**** BEGIN LOGGING AT Sat Jan 30 08:39:17 2010 Jan 30 13:12:19 lars * r19394 /trunk/target/linux/xburst/files-2.6.32/drivers/misc/jz4740-adc.c: [xburst] Dynamically gate adc clock Jan 30 13:47:41 jow * r19395 /packages/utils/strace/ (Makefile patches/ patches/000-linux_socket.patch): [packages] strace: fix compilation against Linux 2.6.32+ (#6604) Jan 30 13:52:10 jow * r19396 /packages/net/transmission/ (4 files in 2 dirs): [packages] transmission: add uci config and init script (#6611) Jan 30 13:55:46 jow * r19397 /packages/net/transmission/Makefile: [packages] transmission: update to v1.83 (#6606) Jan 30 14:15:53 jow * r19398 /packages/net/coova-chilli/Makefile: [packages] coova-chilli: update to v1.2.1 (#6600) Jan 30 14:25:15 stupid question, since ieee80211d in hostapd.conf is a bool, wouldn't it be cleaner to have http://openwrt.pastebin.ca/1771948 ? Jan 30 14:51:24 lars * r19399 /trunk/include/autotools.mk: Jan 30 14:51:24 autotools.mk: Add libstdc++ quirk to libtool_fixup_libdir. Jan 30 14:51:24 Fixes build of packages which use libtool and link against packages which link against libstdc++ Jan 30 15:25:05 juhosg * r19400 /trunk/target/linux/ramips/ (11 files in 6 dirs): ramips: update mips multi-machine stuff Jan 30 15:25:10 juhosg * r19401 /trunk/target/linux/ramips/files/arch/mips/ralink/common/gpio.c: ramips: sanitize return value of ramips_gpio_get Jan 30 15:25:15 juhosg * r19402 /trunk/target/linux/generic-2.6/files/drivers/net/phy/rtl8366_smi.c: rtl8366_smi: sanitize gpio values to a 0/1 boolean Jan 30 15:25:17 juhosg * r19403 /trunk/target/linux/ramips/ (2 files in 2 dirs): ramips: register rtl8366s device on the RT-N15 board Jan 30 15:25:22 juhosg * r19404 /trunk/target/linux/ramips/ (8 files in 5 dirs): ramips: add common gpio-buttons device Jan 30 15:25:27 juhosg * r19405 /trunk/target/linux/ramips/ (4 files in 2 dirs): ramips: register gpio-buttons on the RT-N15 board Jan 30 15:25:30 juhosg * r19406 /trunk/target/linux/ramips/files/arch/mips/ralink/rt288x/mach-wzr-agl300nh.c: ramips: cleanup WZR-AGL300NH board support file Jan 30 15:25:34 juhosg * r19407 /trunk/target/linux/ramips/files/arch/mips/ (8 files in 4 dirs): ramips: rename dev_gpio_leds.h to dev-gpio-leds.h Jan 30 15:25:39 juhosg * r19408 /trunk/target/linux/ramips/files/arch/mips/ (7 files in 4 dirs): ramips: include 'linux/leds.h' from 'dev-gpio-leds.h' Jan 30 15:25:42 juhosg * r19409 /trunk/target/linux/ramips/files/arch/mips/include/asm/mach-ralink/rt288x_regs.h: ramips: fix wrong base addresses for the RT288x Jan 30 15:25:45 juhosg * r19410 /trunk/target/linux/ramips/ (4 files in 2 dirs): ramips: register gpio buttons on the DIR-300 revB board Jan 30 15:25:48 juhosg * r19411 /trunk/target/linux/ramips/files/arch/mips/ralink/rt305x/ (Kconfig mach-v22rw-2x2.c): ramips: register gpio buttons on the V22RW-2x2 board Jan 30 15:25:51 juhosg * r19412 /trunk/target/linux/ramips/files/arch/mips/ralink/rt305x/ (Kconfig mach-whr-g300n.c): ramips: register gpio buttons on the WHR-G300N board Jan 30 15:25:54 juhosg * r19413 /trunk/target/linux/ramips/files/arch/mips/ralink/rt305x/ (Kconfig mach-fonera20n.c): ramips: register gpio buttons on the Fonera 2.0N board Jan 30 15:25:56 juhosg * r19414 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: coding style cleanup Jan 30 15:25:59 juhosg * r19415 /trunk/target/linux/ramips/files/drivers/net/ramips_eth.h: ramips_eth: header file cleanup Jan 30 15:26:01 juhosg * r19416 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: move memset call out from the loop Jan 30 15:26:04 juhosg * r19417 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: fix invalid register writes Jan 30 15:33:24 ping nbd Jan 30 15:33:33 ping cshore Jan 30 15:33:43 ping [florian] Jan 30 15:35:27 pong Jan 30 15:36:00 hauke * r19418 /trunk/target/linux/generic-2.6/config-2.6.31: [kernel] Add missing config option Jan 30 15:46:43 nbd could u please add my patch on the mailing list Jan 30 15:46:46 for devtmpfs Jan 30 15:47:23 on 2.6.32 it is automounted by the kernel but not used ... insted a tmpfs is mount on /dev instead of just moving devtmpfs to the new root Jan 30 15:48:16 the patch checks is devtmpfs is there and if its there it gets used Jan 30 15:48:24 so this wont disturb older kernel versions Jan 30 15:48:35 is/if Jan 30 15:49:07 nbd a save some memory patch because there is only one dev Jan 30 15:49:17 and not one in /rom/dev and one in /dev Jan 30 15:49:19 on 2.6.32 Jan 30 15:51:10 nbd another question too .. does the system bring the services down when i do a reboot/poweroff? as it seems to me the sys just goes down without stopping any service? is this right? Jan 30 15:51:27 lars * r19419 /packages/libs/tiff/Makefile: [packages] Move libtiff c++ bindings into it's own package Jan 30 15:52:41 nbd the mail is called - use devtmpfs if available and dont let it be unused Jan 30 15:53:21 why use devtmpfs if hotplug2 creates all /dev nodes anyway? Jan 30 15:53:38 devtmpfs is useful if the kernel automounts it so that it has initial access to /dev/console Jan 30 15:53:56 but after hotplug2 takes control over user events, it has no purpose anymore Jan 30 15:54:14 nbd yeah but is still mounted and is never unmounted Jan 30 15:54:41 ok, makes sense Jan 30 15:54:53 so the kernel mailing list explains - udev hotplug2 can write into it Jan 30 15:54:56 it doesnt matter Jan 30 15:55:14 maybe a check .. do we have devtmpfs . then hotplug doesnt need to create any devices at all Jan 30 15:55:27 would also decrease bootup time Jan 30 15:56:09 so coldplug can be disabled in hotplug2 Jan 30 15:56:13 it would not decrease boot time at all Jan 30 15:56:20 and coldplug cannot be replaced by devtmpfs Jan 30 15:56:26 okay Jan 30 15:56:32 since we cannot rely on devtmpfs creating devices with the correct name Jan 30 15:56:37 that part is user space policy Jan 30 15:56:42 and it's handled by the hotplug2 ruleset Jan 30 15:56:52 most devices can be created with only their name in the path Jan 30 15:56:54 but not all of them Jan 30 15:57:07 nbd: i booted my sys with this mod and it doesnt seem to create any problems at all Jan 30 15:57:10 hence we should get rid of devtmpfs after we don't need it anymore Jan 30 15:57:56 nbd: here on gentoo i have the same .. mount devtmpfs + udev started by init and it works Jan 30 15:58:58 how about simply doing this somewhere in preinit: grep devtmpfs /proc/mounts | awk '{print $2}' | xargs umount Jan 30 15:59:02 nbd: so when hotplug is run it will maybe create some devices and name it differently ... but i think the only device that devtmpfs is missing is /dev/root Jan 30 15:59:08 and this is not created by hotplug Jan 30 15:59:12 its a link Jan 30 15:59:25 well, xargs -r instead of xargs Jan 30 15:59:48 i don't want devtmpfs around, because it can cause inconsistent device handling in the future Jan 30 15:59:53 awk is a little bit of a big solution... why not use the shell build in functions? Jan 30 16:00:12 it's all in busybox anyway Jan 30 16:00:30 yours would break if accidentally multiple entries show up Jan 30 16:00:32 due to other bugs Jan 30 16:00:45 thats true Jan 30 16:00:49 also, my example is more readable Jan 30 16:01:35 nbd: u can umount devtmpfs when hotplug2 is started the first time in preinit Jan 30 16:01:39 right afterwards Jan 30 16:02:51 ok, i'll change this later Jan 30 16:02:56 i'm busy with some other stuff right now Jan 30 16:03:06 nbd and the other question .. does a reboot/poweroff stop the services befor it actually reboots the sys? Jan 30 16:03:10 or poweroff Jan 30 16:04:02 nbd: as it seems to me ... the reboot is done without stopping any service Jan 30 16:04:58 it only stops services that specify a STOP= value Jan 30 16:05:07 which makes sense, since most services do not need to be stopped on reboot/poweroff Jan 30 16:05:32 okay Jan 30 16:06:18 maybe dropbear and openssh should get a stop value too ... so the sys would stop them and so kill the connection and the connection wouldnt "stuck" forever Jan 30 16:06:39 describind in one ticket Jan 30 16:06:55 that the connections stucks and how this can be avoided Jan 30 16:08:04 nbd * r19420 /trunk/package/hostapd/patches/150-mbss_driver_handling.patch: hostapd: fix segfault in mac80211 wds ap handling Jan 30 16:08:13 makes sense Jan 30 16:08:33 i do this in /etc/profile now but this a dirty hack Jan 30 16:08:52 nbd: stupid question, since ieee80211d in hostapd.conf is a bool, wouldn't it be cleaner to have http://openwrt.pastebin.ca/1771948 ? Jan 30 16:09:00 stintel: yes Jan 30 16:09:03 nbd: also nfsd and portmap and syslog have to be stop so the system can umount -a -r all in /etc/init.d/umount Jan 30 16:10:02 otherwise a writeable disk wont get umounted correctly and get corrupted/checked on next boot (ex ext3) Jan 30 16:10:15 ok, can you send patches for that? Jan 30 16:10:31 yeah ill but u have to make the dropbear patch Jan 30 16:10:33 i dont use it Jan 30 16:10:38 the rest can be done by me Jan 30 16:10:43 ill send it when im ready Jan 30 16:10:53 maybe some other progs are affected too Jan 30 16:11:05 watching a 1080p hd movie over cifs with 2.6.32.7 athk9 right now, works quite well, had a few audio drops but still going, no disassocations or traffic stalls so far, this looks promising :-) Jan 30 16:11:14 which is nearly for shure true for some other daemons Jan 30 16:11:23 stintel: cool Jan 30 16:12:43 nbd: can more then one programm have the same stop value? Jan 30 16:15:26 Any comments on splitting up coreutils? Jan 30 16:17:05 Guest82296: looks good to me Jan 30 16:17:22 nbd: I made some changes to hostapd.sh and Makefile to enable 80211w and allow that and peerkey to be enabled in /etc/config/wireless. are these features too experimental to have it added to trunk, or shall I create a ticket with patch ? Jan 30 16:18:01 stintel: feel free to send the patch to the list Jan 30 16:25:25 nbd: can it be merged to trunk so that we can sync our packages and contribute our repository as a feed to OpenWrt? Jan 30 16:30:15 nbd: i added stop values to some init.d srcipts but it seems they are not recogniced Jan 30 16:30:27 what do i have to do to get them used? Jan 30 16:31:28 ahh enable Jan 30 16:34:11 hi Jan 30 16:34:45 so is it true that trunk mac80211 doesn't support multiple AP networks? Jan 30 16:34:56 no, that changed a while ago Jan 30 16:35:00 it's supported with ath9k Jan 30 16:35:52 huh? Jan 30 16:36:04 does mac80211 support multiple AP networks? Jan 30 16:36:12 yes Jan 30 16:36:15 ok Jan 30 16:36:21 then I should file a but about it not working :) Jan 30 16:36:24 bug* Jan 30 16:36:26 which rev? Jan 30 16:36:41 no need to file a bug yet, i'm the one working on this Jan 30 16:36:43 19232 Jan 30 16:37:21 yes, proper multi-bss support was added after that particular rev Jan 30 16:37:28 doh Jan 30 16:37:33 ok, thanks Jan 30 16:37:54 must be pretty new Jan 30 16:38:07 yes, there were a few things that i had to fix Jan 30 16:38:21 in principle, the support was there for a long time Jan 30 16:38:34 it's just nobody had fixed the bugs that were in the way of having it work Jan 30 16:38:40 cool, thanks Jan 30 16:38:46 I'll give it a try when I have time Jan 30 17:09:01 Guest82296: will take care of that in a few minutes Jan 30 17:10:07 Thanks! Jan 30 17:20:47 stintel: email me a copy ill get some testing on it done Jan 30 17:41:10 nbd * r19421 /trunk/target/linux/generic-2.4/config-default: brcm-2.4: add support for 8-bit flash bus width Jan 30 17:59:24 nbd i've done the patch to stop all progs that need to be stopped to get a working umount/ssh logout when reboot/poeroff is called ... umount -r -a is also not enough ... umount -r -a -l is need ... i send the patch to the mailinglist, could u apply it? Jan 30 17:59:33 sure Jan 30 18:00:14 maybe some more are need .. but i can only stop what i use ... maybe some people point out more progs that need to be stopped Jan 30 18:00:47 nbd also add a stop line to dropbear .. the patch doesnt include this Jan 30 18:00:47 i think it's better to explicitly mark those that need to be stopped than marking those that do *not* need to be stopped Jan 30 18:01:29 what did i do? Jan 30 18:01:38 mark those that need to be stopped Jan 30 18:02:31 nbd: also the network was stopped to early Jan 30 18:02:39 should be last thing befor umount Jan 30 18:02:49 imho Jan 30 18:05:26 i wasn't questioning your patches, just reiterating that even though more packages need changing, it's better than the alternative ;) Jan 30 18:06:14 nbd ohh okay Jan 30 18:06:36 maybe a increase revison on all modified packages is need .. forget about it Jan 30 18:07:12 but as long no one does a /etc/init.d/foo enable ... it wont work ... so it wont make a deffernce Jan 30 18:07:17 difference Jan 30 18:08:47 lars * r19422 /trunk/package/libtool/patches/100-libdir_path.patch: Jan 30 18:08:48 [package] libtool: Don't use the libdir path found in .la files. Jan 30 18:08:48 Fixes the build of various packages. Jan 30 18:09:56 nbd * r19423 /packages/utils/coreutils/Makefile: coreutils: split up package into small individual packages (patch by Daniel A. Nagy) Jan 30 18:09:58 nbd * r19424 /packages/utils/coreutils/Makefile: coreutils: fix a few failed configure tests Jan 30 18:10:01 nbd * r19425 /packages/utils/coreutils/Makefile: coreutils: replace the redundant package containing all coreutils binaries with a metapackage Jan 30 18:10:05 Guest82296: patch applied Jan 30 18:10:09 (with some followup fixes) Jan 30 18:12:59 nbd * r19426 /packages/net/ (3 files in 3 dirs): stop nfsd and sshd on reboot (thx, puchu) Jan 30 18:13:42 nbd: by the way, did you find time to look at the phy stuff? Jan 30 18:14:23 updated openwrt/upstream, http://pastehtml.com/view/5t6qnaq.html Jan 30 18:14:31 nbd but when u stop all eth devices befor u stop sshd its not worth anything Jan 30 18:14:37 the connections will still stuck Jan 30 18:14:48 i know Jan 30 18:14:53 i will commit your other changes as well Jan 30 18:14:57 okay Jan 30 18:15:08 though... what's the point of stopping syslogd after networking? Jan 30 18:15:26 umount didnt work when u log to a file Jan 30 18:15:35 instead of the ram buffer Jan 30 18:15:41 ah, didn't know it supported that, i thought it was only using the ram buffer Jan 30 18:15:49 ok, will commit that as well Jan 30 18:16:02 no when u start it with -s 16000 it will log to a file Jan 30 18:16:20 16000 is the size the logfile should at max have Jan 30 18:16:53 i use it to have a "history" of the last months instead of only the last minutes Jan 30 18:17:15 so stopping it wont hurt anyone :) Jan 30 18:17:50 nbd * r19427 /trunk/package/base-files/files/etc/init.d/network: when rebooting, stop the network interfaces last (just before umount) to allow services to shut down properly (thx, puchu) Jan 30 18:17:55 nbd * r19428 /trunk/package/base-files/files/etc/init.d/boot: stop syslogd before rebooting, otherwise umount might not work properly when logging to a file (thx, puchu) Jan 30 18:17:59 nbd * r19429 /trunk/package/base-files/files/etc/init.d/umount: when rebooting, use lazy umounts as fallback (patch by puchu) Jan 30 18:18:23 nbd dont forget about dropbear .... since most of the user use it (i guess) Jan 30 18:18:48 ok, your patch didn't seem to include that Jan 30 18:18:51 unless i overlooked it Jan 30 18:19:06 no it didnt include it Jan 30 18:19:14 i forgot to include it Jan 30 18:19:43 nbd * r19430 /trunk/package/dropbear/files/dropbear.init: stop dropbear at shutdown as well Jan 30 18:19:53 took care of it now Jan 30 18:19:58 thanks Jan 30 18:20:05 now the umount should work Jan 30 18:20:35 and stuck connections are things of the past ;) Jan 30 18:20:46 great Jan 30 18:50:52 nbd ping Jan 30 18:53:25 pong Jan 30 18:53:50 the fix for devtmpfs is ready Jan 30 18:53:53 do_mount_devfs() { Jan 30 18:53:53 + grep devtmpfs /proc/mounts | awk '{print $2}' | xargs -r umount Jan 30 18:53:53 mount devfs /dev -t devfs Jan 30 18:54:00 in a/package/base-files/files/lib/preinit/20_device_fs_mount Jan 30 18:54:06 could u add this? Jan 30 18:54:16 will do Jan 30 18:54:19 thanks Jan 30 18:54:20 did you test it? Jan 30 18:54:41 yes Jan 30 18:54:42 it works Jan 30 18:54:48 ok, will add it in a few minutes Jan 30 18:56:30 maybe add this line into a Jan 30 18:56:51 if grep devtempfs /proc/mounts > /dev/null; then Jan 30 18:57:11 @nbd Jan 30 18:57:23 but it works without it Jan 30 18:57:43 nbd dont commit it Jan 30 18:57:47 there is problem Jan 30 18:57:49 ok Jan 30 18:57:49 a Jan 30 19:32:02 nbd can u explain it? it works when do this on cmd line but not when i do it in the preinit script ... -o move also works there Jan 30 20:06:28 nbd i have a working version Jan 30 20:06:35 can u commit it? Jan 30 20:26:43 puchu: yes Jan 30 20:26:58 where is the patch? Jan 30 20:36:55 nbd hmm it doesnt work .. no matter where i put the code Jan 30 20:37:05 could it be that the ' from awk somehow dont work? Jan 30 20:37:58 it worked for me Jan 30 20:39:34 nbd where in preinit? Jan 30 20:39:40 or just the code on cmd line? Jan 30 20:39:44 on cmd line it also works for me Jan 30 20:40:25 grep devtmpfs /proc/mounts | awk '{print $2}' | xargs -r umount Jan 30 20:40:30 this is the code Jan 30 20:43:58 nbd? Jan 30 20:47:04 yes Jan 30 20:48:57 how about adding -l to the umount command and doing it just before mounting the tmpfs to /dev? Jan 30 20:50:14 yeah i added this at this point Jan 30 20:50:18 and dev was still mounted Jan 30 20:50:26 with -l? Jan 30 20:50:35 no just umount -r Jan 30 20:50:38 no just umount Jan 30 20:50:41 try with -l Jan 30 20:50:49 it works when the sys is up and running Jan 30 20:50:55 why doesnt it work when im in preinit? Jan 30 20:50:56 i know Jan 30 20:50:59 no idea Jan 30 20:51:11 linux is like windows at this point Jan 30 20:51:19 the output of umount would be nice Jan 30 20:51:29 i pipe this into a file Jan 30 20:51:32 and look at it Jan 30 20:51:58 nbd a better point is after hotplug2 is started Jan 30 20:51:59 i think Jan 30 20:52:04 then /dev is populated Jan 30 20:52:19 imho you should do it before the initial coldplugging Jan 30 20:52:26 okay Jan 30 20:52:37 then where tempfs is mounted to /dev Jan 30 20:52:39 kaloz * r19431 /trunk/target/linux/generic-2.6/patches-2.6.32/ (150-netfilter_imq.patch 205-skb_padding.patch): upgrade IMQ patch to the latest one, refresh patches Jan 30 20:53:14 yes, just before it Jan 30 20:53:22 befor? Jan 30 20:53:25 not after Jan 30 20:53:28 kaloz * r19432 /trunk/target/linux/avr32/ (Makefile config-default patches/001-fix_usart_clash.patch): upgrade avr32 to 2.6.32 Jan 30 20:53:37 yes, before you create a new /dev get rid of the old /dev Jan 30 20:53:54 devtmpfs is only useful for getting initial access to /dev/console Jan 30 20:54:04 nbd dev is not at /dev its at /rom/dev Jan 30 20:54:05 i think Jan 30 20:54:12 initially it's at /dev Jan 30 20:54:20 before / is moved to /rom Jan 30 20:54:44 where is / moved to rom Jan 30 20:54:55 ahh 70 Jan 30 20:54:59 70_pivot_jffs2_root: Jan 30 20:57:31 bbl Jan 30 20:57:35 nbd where is /dev moved after the pivor root? Jan 30 21:01:39 someone want to take a look at #5779? Jan 30 21:02:18 i'll take care of that when i get back **** BEGIN LOGGING AT Sat Jan 30 22:55:46 2010 Jan 30 23:22:21 swalker * r19433 /packages/multimedia/graphicsmagick/Makefile: [packages] GraphicsMagick: update to 1.3.8 Jan 30 23:29:02 cshore: I hereby declare the core as bugfree Jan 30 23:29:12 swalker * r19434 /packages/net/dmapd/ (4 files in 3 dirs): [packages] dmapd: update to 0.0.21 (#6516) Jan 31 00:15:31 cshore: ping Jan 31 00:41:48 rtz: pong Jan 31 00:43:02 cshore: you did some testing with qemu, right? Jan 31 00:43:13 rtz: yes Jan 31 00:43:40 cshore: which platforms? and were you able to emulate flash memory or only harddisks? Jan 31 00:44:31 rtz: I used x86 only, and it uses block2mtd to treat a harddrive as mtd, so you can test jffs, as well as ext2 Jan 31 00:45:05 cshore: any handy guides or recommended documentation? Jan 31 00:45:11 and which version did you use? Jan 31 00:47:05 rtz: hmm...docs....not sure on that, I just looked at the qemu man page Jan 31 00:47:05 version was qemu-kvm in Ubuntu 9.10...I look up the version, just a sec Jan 31 00:47:24 rtz: the main thing that's tricky is networking Jan 31 00:48:02 version 0.11.0 Jan 31 00:49:06 cshore: as I have a working console, i can do without networking Jan 31 00:49:08 cshore: I will take a look at it, thanks for the info Jan 31 00:58:22 rtz: np Jan 31 01:25:59 ping nbd Jan 31 01:29:55 kaloz * r19435 /trunk/target/linux/ixp4xx/patches-2.6.32/ (3 files): [ixp4xx] fix some compile warnings with 2.6.32 Jan 31 02:21:39 marek * r19436 /packages/net/batman-advanced/Makefile: Jan 31 02:21:39 batman-adv: automatically rebuild components when the config changes Jan 31 02:21:39 Signed-off-by: Linus Luessing Jan 31 02:32:11 ping xMff Jan 31 02:59:40 build #3 of octeon is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/3 **** ENDING LOGGING AT Sun Jan 31 02:59:57 2010