**** BEGIN LOGGING AT Thu Feb 23 10:59:56 2006 Feb 23 14:06:16 morning dudes. Feb 23 14:07:32 morning Feb 23 14:11:26 NAiL, is it me or you have a real bizzare sleeping schedule? It seems some days you are sleeping at this time?? Feb 23 14:12:04 heh. I sleep when I have to ;) Feb 23 14:12:31 I'm kinda tired now, since I've been up all night Feb 23 14:12:42 Got the DoC kindaworking though :-D Feb 23 14:12:54 (and my LCD-display) Feb 23 14:13:02 http://www.nslu2-linux.org/gallery/album01/img_1983 Feb 23 14:15:02 (http://www.nslu2-linux.org/gallery/DS101-hardware-mods) Feb 23 14:19:10 hehehe Feb 23 14:19:20 nice. Feb 23 14:19:31 what kind of LCD is that? Feb 23 14:19:58 Ah. the internal one. I see. Feb 23 14:20:07 what kind of interface does it have to the CPU? Feb 23 14:25:40 It's a sparkfun serLCD, which is connected to a pl2303 and soldered to the internal USB-prot Feb 23 14:25:44 s/prot/port/ Feb 23 14:25:44 NAiL meant: It's a sparkfun serLCD, which is connected to a pl2303 and soldered to the internal USB-port Feb 23 14:30:18 It's a hd44780 with a parallell->serial thingo tacked on Feb 23 14:30:28 I can control the brightness on the bugger :-D Feb 23 14:31:44 My DS101 has a couple of hardware mods so far. There's two pl2303's inside the case, one for serial console -> pc usb, and the other one for the lcd-display Feb 23 14:31:56 (And there's *plenty* of room to spare) Feb 23 14:40:56 rwhitby: ? Feb 23 14:41:19 yes? Feb 23 14:49:52 Does the openslug suffer from the same USB2.0 bug as OpenWrt have for a looong time (and still is) - that sometimes the kernel cant read the filesystems, and sometimes even corrupts it...? Feb 23 14:50:44 not to my knowledge, no. Feb 23 14:51:03 hmm..odd Feb 23 14:51:12 scary Feb 23 14:51:15 according to some of the wrt guys, it has to do with a kernel.org bug Feb 23 14:51:23 OpenWrt uses a 2.4-kernel, no? Feb 23 14:51:24 which kernel are you guys using inthe slug ? Feb 23 14:51:27 yeah Feb 23 14:51:30 you guys use 2.6 ? Feb 23 14:51:34 yep. Feb 23 14:51:36 ok Feb 23 14:51:39 explains it Feb 23 14:51:40 2.6.16-rc4 in head, 2.6.15.1 in stable Feb 23 14:52:06 ffs - if only someone could write thoose wireless drivers so that openwrt could move on to 2.6 for real :) Feb 23 14:52:43 I bet Oleg would love to see some 2.6 drivers as well - so he could follow suit with you guys better as well Feb 23 14:53:48 2.6 on the wrt would be nice Feb 23 14:53:57 Remember that there are some broadcom binaries only and that move to 2.6 in not possible! Feb 23 14:54:22 well, it's possible but takes a lot of work Feb 23 14:55:10 Reverse engineering of wireless stuff? No thanks. Feb 23 14:55:25 or write a wrapper Feb 23 15:22:38 wireless drivers are in the works as far as I know Feb 23 15:23:31 do you think the performance-lead on file-transfering that the NSLU2 box have is connected to the 2.6 vs. 2.4 stuff as well - or is the NSLU2 better harware wise than say, the ASUS WL-500G Deluxe? Feb 23 15:23:39 the CPU dosnt seem to be faster Feb 23 15:33:23 How fast is the wl500gx? Feb 23 15:33:33 rwhitby: what's up with the joining/leaving? ;) Feb 23 15:38:37 mine wl500g with loadavg 4 shows the following iostat 42.30tps 2792.07Blk_read/s 54.07Blk_wrtn/s Feb 23 15:40:01 s/500g/500gx/ Feb 23 15:40:01 oleo meant: mine wl500gx with loadavg 4 shows the following iostat 42.30tps 2792.07Blk_read/s 54.07Blk_wrtn/s Feb 23 15:42:34 sorry...was a bit away Feb 23 15:42:50 the wl500gx is a 200mzh cpu....but samba gives me only some 1MB/sec appx Feb 23 15:43:24 with 100% cpu load, caused by samba + USB2 drivers Feb 23 15:43:36 ftp and sftp yields simular speeds Feb 23 16:01:57 I would also like faster wl500gx disk subsystem. But I cannot answer if 2.6 is the cure. USB is eating 50% of CPU. So the question is "Is USB in 2.6.16 faster than 2.4.25?". Feb 23 16:05:43 yeah...this is the real question. You are absolutely right! Feb 23 16:06:00 I am getting quite sick of the fact that my data becomes corrupted Feb 23 16:06:27 almost so sick of it, that I am considering getting a NSLU2 for samba serving Feb 23 17:14:50 03azummo * 10kernel/2.6.16/ (patch-2.6.16-rc4-ide2 series patch-2.6.16-rc4-ide1): upgraded to patch-2.6.16-rc4-ide2 Feb 23 17:15:52 03azummo * 10kernel/2.6.16/leds-class.patch: refreshed Feb 23 17:21:34 03azummo * 10kernel/2.6.16/70-artop-latency.patch: Signed and described. Feb 23 17:22:35 03azummo * 10kernel/2.6.16/ (5 files): Added DSM-G600 and refreshed conflicting patches. Feb 23 17:24:17 03azummo * 10kernel/2.6.16/KERNEL: This file will tell you which kernel version to use Feb 23 17:30:46 03azummo * 10kernel/2.6.16/ (11 files): upgraded to 2.6.16-rc4-git5 Feb 23 19:13:57 03bzhou * 10unslung/sources/tethereal/configure.in.patch: adding configure.in.patch for cross compilation Feb 23 20:27:04 03jp30 * 10unslung/Makefile: optware: demote mini_httpd Feb 23 20:45:01 03jp30 * 10unslung/make/libftdi.mk: libftdi: strip and remove libtool archive from package Feb 23 20:56:52 03jp30 * 10unslung/Makefile: optware: promote libftdi Feb 23 22:16:26 eno, ping? Feb 23 22:17:22 03bzhou * 10unslung/make/lighttpd.mk: lighttpd cross compiled Feb 23 22:17:37 hi jp30 Feb 23 22:18:01 hi, eno: i'm working on promoting librsync. do you want it to generate a shared lib? Feb 23 22:18:47 shared lib usu. is a good idea, right? Feb 23 22:19:34 yeah, and it's bad if a shared library depends on an unshared one (which might happen when your python packages depend on librsync) Feb 23 22:20:12 i'll try to make it generate a shared lib - can you check that my changes don't break anything at runtime? Feb 23 22:20:23 sure, will do Feb 23 22:21:15 ok, expect some changes to librsync.mk to show up soon. i'm also going to make it depend on popt and bzip2, since it picks those up from staging at configure time on the official build system Feb 23 22:25:20 any idea for error below: Feb 23 22:25:22 libtool: link: cannot find the library `/mnt/sdb1/bzhou/devel/unslung/toolchain/armv5b-softfloat-linux/gcc-3.3.5-glibc-2.2.5/armv5b-softfloat-linux/lib/./libstdc++.la' Feb 23 22:25:48 it works on another machine Feb 23 22:26:46 03jp30 * 10unslung/ (make/librsync.mk make/popt.mk sources/popt/control): librsync: make a shared library, depend on popt and bzip2; popt: bring popt.mk into the 21st century Feb 23 22:27:31 eno, what are you building when you see that error? Feb 23 22:27:38 tethereal Feb 23 22:28:19 do you have any .la files in staging/opt/lib? Feb 23 22:29:56 a couple Feb 23 22:31:20 could be the problem... are any of them involved in the failing link? Feb 23 22:31:25 including staging/opt/lib/libpcre.la, which references libstdc++ Feb 23 22:31:51 hah! pcre should not be installing a libtool archive in the staging - that will be the problem Feb 23 22:32:21 so for now just rm it? Feb 23 22:32:33 libtool archives need to be specially patched before being edited in staging - in general, it is easiest to delete them Feb 23 22:32:51 alright Feb 23 22:32:54 i'll look at pcre.mk and see if it needs correcting Feb 23 22:32:55 thx Feb 23 22:33:05 s/edited/installed/ Feb 23 22:33:09 could be an old left-over Feb 23 22:34:02 looks like it is - pcre-stage should have deleted the libtool archive Feb 23 22:34:12 restarting tethereal build ... Feb 23 23:11:06 eno, have you had a chance to check librsync and py-rdiff-backup following my changes? Feb 23 23:16:21 03bzhou * 10unslung/make/ (libpcap.mk tcpdump.mk): upgrade libpcap to 0.9.4, and tcpdump to 3.9.4 Feb 23 23:20:08 jp30-work, am doing the runtime testing right now Feb 23 23:24:10 jp30-work, rdiff-backup works fine, thank you Feb 23 23:29:21 03jp30 * 10unslung/Makefile: optware: many promotions Feb 23 23:31:18 03bzhou * 10unslung/ (make/tethereal.mk Makefile): tethereal ready for testing Feb 23 23:58:46 Hi, where can i find the log for the changes in the cvs? Feb 24 00:04:22 EvilDevil, i'm not sure i understand the question, cvs log? Feb 24 00:05:26 on the web, i use this RSS feed: http://cia.navi.cx/stats/project/nslu2-linux/.rss Feb 24 00:10:13 or http://cvs.sf.net/viewcvs.py/nslu/ Feb 24 00:11:09 eno, you need to specify aclocal and automake versions when using autoreconf... i have fixed this for you in tethereal Feb 24 00:11:57 the official build host defaults to automake 1.4 which is not what you want Feb 24 00:12:31 ic, thx jp30-work Feb 24 00:13:03 eno that's what i meant :-) Thanks Feb 24 00:13:57 03jp30 * 10unslung/make/tethereal.mk: tethereal: specify automake version when autoreconfing Feb 24 00:22:04 can someone help me / point me to some instructions? i need to enable some extra kernel-modules which aren't build by default. i changed the defconfig in openembedded/packages/linux/ixp4xx-kernel/2.6.15.3/ and removed the ixp4xx-kernel-2.6.15.3-r0.0.do_compile in the stamps directory, but make debianslug-image seems to overwrite my changes all the time Feb 24 00:38:40 03jp30 * 10unslung/make/tethereal.mk: tethereal: remove libtool archives Feb 24 00:40:12 03jp30 * 10unslung/Makefile: optware: promote tethereal Feb 24 01:09:18 EvilDevil: change the defconfig, then blow away tmp/{work,stamps}/ixp4xx-kernel* Feb 24 01:09:33 assuming you're changing the right defconfig, then it will pick it up. Feb 24 01:09:47 just removing the compile stamp is not enough, cause the configure will not rerun. Feb 24 01:12:59 rwhitby which defconfig? i found several ones. Is it the defconfig in ~/slug/openembedded/packages/linux/ixp4xx-kernel/2.6.15 ? Feb 24 01:34:17 i changed the defconfig in ~/slug/openembedded/packages/linux/ixp4xx-kernel/2.6.15, did rm -Rf ~/slug/debianslug/tmp/{work,stamps}/ixp4xx-kernel*, and ran make debianslug-image. Feb 24 01:36:42 the defconfig in debianslug/tmp/work/ixp4xx-kernel-2.6.15.3-r0.0/ is now the same as the one in openembedded,but the .config in ~/slug/debianslug/tmp/work/ixp4xx-kernel-2.6.15.3-r0.0/linux-2.6.15.3 is still unchanged (and no extra modules in tmp/deploy/ipk ) Feb 24 02:22:38 03msteveb * 10unslung/make/ftpd-topfield.mk: ftpd-topfield: Updated to 0.6.5 Feb 24 03:52:20 03bzhou * 10unslung/ (Makefile make/py-nose.mk): py-nose ready for testing Feb 24 05:36:30 03bzhou * 10unslung/make/ (py-paste.mk py-pastescript.mk): two more turbogears dependencies Feb 24 06:17:16 morning, anyone here to help me with compiling additional modules for the debianslug-image which are very recalcitrant ? Feb 24 06:32:40 03bzhou * 10unslung/make/git-core.mk: upgrade to 1.2.2 Feb 24 07:11:26 anyone able to help with debianslug-image and enabling additional modules ( dallas 1-wire ) ? Feb 24 08:02:54 jp30-work: minihttpd in optware conforms to ipkg naming scheme from start! Feb 24 08:04:02 * NAiL had to look up recalcitrant Feb 24 08:42:45 03orepvik * 10kernel/2.6.16/defconfig: Add DoC support to defconfig Feb 24 08:53:16 03orepvik * 10kernel/2.6.16/ (97-ds101-doc.patch series): Add DoC support for DS101 Feb 24 10:47:35 Now I'm getting these warnings when trying to compile a 2.6.16-rc4-git5 kernel: Feb 24 10:47:38 .config:15:warning: trying to assign nonexistent symbol CLEAN_COMPILE Feb 24 10:47:40 .config:1029:warning: trying to assign nonexistent symbol SENSORS_RTC8564 Feb 24 10:47:43 .config:1031:warning: trying to assign nonexistent symbol RTC_X1205_I2C Feb 24 10:57:53 and these warnings: Feb 24 10:57:54 drivers/scsi/libata-core.c: In function `ata_pio_data_xfer': Feb 24 10:57:54 drivers/scsi/libata-core.c:3176: warning: passing arg 2 of `__ixp4xx_outsw' from incompatible pointer type Feb 24 10:57:57 drivers/scsi/libata-core.c:3178: warning: passing arg 2 of `__ixp4xx_insw' from incompatible pointer type **** ENDING LOGGING AT Fri Feb 24 10:59:56 2006