=== fta_ is now known as fta === fta_ is now known as fta === bjf is now known as bjf[afk] === fta_ is now known as fta [02:56] jayabharath: don't you ever go home? === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === hrw|gone is now known as hrw [08:08] morning === fta_ is now known as fta [08:15] Already? I really wish I could go to sleep, but I leave for the airport in 3 hours. [08:18] GrueMaster: trip to Prague? === amitk is now known as amitk-afk [08:31] Prague is the hot place to be! [08:35] 33°C there today [08:52] I have a iMX25 with Ubuntu installed. How can I get the touchscreen working? ts_calibrate throws following message: "ts_open: No such file or directory" [09:01] Taalas: which file does ts_open try to open? [09:05] Taalas: and /proc/bus/input/devices lists ts device at all? [09:08] sebjan: Sorry if I've asked this before (if I did, I've forgotten the answer) [09:08] sebjan: Do _any_ of the OMAP chips have Parallel Ports? [09:08] Either ISA or PCI? [09:10] lag: hi! I would not speak about all omap chips, as there are many variants. I am not aware of such ports. [09:11] lag: you could maybe ask this question on #linux-omap? [09:11] I can :) [09:12] And have === fta_ is now known as fta [09:22] #linux-omap [09:29] ukleinek: I don't know. Where can I see it? [09:30] Taalas: ask strace [09:30] hrw: /proc/bus/input/devices lists 2 devices. One is named N: Name="mxckpd" the other N: Name="imx_adc_ts" [09:30] Taalas: did you ever used tslib tools before? [09:30] hrw: yea, Prague. [09:31] so imx_adc_ts is your touchscreen. check which /dev/input/event it got [09:31] ogra: i tried to turn off swap in the kernel. then one of the oops is gone [09:31] tslib is easy [09:31] cooloney, so it apparently cant handle swapfiles properly [09:31] hrw: I never used tslib before. The Handler is event1 [09:31] ogra: i think swap is required for you Ubuntu system, right? [09:32] Taalas: TSDEVICE=/dev/input/event1 ts_calibrate [09:32] cooloney, well, at least with so low memory [09:32] ogra: https://bugs.edge.launchpad.net/ubuntu/+source/linux-ti-omap4/+bug/605739 [09:32] Launchpad bug 605739 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "BUG: Bad page state in process swapper pfn:94d23 (affects: 1) (heat: 6)" [High,In progress] [09:33] ogra: so currently, we are using a swapfile in SD card, how about a swap partition? [09:34] cooloney, tricky but doable [09:34] ukleinek: strace tells /dev/touchscreen/ucb1x00 is not found [09:34] cooloney, though its still a bug [09:34] hrw: same error [09:34] cooloney, swapfiles work fine on other ubuntu arches [09:35] ogra: but without swap, i saw several memory allocation fails when i was trying apt-get dist-upgrade on the board [09:35] sounds more like a general memory prob with the kernel [09:35] The folder /dev/touchscreen is not existing [09:36] ogra: maybe, i need more investigation. [09:36] but i think 512M memory is OK for Ubuntu on omap4, === amitk-afk is now known as amitk [09:37] although i found the system sometime is quite slow [09:37] cooloney, yes, something is wrong, it got a lot slower between two image tests i did [09:37] cooloney, but that also happened on omap3 so i suspect thats rather userapce [09:38] i'll resarch and fix that at the sprint [09:38] ogra: got it. i assume is my SD card's problem. [09:38] no, we all see it on different SD cards [09:38] sometimes my SD card cannot be recognized by my PC [09:39] there is definately something not acting as it should [09:39] Taalas: /dev/touchscreen never existed - maybe ltib (or whatever you used before) created it [09:39] ogra: ok, cool. [09:39] the MMC driver on the panda *has* some speed issues though [09:39] but thats not alone the reason for the slowness [09:40] I created my rootfs with rootstock [09:47] Taalas: so use strace to check why it does not work [09:47] Taalas: check permissions etc === frivas is now known as fjrivash === fta_ is now known as fta === fta_ is now known as fta [10:45] amitk: Hi === fta__ is now known as fta [11:30] lag: hola! [11:31] Hey amitk [11:31] Do you have the ability to search though the kernel mailing list? [11:32] lag: LKML? [11:32] linux-kernel@vger.kernel.org [11:33] yes, why? [11:33] Did Google go bankrupt? ;) [11:33] lol [11:33] http://lkml.indiana.edu/hypermail/linux/kernel/ [11:34] http://marc.info/?l=linux-kernel === cooloney is now known as cooloney-afk [11:34] * amitk subscribes to lkml through gmail to use search [11:35] lag: gmane ;) [11:35] amitk: Search for - Stop ARM boards crashing when CUPS is loaded [11:37] http://marc.info/?l=linux-kernel&w=2&r=1&s=Stop+ARM+boards+crashing+when+CUPS+is+lo&q=b [11:37] lag: ^^ I see your comment there [11:40] apw: Mentioned that Lenaro may have the funding to make the LK better in this way [11:40] lag, are you pitching a project for Linaro? :) [11:40] * apw referred to the 'platformisation' of things like the paralellel port driver ... [11:42] Maybe :) [11:43] apw: Yep [11:44] Similar to the work that was undertaken on the serial driver a little while ago [11:44] Mentioned by Russle [11:44] Mentioned by Russell [11:44] :) [11:45] lag: This is unlikely to be a priority for Linaro (parallel port is not an interface that many people care about anymore). You might try pitching it to to GregKH for his Linux drivers project. [11:45] some student wanting to get their feet wet might want to dive into it [11:47] Okay, I'll ping him an email explaining [11:47] lag: I'll ask around in Linaro, but don't expect too much. [11:47] No problem, thanks [11:48] Can I mention Linaro in an email to Greg and Russell? [11:48] I'm assuming I can if I can mention them here [11:49] lag: mention Linaro in what context? === fta_ is now known as fta [11:50] "I've put in a pitch for Linaro to complete the work, however, this is unlikely .... [11:51] "I was wondering if you'd be interested ... [11:53] amitk: --^ [11:53] Hi mythripk [11:55] lag: sure, but TBH this isn't the kind of work Linaro is geared for. I can almost certainly say that no machine ships with parallel ports anymore (on any platform). So just ask Greg if he knows anybody who might be interested in taking on this challenge [11:55] amitk: np [11:57] amitk: parport_pc is not only LPT1: [11:57] parport_serial which handles multi port PCI serial cards depends on parport_pc [11:57] there are ARM boards with PCI slots... [12:00] hrw: interesting. But are these modern boards? [12:01] I know that pcie multi i/o cards also exists. and pcie is present in modern arms too [12:01] amitk: intel ixp4xx had pci but it is xscale === fta_ is now known as fta [12:03] hrw: that is not really an architecture linaro cares about, is it? The discussion here is whether this is something that is a blocker for Linaro. There are ARM boards doing all sorts of things, but Linaro only cares about a small percentage of those [12:03] amitk: Is it worth mentioning that you suggested to email Greg? I.e. does he know who you are by name? [12:04] I would just disable parport_pc from default configs [12:04] hrw: Check the kernel-team mailing list [12:05] lag: you're being too formal about this (intros and references) :) And no, I'm not that famous ;) [12:05] One must be professional :) [12:08] which one ? [12:08] lag: did not found 'parport' in last 3 months [12:12] hrw: Are you searching in subject lines or main body? [12:13] hrw: Search for "Stop ARM boards crashing when CUPS is loaded" [12:15] found [12:15] :) [12:15] !ARCH_OMAP one [12:15] Factoid 'ARCH_OMAP one' not found [12:15] Yup [12:15] lag, cooloney-afk, "omap_device: mmci-omap-hs0: new worst case deactivate latency 0: 30517" any idea what that cound mean ? [12:15] i wonder if thats the reason for the SD slowness [12:15] Slow read? [12:16] i get that on the console if fsck runs before resizing the filesystem [12:16] (on panda) [12:16] It's not a kernel message [12:17] Wait [12:17] Sorry [12:17] 1 min [12:18] it must be kernel or a driver [12:22] ogra: That's just wake-up latency [12:22] ogra: It's putting the card to sleep - saving power [12:23] while writing to it ? [12:23] wow [12:23] Yeah, that doesn't sound good [12:23] it appears immediately after fsck starts [12:23] Are you sure it't not _after_ it finishes? [12:23] yes [12:24] printf "Checking filesystem before resizing..." [12:24] /sbin/e2fsck -fy /dev/${DISK}${SEP}2 >>${LOG} 2>&1 [12:24] echo "[done]" [12:24] thats the code [12:24] it appears between the ... and done [12:25] immediately after the printf was run [12:26] And in the ${LOG}? [12:26] so either while e2fsck runs or right before it starts [12:26] still waiting for the resizing to finish [12:26] i'll grab the log but usually there is just plain e2fsck babbling [12:30] lag, http://paste.ubuntu.com/464486/ nothing special from e2fsck === fta_ is now known as fta [12:30] do cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed [12:31] hrm [12:31] i have no user on that system yet [12:31] that will take a while [12:33] I don't think that messages is an issue [12:33] It's just showing you the worst case latency of the MMC card [12:33] Once it's awake, it's awake [12:33] How slow is e2fsck? [12:35] takes around a minute [12:35] probably two [12:36] (1.5G filesystem) [12:36] overall the SD I/O seems to be way to slow though [12:40] lag, no cpufreq at all [12:41] Hold on [12:42] (in another note i tried to run iotop yesterday and it seems we're missing config options for it to work) [12:43] Which board is this? [12:43] panda [12:46] From initramfs? === fta_ is now known as fta [12:47] yep [12:47] How can I replicate? [12:48] mythripk: ping [12:48] grab a recent preinstalled image [12:48] and i can give you a uInitrd to replace on the SD card with the jasper changes [12:49] I'm still having monitor issues [12:50] just change the shipped boot.scr and use a serial console for that [12:52] lag: That is in the PSM mode and when it is on ,it is working fine with the patch right ? [12:54] Mythripk: Yes [12:54] Is the delay in the correct place? [12:54] Is it long enough? [12:55] ogra: Okay [12:55] i have sent a single clean patch with this i was not able to reproduce the on/off issue [12:59] lag, http://people.canonical.com/~ogra/uInitrd [13:30] ogra: debconf: DbDriver "config": could not write /var/cache/debconf/config.dat-new: No space left on device [13:30] On a 16GB SD - WTF? [13:30] that was with a fresh image and you only added the console entry to the existing boot.scr ? [13:31] (it didnt resize, thus the no space message) [13:31] I had to mkimage it again [13:31] But yes [13:31] Fresh image, only the boot.scr changed to add console [13:31] looks to me like the image is already configured [13:31] It was brand-spanking [13:31] hmm === fta_ is now known as fta [13:31] I dd'ed myself [13:32] can you check if 7var/loga7jasper.log exsists in the foorfs ? [13:32] *rootfs [13:32] gah [13:32] /var/log/jasper.log [13:32] Nope [13:32] No consle [13:32] console* [13:32] on your PC [13:32] just mount the Sd [13:33] the resizer only doesnt run if root=UUID= is set on the cmdline [13:33] lag: can you try and let me know [13:33] and thats set by the second part of jasper [13:33] mythripk: Doing it right now [13:35] mythripk: Monitor entered PSM/on/PSM/on ... [13:35] mythripk: You want the log? [13:36] mythripk: http://paste.ubuntu.com/464515/ [13:41] lag:looking at the log give me a min [13:43] ogra: /var/log/jasper.log is missing [13:43] lag, very weird [13:43] ogra: I haven't used your initrd yet [13:43] The only thing I changed was the boot.scr [13:43] my initrd just adds more output to jasper [13:44] apart from that it shoudl be identical to whats on the image [13:44] So why isn't it resizing then? [13:44] no idea [13:44] and hard to say without log [13:44] Are there any other logs which may help? [13:44] nope, its all in initramfs, nothing logs there by default [13:45] it works fine here with the latest image and my hacked initramfs [13:45] as well as with the shipped one [13:45] lag: did you try with debug today and with debug prints it works is it [13:46] Not yet [13:46] Wait one [13:47] Yes, it works fine with debug on [13:47] ogra: http://paste.ubuntu.com/464520/ [13:49] lag, how long did it sit between line 243 and 244 ? [13:49] should be like 30min for a 16G card [13:50] Nowhere near that [13:50] so resizing failed for an unknown reason [13:52] How do I debug? [13:53] you cant [13:54] well, you could use break=premount and try to run jasper_growroot manually [13:55] or use break=bottom and inspect /dev/.initramfs/jasper.log [13:55] but i doubt you will get much info from that [13:56] Well I need to fix it [13:56] So I need somewhere to start [13:57] right, on a fresh image break=bottom and checking /dev/.initramfs/jasper.log is probably the best [13:57] Does growroot have a -vvv setting? [13:57] heh, its a script [13:57] Okay [13:57] I'll hack the script [13:57] depends how good you are with ed :) [13:57] there is no editor in the initramfs === fta_ is now known as fta [13:58] I'll mount it on my PC [13:58] you mount the initramfs ? [13:58] :) [14:02] Is that where the script is/ [14:02] ? [14:02] yes [14:03] thats why making or testing changes is so hard [14:03] or debugging [14:04] all commands called by jasper redirect their output to jasper.log though, if you boot with a break= statement you can at least cat that file from the busybox prompt [14:08] Folks, do you know why we don't have ddebs for armel ATM? [14:08] what are ddebs? [14:09] lool, no idea [14:09] lool, ask pitti ? [14:10] i think he fiddled with pkgbinarymangler [14:10] ogra: [14:10] (initramfs) ls /dev/.initramfs [14:10] bootchart.pid [14:10] .. [14:10] where did you break ? [14:10] break=bottom [14:10] try premount [14:11] Do I have to reflash? [14:11] bottom might be to late ... the script in local-bottom has a mv command for the log [14:11] yes [14:11] Oh no, I tried premount [14:11] Should I try bottom? [14:11] whats the last you see on the screen ? [14:12] did it run local-premount already ? [14:12] device-mapper: ioctl: 4.17.0-ioctl (2010-03-05) initialised: dm-devel@redhat.com [14:12] or only init-premount [14:12] udev: starting version 151 [14:12] omap_device: mmci-omap-hs.0: new worst case deactivate latency 0: 30517 [14:12] Begin: Loading essential drivers ... Btrfs loaded [14:12] device-mapper: uevent: version 1.0.3 [14:12] device-mapper: ioctl: 4.17.0-ioctl (2010-03-05) initialised: dm-devel@redhat.com [14:12] done. [14:12] Spawning shell within the initramfs [14:12] k [14:12] So? [14:12] Now what? [14:13] sh /scripts/local-premount/jasper_growroot [14:13] sh shouldnt be necessary though [14:13] then check for the log again [14:14] (initramfs) sh /scripts/local-premount/jasper_growroot [14:14] scripts/local-premount/jasper_growroot: line 104: syntax error: /255/63/512 [14:14] Oooooooooooooooooooooooo [14:14] hmm, seems the mmc isnt there [14:14] check /dev if mmcblk0 exists [14:14] you probably need to break later [14:14] Doh! [14:15] My bad [14:15] Wait one [14:18] Resizing root fiomap_device: mmci-omap-hs.0: new worst case deactivate latency 0: 30517 [14:18] lesystem please wait, this will take about ten minutes ... [14:18] (initramfs) [14:19] heh, thats it ? [14:19] it should have taken 10mins :) [14:20] well, check for jasper.log [14:21] http://paste.ubuntu.com/464529/ [14:23] weird [14:23] given that e2fsck just ran before [14:24] * ogra has no idea how that could happen [14:25] Well learn ;) [14:25] This is userspace :) [14:25] well, that would mean that something in your FS is dirty [14:25] which cant be [14:26] It's a new dd [14:26] resize2fs is the next line after e2fsck ... the disk isnt mounted so there is no way it could get dirty again after it was just checked [14:28] Anything else I can try? [14:28] not really [14:29] i have no clue why your SD appears dirty right after e2fsck -fy [14:29] thats technically impossible [14:30] you could try the same with the other uInitrd [14:30] With your one? [14:30] but i doubt it will change much about the fact that either e2fsck or resize2fs cant handle your Sd properly [14:30] yes, with mine [14:39] you dont happen to have it mounted manually or some such ? === fta_ is now known as fta [14:52] ogra: I flash the card and put it in the Panda [14:52] ogra: I do nothing else [14:53] right, but there is no technical explanation for that error [14:53] ogra: It must do something, cause it trashes the boot partition [14:53] trashes in what way ? [14:55] Won't mount on my PC any more [14:57] like you issue mount /dev/mmcblk0p1 /mnt and get an error ? [14:57] or how does that manifest ? [14:58] * ogra quickly makes some coffee before the call [14:58] It does mount [14:59] It just doesn't auto mount [15:01] thats how its supposed to be :) [15:01] it shouldnt show on your desktop or in places or automount [15:02] mount it manually if you need to manipulate it [15:18] lag, when the repartitioning happens the vfat partition gets recreated and gets a label added that udev explicitly ignores [15:18] k [15:18] So it gets that far [15:18] it runs the whole stuff apart from resizing [15:19] Seemingly so [15:19] i'm pretty sure even the jasper_setup script from local-bottom runs fine [15:19] i could add a force to resize2fs but that scares me a bit [15:20] it might ignore if the FS is dirty [15:20] not sure what checks -f disables i know it doesnt drop all of them [15:20] Well it can't be any worse than it already is [15:20] indeed [15:22] let me roll you another uInitrd with -f [15:22] k [15:24] http://people.canonical.com/~ogra/uInitrd [15:24] try that one on a fresh dd'ed card [15:25] * ogra takes a break === bjf[afk] is now known as bjf [15:51] lag, any change ? === hrw is now known as hrw|gone [16:24] lag, did you have any luck with the other uInitrd ? [16:28] hmm, is there any 4430 SDP compatible images? [16:28] maverick ones [16:29] only for panda [16:30] * XorA|gone suspects has destroyed image trying to upgrade :-) [16:31] http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/ [16:31] oh no, just too a long time for ext3 recovery === JaMa|GoNe is now known as JaMa [16:50] ogra: How do I know if the resize has completed correctly? [16:50] (initramfs) sh /scripts/local-premount/jasper_growroot [16:50] mmcblk0: p1 p2 [16:50] Resizing root filesystem please wait, this will take about ten minutes ... [16:50] (initramfs) [16:51] check jasper.log [16:51] but for my hand rolled initramfs thats not enough output [16:51] so i guess it died again [16:54] http://paste.ubuntu.com/464598/ [16:54] sigh [16:56] Fix it [16:56] buy a better SD :P [16:57] bug 373409 [16:57] Launchpad bug 373409 in gparted (Ubuntu) (and 1 other project) "resize2fs doesn't notice the partition was fsck'ed (affects: 1) (heat: 13)" [Undecided,Confirmed] https://launchpad.net/bugs/373409 [16:57] It's a Kingston [16:57] gparted ? [16:57] lol [16:58] * lag shrugs === fta_ is now known as fta [17:03] lag, hmm [17:04] we actually set the clock to last mount time in omap [17:04] so that such a broken clock setting cant interfere with mounting [17:07] hmm [17:07] * ogra has an ugly idea but one that will surely fix it [17:09] * ogra rolls a new test initrd [17:13] (initramfs) /sbin/resize2fs /dev/mmcblk0p2 [17:13] resize2fs 1.41.12 (17-May-2010) [17:13] ext2fs_check_mount_point: Can't check if filesystem is mounted due to missing mtab file while determining whether /dev/mmcblk0p2 is mounted. [17:13] lag, http://people.canonical.com/~ogra/uInitrd updated, i bet it works now [17:13] What did you do? [17:14] i mount/unmount it once before e2fsck runs :) [17:14] Okay [17:14] just to confirm its the same [17:14] if its the same bug it should have a new last mounted timestamp and stop complaining [17:15] k [17:15] indeed thats not the right solution [17:15] Did you see the output of my /sbin/resize2fs /dev/mmcblk0p2 [17:16] your jasper.log you mean ? [17:16] oh, i missed the lines above [17:17] dont worry, the script creates an mtab [17:17] and removes it at the end again [17:17] you would have to do the same if you want to run it manually ... just link /proc/mounts to it [17:23] k === fta_ is now known as fta === bjf is now known as bjf[afk] === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === bjf[afk] is now known as bjf === fta_ is now known as fta === bjf is now known as bjf[afk]