**** BEGIN LOGGING AT Fri Oct 05 02:59:57 2007 Oct 05 03:38:44 <{Nico}> nbd: ping Oct 05 03:38:51 pong Oct 05 03:39:23 <{Nico}> about the deps issue Weedy spotted : http://pastebin.com/m53c3526c Oct 05 03:41:04 <{Nico}> mac80211 has '+' for the crypto modules, which turns into a depends on kmod-crypto-core Oct 05 03:41:18 <{Nico}> any ideas how we can solve this? Oct 05 03:42:13 make the kmod-crypto-* select instead of depends on -core Oct 05 03:46:00 <{Nico}> ok Oct 05 06:36:58 I'm new here. I posted to openwrt-devel@lists.openwrt.org a couple of patches to diag.c and a bunch of fixes to motorola-bin.c. No reaction. How can I get these considered for adoption? Oct 05 06:39:02 <{Nico}> dhr: use ticket(s) Oct 05 06:41:28 OK. I added the least controversial one to https://dev.openwrt.org/ticket/1936, but so far no response. The others I would expect worth discussing. Oct 05 06:41:47 How does one get a login on trac? Oct 05 06:45:14 <{Nico}> dhr: you can't and don't need one to submit tickets and attach patches Oct 05 06:52:56 On a CFE-base system, can one create arbitrary new nvram variables? One of my suggestions was to change diag.c to look at what I hope is a new nvram variable, wpenwrt_platform. Would it be legal to issue the command "nvram put openwrt_platform=Buffalo WHR-G54S"? Oct 05 06:55:23 <{Nico}> dhr: we're trying to get rid of nvram in kamikaze Oct 05 06:59:48 Kaloz: did you hear anything about those wrt350nv2 ? Oct 05 06:59:56 morning btw Oct 05 07:01:50 OK. My idea is that diag.c might be wrong and it would be good to provide a way to override. I added code to diag.c that looked for the value of this new nvram variable in the table of platforms. It would be fine if there is another place where the user could put this override request. I just used nvram because diag.c already does that a lot. Oct 05 07:02:40 {Nico}: your point is exactly why I hoped for discussion on the mailing list. I don't know enough to be confident that what I did was appropriate. Oct 05 07:03:51 * dhr must go sleep but will look for a reply when back Oct 05 07:04:08 <{Nico}> dhr: the idea has always been not to introduced new nvram variables (already bloated) and try to find the one needed for proper detection Oct 05 07:06:37 <{Nico}> the detection shouldn't involve any user intervention Oct 05 07:52:58 blogic: nope, if I hear anythin, I let you know Oct 05 08:30:01 juhosg * r9132 /trunk/target/linux/adm5120/ (4 files in 3 dirs): [adm5120] remove dashes from profile names Oct 05 10:22:11 [florian], hi, are you here? Oct 05 10:22:51 [florian], have you looked at how webs works? Oct 05 10:23:26 [florian], since only thing I could make working is flashing somehow over it :( Oct 05 10:24:31 [florian], so I'd like to unbixify firmware, change it, and bix it back again, which is somehow horrifies me... :) Oct 05 10:27:01 <[florian]> slapin: pong Oct 05 10:27:12 <[florian]> slapin: I did not have time yet Oct 05 10:27:47 [florian], have you looked at my page - do you have any ideas/comments? Oct 05 10:28:21 <[florian]> slapin: yes I have been looking at your page Oct 05 10:29:49 [florian], I've added a few interfaces, any ideas about them? and I added exact erroro I had with bootp. Oct 05 10:30:49 [florian], btw, is it possible to flash without mtd character devices, using only block ones? Oct 05 10:30:57 * slapin feels terrified... Oct 05 10:31:25 <[florian]> slapin: yes it is possible Oct 05 10:31:40 [florian], any docs on that? Oct 05 10:31:50 <[florian]> slapin: you need a binary for that and it should be something like dd if=file.bin of=/dev/mtdblock/x Oct 05 10:32:28 [florian], and no tools to make sure, which is bad :( Oct 05 10:33:08 [florian], also I managed to get dumps of my bootloader, if you're interested I might share Oct 05 10:33:51 [florian], and what is MULTI they speak about? Oct 05 10:36:00 <[florian]> slapin: no idea about what multi is :) Oct 05 10:36:23 <[florian]> slapin: I see you can enable some sort of tftpd server, did you try sending something, even if rejected afterwards ? Oct 05 10:37:36 florian * r9133 /trunk/target/linux/rb532/files/arch/mips/rb500/ (devices.c prom.c): Finally fix the rb500 boot (#2436) Oct 05 10:38:33 florian * r9134 /tags/kamikaze_7.09/target/linux/rb532-2.6/files/arch/mips/rb500/prom.c: Backport the routerboad detection to 7.09 Oct 05 10:39:00 [florian], well, what I found that it was able to enable ethernet now (when just try to bootp it doesn't, since it does not react on cable insertion/removal), and _all_ ports (including WAN) are bridged. I found no arp similar to what is displayed though and was unable to send it anything. Oct 05 10:40:31 [florian], so it seems that wan is not different from other ports, and it seems to be software-locked to its speed (35mbit), like with shaper. Oct 05 10:41:24 <[florian]> that's possible Oct 05 10:41:34 <[florian]> most devices will bridge everything at boot time Oct 05 10:41:35 [florian], also funny to know - it is not possible to download anything to router through wifi - device accidentally sets its tcp window to 0 and download stalled, while it works over plain ethernet. Oct 05 10:41:43 <[florian]> because it simplifies the network code to be put in the bootloader Oct 05 10:45:55 [florian], I think it is not that trivial, since they should maintain bridge. A problem is that there's no these mac addresses there. I start to think they use some separate interface or PHY for development/debugging, not embedded. Oct 05 10:46:21 <[florian]> the switch default behavior is probably unbridged for every port Oct 05 10:47:27 [florian], also line with memload looks very curious... Oct 05 10:50:16 <[florian]> slapin: yes, that's a bit weird Oct 05 10:52:00 [florian], so I need ideas on where to look farther. Oct 05 10:53:12 <[florian]> slapin: have you tried the 'r' switch ? top run an image ? Oct 05 10:53:31 <[florian]> slapin: no, the 'g' switch is better Oct 05 10:53:43 <[florian]> slapin: I think it would load the image to ram and jump to it Oct 05 10:54:16 <[florian]> slapin: it might accept a bix image Oct 05 11:04:08 juhosg * r9135 /packages/net/uob-nomad/Makefile: [packages] uob-nomad depends on iptunnel4 since [8910] Oct 05 11:35:22 [florian], ok, I'll try tonight, but dunno how it will get it to RAM... Oct 05 11:39:34 florian * r9136 /trunk/target/linux/rb532/ (5 files in 3 dirs): Clean up the prom code / board detection (thanks Gabor), fix some warnings Oct 05 12:28:03 do anybody knows of javascript-capable console web browser? (to control my dlink a bit). Oct 05 12:28:44 maybe elinks Oct 05 12:31:43 loswillios, thanks! Oct 05 13:05:26 nbd * r9137 /trunk/package/busybox/patches/470-insmod_search.patch: fix compile error with 2.4 Oct 05 13:50:24 florian * r9138 /trunk/target/linux/rdc/files/arch/i386/mach-rdc/ (platform.c wdt.c): Led is not active low by default, register the watchdog Oct 05 13:51:24 ejka * r9139 /trunk/target/linux/generic-2.6/patches-2.6.23/190-netfilter_rtsp.patch: fix rtsp patch Oct 05 14:00:37 florian * r9140 /trunk/target/linux/rdc/Makefile: RDC is not broken with trunk Oct 05 14:50:54 juhosg * r9141 /trunk/target/linux/adm5120/ (4 files in 4 dirs): Oct 05 14:50:54 [adm5120] Edimax BR-6104K/KP specific fixes Oct 05 14:50:54 * add profile for Edimax BR-6104KP which contains the experimental USB driver by default Oct 05 14:50:54 * generate CSYS style images which are uploadable via ADMBoot/TFTPBoot (saves 64KB of space) Oct 05 14:50:54 * update kernel configuration Oct 05 14:57:23 ejka, around? Oct 05 14:57:33 yes Oct 05 14:58:46 how do i test 2.6.23 on ar7? edit target/linux/ar7/Makefile? Oct 05 14:58:51 yes Oct 05 14:59:10 what version should i put? 2.6.23.6 Oct 05 15:02:10 * Weedy waits Oct 05 15:04:49 * juhosg think Weedy must wait at least 1,5 months for the 2.6.23.6 Oct 05 15:05:14 Weedy: maybe 2.6.23-rc9 Oct 05 15:10:46 if you wanna try 2.6.23.6 put that version number :P Oct 05 15:17:11 .config:52:warning: trying to assign nonexistent symbol TARGET_ar7_Default Oct 05 15:28:06 arch/mips/ar7/vlynq-pci.c:418: error: conflicting types for 'pcibios_map_irq' Oct 05 15:28:06 include/asm/pci.h:59: error: previous declaration of 'pcibios_map_irq' was here Oct 05 15:28:10 darn Oct 05 15:59:44 juhosg * r9142 /packages/libs/alsa-lib/Makefile: [packages] alsa-lib depends on kmod-soundcore since [8910] Oct 05 16:08:39 [florian], ping Oct 05 17:54:39 ejka * r9143 /trunk/target/linux/ar7/ (17 files in 7 dirs): cleanup vlynq. drop vlynq-pci Oct 05 17:55:16 ejka * r9144 /trunk/package/acx-mac80211/patches/ (002-acx_vlynq.diff 002-no_reset.diff): add rough vlynq support Oct 05 17:56:17 ejka * r9145 /trunk/package/mac80211/ (5 files in 4 dirs): add minimal uci support Oct 05 20:09:59 nbd * r9146 /trunk/package/busybox/patches/470-insmod_search.patch: fix segfault with insmod on 2.4 Oct 05 20:27:44 blogic * r9147 /packages/libs/tiff/ (. Makefile): added libtiff Oct 05 20:29:07 blogic * r9148 /packages/XOrg/ (4 files in 4 dirs): removed duplicate freetype Oct 05 20:34:05 blogic * r9149 /packages/libs/alsa-lib/Makefile: fixed broken alsa-lib depedency introduced by #8910 Oct 05 20:34:15 ejka, 9148 locks up with 2.6.23-rc9 , im installing 2.6.22.4 Oct 05 20:34:18 acx: found TI TNETW1130-based wireless network card at vlynq0, irq:80, phymem:0x Oct 05 20:34:18 4000000, mem:0xa4000000 Oct 05 20:34:18 initial debug setting is 0x000A Oct 05 20:57:05 ejka, stalls at requesting firmware image 'tiacx111c16' Oct 05 21:44:45 nbd * r9150 /trunk/ (include/target.mk scripts/metadata.pl): allow targets to select/deselect config symbols Oct 05 21:52:06 nbd * r9151 /trunk/ (Config.in package/Makefile): add option for cleaning all ipkg info before the image build Oct 05 23:07:15 blogic * r9152 /packages/XOrg/app/fontconfig/Makefile: updated fontconfig version to be compatible with newer freetype Oct 05 23:07:40 blogic * r9153 /packages/XOrg/lib/cairo/Makefile: added missing libpng dependency to cairo Oct 06 00:31:49 blogic * r9154 /packages/libs/glib2/Makefile: make glib2 stage additional pc files Oct 06 00:33:03 blogic * r9155 /packages/XOrg/lib/ (atk/Makefile pango/Makefile): make pango and atk find the correct location of iconv and intl Oct 06 00:33:38 blogic * r9156 /packages/XOrg/lib/gtk-2.8.20/Makefile: fix dependencies of gtk+-2 Oct 06 01:01:21 blogic * r9157 /packages/XOrg/lib/gtk-2.8.20/Makefile: not files from gtk+-2 were staged correctly Oct 06 02:01:33 blogic * r9158 /packages/XOrg/lib/ (gtk-1.2.10/Makefile gtk-2.8.20/Makefile): seds in gtk+-2 were effecting gtk+-1 files **** ENDING LOGGING AT Sat Oct 06 02:59:56 2007