[17:11] <lool> bjf: https://pastebin.canonical.com/22225/ > looks like uboot cant read from IDE which it should; try powering off completely and booting again
[17:12] <ogra> woah, thats niosy
[17:12] <bjf> lool, GrueMaster just reported success
[17:12] <GrueMaster> I've got it working.  Will try to pastebin my minicom log.
[17:13] <bjf> lool, same after full power off
[17:15] <lool> bjf: perhaps it's too slow
[17:15] <lool> bjf: Can you try manually running the bootcmds?
[17:15] <lool> bjf: So break the script, printenv, and try the commands in the script but slowly
[17:15] <lool> ** Unable to use ide 0:0 for fatload **
[17:15] <lool> ** Partition 0 not valid on device 0 **
[17:15] <lool> ** Unable to read "/boot.scr" from ide 0:1 **
[17:15] <lool> Scanning ${interface} ${device}:${partition} ...
[17:15] <lool> reading /bootboot.scr
[17:16] <lool> bjf: hold on
[17:16] <lool> bjf: that's the issue
[17:16] <GrueMaster> I think I see what's happening.  The script is cycling through several usb iterations before booting from sata.
[17:16] <lool> bjf: did you install with a separate /boot?
[17:16] <lool> GrueMaster: bjf: Can one of you guys show me the full bootcmd as returned by printenv?
[17:16] <bjf> lool, I had the installer "use entire disk" and do what it wanted
[17:17] <lool> The issue is likely in NCommander's UBoot
[17:17] <lool> bjf: Ok; I'm not sure what it does on dove
[17:17] <bjf> bootcmd=setenv file_load '${fs}load ${interface} ${device}:${partition} 1000 ${prefix}boot.scr';setenv fat_load 'setenv fs fat; run file_
[17:17] <bjf> load'; setenv ext2_load 'setenv ext2; run file_load'; setenv attempt_load 'run fat_load || run ext2_load'; echo Scanning for boot device
[17:17] <bjf> ...; usb start; ide reset; for i in usb ide; do for j in 0 1 2 3; do for k in 0 1 2 3; do for l in / /boot; do setenv interface $i; seten
[17:17] <bjf> v device $j; setenv partition $k; setenv prefix $l; echo 'Scanning ${interface} ${device}:${partition} ...'; if run attempt_load; then ec
[17:17] <bjf> ho Found boot.scr, checking ...; if imi 1000; then echo boot.scr OK! Executing ...; autoscr 1000; fi; fi; done; done; done; done; echo No
[17:17] <bjf>  boot device found. Restarting; sleep 3; reset;
[17:18] <bjf> lool, i'll pastebin an entire printenv
[17:18] <lool> bjf: thanks
[17:18] <bjf> lool, https://pastebin.canonical.com/22226/
[17:19] <lool> bjf: the issue is bootcmd
[17:19] <lool> ${prefix}boot.scr should be ${prefix}/boot.scr
[17:19] <bjf> lool, I don't see "prefix" defined
[17:19] <lool> or rather "for l in / /boot" should be "for l in / /boot/"
[17:19] <lool> bjf: prefix is set in the loop
[17:19] <lool> setenv prefix $l
[17:19] <GrueMaster> lool: https://pastebin.canonical.com/22227/ is my minicom capture from poweron.
[17:19] <lool> bjf: change your bootcmd to "for l in / /boot/" and saveenv
[17:20] <ogra> bug 431963
[17:20] <ubot4> Launchpad bug 431963 in linux "io/fs errors when launching gdm on imx51 with sata" [Undecided,New] https://launchpad.net/bugs/431963
[17:20] <lool> GrueMaster: You're using the / prefix
[17:21] <lool> GrueMaster: and it works for you, great!
[17:21] <GrueMaster> brb.  household issue.
[17:22] <lool> bjf: Did this help?
[17:22] <bjf> lool, taking me a minute to the the change in
[17:22] <lool> Ok :)
[17:25] <bjf> lool, is there an easy way to edit that line or do I need to cut and paste a new one in?
[17:26] <lool> bjf: You need to cut and paste
[17:26] <bjf> lool, ok that's what I thought
[17:28] <NCommander> lool, that's an old version of th eboot script
[17:30] <NCommander> lool, in the bootscript, it needs to be {prefix}boot.scr
[17:32] <GrueMaster> that's what I have as well.
[17:33] <GrueMaster> I initially thought it was stuck in a loop, but it is just looping through all of the non-existant usb devices before going to sata.
[17:33] <lool> Yes
[17:34] <bjf> NCommander, lool, so do I need to modify mine or not?
[17:34] <NCommander> bjf, no, just update to the latest binaries, and be patient :-0
[17:34] <NCommander> *:-)
[17:34] <bjf> NCommander, did that
[17:34] <bjf> NCommander, at least I think I did that
[17:35] <NCommander> GrueMaster, lool, its unfortunately pretty slow. I need to extend the usb/ide subsystem to throw an environment variable so I can iterate on extactly how many devices there are, and then try to see if I can figure out another method to determine the fs then forcibly load ext2/fat
[17:35] <GrueMaster> bjf: It takes about 20 seconds to run through the loop.
[17:35] <NCommander> GrueMaster, 20?, it only takes about 5 here
[17:35] <bjf> GrueMaster, if you look at my pastebin, you see it goes through the entire loop
[17:36] <bjf> NCommander, I'm checking what I did to see where I went wrong
[17:36] <GrueMaster> I'm not digging up a stopwatch for exact timing.
[17:36] <NCommander> bjf, whats the version string? (its posted at the start)
[17:36] <NCommander> GrueMaster, fair enough. I do know its an issue that I didn't realize quite how bad it was until we were doing full blown installer testing :-/
[17:37] <bjf> NCommander, U-Boot 1.3.4-dirty (Sep 16 2009 - 20:05:29) Marvell version: 4.2.3 NQ.  Canonical Development Version.
[17:37] <GrueMaster> Not a big problem.  If I didn't have a serial port hooked up, I would think it is just doing it's normal bios boot (like x86 systems).
[17:37] <NCommander> bjf, interrupt the autostart, then type ide reset and see if your drive pops up
[17:38] <bjf> Canonical>> ide reset
[17:38] <bjf> Reset IDE:
[17:38] <bjf> Marvell Serial ATA Adapter
[17:38] <bjf> Integrated Sata device found
[17:38] <bjf> [0 0 0]: Enable DMA mode
[17:38] <bjf>   Device 0 @ 0 0:
[17:38] <bjf> Model: ST9250410AS                              Firm: 0002SDM1 Ser#:             5VG06NA2
[17:38] <bjf>             Type: Hard Disk
[17:38] <bjf>             Supports 48-bit addressing
[17:38] <bjf>             Capacity: 238475.1 MB = 232.8 GB (488397168 x 512)
[17:38] <NCommander> bjf, try: ext2ls ide 0:1
[17:39] <bjf> Canonical>> ext2ls ide 0:1
       1024 .
       1024 ..
      12288 lost+found
[17:39] <bjf>          1506745 System.map-2.6.31-204-dove
[17:39] <bjf>           409765 abi-2.6.31-204-dove
[17:39] <bjf>            90063 config-2.6.31-204-dove
         23 vmlinuz
         26 initrd.img
[17:39] <bjf>          3360412 vmlinuz-2.6.31-204-dove
[17:39] <bjf>          3360476 uImage
[17:39] <bjf>          3273248 uInitrd
[17:39] <bjf>          3273184 initrd.img-2.6.31-204-dove
[17:39] <bjf>              356 boot.scr
[17:39] <bjf>          3360476 uImage.bak
[17:39] <bjf>          3273199 uInitrd.bak
[17:39] <bjf>              356 boot.scr.bak
[17:39] <bjf> Canonical>>
[17:40] <NCommander> bjf, ext2load ide 0:1 1000 /boot.scr
[17:40] <NCommander> autoscr 1000
[17:40] <bjf> Canonical>> autoscr 1000
[17:40] <bjf> ## Executing script at 00001000
[17:40] <bjf> Starting Ubuntu...
[17:40] <bjf> Unknown command 'load' - try 'help'
[17:40] <bjf> Unknown command 'load' - try 'help'
[17:40] <bjf> Wrong Image Format for bootm command
[17:40] <bjf> ERROR: can't get kernel image!
[17:40] <bjf> Canonical>>
[17:40]  * NCommander hits head
[17:41] <bjf> NCommander, let me try that again
[17:41] <NCommander> why did I think that was going to work >.<;
[17:41] <NCommander> Sounds like autodetection isn't very happy
[17:41] <bjf> NCommander, same failure
[17:41] <ogra> lool, "Configuring the network with DHCP " heh
[17:41] <lool> ogra: hmm?
[17:42] <ogra> copying /lib/modules from the usb key that had the former install
[17:42] <NCommander> bjf, when the autodetection fires up, what does it do when it gets to ide 0:1? (it should be printing out each device as it scans for it)
[17:42] <ogra> meh, dhcp needs some special love
[17:43] <lool> ogra: Ah it doesn't work
[17:43] <lool> ?
[17:43] <lool> ogra: Check syslog and the like
[17:43] <NCommander> ogra, where's that message from?
[17:43] <lool> d-i
[17:43] <ogra> NCommander, imx51 netinst
[17:43] <NCommander> hrm
[17:44]  * NCommander personally doesn't feel like dying anymore, and grabs dove netinstall
[17:44] <lool> NCommander: I think it's broken
[17:45] <lool> https://bugs.launchpad.net/ubuntu/+source/linux-mvl-dove/+bug/431945
[17:45] <ubot4> Launchpad bug 431945 in linux-mvl-dove "Lacks a bunch of modules for d-i support" [Undecided,New]
[17:45] <ogra> its definately is given it misses all udebs
[17:45] <NCommander> ogra, I did a netinst though w/ dove back in Dublin
[17:45]  * NCommander can' tremember what kernel he did it with though ...
[17:46] <lool> cp: cannot create link `/home/lool/cdimage/scratch/ubuntu/ports_daily/tmp/karmic-armel+dove/CD1/cdrom/uImage.in': No such file or directory
[17:46] <lool> odd first time I get a different error on a manual build
[17:46] <NCommander> lool, anyway, after today, I agree on having boot.scr be self-contained is a good thing, but it should detect if its been run off our boot.scr or not to prevent searching twice
[17:47] <bjf> NCommander, https://pastebin.canonical.com/22227/
[17:47] <NCommander> bjf, looks like it started up just fine?
[17:47] <bjf> NCommander, that's not right
[17:47] <lool> NCommander: Hmm ok
[17:47] <NCommander> oh bugger
[17:47] <NCommander> I see
[17:48]  * NCommander is still not fully there
[17:48] <NCommander> lool, the search loop is simply very slow (I'll be making improvements to extend it), but we'll have even more specific Canonical-functioanlity in the loop.
[17:48] <bjf> NCommander, that was GrueMaster's here is mine: https://pastebin.canonical.com/22225/
[17:49] <NCommander> oh ...
[17:49] <NCommander> crud
[17:49] <NCommander> I think I know what happened
[17:49] <NCommander> GrueMaster, bjf, I think your old envirionmental variables got retained when you upgraded u-boot
[17:49] <NCommander> or at least bjf's
[17:50] <GrueMaster> Mine is ok, I think.  Let me pastebin my prinenv for comparison.
[17:50] <NCommander> bjf, do: bubt u-boot-db88f6781y0bp_hX00_nand_1CS_fast_bank_en-09162009.bin from the Canonical>> prompt (bubt can be used safely once you switch from SPI-NOR to NAND)
[17:51] <NCommander> bjf, hit yes when it asks if you want to zap the environment
[17:51] <GrueMaster> https://pastebin.canonical.com/22229/ is my environment
[17:52] <NCommander> GrueMaster, your got properly reset
[17:52]  * NCommander isn't sure how that worked out
[17:52] <GrueMaster> Probably because I didn't have an older Nand boot bin.
[17:53] <NCommander> Ah, that would explain it
[17:53] <lool> NCommander: make sure you update the instructions to reflect this
[17:53] <NCommander> lool, right, let me just make sure bjf's upgrade goes safely/sanely
[17:53] <GrueMaster> remember, yesterday when I set the dip switches, I was getting bootparam (1.07) > and nothing else.
[17:54] <GrueMaster> Which I would guess ment that there was nothing at that flash location.
[17:54] <bjf> NCommander, it's thinking about it... looks like that worked
[17:54] <NCommander> bjf, \o/
[17:56] <bjf> NCommander, I'm logging in now!
[17:56] <NCommander> bjf, woo
[17:58]  * NCommander notes there's probably a good way to blow away the environment with the nand command ...
[18:00] <GrueMaster> woohoo!  First error on dove.  Already reported, but reproduced on dove none-the-less.  Bug #418300.
[18:00] <ubot4> Launchpad bug 418300 in gnome-disk-utility "gdu-notification-daemon crashed with SIGSEGV in gdu_pool_get_devices()" [Medium,Confirmed] https://launchpad.net/bugs/418300
[18:04] <GrueMaster> Hah!  My usb-nic works on dove.  Broken on imx51.
[18:06] <lool> GrueMaster: Oh cool, compare lsmod?
[18:06] <GrueMaster> this one is the dm9601 driver based nic.
[18:08] <ogra> hmm, even with the whole /lib/modules content copied over i dont get the system on the net
[18:09] <lool> NCommander: So I fixed the alternate CD for dove
[18:09] <lool> NCommander: But I'm pretty sure /cdrom is incorrect
[18:09] <lool> NCommander: http://cdimage.ubuntu.com/ports/daily/current/karmic-alternate-armel+imx51.list uses /install
[18:09] <ogra> not via dhcp nor via fixed ip
[18:09] <ogra> lool, for kernel and initrd ?
[18:09] <ogra> yes, its /install
[18:09] <lool> ogra: yes
[18:10] <NCommander> lool, we can change that, as long as the paths in boot.scr and the actual files match up
[18:10] <lool> NCommander: I think I fixed everything I knew about in dove; I'll leave it to you to fix the actual alternate CD
[18:10] <lool> NCommander: That's up to you
[18:10] <NCommander> lool, I'll hit d-cd once I get a chance
[18:11] <GrueMaster> Interesting.  I unplugged the onboard nic port and plugged in the dm9601.  NetManager crashed (generating bug report now), restarted, and connected through the usb-nic no problem.
[18:11] <ogra> ~ # ip addres show eth0
[18:11] <ogra> 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast qlen 1000
[18:11] <ogra>     link/ether 00:13:3b:04:05:b5 brd ff:ff:ff:ff:ff:ff
[18:11] <ogra>     inet 192.168.2.32/24 brd 192.168.2.255 scope global eth0
[18:11] <ogra> hrm
[18:11] <ogra> i dont get why it doesnt work
[18:11] <GrueMaster> heh.  firefox just crashed waiting for LP.
[18:12] <lool> ogra: What's the error in the log?
[18:12] <ogra> no errors
[18:13] <ogra> ah looking closer (i was looking for NIC probs) ...
[18:13] <ogra> Jan  1 00:32:12 main-menu[701]: DEBUG: resolver (libc6-udeb): package doesn't exist (ignored)
[18:13] <ogra> Jan  1 00:32:12 main-menu[701]: DEBUG: resolver (libnewt0.52): package doesn't exist (ignored)
[18:13] <ogra> meh
[18:14] <ogra> that doesnt explain why fixed IP doesnt work though
[18:15] <lool> ogra: You dont get a dhcp error?
[18:15] <NCommander> lool, thank you for fixing livecd-rootfs for dove yesterday
[18:15] <ogra> nope
[18:16] <ogra> it silently times out
[18:16] <ogra> i get the dhcp failed message from d-i though
[18:17] <lool> ogra: tcpdump?
[18:17] <ogra> i just resetted
[18:17] <lool> NCommander: np
[18:18] <NCommander> lool, did you change the d-cd scripts so alternate boot? (this code looks different then the last time I touched it)
[18:18] <NCommander> er, build
[18:18] <lool> NCommander: I actually fixed a bunch of your stuff *cough*
[18:18]  * NCommander sweatdrops
[18:18] <lool> flash-kernel and debian-installer notably
[18:19]  * NCommander goes to go hide 
[18:19] <GrueMaster> heh.  Firefox just crashed while trying to report a bug on another firefox crash.  sigh.
[18:19] <NCommander> GrueMaster, recursion: n. see recursion.
[18:20] <lool> < NCommander> lool, did you change the d-cd scripts so alternate boot? < NCommander> er, build    >>   19:09 < lool> NCommander: So I fixed the alternate CD for dove
[18:21] <NCommander> right, sorry
[18:21] <NCommander> I'm going to go get food
[18:21] <NCommander> then I'll look at straighting out dove alternates
[18:22] <ogra> lool, http://paste.ubuntu.com/272974/ in case you see something i dont
[18:22] <lool> ogra: No, looks like dhclient doens't work; it doesn't work manually either does it?
[18:23] <ogra> nope
[18:23] <lool> ogra: You could check whether you have CONFIG_PACKET
[18:23] <ogra> i would assume so, the same kernel works fine in the live image and installed systems
[18:23] <lool> http://cdimage.ubuntu.com/ports/daily/current/karmic-alternate-armel+dove.img
[18:23] <ogra> CONFIG_PACKET should be builtin
[18:23] <lool> plars: GrueMaster: The fun never ends!  ^
[18:23] <lool> ogra: Yeah I assume the same, just a desperate attempt
[18:24] <ogra> i was searching for af_packaet for a while half an hour ago :)
[18:24] <ogra> *packet
[18:25] <plars> lool: which thing, kernel in alt image lacks CONFIG_PACKET?
[18:25] <lool> ogra: that's on imx51?
[18:25] <ogra> yeah
[18:25] <lool> plars: The new image
[18:25] <plars> lool: ah, that's dove though
[18:25] <lool> plars: GrueMaster: You dont need to test it though; it's just to have things working that we build it and test from time to time
[18:26] <lool> plars: Yeah; we have one for imx51 though
[18:26] <lool> plars: It's slighlty broken and will be fixed soon
[18:26] <plars> lool: ok, let me know if you think it needs my attention
[18:26] <ogra> does it actually need a d-i uplaod ?
[18:26] <ogra> i would think a respin should suffice for imx51 alternate
[18:26] <lool> plars: No, unofficial image
[18:27] <lool> plars: If you are idle and wish to do some extra QA, I'm happy to fix the image but it's not an important image
[18:27] <lool> ogra: It needs a d-i upload
[18:27] <lool> to refresh initrd
[18:27] <plars> lool: if I'm idle, that's funny
[18:27] <plars> :)
[18:27] <ogra> oh, right
[18:27] <lool> Yeah, #if 0
[18:28] <lool> just a manner of saying that you likely wont ;)
[18:28] <lool> Ok enough debian-cd/cdimage/flash-kernel/redboot-tools for me today; I did that basically the whole day
[18:29]  * ogra tries a different NIC
[18:29] <lool> plars: Note that linux bugs for imx51 should be filed on linux-fsl-imx51 not linux
[18:29] <plars> lool: ah, crap... thanks
[18:29] <lool> plars: and assigned to amitk at his request
[18:30] <lool> plars: For dove it's linux-mvl-dove
[18:30] <plars> lool: as in, assign it to him when he requests it, or as in "he requested that we always assign to him"?
[18:32] <lool> plars: I understood the latter
[18:32] <plars> lool: ok, what about dove kernel? assign to bjf?
[18:32] <lool> plars: I'd guess so yes
[18:32] <lool> bjf: ^
[18:33] <plars> I'll add that to the workflow doc, which reminds me....
[18:35] <ogra> ha !
[18:35] <ogra> works as soon as i connect directly
[18:35] <ogra> the hub was in the way it seems
[18:40] <ogra> Installing the base system :D
[19:01] <bjf> plars, yes, dove kernel can be assigned to me
[19:03]  * ogra cries
[19:03] <ogra> http://paste.ubuntu.com/272997/
[19:04] <plars> bjf: ok, thanks
[19:04] <ogra> Sep 17 17:59:48 in-target: update-initramfs: Generating /boot/initrd.img-2.6.31-100-imx51
[19:04] <ogra> Sep 17 17:59:48 in-target: readlink: missing operand
[19:04] <ogra> Sep 17 17:59:48 in-target: Try `readlink --help' for more information.
[19:04] <ogra> Sep 17 17:59:48 in-target: mkinitramfs: missing  root  /sys entry
[19:04] <ogra> Sep 17 17:59:48 in-target: mkinitramfs: workaround is MODULES=most
[19:04] <ogra> Sep 17 17:59:48 in-target: mkinitramfs: Error please report the bug
[19:04] <ogra> Sep 17 17:59:48 in-target: update-initramfs: failed for /boot/initrd.img-2.6.31-100-imx51
[19:04] <ogra> Sep 17 17:59:48 in-target: Failed to create initrd image.
[19:04] <ogra> :(
[19:33] <kblin> hi folks
[19:35] <kblin> I'm running a sheevaplug with the preinstalled ubuntu jaunty, and it seems like I don't have any modules in my /lib/modules dir
[19:35] <kblin> how would I get those?
[20:12] <lool> kblin: I dont think they are shipped; I suggest you rebuild your own kernel
[20:22] <kblin> yeah, looking into that when I get home
[20:23] <kblin> I'm still at a conference, busy enough with other stuff
[20:44] <NCommander> lool, ogra dove alternates boot and properly come up to d-i, but the ENOMODULES stops anything from working, so other the path issue, they look ok. Once bjf has an opportunity to sort out the udebs, I'll retest off a daily