16:00:51 X-Loader 1.5.0 (Apr 11 2011 - 09:47:05) & U-Boot 2011.03-rc1 (Feb 09 2011 - 01:33:04) Apr 13 16:01:07 sounds fine+ Apr 13 16:01:26 MMC: block number 0x1 exceeds max(0x0) Apr 13 16:01:27 ** Can't read from device 0 ** Apr 13 16:01:27 ** Unable to use mmc 0:1 for fatload ** Apr 13 16:01:41 So I reset and go manual. Apr 13 16:01:53 well, only C4 ... Apr 13 16:02:01 we can document around it Apr 13 16:02:38 as long as xm works fine Apr 13 16:03:32 weird, will try to reproduce at my C4 Apr 13 16:03:38 xm should be fine Apr 13 16:03:56 Yea, xm is fine. No nand to fall back on. Apr 13 16:04:11 Ok, panda netbook worked this time. Apr 13 16:04:32 Gotta go. Dr. appt. Apr 13 16:04:42 good luck Apr 13 16:05:19 Just pulling a few stitches. I'm getting more mobile each day. Currently at ~60%. Apr 13 16:38:17 Hi! Currently the omap4 gfx -dev packages replace, provide, conflict: libgles2-dev, libgles2-mesa-dev Apr 13 16:40:02 Is the 'libgles2-mesa-dev' part because some packages only have libgles2-mesa-dev as a dependency instead of libgles2-dev | libgles2-mesa-dev? Apr 13 16:40:48 alf_: if you check the mesa package, it's not providing libgles2-dev for example Apr 13 16:41:02 it just provides the normal package, not the -dev ones Apr 13 16:41:20 that should probably be fixed too Apr 13 16:41:30 tried to talk with roaf be got no answer Apr 13 16:42:16 alf_: I believe the only package that is depending on the mesa one is the mesa-utils Apr 13 16:42:34 most other packages are depending on libgles2-dev | libgles2-mesa-dev Apr 13 16:43:32 rsalveti: and cairo, too Apr 13 16:43:58 alf_: hm, why cairo is depending directly on the mesa packages? Apr 13 16:44:35 rsalveti: not sure, just noticed it probably just a mistake Apr 13 16:44:42 yeah Apr 13 16:50:02 rsalveti: I will send an email to raof and cc you Apr 13 16:50:33 alf_: cool, thanks Apr 13 16:50:58 would be good if we could get that fixed before the release Apr 13 16:53:16 rsalveti: so just to be in tune, we want libegl/gles...-mesa-dev to provide/conflict/replace libeg1/gles2...-dev, right? Apr 13 16:53:59 alf_: yes, I believe that would be the right solution, but not yet sure, that's why I wanted to talk with raof Apr 13 16:54:13 to understand how it's done with GL, and nvidia/radeon drivers Apr 13 16:55:09 the libegl1/gles2-dev are needed because we want to depend on something that can be used by the packages Apr 13 16:55:49 rsalveti: yes, we need to come up with guidelines of how to package up alternative drivers... so other vendors can package up correctly Apr 13 16:56:20 the only thing it seems weird is that in the end the provided packages for -dev are all the same, even if it's from mesa, sgx or imx Apr 13 16:56:47 as the headers should all be the same, provided by khr Apr 13 16:56:56 alf_: yes Apr 13 16:56:58 rsalveti: well almost, eg eglplatform.h can be different Apr 13 16:58:05 alf_: why? because x11 integration? Apr 13 17:00:51 if they are all the same we could just have a common header package Apr 13 17:01:05 but not yet sure if they need to be the same Apr 13 17:01:22 or if the vendor could also add some other definitions there Apr 13 17:02:15 rsalveti: khronos recommends to use their file (of course ;)) but each vendor may need additional intergration with systems not covered by the default headers Apr 13 17:03:33 alf_: hm, ok Apr 13 17:03:54 rsalveti: or variations of existing systems etc for which the default is not suitable Apr 13 17:05:19 alf_: but then, if we compile a package with a specific set of headers, is it going to work fine with a different set of libraries? Apr 13 17:06:17 if we can't guarantee that then the 'provide' is not necessarily right Apr 13 17:08:52 rsalveti: supposedly, everyone should provide the same (or equivalent) definitions for the same system config Apr 13 17:09:07 rsalveti: Khronos STRONGLY RECOMMENDS that you use the default definitions * provided below, since these changes affect both binary and source * portability of applications using EGL running on different EGL * implementations. Apr 13 17:09:30 ogra_: this is the current postinst script for the extra package: http://paste.ubuntu.com/593647/ Apr 13 17:09:55 ogra_: what do we need to change to fit the unity-2d favorite system? Apr 13 17:10:43 alf_: I see Apr 13 17:11:01 it should be the same bug we cannot guarantee Apr 13 17:12:02 I'd prefer to have a common package set for the headers Apr 13 17:12:19 so we can kind of force the vendors to respect the default definitions Apr 13 17:12:45 but then, would prefer to discuss this with raof Apr 13 17:12:56 to understand if they had the same issues with the normal desktop gl drivers Apr 13 17:13:33 alf_: can you put these questions at your email? Apr 13 17:13:34 rsalveti: the thing is that they may want to just extend them and forcing a single set of headers won't do (for them) Apr 13 17:13:36 rsalveti, unity-2d uses a gconf key for the list of favorites Apr 13 17:14:02 so some gconftool-2 call that removes the .desktop file from that list is needed Apr 13 17:14:07 alf_: yeah =\ Apr 13 17:14:29 though hmm Apr 13 17:14:52 you can just remove the override file that adds this entry and run update-gconf-defaults Apr 13 17:15:03 ogra: first I don't understand why this script was removing the .desktop file Apr 13 17:15:06 probably a hack Apr 13 17:15:14 yes, and still in place Apr 13 17:15:24 but why we need to remove it? Apr 13 17:15:36 because its not comoing from a package Apr 13 17:15:51 oh, sure Apr 13 17:15:56 once installed it should remove the icon Apr 13 17:16:04 as jasper is the one putting it there Apr 13 17:16:08 right Apr 13 17:16:32 ogra_: can you create this new postinst script? Apr 13 17:16:33 rm /usr/share/gconf/unity-2d/default/25_jasper-favorites-override Apr 13 17:16:39 I believe you can easily test at your system Apr 13 17:16:50 update-gconf-defaults --source /usr/share/gconf/unity-2d/default --destination /var/lib/gconf/unity-2d.default --no-signal Apr 13 17:16:57 thats what needs to be added Apr 13 17:17:00 hm, ok Apr 13 17:17:31 for oneric i'll finally put some time into doing it with a package Apr 13 17:17:44 ogra: so /usr/share/gconf/unity-2d/default/25_jasper-favorites-override is now the file that jasper is installing? Apr 13 17:18:34 right, it pulls the original favorites list out of the ubity-2d file Apr 13 17:18:46 and appends the ti .desktop entry Apr 13 17:19:11 because its a higher sequence number update-gconf-defaults then overrides the default entry Apr 13 17:20:41 ogra_: oh, ok Apr 13 17:21:11 ogra_: http://paste.ubuntu.com/593654/ Apr 13 17:21:12 so rm'ing it and running update-gconf-defaults again will restore the original system defaults Apr 13 17:21:36 well, and the .desktop file Apr 13 17:22:16 you should still remove it to not leave unpackaged cruft around Apr 13 17:22:27 yeah, sure Apr 13 17:22:40 but yes, apart from that thats what should happen Apr 13 17:23:10 ogra_: is there any image where I can use to test this? Apr 13 17:23:19 with your latest jasper fixes Apr 13 17:23:28 no Apr 13 17:23:37 unless someone re-rolls Apr 13 17:23:43 hm, ok Apr 13 17:23:48 but you can manually hack the two files in netbook Apr 13 17:24:15 /usr/share/app-install/channels/ has the files Apr 13 17:24:28 you want to modify .list Apr 13 17:24:46 (indeed before clicking the TI icon) Apr 13 17:24:55 sure Apr 13 17:30:43 ogra_: should we also remove the ti-ppa.svg or we can't? Apr 13 17:31:06 rsalveti, that should probably go too, yeah Apr 13 17:31:17 ogra_: ok Apr 13 17:33:59 ok, I'm upgrading my natty rootfs and should be able to test it soon Apr 13 21:06:54 ogra: That oem-config restart issue I saw earlier on panda netbook I am seeing again on my beagleXM. Apr 13 21:07:07 Way more than possible coincidence. Apr 13 21:14:11 rsalveti, we would like to update u-boot but obviously do not want to break you guys in the process. My proposed branch is here:https://code.launchpad.net/~jcrigby/ubuntu/natty/u-boot-linaro/proposed-2011.04.1 slangasek is working on merging it Apr 13 21:14:54 jcrigby: we should be fine with it, did you test with latest x-loader with both omap3 and 4? Apr 13 21:14:55 rsalveti, the reason is device tree support needs to be there for Linaro 11.05 release Apr 13 21:15:02 yes Apr 13 21:15:16 jcrigby: hm, ok Apr 13 21:15:59 jcrigby: will test it with our images here, ping you back once it's done Apr 13 21:16:40 thanks, I realize I may not have latest x-loader. An update happened but flash kernel didn't like my arch name Apr 13 21:16:56 not sure if latest was in the last hwpack I downloaded Apr 13 21:17:07 I will test again Apr 13 21:17:44 jcrigby: ok, latest one was uploaded at the beginning of this week Apr 13 21:18:01 ok, my hwpack was from last week Apr 13 21:48:13 Hey Apr 13 21:48:53 Are there any good opengl examples in the ubuntu 10.10/TI ppa for testing OpenGL (ES) support on the panda board? Apr 13 21:49:01 any help would be greatly appreciated Apr 13 22:07:37 dcarr_home: if you build the mesa-utils package available at natty you'll be able to run es2gears Apr 13 22:07:50 a nice and simple example Apr 13 22:23:24 rsalveti: Is the edid detect code in the omap kernel yet? Apr 13 22:25:22 GrueMaster: nops Apr 13 22:25:34 ok. Apr 13 22:33:44 rsalveti: thanks Apr 13 23:17:16 rsalveti, just to let you know. I have now tested the proposed u-boot with the latest x-load on panda and beagle Apr 13 23:17:21 beagle c4 Apr 13 23:18:12 jcrigby: yeah, just tested fine on panda, xM and at my c4 Apr 13 23:18:20 I'm now booting at my b5 Apr 13 23:18:58 rsalveti, thanks for doing that Apr 13 23:28:24 jcrigby: http://paste.ubuntu.com/593805/ Apr 13 23:28:26 at my B5 Apr 13 23:30:34 jcrigby: same sd card works fine at my C4 Apr 13 23:30:58 jcrigby: could be a regression, just don't know exactly which upload could be the culprit one Apr 13 23:31:06 quite a while that I don't test it with my B5 Apr 13 23:31:33 hmm Apr 13 23:31:48 let me flash a maverick sd card and we can start tracing it Apr 13 23:31:55 want also to make sure it's not the x-loader Apr 13 23:31:58 ok Apr 13 23:31:58 or not a hardware issue Apr 13 23:42:01 jcrigby: yup, 2010.09-rc1 from maverick worked fine Apr 13 23:42:17 so it's a regression Apr 13 23:43:43 ok, thanks Apr 13 23:44:02 will try with the one we're using, then I can just open a bug for it Apr 13 23:44:12 using at the image Apr 13 23:44:47 also, let me first update just the x-loader Apr 13 23:45:13 ok, so latest in archive is also broken? Apr 13 23:45:19 for B5 Apr 13 23:50:21 jcrigby: trying now Apr 13 23:50:29 jcrigby: worked fine with latest x-loader Apr 13 23:50:37 so it's probably a regression at u-boot Apr 13 23:51:15 ok Apr 13 23:52:34 jcrigby: yup, broken Apr 13 23:57:50 jcrigby: https://bugs.launchpad.net/ubuntu/+source/u-boot-linaro/+bug/760350 Apr 13 23:58:18 so 2011.02.1 is also broken Apr 13 23:59:06 rsalveti, I believe that is what is the latest in the archive right? Apr 13 23:59:11 jcrigby: yes Apr 14 00:00:06 and we don't know when the regresson happend. Just sometime between 2010.09 and 2011.02.1 Apr 14 00:00:26 rsalveti, this is going to be fun Apr 14 00:00:37 jcrigby: haha, yeah :-) Apr 14 00:00:59 there's also a new error with latest u-boot: timed out in wait_for_pin: I2C_STAT=0 Apr 14 00:01:09 but can be normal, probably trying to identify the expansion board Apr 14 00:01:14 but would be good to check Apr 14 00:01:43 yes, I saw that too, I will make sure it is expected with no expansion board Apr 14 00:02:43 ogra_: FYI, threw in another patch from upstream for UCM. It basically enables UCM only for devices that have a UCM argument setin the udev rules, which should now cover OMAP4. Apr 14 00:03:01 jcrigby: will be out for dinner, should be back in one hour, then we can try to bisect it Apr 14 00:03:28 ogra_: So source should b epublished by the time you get online. Apr 14 00:04:52 rsalveti, I will be out then so may need to wait until tomorrow. I will be back later tonight I will ping you. Apr 14 00:05:32 jcrigby: sure, np Apr 14 00:10:46 sakoman, ping? Apr 14 00:16:11 TheMuso: Does that also cover omap? It will have UCM support as well, hopefully by next week. Apr 14 00:16:25 GrueMaster: Yes it does. Apr 14 00:16:30 Excellent. Apr 14 00:16:36 As I said, the udev stuff is to cover omap4 Apr 14 00:18:13 OK. That looks like jack detection, which isn't needed for beagle/beaglexm but may be needed for other omap based systems. Not worried about it now. Apr 14 00:19:17 Its not jack detectino. It basically uses a definition in the pulse udev rule to tell pulse to enable UCM only for omap4, which reduces regression potential for other hadrware. Apr 14 00:19:59 Oh. Hmmm. I'll have to look over the patches again. Apr 14 00:21:18 This is from PA_UCM_patches_20110407.tar.bz2 right? Apr 14 00:21:40 No, its another one that was sent to david and myself. Apr 14 00:21:58 Not attached to the bug? Hate when that happens. Apr 14 00:22:00 Its in teh package I just uploaded to my PPA. Apr 14 00:22:05 ok Apr 14 00:22:11 I'll attach it to the bug if you'd like. Apr 14 00:22:28 It makes tracking easier, but I can find it. Apr 14 00:22:42 ok Apr 14 00:41:16 rsalveti: http://comments.gmane.org/gmane.linux.distributions.gumstix.general/57467 Apr 14 00:41:38 looks like multiblock reads are broken in older omap silicon Apr 14 02:07:55 jcrigby: yeah, that makes sense Apr 14 02:08:58 jcrigby: will try that patch Apr 14 02:16:59 rsalveti: re: IC2_ERROR, almost certian that error has to deal with the expansion port. Apr 14 02:17:16 * NCommander is catching up on backscroll) Apr 14 02:17:27 yup, probably Apr 14 02:29:12 rsalveti, I'm going to do a patch that would be acceptable to upstream. Something like what gcl suggests in the thread. Apr 14 02:29:22 rsalveti, that is if it works for you Apr 14 02:29:38 jcrigby: cool, sure Apr 14 02:44:56 jcrigby: yup, worked fine Apr 14 02:45:15 great!, thanks Apr 14 02:49:30 GrueMaster: ogra_: https://bugs.launchpad.net/ubuntu/+source/u-boot-linaro/+bug/760350 Apr 14 02:49:41 this should be at the beta-2 release notes Apr 14 02:49:59 our image doesn't work with older beagles, probably <= rev B5 Apr 14 02:50:15 jcrigby: just updated the bug with the test result Apr 14 02:50:49 Ok. Very odd, but I'm not sure we really even support hw <512M (except maybe headless). Apr 14 02:51:47 GrueMaster, that might explain why this problem has never been reported Apr 14 02:52:22 I've seen my own issues with u-boot on beagle C4. Currently, when I reboot I hold the user button to load x-loader & uboot from SD, but my bootcmd fails. Apr 14 02:54:56 bootcmd=bootcmd=mmc init;mmcinfo;if fatload mmc 0 0x82000000 boot.scr;then echo boot.scr found! Executing ...;source 0x82000000;fi Apr 14 02:55:12 It would fail without the mmcinfo. Apr 14 02:56:29 GrueMaster: headless image can be useful for those old boards Apr 14 02:56:47 GrueMaster: can be be used to be a small server, or something like that Apr 14 02:56:55 yes, I realize that. That's why I dusted mine off and have it hooked up. Apr 14 02:57:26 I usually install openssh, and basic server. Apr 14 02:57:28 we don't expect it to work fast, but it should at least boot :-) Apr 14 02:58:22 Fast is relative to the bottlenecks in the environment. For a printerver, it should be great (haven't tested yet). Apr 14 02:58:52 yeah Apr 14 02:59:36 Unfortunately, my office is in flux and need of an overhaul (hopefully this weekend). **** ENDING LOGGING AT Thu Apr 14 02:59:58 2011