/srv/irclogs.ubuntu.com/2009/09/18/#ubuntu-arm.txt

=== bjf is now known as bjf-afk
sijihow to create *kernel.jffs209:39
loolI have no idea what that is09:44
sijilool, am tryng to put both kernel and rootfs inside the nand memory09:48
sijiAnd i think nand memory will support only jffs2 file system09:49
sijiso need to convert both kerenl and rootfs to .jffs2 format09:50
loolsiji: What board is this on?09:51
sijiCompulab,CMX 30009:51
loolWhat SoC is that?09:52
sijiSoc means ??09:52
sijichip?09:52
loolYes09:52
sijimarvel pxa 30009:52
loolsiji: that's v5 right?10:04
loolsiji: Only jaunty supports that10:04
loolsiji: What bootloader does it use?10:04
sijiit's using compulab's customised u boot10:10
loolsiji: So I understand you can either partition the flash space to have a kernel and a rootfs zone and load them from there or you can put the kernel in the rootfs and tell uboot to load it from there (kernel in rootfs in flash)11:09
loolsiji: Did you look in the two options?11:09
sijilool, yes11:09
sijiit's taking the rootfs too11:10
ogralool, do you have any idea about http://paste.ubuntu.com/273349/ ?11:10
ograSep 17 17:59:48 in-target: update-initramfs: Generating /boot/initrd.img-2.6.31-100-imx5111:10
ograSep 17 17:59:48 in-target: readlink: missing operand11:10
ograSep 17 17:59:48 in-target: Try `readlink --help' for more information.11:10
ograSep 17 17:59:48 in-target: mkinitramfs: missing  root  /sys entry11:10
ograSep 17 17:59:48 in-target: mkinitramfs: workaround is MODULES=most11:10
ograSep 17 17:59:48 in-target: mkinitramfs: Error please report the bug11:10
loolsiji: Sorry can't parse what you said11:10
ogra        if [ -L /initrd.img -a -e /initrd.img ]; then11:10
ogra                linktarget="$(basename "$(readlink /initrd.img)")"11:10
ogra        fi11:10
ogra        if [ -L /boot/initrd.img -a -e /boot/initrd.img ]; then11:10
ogra                linktarget="$(basename "$(readlink /boot/initrd.img)")"11:10
ogra        fi11:10
ograneither of the initrd.img files exist11:11
sijilool,it's giving error kernel panic11:11
ograthe if statement is never fulfilled11:11
ograbut still update-initramfs calls readlink11:11
loolsiji: So your kernel gets loaded then11:12
sijiyes11:12
loologra: Are you sure it's failing on that spot?11:14
loologra: Can you run in-target update-initramfs -c manually with set -x?11:14
ograno, just moved forward and i see outfile="$(readlink -f "$outfile")"11:14
ograin mkinitramfs11:15
loolThere are more uses of readlink in update-initramfs than the one you quoted and the one you quoted seem sufficiently protected that they wouldn't fail11:15
ograand i think i fell for a red herring anyway11:15
ogra<ogra> Sep 17 17:59:48 in-target: mkinitramfs: missing  root  /sys entry11:15
ograthat must be the actual error11:15
loolsiji: So your issue is perhaps with your root= cmdline?11:15
sijican u just give me the cmd line11:16
loologra: Yes11:16
sijito boot frm nand11:16
loologra: hook-functions has for instnace sys_walk_mod_add which looks like it would fail if /sys isn't mounted11:16
ogra if [ -z "${block}" ] || [ ! -e /sys/block/${block} ]; then11:17
ograecho "mkinitramfs: missing ${block} root ${root} /sys entry"11:17
ogracrap so d-i (or rather base-installer) didnt mount sys11:17
loologra: well that error is afterwards11:17
loologra: Run with -x11:17
loolsiji: I'm not sure11:18
loolsiji: What's your kernel version?  I think it's something likeroot=jffs2/rootfs or something11:18
ogralool, well, /sys isnt mounted in /target11:18
ograso update-initramfs cant find out which modules it needs11:18
loolsiji: http://www.denx.de/wiki/DULG/RootFileSystemOnAJFFS2FileSystem11:18
loologra: Did you verify that was the issue or is it an hypothesis?11:19
ogralool, i see that /sys isnt mounted11:19
ograi'm still in the d-i shell directly after the failure11:19
ogra~ # mount11:19
ograrootfs on / type rootfs (rw)11:19
ogranone on /proc type proc (rw,relatime)11:19
ogranone on /sys type sysfs (rw,relatime)11:19
ogratmpfs on /dev type tmpfs (rw,relatime)11:19
ogradevpts on /dev/pts type devpts (rw,relatime,gid=5,mode=620,ptmxmode=000)11:20
ogra/dev/sda1 on /target type ext4 (rw,relatime,errors=remount-ro,barrier=1,data=ordered)11:20
ogra/dev/sda1 on /dev/.static/dev type ext4 (rw,relatime,errors=remount-ro,barrier=1,data=ordered)11:20
sijilool,yes the kernel is in jffs2 format11:20
ogratmpfs on /target/dev type tmpfs (rw,relatime)11:20
loologra: That might be the issue; I wonder if it's a recent change and whether it affects other arches11:21
ogralool, i doubt that, its the A6 candidate image11:21
ograthat would have shown up during testing on other arches11:21
loologra: I dont see the architecture specific issue though11:22
ograme neither but clearly neither /proc nor /sys are actually mounted11:22
ograwhich is the cause for the error i see11:23
loologra: So you verified it's enough to mount /sys to fix the install?11:25
ograto bad i'm on serial, else i could repeat the whole thing and watch it from a second console11:25
ograno, i cant easily do that11:25
ograif i start over now, it will reformat /target and start over with a base install11:26
ograand i dont have a second console to manually mount /sys11:26
loologra: You could go to target, dpkg -C and see if mounting /sys resolves it11:26
ogra-C ?11:28
ograyou mean --configure -a11:28
* ogra tries so ... i didnt touch /target yet to not taint the status of it 11:28
ograits hard to get to the same point again (having to copy around firmware and modules for the netinstall to even work)11:29
loologra: Perhaps it would be best to defer until next debian-installer upload?11:29
ogra# dpkg --configure -a11:30
ograSetting up linux-image-2.6.31-100-imx51 (2.6.31-100.7) ...11:30
ograRunning depmod.11:30
ograupdate-initramfs: Generating /boot/initrd.img-2.6.31-100-imx5111:30
ograreadlink: missing operand11:30
ograTry `readlink --help' for more information.11:30
ogramkinitramfs: missing  root  /sys entry11:30
ogramkinitramfs: workaround is MODULES=most11:30
ogramkinitramfs: Error please report the bug11:30
ograupdate-initramfs: failed for /boot/initrd.img-2.6.31-100-imx5111:30
ograFailed to create initrd image.11:30
ogra:/11:30
ograsys is mounted11:30
ograand i can list /sys/block11:30
ograhrm11:30
ograerr11:30
loolWasn't that obvious that it was /sys   ;-)11:31
loologra: How about you set -x ?11:31
ograone sec11:32
ograthere are supposed to be filled variables before root and after /sys in "missing  root  /sys entry"11:32
loolHow about you set -x ?11:33
lool:-P11:33
ograhttp://paste.ubuntu.com/273378/ ...11:37
* ogra tries in mkinitramfs11:37
loologra: in mkinitramfs11:37
loolright11:37
ograhttp://paste.ubuntu.com/273380/11:38
ogra# ls /dev/root11:39
ograls: cannot access /dev/root: No such file or directory11:39
ograaha11:39
sijihey lool,11:40
sijimtd0: 00040000 00020000 "OBM"11:40
sijimtd1: 00040000 00020000 "U-Boot"11:40
sijimtd2: 00040000 00020000 "Environment"11:40
sijimtd3: 00140000 00020000 "reserved"11:40
sijimtd4: 00400000 00020000 "kernel"11:40
sijimtd5: 1fa00000 00020000 "fs"11:40
sijithis is the nand partition table11:40
ogramust be somewhere between line 218 and 22611:41
ograhmm, it uses the output from "mount"11:42
ograwhich is empty11:42
ogra(well empty beyond sys and proc)11:42
loologra: $(mount | awk '/\/dev\// {if ($3 == "/") {print "root=" $1 "\nFSTYPE=" $5; exit}}')11:42
ograyeah11:43
loolsiji: Read that wiki page I mentionned?11:43
sijiok11:43
sijioh,it was for me :)11:43
ogra# mount11:43
ograproc on /proc type proc (rw)11:43
ograsysfs on /sys type sysfs (rw)11:43
loologra: see dep_add_modules() in hook-functions11:43
sijii thought that it's for ogra :)11:43
loologra: so where are you installing to?11:43
ogra/dev/sda111:44
sijisrry11:44
loologra: can you compare with /proc/mounts?11:44
ograUSB key11:44
ogra# mount11:44
ograsysfs on /sys type sysfs (rw)11:44
ograproc on /proc type proc (rw)11:44
ogra# cat /proc/mounts11:44
ograrootfs / rootfs rw 0 011:44
ogranone /proc proc rw,relatime 0 011:44
ogranone /sys sysfs rw,relatime 0 011:44
ogratmpfs /dev tmpfs rw,relatime 0 011:44
ogradevpts /dev/pts devpts rw,relatime,gid=5,mode=620,ptmxmode=000 0 011:45
ogra/dev/sda1 / ext4 rw,relatime,errors=remount-ro,barrier=1,data=ordered 0 011:45
ogra/dev/sda1 /dev/.static/dev ext4 rw,relatime,errors=remount-ro,barrier=1,data=ordered 0 011:45
ogratmpfs /dev tmpfs rw,relatime 0 011:45
ograproc /proc proc rw,relatime 0 011:45
loologra: So your mtab isn't matching /proc/mounts; that code would work with /proc/mounts11:45
loologra: Now this has no reason of being architecture specific11:45
ograright11:45
loologra: Is /etc/mtab different outside the target?11:45
ograunles mount has a bug11:45
ogramtab is identical to /proc/mounts outside target11:46
ograand ot the mount output inside11:46
loolSo this actually looks like an installer bug but I cant figure why it's only hitting us11:47
ograi initially starte because my /vmlinuz symlink isnt created11:47
lool?11:47
ograand i think i heard NCommander say something like that a wek ago11:47
loolI dont get what you mwan11:47
ograi saw the readlink error11:48
ograand started debugging the kernel postinst, since i was suspecting the missing /vmlinuz and /initrd.img to be at fault11:48
ograthen moved forward to update-initramfs11:48
ograto then get to mkinitramfs and the missing /sys ...11:49
ograwhich now got me to the hooks11:49
ograthe point is that the obvious thing seems to be the missing symlinks11:49
loolOh ok if you sya so11:49
ograthats the error on the surface ...11:49
ograand thats what NCommander dug in a week or two ago in his manually spun dove images11:50
ogracrawling through all of d-i if i understood that right, to find why his symlinks were missing11:50
ograso it seems to happen on dove as well11:51
loollunch11:55
=== cbrake_away is now known as cbrake
Stskeepshmm, is there a typical ubuntu rootfs tar.gz laying about that just uses fbdev?12:16
loolStskeeps: I dont think we publish any regularly; we only build d-i images regularly; but you could easily build your own with rootstock12:43
loolIt might be a bit borken since some days due to upstart12:43
Stskeepsk12:43
loologra: 427199 isn't on the report but is milestoned13:48
lool(from https://bugs.launchpad.net/ubuntu/karmic/+bugs?field.milestone=12715 )13:48
ograah, thanks, missed that one+13:49
ogralool, thats all ?13:52
loologra: You can say "UNR theme" for the title of 43027713:52
ograplars, can i strike mobile-unr-karmic-compliance-autotesting  from https://wiki.ubuntu.com/MobileTeam/ReleaseStatus/Karmic ?13:53
loologra: I'm slow because I'm chatting with telepathy people too13:53
ograyep13:53
ograsaw that in the other chan.13:53
loologra: 430344 isn't on the report either13:53
loolbut is on the milestoned bugs list13:53
loologra: https://bugs.launchpad.net/ubuntu/+source/usplash/+bug/42334613:57
ubot4Launchpad bug 423346 in usplash "[armel+dove] usplash does not currently work on Marvell Dove Y0 boards" [High,New]13:57
loologra: NCommander forgot to track in karmic...13:57
ograok13:57
loolNCommander: It's really CRUCIAL that you get that right13:57
ogralool, do we really belive its still worth to track banshee ?14:02
* ogra thinks that can be dropped from the report14:02
ograwhile it's nice to have and i still hope dyfet finds the issue its really not release critical or even intresting for the release meeting14:03
ogralool, wrt bug 424400 it might also be firmware thats missing ... though only GrueMaster can actually debug and confirm14:07
ubot4Launchpad bug 424400 in linux-fsl-imx51 "DM9601 or Pegasus based USB NICs dont find their MAC address under 2.6.31-100-imx51" [Low,Incomplete] https://launchpad.net/bugs/42440014:07
loologra: That's an idea14:07
ogra[ubuntu/karmic] linux-fsl-imx51 2.6.31-101.8 (Accepted)14:07
* ogra dances !!!14:08
ogramy second kernel this release cycle, yay !14:08
ograhmm, based on rc814:08
ograwhat made me think i get something based on final :P14:08
sijihey how to get the terminal/login window in both lcd and minicom14:09
plarsogra: I marked it complete14:10
ograplars, so i can drop it, right ?14:10
plarsogra: yeah14:10
ogra(i saw its marked implemented, but its still on the release team report)14:10
ogracool14:10
ogra * ARM: IMX51: ENGR00103310 Initial imx_nfc NAND Flash MTD driver14:12
ogralool, ^^^^14:12
ogra!14:12
ogra* [Config] Run updateconfigs after rebasing to 2.6.31 final14:13
ograah14:13
loologra: Cool14:15
ograha, and i can drop at least one bug from the report again :)14:15
ograBug 359049 just closed itself14:15
ubot4Launchpad bug 359049 in linux-mvl-dove "imx51 udeb hardcodes linux version in vmlinuz binary name" [High,Fix released] https://launchpad.net/bugs/35904914:15
loologra: I usually keep the bugs which are mentionned on the email or reports and mention that they are closed on the mobile team report14:20
ograwhat are you referring to ? banshee ?14:20
loologra: 422101 > removing the milestone?14:21
loologra: 15:15 < ogra> ha, and i can drop at least one bug from the report again :)14:21
loologra: I dont drop closed bugs14:21
loolI just mark them closed in the report14:21
ograah, k14:23
* ogra sighs, i dropped them all already14:23
* ogra tries to find them again and re-adds14:23
loologra: https://bugs.launchpad.net/bugs/42546414:24
ubot4Launchpad bug 425464 in unr-meta "Favicon for site being bookmarked should be used as part of the icon in the Launcher" [Medium,Triaged]14:24
loologra: updating milestone14:25
loologra: *I* forgot to track https://bugs.launchpad.net/ubuntu/+source/unr-meta/+bug/425464 in karmic *sigh*14:27
ubot4Launchpad bug 425464 in unr-meta "Favicon for site being bookmarked should be used as part of the icon in the Launcher" [Medium,Triaged]14:27
loolActually i recall killing the browser before launchpad loaded on this bug14:27
loolbut well14:27
ograadded14:27
loologra: Hmm did you update the milestoned features??14:28
ograi removed them14:28
ograthey are all set to implemented14:28
loologra: https://blueprints.launchpad.net/ubuntu/+spec/mobile-qa-karmic-unr and https://blueprints.launchpad.net/ubuntu/+spec/mobile-qa-karmic-arm are milestoned14:28
loologra: I'm not sure you're looking at the correct data14:29
loologra: Goal is to a) move stuff from previous milestone to next milestone and b) comment on everything on the next milestone14:29
loolhttps://bugs.launchpad.net/ubuntu/+milestone/ubuntu-9.10-beta14:29
loolYou can see two specs of plars there14:29
ograwell, i looked at the three specs that were left from last week in the report14:29
looland the moblin remix one14:29
ograall three were marked implemented14:30
loologra: The ones from last week were for A614:30
loologra: Now you need to also list stuff for next milestone14:30
ograand marked implemented when i looked 1h ago14:30
loologra: the three specs for ubuntu-9.10-beta are not marked implemented no14:30
ograno, the three that were on the report before are14:31
loologra: Actually there's even your spec on it14:31
loologra: Well that's good14:31
ograwtf ... how can they show up if they are implementd14:31
loologra: If the specs we intended to be implemented in A6 were implemented then we can remove them from the report and not move the milestone on them; that's considered normal if you like14:31
loologra: I think there's a misunderstanding14:31
ograright14:31
ograi think there is a misconception in LP14:32
loologra: You need to drop everything from the report which was targetted at A6 and implemented14:32
ograright, that i did14:32
loologra: You need to move everything which was targetted at A6 and was NOT implemented to ubuntu-9.10-beta14:32
loologra: And you need to report on everything which is targetted at ubuntu-9.10-beta14:32
ograok, i thought the spec owners do that14:32
loologra: The spec owners give you a one line summary in their blueprint which gets agregated in the burndown report14:33
ogralool, how did you get any work done on fridays in the past14:33
loologra: It takes me about 1.5 hours to get the report written; I'm glad you see it's a lot of work   :-)14:33
ograthats eating half the day14:33
loolI didn't do much on fridays indeed14:33
loologra: That's also why I insist on getting this blueprint stuff in shape, the bugs properly milestoned and implemented the status in blueprints in the burndown chart14:34
ograyeah, i understood that before even assembling this report :)14:35
ograbut i see how painful it is if they arent up to date14:35
loologra: So I think I walked on the A6 stuff; you can just check the https://bugs.launchpad.net/ubuntu/+milestone/ubuntu-9.10-beta stuff14:36
loologra: You want to mention status of the 4 specs14:36
loologra: And poke plars since his specs are Started only14:36
loolplars: ^14:36
loolplars: What's the progress on the testcases specs?  Should we say slow progress?  Good progress now?14:36
loolplars: they are currently "Started"14:37
loolplars: https://blueprints.launchpad.net/ubuntu/+spec/mobile-qa-karmic-arm says you are bringing up your babbage board but that was a while ago14:37
plarslool: well14:41
loologra: So what I do is that I "union" bugs and blueprints from: A6 which need to be moved, beta, the ones mentionned in the report, the bug lists we review during our mobile irc meetings14:41
plarslool: I don't feel right about saying beta available14:42
loologra: I spotted one more blueprint in the agenda: UbuntuSpec:mobile-karmic-marvell-desktop; you might want to comment that it was implemented in A614:42
ograok14:42
plarslool: because I have the wiki testplan up, but not a bzr repo with the tests integrated at this point14:42
loolUbuntuSpec:mobile-unr-karmic-application-res also shows up not sure why14:42
loolplars: That's ok but inbetween we could say Slow progress or Good progress14:43
plarslool: I can change the status to something else if you'd like... what does it affect?14:43
loolplars: I see it like: Started, Slow progress, Good progress, Beta, Implementeed14:43
ogralool, yeah, i was wondering what to do with app-res14:43
loolplars: It will allow answering the release team's concerns if they ask where we stand14:43
loolplars: Since "started" might worry them14:43
plarsok14:44
ogramake it good progress ;)14:44
ogralool, think i can drop app-res ?14:44
plarsI'll mark them both good progress, I've done more work on the unr one than the arm one, but a lot has gone into checkbox already (as expected) that benefits both of them14:46
plarsI wanted to get my A6 one finished up also, which also benefits UNR in particular.  Now that that's done, I'm going to be focused on these two14:47
ogralool, any suggestion what to do with application resolution ? i'm not sure what to say about it or why its on the report at all14:50
* ogra gets some more coffee14:53
loologra: sorry had a phone call14:55
loologra: I recommend you keep app-res and I'll clear things with paulliu when it's daylight time for him14:56
loologra: I think the remaining bugs are upstream ones but still we need to update its state; sorry I didn't do that earlier this week14:56
ograremaining bugs ?14:59
ograso what do i write for app-res ?15:00
loologra: in the app res spec15:00
ograah15:00
loologra: That most bugs are upstream one and we will look at reducing its scope next week15:00
ograwell, i'm happy with a summary line to add to the report ...15:00
ograok15:00
loologra: you can take summary lines from http://piware.de/workitems/mobile/karmic/report.html when that's up-to-date15:00
ograwhich it isnt :P15:01
loolWell the report is, the spec wasn't kept up-to-date15:01
ograso https://wiki.ubuntu.com/MobileTeam/ReleaseStatus/Karmic looks ok to you now ?15:02
loologra: status for #423767 seems out of date15:03
ograwell, who is pushing to linux-firmware then ?15:04
loologra: It says "pending push from amitk but driver doesn't work and needs porting"15:04
ograyeah, i looked at the bottom15:04
ograits irritating that we have them twice15:04
loolI think it should say something like "Needs kernel love to add firmware from upstream" or smth15:04
loologra: agreed15:05
* ogra moves description from the bottom one to the top one as well15:05
ograDove squahfs's ...15:05
ograhhe15:05
* ogra fixes NCommanders typos i copy/pasted15:05
loologra: 425547 > we're blocked on telepathy-mission-control-5 port of moblin stuff but I just clarified the plan with upstream15:06
ograok15:07
loologra: there's a typo in the bug number 427199 for dove squashfs15:08
loolin the report it says #2719915:08
ografixed15:08
loologra: lool started testing on the ffmpeg bug if you like15:09
loologra: status of #420555 is out of date15:09
loolYou might want to mention FSL is looking into it15:10
ograi updated it though15:10
loologra: Ok that's all I had15:10
ogramight not be massively obvious, i just mangled your sentence15:11
ograthanks, will add the changes15:11
loologra: Looking at general status stuff15:11
lool"inmportant" typo15:11
loologra: I would mention addition of installer support in dove and the oo.o issue in the sections at the bottom15:12
loologra: and moblin remix being blocked on the TMC 5 port15:12
loologra: Perhaps the ooo issue can go in armel general status15:12
loologra: As a reward I'll share something with you in some minutes15:15
ograupdated15:16
loologra: email sent15:18
ograthanks :)15:18
=== bjf-afk is now known as bjf
GrueMasterSo, I heard my name, metaphorically speaking.16:10
=== Meiz__n810 is now known as Meiz_n810
=== cbrake is now known as cbrake_away

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