/srv/irclogs.ubuntu.com/2010/08/03/#ubuntu-arm.txt

dcordesGrueMaster, awesome. I didn't know about that option, thanks01:48
dcordesrsalveti, any news :) ?01:49
rsalvetidcordes: worked fine :-)01:52
rsalvetidcordes: installed the ubuntu-minimal on my sd card01:53
rsalvetiwith maveric01:53
rsalveti*maverick01:53
rsalvetiand then just installed the ubuntu-netbook package, with tons of dependencies01:53
rsalvetimono was included and did work fine01:53
dcordesrsalveti, oh nooo01:57
dcordes^^01:58
dcordesrsalveti, hm my first thought is something went wrong during my ubuntu-minimal image install01:58
rsalvetidon't know, could be, try it again01:59
dcordesno I won't01:59
dcordesit would be the third time01:59
dcordesand I will see the same error01:59
dcordesI must do something wrong01:59
dcordesI will boot the image now and see if I can work around it somehow01:59
rsalvetitry giving strace on the process that got hang02:01
dcordesapt-get02:01
rsalvetior running the update by hand, to find the problem02:01
rsalvetiI got the correct line that was giving problem on the bug, let me find it02:01
rsalvetidcordes: just /usr/bin/mono /usr/lib/mono/2.0/gacutil.exe was enough to generate the error with qemu02:02
rsalvetisee bug 61071902:02
ubot2Launchpad bug 610719 in mono (Ubuntu) "Mono hangs while running with qemu ARM user emulation (chroot) (dup-of: 530000)" [Undecided,New] https://launchpad.net/bugs/61071902:02
ubot2Launchpad bug 530000 in qemu-kvm (Ubuntu) "mono assembly installation under qemu-arm-static hangs (affects: 2) (dups: 1) (heat: 30)" [Medium,Won't fix] https://launchpad.net/bugs/53000002:02
rsalvetiI think you're probably getting another bug02:03
dcordesrsalveti, it is not the same mono bug I see in qemu.02:03
rsalvetidcordes: what arm machine are you using at your installation?02:03
dcordesthe native mono bug is a different one.02:03
dcordesrsalveti, cortex-a8 htcleo02:03
dcordesrsalveti, and host is x8602:04
rsalvetiit did work fine on my beagle, weird02:04
rsalvetihost should not be a problem02:04
rsalvetiubuntu-minimal is really minimal, quite small image02:04
dcordesrsalveti, how big is your compressed minimal-image ? could you upload the one you used in the test ?02:04
rsalvetilet me check02:04
dcordesok here is the line02:05
dcordesI ran dpkg --configure -a02:05
dcordes(cause apt tells me to)02:05
rsalvetiuploading it, will take some minutes02:05
rsalvetidcordes: do you have the logs?02:06
dcordes* Installing 2 assemblies from libappindicator0.1-cil into Mono02:06
dcordesthis the same line it froze during the initial 'apt-get install ubuntu-netbook02:06
dcordes'02:06
dcordesand I'm almost sure it is the same thing I got when I tried same procedure a week ago02:06
dcordesall of the installing n assemblies processes fail02:07
dcordesit is the mono program that freezes02:08
dcordesnow I remember...02:08
dcordesI debugged it before02:08
dcordesit eats all cpu02:08
dcordesis there no 'assume provided' kind of thing for apt ?02:10
dcordesso mono must be child of apt-get02:11
dcordesor dpkg in the current case right ?02:11
dcordesi.e. I should be able to debug it with dbg02:11
rsalvetiyep02:11
rsalvetithere is a command that installs the assemblies02:11
rsalvetiit's on the post install of some mono packages02:12
rsalvetiso you could debug it directly instead of debugging apt02:12
dcordescan you tell me how to figure the command that's ran ?02:12
rsalvetiI can try to find it here, give me a minute02:15
dcordesthx02:15
dcordesrsalveti, maybe my kernel is more similar like qemu :)02:16
dcordesrsalveti, lol I can't install strace02:19
dcordesguess why02:19
rsalvetidcordes: hahah, why?02:20
dcordesif you answer correct you get a free mono.deb02:20
rsalvetihaha, no idea, and I don't want a free mono.deb :P02:22
rsalvetithanks02:22
dcordesdpkg was interrupted. you must manually run 'sudo dpkg --configure -a' to correct the problem.02:22
dcordes:>02:23
rsalvetiouch :-)02:24
rsalvetiguess you can still install it by hand, giving dpkg -i02:24
rsalvetiafter downloading the package02:24
dcordesapt-get --download-only install strace to download ?02:25
rsalvetidcordes: check file /usr/share/cli-common/runtimes.d/mono02:26
rsalvetiit's a perl script that probably installs the assemblies02:26
rsalvetiget's called by /usr/share/cli-common/gac-install02:27
rsalvetiduring a post-inst02:27
dcordesI think I make it a hello world.02:27
dcordesrsalveti, how to download package with broken apt ?02:28
rsalvetidcordes: you can download it by hand, with wget :P02:29
rsalvetigenerally apt keeps complaining all the time if you canceled a previous run02:29
dcordesrsalveti, do you have url for the maverick arm debs ?02:30
dcordesI download it to usb mass storage device on host02:30
dcordesrsalveti, are you still uploading your minimal ?02:30
rsalvetidcordes: yep, 98%02:31
dcordesnice thx02:31
rsalvetidcordes: http://ports.ubuntu.com/ubuntu-ports/pool/main02:32
rsalvetidcordes: wget http://rsalveti.net/pub/armel-rootfs-201008021722.tgz02:33
dcordesrsalveti, something is definetly going wrong with my minimal then02:34
dcordesrsalveti, mine is t60 meg02:34
dcordes6002:35
dcordesrsalveti, do you specify extra packages with the roostock script ?02:35
rsalvetidcordes: this is just ubuntu-minimal with kernel and kernel modules02:37
rsalvetithat's why it's probably bigger than yours02:37
dcordesaha ok02:37
dcordesrsalveti, ok now I got the strace output on screen02:57
dcordesrsalveti, and it doesn't give me full enlightenment yet :D03:04
dcordesI'm still starring03:05
dcordesrsalveti, how long does such a mono install process take for you ?03:11
dcordesrsalveti, I made the script do nothing03:12
dcordesrsalveti, now some installing mono tasks succeed03:12
dcordessome still run and fail03:12
dcordesah it found the renamed original script!03:13
dcordesevil03:14
dcordesdone!03:14
rsalvetihaha :-)03:14
rsalvetigenerally it takes a minute or so03:15
dcordesmaybe I'm to impatient03:15
dcordesbut 20 minutes are pretty much foul03:15
dcordesmethink03:15
dcordesapt-get install netbook-launcher-efl03:16
rsalvetiGrueMaster: yep, I'm getting the same error lag is facing with his lg monitor04:04
rsalvetiprobably this is fixed already with robclark patches04:05
GrueMasterInteresting.04:05
GrueMasterI don't have an HDMI monitor, but will test this when I am at the QA sprint in a couple of weeks.04:06
rsalvetiGrueMaster: when are you going to be off for the qa sprint?04:06
GrueMasterI leave next Friday and will be in Oxford until the 22nd.04:07
rsalvetiGrueMaster: oh, ok :-)04:07
GrueMasterFood's cooking, gotta run.04:07
rsalvetisee ya04:07
robclarkfyi, for DVI monitors, I pushed a patch: http://gitorious.org/~robclark/pandaboard/robclarks-kernel-omap4/commit/37050de250e570d67a435cb36ecfa9763a98e5ff04:17
robclarkit won't make any monitor that wasn't working at all start working.. but it will make some DVI monitors that were defaulting to 640x480 pick a better resolution04:18
robclark(this at least helps with my DVI monitor at home ;-))04:18
=== michaelh` is now known as michaelh1
mythripklag:ping . do you see it always ?06:50
mythripkoops : *do you see it always ?06:50
lagmythripk: Morning07:41
mythripkmorning lag07:48
mythripkdid you see it again ?07:48
mythripki tried and could not reproduce07:48
lagI see different things depending on my cmdline07:48
mythripkcan you send your cmdline ?07:49
lagomapdss DISPC error: SYNC_LOST_DIGIT07:49
lagomapdss HDMI error: Failed to lock PLL07:49
lagquiet splash ro elevator=noop console=ttyO2,115200n8 vram=32M mem=463M fixrtc07:49
mythripkfailed to lock pll is because the first block is 1920 1080 with 138.5Mhz07:49
lagOkay07:50
mythripkthat will be fixed with the patches i sent you  , which will be part of out L24.9 release07:50
lagOkay07:50
lagSo they're not out yet?07:50
mythripknot yet! , will be by this friday07:50
lagHave they been approved?07:50
mythripkyes07:50
lagExcellent, well done07:51
mythripkthat is anyways our internal tree :) though07:51
lagI guess I'll wait for those before doing anymore07:51
lagI'll try and get them into our tree asap07:51
mythripkbut this is a strange error you see , it is failing in dispc_set_digit_size which would mean   x and y res are goofed up, in case you see that again a full log with debug enabled would be great07:52
lagI'm sure I can reproduce07:53
lagWait one07:53
lagmythripk: I don't get the whole log (I guess most of it is printed to the monitor (which isn't working)07:57
laghttp://paste.ubuntu.com/472531/07:57
lagmythripk: setenv bootargs vram=32M mem=463M console=ttyO2,115200n8 console=tty2 root=/dev/mmcblk0p2 fixrtc07:58
lagIf I remove the console=tty2 it does something different07:59
lagI correct myself08:02
lagIts actually this cmdline: setenv bootargs quiet splash ro elevator=noop console=ttyO2,115200n8 vram=32M mem=463M fixrtc08:02
lagThat's more like it08:04
lagWhen I remove the console=tty2 I get this:#08:04
laghttp://paste.ubuntu.com/472534/08:05
=== hrw|gone is now known as hrw
hrwalo08:06
lagogra: ping08:17
mythripklag : I suspect that it is becasue of the wrong x and y timing value. which ideally shouldnt , you are using console = ttyo2 which is ok  , this would be needed to redirect your prints to TV when enabled. Can you wait for the patch set to be pushed ?08:33
lagttyO2 won't push the prints to the TV/Monitor, only to the serial console08:34
lagtty2 will push them to the TV/Monitor08:34
mythripklag: my bad . "lag : When I remove the console=tty2 I get this:#" that would when it is trying to redirect the contents  to TV where you get the pll_not locked  state08:40
asacpersia: could you check http://revu.ubuntuwire.com/p/glmark2 for alf`?08:42
lagmythripk: If you read on you'll find that I was incorrect08:43
lagmythripk: In fact, the only think that differs when I removed console=tty2 is that I receive more output on the serial08:43
lagI'll wait for the new patches08:43
mythripklag: I shall update you when its pushed to our tree08:46
lagmythripk: Great thanks08:46
ukleinekericm|ubuntu: I would have expected from me to notice the breakage even though it was your patch :-)08:52
ericm|ubuntuukleinek, nah - you are fine09:16
amitkericm|ubuntu: have you worked with the Freescale guys in Shanghai?09:17
ericm|ubuntuamitk, a bit09:17
ericm|ubuntuamitk, but not much09:17
ericm|ubuntuamitk, knows some guys there as well as cooloney09:18
amitkericm|ubuntu: ok, one of them joined Linaro today (just introducing around)09:18
ericm|ubuntuamitk, in #linaro?09:18
ericm|ubuntuamitk, I missed that part09:18
amitkericm|ubuntu: yes09:18
cooloneyamitk: is he/she based in Shanghai?09:36
amitkcooloney: yes, come to #linaro to meet him09:36
ukleinekamitk: I have a mx51evm here and it seems to be broken with CONFIG_FIXED_PHY=y; CONFIG_MDIO_BITBANG=y; CONFIG_MDIO_GPIO=y09:51
ukleinekamitk: Did you see this already, too?09:52
cooloneyukleinek: hi, how's broken?09:53
amitkukleinek: unfortunately no, haven't had a chance to test mx51 for a few -rcs since my board broke. :-/ I just got a new one now so should be able to test.09:53
cooloneyukleinek: in our ubuntu fsl-imx51, those configs are all off,09:53
ukleinekamitk: the fec doesn't work so I cannot nfsrootboot09:54
amitkcooloney: you fixed the fec driver for MDIO support, right? Do you have time to look at this today?09:55
cooloneyukleinek: yeah, my patch was merged sometime ago and there are some updates in Dave's netdev-next tree09:56
cooloneyamitk: yeah, ^^09:56
cooloneybut too bad, i don't have the hardware now09:56
cooloneymy babbage is broken09:56
cooloneybricked09:56
amitkukleinek: I'll try to look at it as soon as I setup the new board09:58
ukleinekamitk: this issue is in my way, I can do it myself, too.09:58
amitkukleinek: ok, I can't help you right away (so you might consider looking at it)10:00
* ukleinek just wanted to check if it's a known (or even fixed) issue10:00
amitkukleinek: no it isn't (we've had HW availability issues). Thanks for the report10:01
ukleinekCONFIG_FIXED_PHY=y is the trigger10:04
ogralag, yes ?10:06
lagogra: Do you have a uboot which omits the memory error on XM?10:07
ograi have an yet untested patch http://cgit.openembedded.org/cgit.cgi/openembedded/diff/?id=b4c5ef7e0e06890b1369bfbd5c767820024adb21&id2=b738634ead43d9ebcc8f8a4840366528ec91045a10:08
cooloneyukleinek: pls take a look at this http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-lucid.git;a=commitdiff;h=0e5e6e2a981eeab61dcc184d51ab769a33af658910:10
cooloneyukleinek: and this http://bugs.launchpad.net/bugs/45787810:10
ubot2Ubuntu bug 457878 in linux-fsl-imx51 (Ubuntu Lucid) (and 1 other project) "imx51 on board ethernet plug/unplug events not detected (affects: 2) (heat: 15)" [Medium,Fix released]10:10
ukleinekcooloney: is this a fix or a work around?10:11
cooloneyukleinek: i simply turned that off.10:12
ukleinekcooloney: the plug/unplug event thing is fixed upstream10:13
cooloneyukleinek: CONFIG_FIXED_PHY will cause a sysfs conflict when I was trying to add phylib support in fec driver10:14
=== fta_ is now known as fta
=== ApOgEE__ is now known as ApOgEE
lagogra: What SD card are you using for your XM?10:41
=== amitk is now known as amitk-afk
=== fta_ is now known as fta
ogralag, traxdata 4G currently10:50
lagogra: Have you seen the "mmc0: USB HID whilst initialising SD card" issue?10:51
lagogra: Is the Traxdata a High Capacity (SDHC) card?10:53
cwillu_at_workhands out free samples of his slc 4gb flash crack :)11:02
lagcwillu_at_work: ?11:14
cwillu_at_worklag, I buy silly sd cards11:14
cwillu_at_workthey don't break when you pull power from them mid-write11:15
lagcwillu_at_work: The XM's kernel doesn't like my SanDisk Micro SDHC card :(11:15
cwillu_at_work!info ttf-larabie-straight11:16
ubot2cwillu_at_work: ttf-larabie-straight (source: ttf-larabie): Straight fonts from www.larabiefonts.com. In component multiverse, is optional. Version 1:20011216-1.1 (lucid), package size 3462 kB, installed size 7860 kB11:16
cwillu_at_workthat doesn't want to install on arm11:17
cwillu_at_workreports that it's a virtual package11:17
cwillu_at_workmakes me want to cry11:17
cwillu_at_work... or I just don't have multiverse enabled....11:18
cwillu_at_workif I give you ssh to my server, can you call my cell if the build breaks again?11:19
cwillu_at_workI need to go for a walk :/11:19
=== amitk-afk is now known as amitk
=== asac_ is now known as asac
lagogra: Am I correct in assuming you have fully booting XM and Panda boards?12:26
ogralag, no, XM doesnt12:30
ograi see some mmc issues but have lost the dmesg data for it before i could look12:31
ograit finds the mmc though, but the device is readonly12:31
ogra(sfdisk complains it cant open the device for writing)12:31
lagOkay12:33
lagI'm just compiling a kernel which should eradicate my -110 error12:33
lagOnce that's gone, I'll look into any other errors12:33
lagCan you send me the kernel logs you have?12:33
lagFor XM and Panda would be helpful12:34
ograwill do after the next image test (i just triggered a build)12:35
ograthe XM is totally trashed atm12:36
lagKernel error, or userspace?12:36
ograthe filesystem dissolved itself at some point after i did reset the system several times12:36
ograi think the root cause is a kernel or bootloader issue with the SD bus though12:37
lagWell if you can get me some logs, I can try to do something about it12:38
rcn-eeogra.. just fixed the ro bit last night.. http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/2.6.35-devel/annotate/head:/patches/rcn/xm-wp-testing.diff12:41
rcn-eeverified by another user, i'll clean up the patch a little.12:41
ogralag, ^^^12:42
ogracan i has that ? pleeeease12:42
rcn-eebasicly, the xm schematic has no wp, the current kernel defaults to a bx or cx non-existent wp line on the beagle.. (bad)12:42
ograhrm, yeah12:43
rcn-eebut there's another weird bug in 2.6.35 i'm going after.. on both my Bx and Cx boards, the mmc card is defaulting to 'ro' on boot.  Have you guys seen that too on your kernel?12:44
ograyes, thats what i described above12:45
ograit appears like its locked12:46
rcn-eeyeap, well that patch fixes the xm..  the Bx and Cx have a gpio issue, playing around with the write protect lever it seems to be ignored..12:46
rcn-eethis brings ti back to 2.6.34/2.6.33 behavor in my testing: http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/2.6.35-devel/annotate/head:/patches/rcn/regression-disable-wp.diff but i think there's bigger issues in the mmc write protect detection as it doesn't register the write protect lever in my testing..12:48
rcn-eeit's basicly revert ed8303fc111e58530e22bd29b0d7e08dced75999 introduced in 2.6.35-rc112:49
dyfetping: ogra13:05
lagrcn-ee: Have you seen the -110 error?13:10
lagrcn-ee: On XM?13:10
rcn-eelag, probally.. i see lots of errors on mine.. 'sudo aptitude' opps it for me.. (i really need to get production hardware.. very early proto "P7" full of extra solder wires for traces..)13:15
rcn-eei think that's the usb -110 error right?13:16
lagrcn-ee: "mmc0: USB HID whilst initialising SD card"13:16
lagYou'd know if you had it, because it dies on the way up13:16
lagIt doesn't find the SD card13:16
rcn-eethat's very weird.. mine finds the SD: (last saved dmesg) http://pastebin.com/4Dtysi8W13:18
lagIt must like your MicroSD card13:19
lagWhich one are you using?13:19
rcn-eejust generic sandisk..13:19
lagSDHC?13:19
lagOr SC?13:19
rcn-ee4Gb/2Gb sdhc..13:19
lagThat's interesting13:19
lagAre you using our kernel?13:19
rcn-eebut i almost think they are clones.. the adapters don't work and i had to use another one..13:20
rcn-eeyeap that's mine kernel..13:20
lagOkay13:20
lagOur kernel doesn't like SDHC cards13:20
rcn-eeweird...13:20
lagWe have to turn off preemption to get them to work13:20
rcn-eedoes it also fail with the Bx/Cx's?13:21
rcn-ee(sdhc)13:21
lagNo idea13:21
lagI don't have either of those13:21
lagI haven't heard of it13:21
lagOnly XM13:21
lagrcn-ee: bug 59194113:21
ubot2Launchpad bug 591941 in linux (Ubuntu Maverick) (and 1 other project) "SDHC card not recognized (affects: 2) (dups: 1) (heat: 80)" [High,In progress] https://launchpad.net/bugs/59194113:21
rcn-eethat actually might be 'too' cheap cards......  cat-ing my config, wondering if? # CONFIG_MMC_SDHCI is not set13:23
lagIt's a SanDisk13:24
rcn-eeyou guys don't off hand happen to know where the mmc keeps there error code list.. (looking 110)13:27
lagNot off hand, sorry13:28
ukleinekrcn-ee: 110 is usually ETIMEOUT13:29
rcn-eei found another reference in the plug computer forums, looked like a bad sandisk card..13:29
rcn-eethanks ukleinek13:30
lagI've seen it happen with others13:30
rcn-eewhich isn't good...  reading the bug report, it's pretty consistent, but is there a specific image i should test with my xm and collection of sd cards?13:31
lagPlease :)13:33
lagrcn-ee: http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/13:33
laggunzip it and dd it to your MicroSD card13:33
ukleinekcooloney: CONFIG_FIXED_PHY is enabled in mx51_defconfig13:35
rcn-eegot it..  i'll pull it down at work in a bit, just getting ready at the moment..  i'll log back in a couple hours with my findings..13:35
lagrcn-ee: Great, thanks13:35
rsalvetilag: any news regarding the problem you had yesterday with your lg monitor?13:43
rsalvetilag: I got the same problem with mine :-)13:43
rsalvetilag: with panda13:43
lagWhat's it doing?13:44
rsalvetilag: http://paste.ubuntu.com/472128/13:45
cooloneyukleinek: hmmm, but do we use CONFIG_FIXED_PHY in ARM system13:45
lagmythripk: ping13:46
lagrsalveti: Wait one13:46
lagrsalveti: What monitor are you using?13:46
rsalvetilag: lg with hdmi13:46
ukleinekcooloney: mx51 is the only defconfig that has it13:47
mythripklag: is this regarding the dump from rsalveti ?13:47
lagmythripk: Yep13:49
lagMine is LG with HDMI too13:49
lagrsalveti: Which model?13:49
rsalvetilag: W2253V13:50
cooloneyukleinek: if we don't need it, we can disable it in mx5113:51
lagrsalveti: Okay mine is W2261VP13:51
cooloneyukleinek: it looks like just for x8613:51
ukleinekcooloney: then it should depend on X86, no?13:51
rsalvetilag: any news regarding this bug? still didn't look for patches into other trees13:53
cooloneyukleinek: from the code, 'Fixed MDIO bus (MDIO bus emulation with fixed PHYs)'13:54
cooloneyukleinek: i am not sure about what is Fixed PHY13:54
* ukleinek isn't either13:54
lagrsalveti: Ask mythripk13:55
rsalvetimythripk: were you debugging this lg monitor issue?13:55
mythripkrsalveti: yes i was and the issue is fixed now but then the patches will be pushed to our tree only by this friday.13:58
rsalvetimythripk: and where can I find these patches? is it just an internal tree?13:59
mythripkrsalveti : The issue is because in the lg monitor block 0 has 1080P timing with 138.5Mhz which is not a standard13:59
mythripkits still in the internal tree13:59
rsalveti=\13:59
mythripkoh wait let me check robclark's tree14:00
mythripkhe must have pushed14:00
rsalvetiok14:00
rsalvetinice to know that my monitor doesn't follows the standard :-)14:02
mythripkhttp://gitorious.org/~robclark/pandaboard/robclarks-kernel-omap4/commit/f2fa54fcfe8fa09ad14f104ae64d1bb5c93982bc and this http://gitorious.org/~robclark/pandaboard/robclarks-kernel-omap4/commit/5553031b56322a7dbaa2f57b8f773be9ae2baaff14:02
mythripkrsalveti : other block timings are still ok :)14:03
rsalvetimythripk: nice, thanks for the links14:03
mythripkrsalveti : you dont have to give any bootargs with this if you were giving any for HDMI , just try and let me know.14:07
rsalvetimythripk: sure, will do14:07
rsalvetithanks14:07
ukleinekamitk: mx51 doesn't use IMX_IO_ADDRESS.  Do you prefer the mx51's current approach or did this just slip through?14:22
lagrsalveti: What are you going to do with that patch/14:25
lag?14:25
rsalvetilag: test it? :-)14:25
lagAnd then/14:26
lag?14:26
rsalvetilag: will let you all know if it worked or not14:26
lagOkay14:26
rsalvetiand then we can think if we're going to merge or wait for friday's tree14:26
lagThose patches should all be coming to us sooner or later anyway14:26
rsalvetiyep14:26
lagLet me know how you get on and we'll have a chat14:27
ograthe sooner the better14:27
rsalvetiyep14:28
lagogra: It matters not14:36
ogralag, it does14:36
ograwe'Re way behind on the omap arches and need the HW to work14:36
lagI'm still getting OOM14:40
lagon XM14:41
lagI thought this was a HW issue14:41
lagWe can only work with what we've got14:41
mythripklag: i guess mdelay is missed in the patch from rob's tree can you please point that out to rsalveti ?14:49
ogralag, yes, and there are a ton of patches for XM that surely arent applied to our tree14:50
ograsome crappier some not14:50
ograi know that people in #beagle do some work on booted XMs so we have to be missing a lot14:50
lagrsalveti: ping14:53
rsalvetilag: pong14:53
laghttp://kernel.ubuntu.com/git?p=lag/ubuntu-maverick.git;a=blobdiff;f=drivers/video/omap2/dss/hdmi.c;h=0830bbc6baed30e04afca7b1c1a5b83ec298c890;hp=a8bedb1facd324eb6d2f07767ae6399f7c24ec3a;hb=76dd0768b77f731bbe6ad4df55379734f3a38770;hpb=f6c95ac85cfa087a84cd82564dcaea8ce4a6c86714:55
lagmythripk says this is missing from robclark's patches14:55
lagInsure you add it, or your monitor won't work14:55
robclarkyes it is14:55
lagrobclark: Don't shoot the messenger14:55
amitkukleinek: MX51_IO_ADDRESS should be replaced with IMX_IO_ADDRESS at some point14:55
lag;)14:55
ukleinekamitk: ok, will do14:56
robclarkyou need that one from mythripk's patches :-)14:56
robclarkno shooting involved :-)14:56
robclark("yes it is" == "yes it is missing")14:56
robclarksorry, I noticed my wording wasn't so clear14:56
rsalvetilag: ok, will apply the patches and test it here, thanks for the link14:56
amitkukleinek: thanks14:56
lagrobclark: Ah, lost in translation (from US to English) ;)14:56
robclarkno.. I just haven't had my coffee yet this morning ;-)14:57
ogratexan vs australian you mean ?14:57
robclarkheheh14:57
* lag kicks ogra in the nether-regions14:57
ograouch14:58
ogrageez !14:59
lag:)14:59
ogralogging out on the panda steals my display14:59
ograno signal anymore15:00
ograah, its just slow15:00
=== bjf[afk] is now known as bjf
=== fta_ is now known as fta
=== fta_ is now known as fta
GrueMasterwhat adds flash-kernel to the /etc/kernel-img.conf file?  It is currently not enabled in the 20100802 image, so updating to the latest kernel doesn't run flash-kernel.16:28
loolGrueMaster: Usually, it's flash-kernel-installer, but it might be the image generation script16:29
GrueMasterThanks. I see it in the changelog as being added in May, but I see nothing in the code now.  Looking at actual diffs.16:34
GrueMasterlool: It looks like this was heavily discussed in the last few cycles, but never really resolved (that I can see).  See bug 365053.16:53
ubot2Launchpad bug 365053 in initramfs-tools (Ubuntu) (and 1 other project) "On armel (Babbage platform), kernel image upgrading breaks if Ubiquity is instructed not to install a bootloader (heat: 12)" [Medium,Triaged] https://launchpad.net/bugs/36505316:53
GrueMasterOf course this bug is moot with preinstalled images as we don't use ubiquity (except for oem-config).16:54
GrueMasterI would think that it should be up to the flash-kernel-installer.postinst script to ensure that it is added to /etc/kernel-img.conf.16:55
lagrcn-ee: Hey Robert16:56
=== fta_ is now known as fta
=== fta_ is now known as fta
=== hrw is now known as hrw|gone
ogra_cmpcGrueMaster, lool, kernel-img.conf isnt used anymore by flash-kernel (debian dropped it) flash-kernel is always run by update-initramfs now17:27
GrueMasterHmmm.  It didn't run on this update.17:27
ogra_cmpcdoes /etc/flash-kernel.conf exist ?17:28
GrueMasteryes17:28
ogra_cmpcthen there is nothing that could block it from being executed at least17:29
ogra_cmpcare you sure you got a new kernel ?17:29
ogra_cmpcwith this update i mean17:29
GrueMasterI manually ran flash-kernel and it updated the boot partition with the new kernel.  It just didn't run when installing the new kernel.17:29
GrueMasterYes.17:30
ogra_cmpcthats pretty weird since our kernel packages usually run update-initramfs from their postinst17:30
ogra_cmpcwhich in turn should run flash-kernel17:31
GrueMasterthe 20100802 image has 2.6.35-13-omap kernel installed.  Updating pulled in 2.6.35-14-omap.17:32
GrueMasterFlash-kernel should have run, but it didn't.17:33
ogra_cmpcreally the kernel or just meta ?17:33
GrueMasterkernel.17:33
ogra_cmpci didnt actually see a linux upload17:33
GrueMasterI'm looking at /boot17:33
ogra_cmpcbut there was a meta upload yesterday17:33
ogra_cmpcthats on a panda ?17:34
GrueMasterbeagle.17:34
ogra_cmpcah17:34
GrueMaster(hence th 2.6.35-14-omap kernel) ;P17:34
ogra_cmpcif you run update-initramfs, does it run flash-kernel ?17:34
GrueMasterchecking...17:37
GrueMasterdmesg17:38
GrueMasteroops17:38
ogra_cmpcheh17:38
GrueMasterI didn't see that it ran.  is there a log I can check?17:39
ogra_cmpci dont think so17:40
ogra_cmpcit should tell you it runs, its actually still very noisy17:40
GrueMasterNevermind.  timestamp on mmcblk0p1/uInitrd is an hour old.17:40
GrueMasterSo no, it isn't running.17:40
ogra_cmpchmm17:41
ogra_cmpcif you run: sudo flash-kernel --supported; echo $?17:42
GrueMasterGrrrr.  terminal reset.17:42
ogra_cmpcwhats the return value ?17:42
GrueMasterrunning flash-kernel manually works fine.  I already tested that to make sure (and rebooted to verify new kernel loads).17:42
GrueMasterThe return value is 0 (which I assume is good).17:43
ogra_cmpcyep17:43
ogra_cmpcmeans your HW is supported17:43
ogra_cmpcif flash-kernel --supported >/dev/null 2>&1; then17:44
ogra_cmpcflash-kernel17:44
ogra_cmpc...17:44
ogra_cmpcthats what update-initramfs executes17:44
ogra_cmpcurgh17:45
ogra_cmpci think i see the issue17:45
=== fta_ is now known as fta
GrueMasterIt doesn't even run that test.17:49
ogra_cmpcno, because there is still old code lool added once to check for the postinst_script value in kernel-img.conf17:51
ogra_cmpci thought i had dropped that when i merged the new flash-kernel17:51
GrueMasteroops17:51
ogra_cmpccomitted and pushed17:52
ogra_cmpci dont think its critical to upload it now though, we'll get just in the way of other arches ...17:52
GrueMasterok.17:54
GrueMasterAs long as it makes beta17:54
ogra_cmpcas a quick fiox you can just edit update-initramfs17:54
GrueMasterPlanned on it.17:54
ogra_cmpcjumpo to flash-kernel and remove the outer if statement17:54
ogra_cmpcthe one that checks for kernel-imf.conf contents17:55
=== fta_ is now known as fta
GrueMasteryep, that fixed it.17:58
ogra_cmpcgreat17:58
GrueMasterShould I post a bug against initramfs-tools to track this?17:59
ogra_cmpcnah17:59
ogra_cmpcthe fix is already in the tree17:59
GrueMasterIt's in the debian git tree.  Will we pull an update prior to Beta?18:00
ogra_cmpcnext image will have an up to date kernel and before the next kernel comes we'll have the update in18:00
GrueMasterok18:00
ogra_cmpcits in the ubuntu bzr branch18:00
GrueMasterOh.  Didn't find that.  Only found the debian git tree.18:00
GrueMasterDidn't look too hard either.18:01
lagrcn-ee: And now?18:07
=== fta_ is now known as fta
ogra_cmpcGrueMaster, http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/maverick/initramfs-tools/maverick/revision/210 btw18:14
GrueMasterYea, got it.18:15
GrueMasterogra_cmpc: Are we supposed to be generating netboot images for dove & imx?18:49
=== fta_ is now known as fta
=== bjf is now known as bjf[food]
ogra_cmpcGrueMaster, no, but i wont switch them off unless they cause failures19:46
GrueMasterJust emails from iso.tracker.19:46
ogra_cmpcthey are built automatically if debian-installer is built19:47
ogra_cmpcoh, they should be removed19:47
=== fta_ is now known as fta
=== hrw|gone is now known as hrw
=== fta_ is now known as fta
dcordeshi my friends20:35
rcn-ee_lag, same mmc bug as your guys xm boards: http://pastebin.com/R0tTmXjE20:38
rcn-ee_however this sandisk card works fine with my 2.6.35 kernel, so it's got to be a patch/config difference...20:39
=== bjf[food] is now known as bjf
dcordesrcn-ee, did you get the mmc from GrueMaster  ??20:40
rcn-ee_nope, off amazon. ;)20:41
rcn-ee_back in april when i realized i should stock up before everyone found out the xm's used micro sd cards. .;)20:42
GrueMasterWhat is it?  (class/size)20:42
rcn-ee_Sandisk 4GB SDHC20:42
GrueMasterShould work, although you have oops that I haven't seen yet.20:43
rcn-ee_Yeah, this XM rev P7 (256Mb) isn't 100% anyways.. ;)20:44
rsalvetircn-ee: what filesystem are you using by default?20:44
rcn-ee_oh this was lag's image he wanted me to test this morning.. just a dd of http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/20:45
rcn-ee_normally i rootstock my own over ext420:45
rcn-ee_for some reason my kernel 2.6.35-dl12 works fine with these sd cards, but ubuntu's kernel can't load the mmc.. since most of the xm patches are similar we are left with the config...20:46
GrueMasterThere is an updated kernel, but I am not sure how you would install it w/o booting.20:46
rcn-ee_that's what the bx is for.. ;) give me 5mins..20:46
rsalvetiwith a rootstock rootfs you can generate it, with qemu user emulation20:47
rsavoyercn-ee: I had to drop using your patched kernel :-(20:47
rsalvetior just mount the dist, copy qemu-arm-static there and install it with qemu user emulation20:47
rsalveti*disc20:47
rcn-ee_yeap.. i read that...20:47
rsavoye500Mhz sucks,,,20:48
rcn-ee_blah.. i've built kernels on a Bx at 400Mhz.. (debian etch 'arm')..20:49
rsavoyeI'm doing a build now, almost to the part that kept causing trouble, so I";ll let you know how it goes20:49
rsavoyebeats my 200Mhz Sharp Zarus...20:49
dcordesrsalveti, may I ask which zaurus you have ?20:50
rsalvetirsavoye: ^20:51
rsavoyeme, I've got several, the 3200 clamshell being my favorite20:51
rcn-ee_crap there goes the upgrading idea, my Bx's musb port isn't coming alive... ;)20:51
dcordesrsalveti, :)20:51
dcordesrsavoye, :)20:51
dcordesrsavoye, the akita is a lovely device20:52
rsavoyeI wound up doing a project a bunch of years ago for NASA, and we used 6000Ls, cause I could install ipsec20:53
rsavoyeI was pumping air traffic data over the network to it for monitor alerts20:54
slangasekdyfet: are you at DebConf?  They have you on the schedule giving a talk?21:06
=== hrw is now known as hrw|gone
dcordesrsalveti, nice. I love the nasa21:18
dcordesrsalveti, 6000 are the potrait orientation ones right ?21:20
dcordesoops21:20
dcordesrsalveti, sorry that was for the other guy also starting with rs21:20
rsalveti:-)21:20
rsalvetimythripk: now I'm finally able to test the patches, but didn't work that well :-(21:53
rsalvetinow I'm able to get something on the screen but the image is really big and distorted21:53
rsalvetiprobably still missing some patches from robclark21:53
rsalvetiand I don't have a console because of the oem-config, needs to generate a minimal image21:55
rsalvetianother weird thing is that I need to recreate the first partition every time I need to update uImage or uInitrd21:57
rsalvetiotherwise u-boot doesn't like it21:57
rsalvetiGrueMaster: did you ever make any kernel update with panda?21:57
prpplaguersalveti: how are you partitioning and mounting the sd card?21:57
rsalvetithis is the pre-installed image generated by ogra21:58
rcn-ee_that's a little weird..21:58
rcn-ee_standard fat16/32 partition?21:58
rsalvetiprpplague: mounting without any extra options at my desktop21:58
rsalvetiyep21:58
rsalvetifat 3221:58
GrueMasterrsalveti: not yet, but I found a bug in the initramfs-tools.  It won't run flash-kernel unless it is in /etc/kernel-img.conf Fix committed).21:58
rsalvetiif I just copy the new uInitrd or uImage file on top of the older one I'm not able to boot21:58
prpplaguersalveti: interesting21:58
rsalvetiu-boot will complain about it21:58
rsalveti** Unable to read "uImage" from mmc 0:1 **21:59
rsalvetimmc read: Invalid size21:59
rcn-ee_after updating..  then rebooting.. is "fatls mmc 0:1" still show a clean fat?21:59
rsalveti## Booting image at 80000000 ...21:59
rsalvetiBad Magic Number21:59
rsalvetiif I create the vfat fs again and copy the files, everything works21:59
GrueMasterI'll try here.  I thing I pulled an updated kernel during the massive updates I've done today.21:59
prpplaguersalveti: i normally only see that problem when the FAT index tables are corrupted or it contains data that can't be parsed by u-boot21:59
rsalvetircn-ee_: didn't try this yet22:00
rsalvetiprpplague: the weird thing is that once I created it again and try for the second time, I get the same error22:00
rsalvetiso it's not a problem on how ogra is creating it22:00
rcn-ee_rsalveti, it just seems very weird, i've been doing stuff like: http://pastebin.com/ccM2tDH5 for a good 2 years on my beagle..22:01
prpplaguersalveti: yea, i suspect that is some bug, i'll see if i can recreate the issue here and have a look22:01
rsalvetiargh, my screen size is totally wrong22:01
rsalvetircn-ee_: yep, but I'm doing this on panda, with beagle it works fine22:02
rsalvetiGrueMaster: try updating and running flash-kernel to see how it goes22:02
rcn-ee_ah sorry...  wrong board.. i still haven't got my panda's usb port to work yet for any upgrade.. ;)22:02
rsalvetiif flash-kernel is not called by default22:03
rsalvetiprpplague: could be something wrong with our u-boot version too22:04
* rsalveti still needs to get the u-boot hash we're using for omap422:08
prpplaguersalveti: i suspect there is a minor fat bug in the u-boot we are using22:12
prpplaguersalveti: it is rather old22:12
rsalvetiprpplague: oh, ok, do you know if we already have the fix somewhere?22:13
prpplaguersalveti: i've only seen the problem occur a couple of times, and i just assumed i had damaged the partition with my testing22:13
prpplaguersalveti: so, no, i doubt there is a fix22:14
prpplaguersalveti: i'll put it on my bug list to see if i can get to the bottom of it22:14
rsalvetiprpplague: do you know where can I find the "upstream" development of the uboot for panda?22:15
* rsalveti is just new regarding panda work22:16
rsalvetiGrueMaster: you may also know it22:16
GrueMasterNot sure, but it is part of the image, so the packaging tools should know.22:17
rsalvetiGrueMaster: L24.7git2010062422:20
prpplaguersalveti: sakoman is doing some contracting for TI to upstream all the new support to u-boot22:22
rsalvetihttp://git.omapzoom.org/?p=repo/u-boot.git;a=summary22:22
prpplaguersalveti: negative22:22
prpplaguersalveti: he has a repo somewhere22:23
rsalvetiprpplague: this link I guess is the one used for our package22:23
prpplaguersalveti: basically yea22:23
prpplaguersalveti: one sec, getting you some urls22:23
rsalvetibut I guess I have sakoman's repo too22:23
robclarkrsalveti: you need the patches related to framebuffer supporting resizing22:24
robclarkand you need to plugin monitor before xserver starts..22:24
robclark(because no KMS yet)22:24
rsalvetirobclark: the patches I got from mythripk made my hdmi work, but with a very wrong size22:24
robclark640x480?22:24
rsalvetirobclark: but I'm not even using X11 :-)22:24
robclarkis it a DVI monitor?22:25
rsalvetirobclark: probably something like that22:25
rsalvetihas dvi and hdmi22:25
rsalvetibut I'm using just the hdmi22:25
robclarkso there were a few patches that I think might be interesting to you...  hang on22:25
rsalvetirobclark: the fonts are huge and distorted (from console)22:25
robclarkis the picture scrambled, or just distorted?22:26
prpplaguersalveti: my core development has been to get the hardware verification done, now that it _is_ done, i'll be looking at getting all the code consoladated22:26
robclarkrsalveti: from here: http://gitorious.org/~robclark/pandaboard/robclarks-kernel-omap4/commits/L24.7_panda-hdmi-patches  I think you might want22:27
robclark"add callback to notify client of resolution change"22:27
prpplaguersalveti: do you care to get his ppt with slides describing the status, or just the url?22:27
robclark"register callback to get notified of resolution change"22:27
rsalvetiprpplague: could be the whole thing, np22:27
robclarkthat should at least give you an ok picture..22:27
robclarkand then for some monitors that are falling back to 640x480, when a higher resolution would be possible, "better support for DVI monitors"..  although I think that will mainly matter when you plug in over DVI..  it is basically adding support for parsing parts of EDID that HDMI screens don't seem to use22:28
prpplaguersalveti: you've got spam22:29
rsalvetiprpplague: nice, so the idea is to get most of this support directly on u-boot upstream22:30
rsalvetiprpplague: nice, thanks22:30
rsalvetirobclark: http://www.flickr.com/photos/rsalveti/4857717987/22:31
GrueMasterrsalveti: Ok, I am finally finished with the updating.  flash-kernel updates the boot partition, but somehow it is getting clobbered.  I think it may be a file order thing.  Will do some more testing.22:32
rsalvetiGrueMaster: ok22:32
prpplaguersalveti: that's a "feature" for people who have visual handicaps22:32
robclarkrsalveti: I suspect you need the patches related to resizing22:33
rsalvetifor sure :-)22:33
robclarkrsalveti: what you might want to do before spending much time, is just pull my branch and build it.. or send me an email addr and I can send you my uImage..22:34
rsalvetirobclark: will try some of your patches22:34
robclarkjust to try and see if that solves22:34
rsalvetirobclark: yep, that's why I'm doing now22:34
rsalveti*what22:34
robclarkthen you can try and merge on top of ubuntu kernel22:34
robclarkok, cool22:34
robclarkif that still doesn't work, email me the contents or /sys/devices/omapdss/display0/edid22:35
GrueMasterInteresting.  The fat partition has a date stamp of 1969-12-31 16:0022:35
rsalvetiGrueMaster: ops22:35
GrueMasterThe files are ok and have the correct timestamp.22:35
rsalvetiGrueMaster: when I tested the md5 was ok, but u-boot didn't like it22:36
rsalvetihad to recreate the fs22:36
GrueMasterDon't tell me we found another mkfs glitch.22:37
rsalvetiu-boot just didn't like it, probably22:39
GrueMasterWhat I am seeing is that when I mount the filesystem on my x86, the date of the mount directory becomes 12-31-1969.  That is a glitch somewhere in the filesystem, and I am willing to bet it is during fs creation.22:40
GrueMastermount22:40
GrueMastergah.22:40
GrueMasterNevermind.  I have a new SD that still has the factory format.  same issue there.22:42
rsalvetiyep, the bug happens every time with me22:43
GrueMasterback to my earlier theory.  It is a location issue on the drive.  testing now.22:43
GrueMasterrsalveti: I'm thinking it is an issue with the uboot fat code.  I don't think it can read beyond a certain point or something, as flash-kernel renames the existing files and creates new files.  That or uboot can't handle file fragmentation.22:52
rsalvetiGrueMaster: it's also happening when I just copy a new file on top of an older one22:53
rsalvetibut yes, probably on u-boot22:53
rsalvetistill needs more debugging22:53
GrueMasterCopying a new file over an old one without syncing could be a separate issue.22:54
rsalvetiGrueMaster: so currently if we get a new kernel on a pandaboard the panda is not going to boot anymore :-)22:54
GrueMasterYep.  That's what my test currently shows.22:55
GrueMasterI have another part of the test to run.  Give me a few.22:56
rsalvetiGrueMaster: ok22:56
rsalvetiGrueMaster: would you mind to open a bug later on about this issue?22:56
GrueMasterwill do after this reboot.22:58
rsalvetinice, thanks :-)22:58
GrueMastervery interesting.  As a test, I renamed the new kernel & initrd as uI*.old and renamed the old kernel & initrd from uI*.bak to uI* (shortened for explanatory reasons).  It still loaded the new kernel.23:08
rsalvetiinteresting23:10
GrueMasterI'm not even sure what that suggests.  iirc, the fat system should just change the entry in the directory table, not the order or the fat table starting point.23:10
GrueMasterOk, filing a bug, then doing more research.23:11
=== fta_ is now known as fta
=== bjf is now known as bjf[afk]
GrueMasterrsalveti: Bug #61323023:27
ubot2Launchpad bug 613230 in u-boot-omap4 (Ubuntu) "u-boot fails to load new kernel fromm boot partition after kernel update (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/61323023:27
rsalvetiGrueMaster: interesting, after renaming the old files it's still trying to load the newer ones23:30
GrueMasteryep23:30
GrueMastersomething is trashing the fat directory tree in uboot.  Try running "mmcinit 0;fatls mmc 0".23:39
GrueMasterugly23:39
rsalvetiprobably u-boot itself is trashing the fat directory tree23:40

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