**** BEGIN LOGGING AT Tue Feb 03 02:59:57 2009 Feb 03 03:21:34 kupesoft: I'll get to it in a few minutes, thanks for the reminder. Feb 03 03:21:40 :) Feb 03 03:55:45 I've enjoyed writing the few packages I've got in OpenWrt, and ultimately down the road would like to become a maintainer Feb 03 04:05:46 that is definitely a possibility, just keep contributing good code :) Feb 03 04:41:16 Like: main() { while (1) {malloc(1); fork();} } Feb 03 04:45:18 ;/ Feb 03 05:00:28 evil != good Feb 03 05:44:57 bah, I've spent far too long trying to figure this out. the package itself is fine. Feb 03 05:45:18 agb * r14380 /packages/mail/msmtp/ (. Makefile): (log message trimmed) Feb 03 05:45:18 New Package -- msmtp Feb 03 05:45:18 Here's a patch for msmtp, an advanced but still lightweight sendmail Feb 03 05:45:18 replacement that forwards email upstream to an SMTP server. This Feb 03 05:45:18 software is actively developed, unlike ssmtp which is out of date and Feb 03 05:45:19 only seems to be linked on the Debian repository. I was able to easily Feb 03 05:45:23 forward email upstream using GMail/Google Apps with this package, but Feb 03 05:45:27 agb * r14381 /packages/mail/msmtp/Makefile: [packages] mail/msmtp: mark /etc/msmtprc as a conffile Feb 03 05:50:06 danke Feb 03 05:51:10 :) Feb 03 05:51:23 Bartman007: what was so confusing? Feb 03 05:53:20 kupesoft: the conffile section I added does not appear to work correctly when switching to/from the -nossl variant Feb 03 05:54:10 I didn't mark it as a conffile because I didn't think it'd be useful on removal, I dunno Feb 03 05:56:21 if it is not marked as a conffile it will be overwritten with the distro version when upgrading the package Feb 03 05:58:12 Ah, okay Feb 03 06:47:21 allo Feb 03 07:13:39 hmm Feb 03 07:26:35 juhosg * r14382 /trunk/target/linux/ar71xx/ (4 files in 3 dirs): [ar71xx] fix handling of invalid arguments passed by some bootloaders (thanks to Jeff Hansen and to Jonas) Feb 03 07:41:25 hi juhosg Feb 03 07:41:31 blogic: huh? Feb 03 07:45:48 hi blogic Feb 03 07:45:55 hello Feb 03 07:53:49 juhosg * r14383 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-mzk-w04nu.c: [ar71xx] mzk-w04nu: fetch MAC address from the flash Feb 03 08:45:10 juhosg * r14384 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/prom.c: [ar71xx] parse RedBoot environments as well (patch by Jonas ) Feb 03 09:18:32 juhosg * r14385 /trunk/package/mac80211/ (2 files in 2 dirs): [package] mac80211: update compat-wireless to 2009-02-02 Feb 03 10:24:45 Debugging interface (SKY2_DEBUG) [N/y/?] (NEW) Feb 03 10:25:21 x86-alix compile of r14385 Feb 03 10:29:23 russell_: can you try this patch? http://openwrt.pastebin.com/d44abad99 Feb 03 10:37:08 * russell_ tries Feb 03 10:39:47 that fixes the config query, as anticipated Feb 03 10:40:12 committed. Feb 03 10:40:41 (once the server finishes) Feb 03 10:40:46 agb * r14386 /trunk/target/linux/generic-2.6/config-2.6.28: [kernel] generic-2.6: sync 2.6.28 config Feb 03 10:56:09 In file included from dsl_qos_queue.c:15: Feb 03 10:56:32 /aux/src/openwrt/staging_dir/target-i386_uClibc-0.9.29/usr/include/linux/netfilter.h:44: error: field 'in' has incomplete type Feb 03 10:56:32 /aux/src/openwrt/staging_dir/target-i386_uClibc-0.9.29/usr/include/linux/netfilter.h:45: error: field 'in6' has incomplete type Feb 03 10:56:55 make[4]: *** [dsl_qos_queue.o] Error 1 Feb 03 10:56:55 make[4]: Leaving directory `/aux/src/openwrt/build_dir/target-i386_uClibc-0.9.29/dsl_qos_queue' Feb 03 11:43:32 hmm. The xorg-kdrive package is build-fail again (maybe i'd just patched locally) Feb 03 11:43:51 seems like there was a ticket for that Feb 03 11:44:12 * russell_ vaguely remembers added a patch to that ticket Feb 03 11:45:24 https://dev.openwrt.org/ticket/4473 Feb 03 13:28:48 is there a shortcut to build a package in a feed? Feb 03 13:29:00 like, make package/samba3/install Feb 03 13:30:30 ssssyes Feb 03 13:30:37 chihchun: make package/samba3/install Feb 03 13:30:44 :) Feb 03 13:31:05 chihchun: make package/samba3/{clean,prepare,compile,install} V=99 QUILT=1 DEVELOPER=1 Feb 03 13:31:11 would be the full line Feb 03 13:31:13 mirko * r14387 /feeds/efl/ (15 files in 15 dirs): upgraded enlightenment, efl libraries and its python bindings to svn rev. 38886 Feb 03 13:31:31 can I do this ? - make package/feeds/samba3/install Feb 03 13:31:41 there are some packages in different feeds Feb 03 13:32:06 and I only want to build the package in the feeds Feb 03 13:32:29 err Feb 03 13:32:33 it works, Feb 03 13:34:04 ok, I see the problem, I can "make V=99 package/luci-uci/install", but not "package/luci-uci/install" Feb 03 14:11:36 florian * r14388 /packages/libs/zaptel-1.4.x/patches/330-remove_m64_switch.patch: [package] fix zaptel compilation on non-64bits platforms, should not break x86_64 Feb 03 14:19:43 florian * r14389 /packages/libs/zaptel-1.4.x/patches/ (330-remove_m64_switch.patch 330-uname_m.patch): [package] rework 330-remove_m64_switch patch to be less intrusive, thanks nbd Feb 03 14:20:15 <{Nico}> chihchun: make package/feeds/packages/samba3/install Feb 03 14:20:34 {Nico}: thanks! Feb 03 14:21:08 <{Nico}> clean,prepare,compile should work also Feb 03 14:26:28 florian * r14390 /trunk/package/kernel/modules/netdevices.mk: [package] fix tg3 libphy dependency, thanks Acinonyx Feb 03 14:39:19 florian * r14391 /packages/libs/zaptel-1.4.x/Makefile: [package] also pass ARCH to the user-space applications to compile properly Feb 03 15:26:16 mirko * r14392 /trunk/target/linux/s3c24xx/config-2.6.28: enable inotify kernel support Feb 03 16:23:54 nico * r14393 /trunk/toolchain/binutils/ (7 files in 2 dirs): [toolchain] binutils: add support for v2.19.1 (from #4492, thanks to dwrobel) Feb 03 18:18:17 mb * r14394 /trunk/target/linux/brcm47xx/patches-2.6.28/810-ssb-add-pmu-support.patch: bcm47xx: Implement the remaining parts of PMU init. Feb 03 18:38:04 FYI - I got client mode working fine on the ath9k - did indeed forget wpa_supplicant, oops. Feb 03 18:38:14 although the performance is horrible, need to do some more tests Feb 03 18:39:31 RoundSparrow: what performance do you gfet ? Feb 03 19:25:12 /aux/src/openwrt/build_dir/linux-x86_generic/linux-uvc-r261/uvc_v4l2.c:986: error: too many arguments to function 'v4l_compat_translate_ioctl' Feb 03 19:25:12 make[5]: *** [/aux/src/openwrt/build_dir/linux-x86_generic/linux-uvc-r261/uvc_v4l2.o] Error 1 Feb 03 19:25:12 make[4]: *** [_module_/aux/src/openwrt/build_dir/linux-x86_generic/linux-uvc-r261] Error 2 Feb 03 19:28:13 blogic: not getting 1MB/s (10Mb) Feb 03 19:28:28 cpu usage is normal, this is from an OS X atheros laptop Feb 03 19:29:14 cool, nice work Feb 03 19:29:21 1MB/s is not bad Feb 03 19:29:37 erm... i assume the power calibration code is still WIP Feb 03 19:30:48 yha, from wht I see on OS X side the link is retraining too much, jumps from 130 down to 54 and sometimes I even see 0 for a brief second Feb 03 19:31:07 ouch :) Feb 03 19:31:10 but still Feb 03 19:31:29 too much retrain, agreed this driver is new Feb 03 19:59:16 florian * r14395 /trunk/target/linux/brcm63xx/config-2.6.27: [brcm63xx] resync kernel configuration Feb 03 20:26:10 florian * r14396 /trunk/package/base-files/files/etc/init.d/boot: [base-files] allow specifying port where to log (#2019) Feb 03 20:36:33 juhosg: on the Planex ap81, what's the prospect of flashing it with OpenWRT? Not to rush you, just curious. Have you found any way to recovery load that ... once we flash OpenWRT - no way to go back to factory firmware? Feb 03 20:47:25 RoundSparrow: the u-boot has no recovery mode, and i have not analyzed the firmware format yet Feb 03 20:47:37 but if something goes wrong we can reflash the frimware either from u-boot or from a tftpboot-ed kernel Feb 03 20:48:32 ok, cool, didn't know we could flash from u-boot Feb 03 20:49:07 so I presume from u-boot there would be a way to load/restore the vendor firmware? Feb 03 20:51:40 so far on the TPLink there is no discovered recovery mode.... the netgear wnr2000 with u-boot has a tftp recovery mode it goes into if the main firmware fails Feb 03 20:54:54 on the planex the u-boot has standard commands to erase/write the flash, so it should be possible to write any code into it, although i did not try it yet Feb 03 21:02:10 juhosg: someone mention this on the main #Openwrt channel about how luci lets you pick 5ghz channels... someone did this on the forum with ath9k and they said it wrote to the eprom of the adapter and would not work even after cold boot Feb 03 21:02:22 they said eventually they got it working again, but did not detail how Feb 03 21:08:40 RoundSparrow: really? that's bad... well luci only writes uci which is translated into iw commands later... Feb 03 21:08:51 well, these boards have no eeprom at all, and the ath9k driver can't write to the flash Feb 03 21:09:09 apart from this, the driver can write invalid values to some registers, and this may cause problems Feb 03 21:09:53 if this is the case and there's invalid stuff in uci then the device might got stuck in a reboot-oops-llop Feb 03 21:09:57 *loop Feb 03 21:14:38 yes, probably Feb 03 21:17:12 yes, user reported this on the trendnet 652 with ath9k, he set luci to 5ghz channel (only 2.4ghz hardware) and then even flash back to factory firmware would not work - router reboots on wireless access Feb 03 21:17:37 I'm thinking it set registers on the internal eprom of the ar9102 Feb 03 21:17:47 he did not detail how he recovered it Feb 03 21:18:27 the register setting would have to occur during init for example, as a result of wrong config settings Feb 03 21:18:42 so maybe the openwrt recovery mode would work in such a case Feb 03 21:19:08 message #13 and message #15: http://forum.openwrt.org/viewtopic.php?pid=79626#p79626 Feb 03 21:19:20 juhosg said there is no way to write to the eeprom through ath9k, so it's highly unlikely Feb 03 21:21:22 well this is good, fear it was like situation with the Intel e1000 eprom ;) Feb 03 21:21:39 anyway, i ask the person to detail how they got it working again in case someone else does this. Feb 03 21:25:57 nico * r14397 /trunk/toolchain/ (7 files in 7 dirs): [toolchain] cosmetic & coherency fixes Feb 03 21:45:05 juhosg: FYI - I try client mode against a netgear broadcom chipset router, I see behavior like this.... pings go up and data rate drops. This is after a cold boot of trendnet tew-652brp with newest compat-wireless (openwrt svn) from today http://pastebin.com/f1b220180 Feb 03 21:46:46 this is with no LAN cables connected to router, I am strictly controlling from rs232 serial port... Feb 03 22:01:31 this is even more dramatic of the issue with ath9k on svn 14394 http://pastebin.com/d5b257bfa Feb 03 22:07:18 anyone know how to add timestamp to my dmesg logs? Feb 03 22:27:52 florian * r14398 /trunk/package/dnsmasq/Makefile: [package] add dnsmasq config file to the conffiles section Feb 03 22:28:19 florian * r14399 /trunk/package/dropbear/ (Makefile files/dropbear.init): [package] add dropkey ssh keys and config files to the conffiles section (#2014) Feb 03 23:46:51 thepeople * r14400 /packages/net/ntpclient/Makefile: ntpclient depends on librt Feb 03 23:49:25 Hello ;) Feb 03 23:49:32 hey Feb 03 23:50:38 i have coded a program that allows you to access nvram(set/unset/show/commit etc..) in 2.6 as in 2.4 . Feb 03 23:50:59 do share Feb 03 23:51:18 i am not done at 100% yet, must betatest it =) Feb 03 23:52:01 do you think it's usefull? Feb 03 23:52:04 yes Feb 03 23:52:47 you may want to submit a patch to either trac or the mailing list for testing Feb 03 23:53:48 yeah, will do that. just have some coding cleanup's and want to test it on some of my other linksys routers. Feb 03 23:54:03 ok Feb 03 23:54:18 heh, the first one got bricked :D but i got a jtag cable so it was kinda easy to fix. Feb 03 23:54:28 :-) Feb 03 23:54:33 =) Feb 04 00:07:05 nbd: ping Feb 04 00:07:30 nbd: sehr sehr coole stats :) Feb 04 00:07:45 hast du erklaerungen fuer die parameter? Feb 04 00:55:41 Worth anyone's time bringing imagemagick to openwrt? Feb 04 00:57:05 I mean for me to bring it ;/ Feb 04 00:59:42 if you are packaging it anyhow it could be useful for someone Feb 04 01:01:24 I wanted to experiment with the "upside-down-ternet" ;/ Feb 04 01:01:41 But packaging imagemagick is going to be nontrivial Feb 04 01:02:53 I was looking for basic image processing Feb 04 01:16:06 Any suggestions of alternatives? Feb 04 01:16:17 I could just offload the work, Feb 04 02:00:44 florian * r14401 /trunk/package/kernel/modules/i2c.mk: [package] correct kernel modules dependencies against kmod-i2c-core, they already depend on kmod-i2c-algo-bit, which depends on kmod-i2c-core (#4330) Feb 04 02:05:27 florian * r14402 /branches/8.09/package/kernel/modules/i2c.mk: backport r14401 to 8.09 **** ENDING LOGGING AT Wed Feb 04 02:59:58 2009