**** BEGIN LOGGING AT Mon Dec 12 02:59:57 2005 Dec 12 05:21:16 join #elinux Dec 12 06:50:09 hi to all Dec 12 06:50:29 does anyone knows when openslug 3.0 is due ? Dec 12 07:13:24 does anyone knows when openslug 3.0 is due ? Dec 12 07:14:05 when it's done, I guess Dec 12 07:14:17 rwhitby sent a update email to the mailinglist a while back Dec 12 07:20:19 ok koen Dec 12 07:20:32 but when u suppose it will be ready ? Dec 12 07:20:52 i see that some times has passed since the last 2.7 relese.. Dec 12 07:27:00 I was told that aside from the new intel drivers work there is the interfacing the new led system to do. **** BEGIN LOGGING AT Mon Dec 12 10:50:11 2005 Dec 12 13:08:40 how can i build a single package? ./bitbake complains about importing the bb library.. Dec 12 13:15:56 source ./setup-env ; bb foo Dec 12 13:16:41 tnx Dec 12 13:17:06 rwhitby: did you have much trouble rebuilding the monotone db? Dec 12 13:17:38 s/building/pulling/ Dec 12 13:23:34 no, I didn't notice anything Dec 12 13:23:53 what happened thart I should have noticed? Dec 12 13:28:03 nothing should have happened Dec 12 13:28:11 as in, no errors Dec 12 13:28:40 I wasn't aware it was even rebuilt ... Dec 12 13:29:32 hmmm Dec 12 13:29:40 so nobody of the nslu2 team has read http://handhelds.org/hypermail/oe/47/4772.html ? Dec 12 13:30:47 .. Dec 12 13:35:15 koen: must admit that I hadn't read that. Dec 12 13:36:27 but I expect that jbowler handled it Dec 12 13:37:02 yeah Dec 12 13:42:10 I guess we should upgrade our server to 0.24 sometime ... Dec 12 13:43:05 * koen dreads to upgrade to 0.25 Dec 12 13:54:18 oh, is 0.25 incompatible? Dec 12 13:54:37 yeah Dec 12 13:54:53 they figured out how to make everything at least twice as fast Dec 12 13:55:05 which needs some new datastructures Dec 12 13:55:32 0.25 seems very promising, but the rebuild will be annoying Dec 12 13:56:19 but a rebuild will get rid off all those 'strange' stuff we've encountered Dec 12 13:56:26 well the speed will be a good thing Dec 12 13:56:42 a badly needed improvement IMHO Dec 12 13:56:47 indeed Dec 12 13:57:12 although it isn' too bad compared to bk if you use it everyday Dec 12 13:57:28 bk's integrety check was ueber-slow on my old disks Dec 12 13:58:35 if you don't pull everyday bk is loads faster Dec 12 14:05:44 So is there anything else which should be fixed in OpenSlug 3.x ? 2.6.15 should be out soon, and we're tracking that, and the latest IAL seems to work, and everything builds both BE and LE. What else do we need to do for 3.x ? Dec 12 14:08:36 I guess leds needs integration ... Dec 12 14:09:09 is the powerbutton working as supposed? Dec 12 14:09:18 rwhitby: thumb support? Dec 12 14:09:18 * rwhitby kicks off an openslug build to see what the current status is ... Dec 12 14:09:28 koen: already in there I think Dec 12 14:09:38 of course Dec 12 14:09:38 senneth: define "as supposed" :-) Dec 12 14:09:43 * koen forgot the slugos merge Dec 12 14:09:47 (propper shutdown;) Dec 12 14:10:11 senneth: that's been on HEAD for some time hasn't it ? Dec 12 14:10:25 have you tested latest head ? Dec 12 14:10:38 not really, no time for building :) Dec 12 14:10:43 my slug running 2. does proper shutdown Dec 12 14:10:47 i've just installed 2.7 Dec 12 14:11:01 uptime Dec 12 14:11:01 15:04:05 up 73 days, 19:22, load average: 1.00, 1.00, 1.00 Dec 12 14:11:09 ~rarr~ Dec 12 15:18:07 hmm Dec 12 15:18:35 Can't remember any bugs that are "must-fix" before 3.0. Dec 12 15:18:57 Samba would be nice to get large file support to work though Dec 12 15:20:09 if i can propose..i want a php apache2 module :P Dec 12 15:20:47 php5 should build with the libxml2 I added to OE Dec 12 15:20:48 yeah, don't we all ;) Dec 12 15:20:50 oh Dec 12 15:20:52 cool Dec 12 15:22:07 but anyway, fixes like that aren't necessary before a new release, since that isn't in the base firmware Dec 12 15:22:46 hmm Dec 12 15:23:54 cherokee + fastcgi + php would be faster anyway :) Dec 12 15:24:54 There's 8 reported bugs. Two of them might've been fixed in later kernel releases Dec 12 15:25:04 one is about slugbug itself, so that doesn't count Dec 12 15:26:53 NAiL: what are the other 7 ? Dec 12 15:27:53 no kernel config in /proc, ip_dst_cache overflow, unpredictable USB mappings to scsi device mounts Dec 12 15:28:22 no sticky bit on /tmp, memory alloc on FATslug, usb-drives not mounted/visible on boot Dec 12 15:28:28 and NFS lockups Dec 12 15:28:51 IKconfig is easily solved Dec 12 15:29:00 yeah Dec 12 15:29:06 although the ucslugc people might get upset ;) Dec 12 15:29:27 can't ucslugc use a different defconfig? Dec 12 15:29:57 BTU, or whatever ucslugc became ;) Dec 12 15:30:45 the NFS lockup thingy seems to be "fixable" with 4 threads instead of 8. Dec 12 15:30:48 perhaps we can just copy the defconfig into the rootfs instead of bloating the kernel with it ... Dec 12 15:31:02 not a bad idea Dec 12 15:31:09 in /boot on flash Dec 12 15:31:11 usb mappings people can solve themselves via labels - it's not a firmware bug Dec 12 15:31:13 the kernel stores it compressed anyway iirc Dec 12 15:31:45 koen: is the lack of a sticky bit in /tmp a generic OE problem ? Dec 12 15:31:55 FATslug memory alloc is out of scope Dec 12 15:31:58 hmm Dec 12 15:32:24 usb drives not mounted is usually user error or incompatible hardware. unless there is a repeatable test case we cannot fix it Dec 12 15:32:27 cat defconfig |grep -v ^#|gzip - > /boot/config.gz Dec 12 15:32:53 rwhitby-away: could, never payed attention to it Dec 12 15:33:16 I'm just about to flash the latest openslug, so I'll have a look and see if it's factual Dec 12 15:33:27 rwhitby-web: I'll quote you on that and close the bug ;) Dec 12 15:33:51 which one? Dec 12 15:34:11 usb drives Dec 12 15:34:21 ack Dec 12 15:34:44 looking at fatslug bug now Dec 12 15:35:03 oh Dec 12 15:35:06 it's a 256mb slug Dec 12 15:35:27 not exactly our problem, is it? ;) Dec 12 15:36:20 oooh Dec 12 15:36:22 256mb Dec 12 15:37:49 down to 6... should I just close the fatslug bug? Dec 12 15:39:30 would a kernel panic be sent over the network with netconsole? Dec 12 15:40:15 I wonder if the NFS lockups are network related.... Dec 12 15:40:31 in that case, the nfs stuff might be fixed with the new drivers. Dec 12 15:40:37 NAiL: re fatslug bug, unless there is a fix included, then we close it asking for the submitter to propose and test a fix. If someone has a fatslug, but cannot build and test a solution for it's firmware, then they really shouldn't be making fatslugs ... Dec 12 15:40:53 good point Dec 12 15:40:55 * NAiL closes Dec 12 15:45:06 I'm not at home, so I can't commit any changes to fix the defconfig stuff right now. But does dumping a filtered, gz'ed defconfig in /boot sound ok? Dec 12 15:45:32 cat defconfig |grep "="|gzip - > /boot/config.gz, actually... Dec 12 15:45:48 note that any change you commit before fixing the nslu2 db will have be added manually after fixing Dec 12 15:46:00 ah Dec 12 15:46:34 with a working viewmtn or monotone-viz that isn't a big problem Dec 12 15:46:38 just slightly annoying Dec 12 15:50:00 i've added a package in openembedded/conf/distro/openslug-packages.conf but bb _packagename_ still doesn't see it.. Dec 12 15:50:12 i've added it and then source setup-env Dec 12 15:50:21 on the wiki i don't see others steps.. Dec 12 15:50:30 anyway, the /tmp bug should be easy to fix. config should be easy to fix if my fix is acceptable ;-), Dec 12 15:51:03 hmm Dec 12 15:51:15 re: usb mappings: Dec 12 15:51:17 That solution won't work when you have a primary root filesystem on the external hard drives and skip the flash filesystem (because there isn't enough room for samba). Dec 12 15:51:19 I'm basically stuck with having the primity root file system ONLY in flash (then I can use labels or RAID UUID's to mount further filesystems). Dec 12 15:51:41 NAiL: config fix is good Dec 12 15:52:10 ok, I'll look at that tomorrow Dec 12 15:52:38 hmm Dec 12 15:52:40 oh Dec 12 15:52:45 yes... Dec 12 15:52:56 udev should go back in the flash fs Dec 12 15:53:05 for some reason it was dropped for 2.7-beta Dec 12 15:53:32 with udev the usb mount thingy can be avoided easily Dec 12 15:54:02 ie. turnup can write the disk ID to the udev config stuff Dec 12 15:55:48 * NAiL should probably look at that too tomorrow Dec 12 15:57:39 Hmm - a ludeos build somehow neglects to put libz.so.1 in the flash image, which stops dropbear from working. Will find out in about an hour whether openslug build does the same ... Dec 12 19:14:35 root@openslug:~# ls -ld /tmp Dec 12 19:14:43 lrwxrwxrwx 1 root root 8 Jan 1 00:00 /tmp -> /var/tmp Dec 12 19:14:51 root@openslug:~# ls -ld /var/tmp Dec 12 19:14:58 drwxrwxrwt 3 root root 60 Jan 1 00:01 /var/tmp Dec 12 19:15:15 NAiL: sticky bit is set on /tmp for virgin Openslug 3.x-alpha Dec 12 19:18:13 did you close the bug? Dec 12 19:18:22 nm Dec 12 19:18:28 * NAiL closes the bug Dec 12 19:19:35 That leaves four bugs Dec 12 19:19:49 I'll fix the config stuff and have a go at the usb mapping stuff Dec 12 19:20:14 That leaves the ip_dst_overflow bug, and the NFS hang Dec 12 19:20:28 both which *might* be fixed in 2.6.15 Dec 12 19:20:28 are they repeatable ? Dec 12 19:21:05 oh Dec 12 19:21:12 that's what causing my slug to crash actually Dec 12 19:21:13 yes Dec 12 19:21:16 it's repeatable Dec 12 19:21:24 the ip_dst_overflow Dec 12 19:22:18 the NFS stuff is only reported by one user. I don't use NFS, so I don't know. But it seems there's a semi-workaround by reducing the amount of threads from 8 to 4 Dec 12 19:23:41 so I guess I should try triggering the ip_dst_overflow in a later kernel with the latest drivers soon Dec 12 19:24:08 if my problem goes away, that only leaves the NFS one. And I don't do NFS. Dec 12 19:24:44 oh, btw, does your current openslug build have udev? Dec 12 19:27:00 yes, udevd is running Dec 12 19:27:32 ah, good Dec 12 19:27:45 there seems to be a problem with hwclock with the latest OpenSlug Dec 12 19:28:14 hmm Dec 12 19:29:33 I think I've closed some 10-ish bugs today Dec 12 19:30:48 I've built and flashed Unslung 6.x, LudeOS 1.x and OpenSlug 3.x :-) Dec 12 19:31:24 Now there's four bugs open on openslug firmware and 11 on packages. 6 of these are assigned to me... Dec 12 19:31:44 * NAiL goes off to bed Dec 12 19:32:10 nite all Dec 12 22:38:47 Hello. Dec 12 22:39:22 is it possible to boot openslug from a memory stick in a usb2 hub? Dec 12 22:40:04 The memstick gets recognised, etc properly and I can see the copied boot filesystem but it just won't boot to it. I can even mount it no problem. **** ENDING LOGGING AT Tue Dec 13 02:59:56 2005