**** BEGIN LOGGING AT Thu Sep 09 02:59:58 2010 Sep 09 06:56:27 cooloney: nice, were you able to build the kernel with highmem and 1gb at your es2 board? Sep 09 06:58:08 sebjan: http://kernel.ubuntu.com/git?p=roc/ubuntu-maverick.git;a=commit;h=68e4544fcd736e6aa00354da15dd48a92c15750c Sep 09 06:58:22 didn't test, but it seems cooloney got it working fine Sep 09 07:08:50 rsalveti: i'm going to test it soon, but you know it will take hours Sep 09 07:09:59 cooloney: but at least it seems we have a fix Sep 09 07:10:17 ndec: http://lkml.org/lkml/2010/9/8/425 Sep 09 07:27:03 well, time to get some sleep Sep 09 07:27:08 * rsalveti out Sep 09 07:30:34 Hello.. am new around here.. I am working on cross-compiling an application to ARM enviroment, so I was wondering if someone can help me with some doubts around it? Sep 09 07:31:31 Better to ask a specific question, generally. In practice, we don't cross-compile as a matter of course, but some subset of us cross-compile for other projects. Sep 09 07:34:31 Well the thing is this.. am trying to cross-compile and application that uses the library "libusb", but when I try to do it it gives me errors that doesn't find the library.. So I checked the toolchain and it doesnt have the libusb packages, so I was wondering on How to add the library into the toolchain, and if that would fix the error? Sep 09 07:39:09 morning Sep 09 07:50:23 cooloney: is this patch supposed to fix the instabilities we observed with highmem? #633227 Sep 09 07:50:59 sebjan: yeah, i hope so Sep 09 07:51:30 sebjan: it is quite slow for me to setup the building environment on my es2.0, ports.ubuntu.com connection here is too slow Sep 09 07:51:45 sebjan: so please help to test on your es2.0 firstly, thx Sep 09 07:51:50 ok, I'll make a test with it then Sep 09 07:52:23 sebjan: great, thx Sep 09 07:59:22 acorvil, Apologies, but it seems that nobody has the answer just now. Have you tried native or emulated compilation? In either case, simply installing libusb-dev ought allow the build to proceed. Sep 09 08:02:54 cooloney, ogra: in one of my last patches, I changed the HID support to be in module (instead of statically linked). So without the kernel modules, there is no mouse/keyboard support... Sep 09 08:03:19 would it be easier to re-integrate HID support statically? Sep 09 08:05:51 persia, yeah I already have the libusb-dev installed in my host PC, so if I try native compilation it goes well.. The thing is if I cross-compile it that doesnt find the library Sep 09 08:06:20 I was guessing that the library most be into the toolchain so the cross-compiler can find it.. but I dont know how to get it into the toolchain Sep 09 08:06:48 sebjan, our initramfs tools should care automatically to have the modules available. compiling them in statically might be a speed advantage though Sep 09 08:07:02 You'd need to have the ARM libusb-dev installed. If you're native ARM, or emulated, it would just be a normal package. I've absolutely no idea how to integrate it in a cross-compiler. Sep 09 08:07:20 sebjan: CONFIG_USB_HID=m is the same in master branch Sep 09 08:07:50 If possible, please set modules to match those for the primary kernel: we've had all sorts of annoying ports bugs in the past for different things being modules vs. built-ins Sep 09 08:08:03 sebjan: so i think udev or other foundations stuff in ubuntu will load right modules when we insert the usb device Sep 09 08:08:17 persia: yeah, i agree Sep 09 08:08:26 * persia has a special desire to see mod_dm has a built-in, for instance, which doesn't seem required for some targets, but breaks boots if not done that way Sep 09 08:09:11 * ogra wouldnt mind to have some essential things statically compiled in since it lowers the boot tima Sep 09 08:09:13 *time Sep 09 08:09:19 cooloney: yes, the modules get loaded and the mouse/keyboard work this way. I was just wondering if it would be easier to have HID statically in case we want to just update the ui Sep 09 08:09:41 cooloney: ... just update the uImage without installing the whole kernel package... Sep 09 08:10:17 oh, sweet, we finally got images again Sep 09 08:10:24 * ogra zsyncs Sep 09 08:12:50 ogra, The kernel team spends a long time each cycle determining the set of stuff to do statically and modularly, and then lots of userspace stuff expects that. We've had ports breakages before on both sides (stuff compiled in that oughtn't be, and so didn't initialise properly, and stuff as modules that oughtn't be that broke booting for some folks) Sep 09 08:13:13 cooloney, did anyone talk to you about bug 605042 yet ? seems we need an imx51 kernel specialist there Sep 09 08:13:16 I agree that more in makes faster boots which is nifty, but I strongly argue that this needs to be coordinated with everyone. Sep 09 08:13:29 ogra: Bug 605042 on http://launchpad.net/bugs/605042 is private Sep 09 08:13:36 sebjan: as persia said, normally we will do that config review during UDS Sep 09 08:13:57 persia, Keybuk does that too and often there is disagreement, i'm most of the time in the Keybuk camp ;) Sep 09 08:14:07 sebjan: and omap4 branch will be merged into master in the future, we gonna follow the config setting of master Sep 09 08:14:21 ogra: yeah, apw pointed me out that kindly Sep 09 08:14:29 I'm in the Keybuk camp, based on long arguments about lucid not working on some of my machines. The kernels *need* to be aligned. Sep 09 08:14:32 ogra: i talked with Yao Qi last night. Sep 09 08:14:42 So don't tell folks to fuss with them :) Sep 09 08:14:43 cooloney, got any idea about it ? Sep 09 08:14:45 persia, Thanks for your answers am going to see if getting the ARM libusb-dev installed and see if it fix the error Sep 09 08:14:46 ogra: just wanna isolate the issues, Sep 09 08:14:51 sweet !! Sep 09 08:15:03 ogra: As Yao Qi said, he found the issue on imx51 with 2.6.31 kernel Sep 09 08:15:27 while other 2 omap3 boards with 2.6.33 kernel don't have such issue Sep 09 08:15:32 cooloney, right which is what we run on the buildds Sep 09 08:16:15 ogra: i wanna know the root cause is the kernel version 2.6.31 misses some key patches which are in 2.6.33 Sep 09 08:16:24 or it's imx51 specific issue. Sep 09 08:16:25 ah Sep 09 08:16:35 yeah, could be either Sep 09 08:16:53 but unforunately, i don't have fsl-imx51 babbage now Sep 09 08:17:04 mine is bricked, Sep 09 08:17:11 ogra, persia, cooloney: ok, I wanted to get your advice on this HID setting. Will follow your recommendation - let's keep the current configuration. Sep 09 08:17:20 yeah, i have a 2.5 here and could test Sep 09 08:17:59 cooloney: my memtester test is running for few minutes now (with your patch), which I have never seen before with 1GB mem! Sep 09 08:18:12 ogra: thanks a lot, so i think is that possible for you to run upstream 2.6.35 or 2.6.33+ kernel to test that. Sep 09 08:18:41 sebjan: good news, man Sep 09 08:19:02 sebjan: actually, it is a very subtle issue, Gary King from NV helped a lot Sep 09 08:20:47 ogra: or do you have fsl contact person who can help to test with their lastest bsp on imx51? Sep 09 08:21:01 cooloney, not anymore ... Sep 09 08:21:14 cooloney, but i know the upstream maintainer *g* Sep 09 08:21:18 * ogra looks at amitk Sep 09 08:22:07 anybody know if ac100 source was released anywhere? Sep 09 08:23:36 tmzt_, I've not heard, and I've asked in a few fora. Might help if someone made a request (if this hasn't already been done). Sep 09 08:23:39 tmzt_, i tried to find out on the weekend, but beyond the tegra2 stuff you can pull directly from nvidia there doesnt seem to be anything yet Sep 09 08:23:57 okay, I don't have hardware or anything Sep 09 08:24:00 ogra: let's welcome amitk, heh Sep 09 08:24:01 Does anyone own one yet? I've heard lots of folks say they are hard to find. Sep 09 08:24:07 I just hope they make enough before market forces kill it Sep 09 08:24:11 and its a) not really clear if you can even get serial access to the netbook Sep 09 08:24:23 serial access may not be required. Sep 09 08:24:28 ogra: so i can try to rebase fsl's latest 2.6.31 based bsp on our lucid kernel Sep 09 08:24:28 and b) i doubt all toshiba patches are in the nvidia source Sep 09 08:24:35 and build a kernel for you to test. Sep 09 08:24:44 I bet they have a EC in it that needs special support Sep 09 08:25:13 cooloney, well, i'd like to hear a statement from amitk about the linaro binaries, they build imx51 already, if we could use that without regression for the buildds that would be the best solution Sep 09 08:25:24 Let's not speculate. Source will be available if anyone purchases one (nature of the license). It purportedly ships with an OS that (for some configurations) allows one to rewrite the flash. Sep 09 08:25:35 tmzt_, the LCD probably too Sep 09 08:25:40 that's good Sep 09 08:25:48 ogra: not ion/nv lcdc? Sep 09 08:25:48 Better to assume that it will be all good, and we can work around anything confusing later. Sep 09 08:25:57 ogra: yeah, good point Sep 09 08:25:58 persia, i saw the unboxing videos, there isnt any CD in the box Sep 09 08:26:01 ion'd be no issue: we have working free drivers for ion. Sep 09 08:26:11 how do you define free here? Sep 09 08:26:22 ogra, So what? If you own it, you can request the source. Sep 09 08:26:45 tmzt_, well, as persia says, i'm only speculating, they sell ac100 around the corner here and i took a look on the weekend, it didnt look easy to unlock it Sep 09 08:26:45 tmzt_, Mostly DFSG+CC-SA2.0+GFDL+a couple other similar licenses. Sep 09 08:27:19 persia, the graphics drivers are definitely closed (even the ones you can download from the tegra site are) Sep 09 08:27:28 ogra, unlocking Android isn't often complicated. Sep 09 08:27:39 if you get serial access, yeah Sep 09 08:27:48 persia: you saying there's an open 3d driver or just framebuffer? Sep 09 08:27:51 That's why I'd like it to be Ion. Run nouveau (which builds for ARM: dunno how tested it is) Sep 09 08:27:52 ah right Sep 09 08:28:07 tmzt_, I run nouveau on an Ion on my desktop. Sep 09 08:28:19 cool, does that work for video playback as well? Sep 09 08:28:23 ogra: cooloney: I'm trying to get linaro to build imx51 images Sep 09 08:28:23 there is only one otg port and that registers as a cardreader on a host PC immediately after powering up Sep 09 08:28:38 3D works (although I won't say it's *great* - some games are good, others not so good. Videos are fine. Sep 09 08:28:43 amitk, i think you already build them (i see them mentioned in the recent upload) Sep 09 08:28:53 ogra: so firmware flash can be confirmed (essentially)? Sep 09 08:29:17 ogra: we build the kernels, the images were only discussed yesterday (dunno if asac already did them) Sep 09 08:29:27 tmzt_, it registers as a cardreader for the SD slot, i didnt see a way to access serial or any flash mode Sep 09 08:29:45 oh okay Sep 09 08:29:47 amitk, i dont care about images, just about binary kernel packages Sep 09 08:29:50 amitk: is the linaro kernel 2.6.31 based? or mainline version? Sep 09 08:29:51 that sounds like what android provides Sep 09 08:30:01 cooloney, linaro only uses upstream Sep 09 08:30:13 thats why i said "lets see if there are regressions) Sep 09 08:30:13 ogra: but as I commented on the bug, if all you need is serial + ethernet, then we might be able to support that Sep 09 08:30:26 cooloney: 2.6.31?!! yeeeccccck! Sep 09 08:30:27 amitk, we need USB at least Sep 09 08:30:29 Android allows one to write to flash from the booted system (as does linux). Not really hard to change what happens on boot. Sep 09 08:30:41 ogra: there is USB support too, but I haven't tested it all Sep 09 08:30:50 persia, that requires that you can install apps to access it Sep 09 08:31:05 persia, toshiba controls the appstore for the device Sep 09 08:31:09 persia: if the android image allows root, this has been contenious in #android since the g1 was relased Sep 09 08:31:16 there is no access to the android stor Sep 09 08:31:16 ogra, There exist Android devices to which you cannot install applications? I thought that installing apps was the entire point of Android. Sep 09 08:31:25 but mostly I'm wanting to replace kernel and remove android entirely Sep 09 08:31:33 you can install apps, but only toshiba selected ones Sep 09 08:31:33 * persia gets confused, and tries to go back to ignoring Android Sep 09 08:31:58 persia, in android you cen define which appstore the device uses ... Sep 09 08:32:02 persia: only some AT&T models, the issue isn't applications (java code in euid/egid sandbox) but the kernel and firmware themselves Sep 09 08:32:05 as a manufacturer Sep 09 08:32:15 persia: local apk install is almost always supported Sep 09 08:32:38 apk requires a console, no ? Sep 09 08:32:50 no, adb works as well Sep 09 08:32:56 there is none preinstalled and i dont know if the toshiba store has one Sep 09 08:32:58 usually has to be enabled in settings Sep 09 08:33:03 ah Sep 09 08:33:11 the review I read mentioned 'sideloading' a wordpress client Sep 09 08:33:27 if i wouldnt be so busy with omap4 i'd really buy an ac100 just to find out Sep 09 08:33:41 point though, very thin full size keyboard debian/ubuntu/linaro device Sep 09 08:33:42 but i'm short on play-around-time currently Sep 09 08:34:06 and 3g options available Sep 09 08:34:20 the one here only comes with 3G Sep 09 08:34:35 1€ with contract, 380€ without Sep 09 08:34:59 amitk: cool, is that possible to verify such issue with Linaro upstream based kernel Sep 09 08:36:03 cooloney: do you have an easy way to reproduce the problem? Sep 09 08:36:15 cooloney, i just need to dpkg -i it i guess Sep 09 08:36:21 amitk, bug 605042 Sep 09 08:36:26 ogra: Bug 605042 on http://launchpad.net/bugs/605042 is private Sep 09 08:37:03 seems a simple "java -version" shows it Sep 09 08:41:10 okay, I need to go, I'd love to be updated on ac100 if you guys find anything (or get one) Sep 09 08:46:11 amitk: yeah, Yao Qi told me that we have to apt-get install eglibc and open-jdk package Sep 09 08:46:19 and run java --version Sep 09 08:48:18 cooloney: your patch solves the memtester issue, but not the problems during the build: I just reproduced it. So we still have another issue... Sep 09 08:49:17 sebjan: oh, is the memtester issue in launchpad? Sep 09 08:49:42 * cooloney just found launchpad is down Sep 09 08:50:53 cooloney: I flagged 1 bug and described both tests: bug 633227 Sep 09 08:50:58 sebjan: Bug 633227 on http://launchpad.net/bugs/633227 is private Sep 09 08:52:04 ndec, http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100909/ works fine on my ES2 6 layer, please spread the word at your team ;) (didnt test on 8 layer yet) Sep 09 09:00:57 persia, wuld you like to work on Bug 613612 ? Sep 09 09:01:02 ogra: Bug 613612 on http://launchpad.net/bugs/613612 is private Sep 09 09:01:12 (i know you tackled that issue before) Sep 09 09:01:28 bah, silly launchpad ... Sep 09 09:01:33 being down doesnt help Sep 09 09:01:42 title is "Favorites missing from netbook-launcher-efl" Sep 09 09:07:50 wohoo ! Sep 09 09:07:57 8 layer ES2 works too ! Sep 09 09:08:31 Yeah, that's in my list of things to look at, although something like 4th or 5th place in the list of bugs I'd like to fix for ARM. Sep 09 09:17:51 * ogra_omap4 hugs his ES2.0 Sep 09 09:18:12 ogra: i like this ;-) Sep 09 09:19:15 strangely the screen turns black from time to time Sep 09 09:19:33 i have to switch consoles to get my X back Sep 09 09:20:00 [ 650.169860] omapdss DISPC error: GFX_FIFO_UNDERFLOW, disabling GFX Sep 09 09:20:00 [ 650.219543] omapdss DISPC error: SYNC_LOST_DIGIT Sep 09 09:20:01 aha Sep 09 09:20:22 hello, anyone know (yet) if this can run/boot ubuntu ? http://www.archos.com/products/ta/archos_101it/specs.html?country=gb&lang=en (and if there are drivers for the wlan, opengl, etc) Sep 09 09:20:41 and the asoc driver complains a lot in dmesg Sep 09 09:21:30 oskude, if you find a kernel, the ubuntu userspace will likely work Sep 09 09:21:47 check the channel topic for rolling a rootfs Sep 09 09:22:07 ogra_omap4: roger. i might just mail them and ask (the device is not yet available, at least not in my country) Sep 09 11:35:34 NCommander, http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/20100909/ FYI Sep 09 11:45:57 WIN! Sep 09 11:46:38 we should probably enable it in the crontab again :) Sep 09 11:46:49 semms colin disabled it during beta Sep 09 11:48:23 ogra_cmpc: yeah, that probably would be good Sep 09 11:49:17 NCommander, is the menifest wikipage alredy updated ? Sep 09 11:49:27 for dove Sep 09 11:52:16 manifest Sep 09 11:52:41 festiman ? Sep 09 11:53:01 festiVAL Sep 09 11:53:08 PARTY !!! Sep 09 11:53:12 Yep :) Sep 09 11:53:15 heh Sep 09 12:43:04 sigh, the XM takes a century to boot Sep 09 12:55:50 ogra: do you mean that Ubuntu takes a century to boot the XM? ;-) Sep 09 12:56:14 he's exaggerating a bit... Sep 09 12:56:18 ogra: slow SD card? Sep 09 12:56:21 It's just slower than for some other platforms. Sep 09 12:57:21 morning Sep 09 12:59:22 Morning rsalveti . Already got coffee? Sep 09 12:59:24 mopdenacker: well, when it boots, its 1907 :-) Sep 09 12:59:48 mopdenacker, well, the whole install in the panda takes between 10 and 12min ... the same thing on the XM takes about 45min Sep 09 12:59:57 EPREEPOCH Sep 09 13:00:01 SD cards slow? Sep 09 13:00:04 :-) Sep 09 13:00:21 i use a class 2 on the panda, and a class 6 on the XM currently Sep 09 13:00:40 if definition of classes recently swapped i missed that ;) Sep 09 13:00:53 that's still kinda slow. it's a damn shame that new SDHC-I thing probably needs a new controller Sep 09 13:01:14 ouch. You've got faster DDR access on OMAP4. Could it be the reason why? Sep 09 13:01:15 mopdenacker: not yet Sep 09 13:01:16 (and I guess on most platforms twice the clock, mx51 esdhc unit runs at 66MHz..) Sep 09 13:01:26 rsalveti: Panda patch is now staged for mainline merge: http://git.denx.de/?p=u-boot/u-boot-ti.git;a=shortlog Sep 09 13:02:07 mopdenacker, well, i have dual core 1GHz and 1G on the panda ... single core 600MHz and 512M on the XM ... Sep 09 13:02:33 sakoman: cool :-) Sep 09 13:02:52 surely the SD plays a role but i guess ram and CPU are the more important parts here Sep 09 13:03:17 the panda is real fun to use now Sep 09 13:03:26 ogra: yep, real fast Sep 09 13:04:15 i just wish it was a but thinner ... i have a bunch of old netbook shells around :) Sep 09 13:04:22 s/but/bit/ Sep 09 13:04:50 Are the clock speeds directly comparable, or have there been scheduler improvements? Sep 09 13:04:56 * ogra wouldnt mind using it for day to day work with a fast usb disk Sep 09 13:05:12 ogra: the XM's OMAP 3730 is supposed to run at 1GHz, according to vstehle ... Sep 09 13:05:14 You might use it as a desktop :) Sep 09 13:05:21 persia, since we run from SD both images use elevator=noop Sep 09 13:05:40 I meant on-chip dispatch scheduling, pipelining, etc. Sep 09 13:05:41 mopdenacker, well, it currently doesnt with our omap3 kernel Sep 09 13:06:02 dmesg shows 600MHz Sep 09 13:06:17 xM should be faster, I guess Sep 09 13:06:17 and /proc/cpuinfo 511 bogomips Sep 09 13:06:37 the panda goes up to 2000 here ... Sep 09 13:06:43 per cpu core ;) Sep 09 13:06:50 http://slashdot.org/submission/1329038/ARM-Eagle-preys-on-smartphones-and-servers Sep 09 13:06:52 Ah, so the clockspeeds *aren't* directly comparable. Sep 09 13:06:57 ^ that will be nice when it hits :) Sep 09 13:07:50 The hardware virtualisation is the bit that is most exciting to me. I don't need 2.5GHz even on my desktop (be nice on a buildd though) Sep 09 13:08:40 28nm? crazy Sep 09 13:08:51 indeed :) Sep 09 13:08:59 Why? Isn't RAM there already? Sep 09 13:10:34 only depressing thing I read there was " it will be at least late 2012 before products featuring the processor arrive on the shelves" Sep 09 13:11:15 until then the pandaboard will serve :) Sep 09 13:11:17 yeah but there'll be a bunch of nice A9s around before then Sep 09 13:15:48 is there a quad A9 on the market yet? :/ Sep 09 13:16:41 probably prototypes at smoothstone Sep 09 13:17:04 surely nothing public yet Sep 09 13:17:11 ogra: you can tweak the xM clock rate with the mpurate u-boot environment variable Sep 09 13:17:40 devilhorns, That's normal, given the design and certification issues. Often at least 12 months between having demo HW and retail HW. Sep 09 13:17:40 sakoman, ah. cool i'll check if we can do that from jasper on firstboot Sep 09 13:17:48 you could try 'setenv mpurate 720' or even 1000 if your kernel will accept it Sep 09 13:17:56 persia, indeed ... but dang it, I want one NOW ! :) Sep 09 13:18:01 sakoman: ogra: I'm just not sure if the kernel supports it Sep 09 13:18:09 well, our kernel still has a bunch of oppses it seems Sep 09 13:18:10 but never tried Sep 09 13:18:12 devilhorns, Be an ODM or IHV :p Sep 09 13:18:22 i see clock related ones :P Sep 09 13:18:27 rsalveti: well, my linux-omap should accept up to 720 Sep 09 13:18:31 persia, hehehe yea...not gonna happen tho :( Sep 09 13:18:51 sakoman: but we're currently using just upstream, could be missing some patches Sep 09 13:19:07 then 720 should work, but not 1000 Sep 09 13:19:08 but didn't check, can do that later Sep 09 13:19:12 i think we use a few cherrypicks from -omap Sep 09 13:19:24 so if its there we might be able to pull it in Sep 09 13:19:35 but be warned running above 500 lessens the life of the processor Sep 09 13:19:51 like ... from 10 years to 8 ? Sep 09 13:20:10 whats the estimated processor life of an omap3 ? Sep 09 13:21:06 cool, seems we got some patches for mpurate support, so it seems it should work fine Sep 09 13:21:39 great Sep 09 13:21:49 * ogra would like to see the oopses go away though Sep 09 13:21:58 seems we still probe for uart3 Sep 09 13:22:05 ogra: yep Sep 09 13:22:13 where is lag? Sep 09 13:22:15 :-) Sep 09 13:22:31 ogra: I think it goes from 100K hours to 44K hours. Let me check Sep 09 13:22:37 lag'ing through his vacation until monday i think Sep 09 13:22:47 hm, ok Sep 09 13:23:21 ogra: https://e2e.ti.com/support/dsp/omap_applications_processors/f/447/p/62063/223487.aspx#223487 Sep 09 13:23:35 well, thats still 5 years ... :) Sep 09 13:23:59 sure, just wanted to make sure you were making an informed decision Sep 09 13:24:07 thanks ! :) Sep 09 13:25:08 after 5 years its a good idea to upgrade to omap5 anyway imho ... and TI will earn some money too :) Sep 09 13:26:03 Depends on the use case. Given how often you like to accumulate new toys, you might want to try 1200 :) Sep 09 13:26:14 (or maybe that requires active cooling) Sep 09 13:27:12 well, given that our images are using a netbook UI you surely want to upgrade relatively soonish :) Sep 09 13:27:29 ogra: where's the manifest webpage? (I can't find it :-/) Sep 09 13:28:25 NCommander, https://wiki.ubuntu.com/MaverickMeerkat/ReleaseManifest davidm should update it though (you had an action item from last meeting) Sep 09 13:30:11 ogra: I know, I spoke with davidm and got a followup action item :-) Sep 09 13:30:18 good :) Sep 09 13:30:44 it needs to be up to date for next milestone so the release team knows what to do Sep 09 13:34:36 * ogra takes a break Sep 09 14:21:48 ogra, mopdenacker: http://media.digikey.com/pdf/Data%20Sheets/Circuitco%20Elect/Beagle_Board_Flyer_5-21-10_2.pdf?cshift_ck=null&client_id=5042&cshift_ck=null&client_id=5042 Sep 09 14:22:09 sebjan: Sep 09 14:22:24 rsalveti: yes? Sep 09 14:22:28 sebjan: the behavior is still weird with the supposed mem fix Sep 09 14:23:07 rsalveti: this patch fixes the memtester issue for me, but not the build issues Sep 09 14:23:31 tried to run memtest but got disconnected from ssh and then the system started to behave in a weird way Sep 09 14:24:03 trying to test again, but now I got usb ehci resets again Sep 09 14:24:24 this is another weird bug that doesn't happen all the time Sep 09 14:24:41 * vstehle can run 2x memtesters just fine since ~1h30... Sep 09 14:25:51 vstehle, ah, thanks Sep 09 14:26:40 ogra_cmpc: You can use 'devmem' to lock the PLL @ 1GHz if you booted @ 600 MHz :) Sep 09 14:27:06 ogra_cmpc: But that is not for the faint of heart... :) Sep 09 14:28:38 yeah, i rather use the u-boot parameter Sep 09 14:29:03 so we can boot at full speed ;) Sep 09 14:34:15 sebjan: http://paste.ubuntu.com/490990/ the boot with usb ehci resets Sep 09 14:35:41 hey, why do you have working asoc ? Sep 09 14:36:42 sebjan: vstehle: got again Sep 09 14:37:04 tried just running memtester, went fine for some seconds Sep 09 14:37:18 then I stopped, started building the kernel, and run memtester on another shell Sep 09 14:38:18 rsalveti: you have this reset issue all the time? Sep 09 14:38:19 and then http://paste.ubuntu.com/490999/ Sep 09 14:38:30 sebjan: nops, most of the time it works fine Sep 09 14:42:13 then test fail when running memtest again: http://paste.ubuntu.com/491000/ Sep 09 14:42:31 so it seems it's better somehow with the patch cooloney found, but still not fixed Sep 09 14:44:27 rsalveti: yes, there i still another issue Sep 09 15:34:33 ndec: ping Sep 09 15:36:54 all... I downloaded the OMAP4 maverick beta release and switch out the MLO , uboot and uImage for panda ES2.0 and I am able to get to the log in screen but I cannot seem to log in :( does anyone know the l/p for this? Sep 09 15:38:10 orbarron: http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100909/maverick-preinstalled-netbook-armel+omap4.img.gz should work fine for es2.0 Sep 09 15:38:23 the latest daily image Sep 09 15:38:40 guess ogra tested already, I'm still downloading it Sep 09 15:39:11 * orbarron have that... but can't log in Sep 09 15:40:12 ogra: ping Sep 09 15:40:19 orbarron: did you run oem-config? Sep 09 15:40:23 at your first boot? Sep 09 15:40:43 or was the login screen the first screen you saw? Sep 09 15:40:55 login was the firs thing I saw Sep 09 15:41:18 hm, ok, so you got the racing condition between oem-config and X Sep 09 15:41:19 followed --> https://wiki.ubuntu.com/ARM/OMAPMaverickInstall Sep 09 15:41:27 try to reboot it Sep 09 15:41:34 sometimes I'm also facing this issue Sep 09 15:41:56 happens when it tries to load but X11 is not ready yet Sep 09 15:42:38 same issue Sep 09 15:44:04 let me try to boot it Sep 09 15:44:43 * orbarron took the img from --> http://cdimage.ubuntu.com/ubuntu-netbook/ports/releases/maverick/beta/ Sep 09 15:46:39 orbarron: That is the beta image. Use today's daily build from http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/20100909/maverick-preinstalled-netbook-armel+omap4.img.gz Sep 09 15:46:56 * orbarron downloading it atm Sep 09 15:46:59 Beta does not support ES2.x Sep 09 15:46:59 thanks Sep 09 15:47:17 the image I pointed is the latest daily image Sep 09 15:47:25 seems you're having problem with it Sep 09 15:47:39 if you want to test beta image then you need to change mlo and kernel Sep 09 15:47:53 GrueMaster: were you able to test current daily already? Sep 09 15:47:56 for omap4 Sep 09 15:47:58 * orbarron switched out uImage, MLO, and uboot with ES2.0 Sep 09 15:48:19 Booting it now. It is working through OEM-Config on my 8L board. Sep 09 15:48:40 GrueMaster: hm, this racing condition doesn't happen all the time Sep 09 15:48:51 orbarron: try latest image and reboot it a couple of times Sep 09 15:48:53 * GrueMaster likes that he no longer needs to add parameters for hdmi<>DVI. Sep 09 15:49:09 gm prpplague Sep 09 15:49:15 I have seen it 1:10 boots. Sep 09 15:49:22 orbarron: boarding in a few minutes Sep 09 15:49:27 It appears more often on XM. Sep 09 15:49:48 orbarron: you get my picture? Sep 09 15:49:51 yeah Sep 09 15:49:54 hehe Sep 09 15:49:58 nice... Sep 09 15:50:00 nothing like a good breakfast Sep 09 15:50:01 still need to debug that, probably an easy fix Sep 09 15:50:49 prpplague: yeah if that is what you call it :p Sep 09 15:51:02 orbarron: hehe Sep 09 15:51:12 crap... no I am seeing omapdss DISPC error: SYNC_LOST_DIGIT Sep 09 15:51:16 * prpplague has been asleep Sep 09 15:51:20 hasn't Sep 09 15:51:37 orbarron: is that before or after the GFX error Sep 09 15:51:43 after Sep 09 15:51:54 saw this on lucid as well Sep 09 15:51:56 trying to start the webcam stuff? Sep 09 15:52:07 not yet... just trying to boot Sep 09 15:52:48 odd Sep 09 15:53:07 ok boarding, i'll be back online after i get to miami Sep 09 15:53:11 orbarron: later Sep 09 15:53:14 prpplague: laters Sep 09 15:55:24 ogra_cmpc: I have one good news and one bad news about the daily images for OMAP4. Which one do you want first? ;) Sep 09 16:06:15 orbarron: went to the oem-config screen fine here Sep 09 16:06:21 with latest daily image Sep 09 16:06:37 * orbarron testing daily image Sep 09 16:15:15 rsalveti: have you seen this error before? --> http://pastebin.com/Dn0thv1D Sep 09 16:15:34 * orbarron still using beta Sep 09 16:15:50 yep, couple of time with different monitors Sep 09 16:16:13 test with latest daily, should have latest kernel Sep 09 16:16:25 then if you still faces this issue, we need to debug the display driver Sep 09 16:16:30 will do Sep 09 16:29:00 GrueMaster: ogra: and about the fixed host name (based on your name) for oem-config? Sep 09 16:29:16 are we going to stay with this "feature" or should we add a bug for it? Sep 09 16:29:29 I think there is a bug already filed. Sep 09 16:30:26 I know we discussed it yesterday. Sep 09 16:31:16 cool Sep 09 16:33:08 * rsalveti out for lunch Sep 09 16:33:32 It is very aggravating for me. I currently have 8 potential "ubuntu-laptop" hosts. Trying to ssh into them based on avahi is no good. Sep 09 16:35:59 yeah it would be really nice if it pulled the mac address of the first ethernet port or whatever it finds Sep 09 16:36:44 at least the default can be ubuntu-laptop-last-6-digits or so Sep 09 16:37:00 interested parties will just press delete 7 times and go on their merry way :D Sep 09 16:39:38 vstehle, the bad ones Sep 09 16:40:09 Actually, having it default to - is nice for noobs, but it would also be nice to have the ability to change it. Sep 09 16:40:13 ogra_cmpc: On my Dell monitor, I have the "resize" step outputs, then reboot, then monitor goes black. Sep 09 16:40:28 vstehle, DSS2 bug i guess Sep 09 16:40:41 ogra_cmpc: Maybe... Sep 09 16:40:44 we dont touch any graphics config at that point Sep 09 16:41:10 GrueMaster, it's being worked on Sep 09 16:41:42 vstehle, dont forget the good news ! Sep 09 16:41:43 ogra_cmpc: ok. rsalveti was asking, that's why the thread. Sep 09 16:42:02 ogra_cmpc: The good news are that your image boots "untouched" :) Sep 09 16:42:14 ogra_cmpc: No bootloaders changes, no kernel replacement Sep 09 16:42:17 ogra_cmpc: No bootargs Sep 09 16:42:20 heh, yeah, here too Sep 09 16:42:32 and i dont have display issues with my samsung Sep 09 16:42:51 ogra_cmpc: I see what you mean :) Reminds me on Prague... Sep 09 16:42:58 heh Sep 09 16:43:15 well, i think there is still one DSS2 patch pending Sep 09 16:43:59 ogra_cmpc: Did you notice the boot partition "disappears" ? I mean that gnome lucid is not able to see it... Sep 09 16:44:06 ogra_cmpc: But a manual mount will do Sep 09 16:44:52 thats on purpose Sep 09 16:45:25 users should never touch it at all, all files you need are maintained in /boot Sep 09 16:45:56 ogra_cmpc: Oh? That's a nice "trick" then, if done on purpose. Sep 09 16:46:01 if you make changes to any file in /boot (i.e. boot.script to change the cmdline) you should use sudo flash-kernel Sep 09 16:46:16 it will update the hidden partition Sep 09 16:46:45 if we wouldnt do that, all users would always have an SD card on their desktop/disk list Sep 09 16:46:57 ogra_cmpc: I replaced the kernel and still black screen; I start to suspect the boot.scr... Sep 09 16:47:18 well, take a look at it, all we set there is vram=32M Sep 09 16:47:33