**** BEGIN LOGGING AT Mon Jul 09 02:59:59 2007 Jul 09 03:05:15 nbd * r7895 / (2 files in 2 dirs): add busybox printf patch from #2021 Jul 09 03:32:22 <[mbm]> the workaround for the printf issue is printf "%d\n" $((0x10)) Jul 09 05:44:12 nbd * r7896 / (2 files in 2 dirs): fix the cpu_wait patch for 2.6.22 Jul 09 07:09:42 nbd * r7897 / (4 files in 4 dirs): Jul 09 07:09:42 add second package source for packages built from the /packages repository Jul 09 07:09:42 fix up first package source (should be release, not snapshots on 7.07) Jul 09 09:33:26 nbd * r7898 /packages/net/nstx/files/nstx.init: fix a typo Jul 09 09:44:50 nbd * r7899 /packages/net/nstx/ (Makefile files/nstx.init): nuke the nstx init script. it is unusable and needs a rewrite Jul 09 16:06:42 florian * r7900 / (8 files in 6 dirs): au1000 now runs .22, mtx-1 watchdog is mainline and i2c patch is planned for rewriting Jul 09 16:11:18 mbm * r7901 / (2 files in 2 dirs): restart wds links when wifi is restarted Jul 09 16:36:17 juhosg * r7902 / (5 files in 5 dirs): disable image generation for boards which we still have problems with (at least if CONFIG_DEVEL is not selected) Jul 09 20:23:25 rwhitby: are you here? I've compiled a custom openwrt (latest svn), and it's not booting correctly, attached a serial cable and the last thing it says is Warning: unable to open an initial console. Jul 09 20:23:31 any ideas/pointers? Jul 09 20:38:15 <[mbm]> that's kind of a normal message, given that it boots with an unpopulated /dev Jul 09 20:38:37 aha, but stopping there is not normal i guess Jul 09 20:38:54 it's basically the same story as in this kernel thread: Jul 09 20:38:55 http://forum.openwrt.org/viewtopic.php?pid=47717 Jul 09 20:38:56 <[mbm]> right, but until the console is created you're somewhat blind Jul 09 20:39:05 which offers little help Jul 09 20:39:27 <[mbm]> like I said, you're somewhat blind Jul 09 20:41:07 indeed, so I'll basically go and have to find out why it can't create a console? Jul 09 20:41:29 <[mbm]> yep, or create an image that already has a /dev/console Jul 09 20:43:10 and that would need to be linked to a serial port, say /dev/ttyS0 Jul 09 20:43:10 ? Jul 09 20:43:25 <[mbm]> no. Jul 09 20:43:27 because according to the kernel booting, it does recognise 2 serial ports Jul 09 20:43:54 serial port is not linked to dev/console ? Jul 09 20:45:00 <[mbm]> /dev/console is not the same as /dev/tts/* (or /dev/ttyS*) Jul 09 20:46:47 indeed, it's a seperate device Jul 09 20:47:11 but if that device isn't created in the rootfs, that would be a bug in the build system, right? Jul 09 20:47:12 i mean /dev/ttyS0 is doublepiped with /dev/console Jul 09 20:47:26 <[mbm]> vlabakje: no Jul 09 20:47:43 <[mbm]> we boot up without a /dev and then create it on boot Jul 09 20:48:22 <[mbm]> (similar to devfs, but with much more headaches since we use udev and hotplug) Jul 09 20:48:32 aha Jul 09 20:48:59 and if that stuff doesn't get launched, you don't have a /dev/console, and you don't get anything on the /dev/ttyS0 output, correct? Jul 09 20:49:05 vlabakje: have you missing somethink regarding hotplug? Jul 09 20:49:14 <[mbm]> vlabakje: yep Jul 09 20:49:25 DBIUOG: not that I know of Jul 09 20:49:58 [mbm]: okay, thanks for the help, it's getting late over here, I'll get cracking on this again in the morning :) Jul 09 20:50:20 CONFIG_DEVFS_FS ? Jul 09 20:50:43 <[mbm]> devfs was obsoleted some time ago Jul 09 20:51:03 <[mbm]> udev is the crappy replacement we got stuck with Jul 09 20:52:03 What is the adavantage of udev against creating devices by makedevs? You have to maintain a list. Okay... Jul 09 20:52:34 <[mbm]> dynamic device creation Jul 09 20:53:03 This only works when the driver registers in sysfs? Jul 09 20:53:12 <[mbm]> no. Jul 09 20:54:08 I tried udev but the devices were not made in the right place. Some were even not created. Jul 09 20:54:27 <[mbm]> udev uses a different naming scheme from devfs Jul 09 20:55:11 <[mbm]> (udev is a compelte mess and I really don't want to continue this conversation) Jul 09 20:55:47 Do you have other plans for openwrt? Or will you leave things as they are? Jul 09 20:56:55 <[mbm]> let's talk about something other than why udev sucks :P Jul 09 20:57:17 Okay. :-) Jul 09 21:27:14 mbm * r7903 /trunk/package/wlcompat/ (Makefile src/wlcompat.c): fix channel/frequency returned by scan **** ENDING LOGGING AT Tue Jul 10 02:59:57 2007