**** BEGIN LOGGING AT Wed Oct 08 02:59:57 2008 Oct 08 05:57:55 Bartman007: i managed to rube-goldberg together a line of binary communication with the ar360w3g's original firmware. i estimate that i will have the flash contents dumped within 15 min Oct 08 05:59:12 any paperclips involved? Oct 08 05:59:58 it involves awk, shell, base64, base91, telnetd, and exec Oct 08 06:00:45 MacGuyver wouldn't be pleased, no paperclips Oct 08 06:01:39 but neat Oct 08 06:01:41 i encased a max3232 in a ball of candle wax Oct 08 06:04:42 of course, rather than being shaped like a ball, it's shaped like a chocolate easter bunny that's been sitting in a car under the sun with no shades all day Oct 08 06:13:30 * Weedy stabinates blogic Oct 08 07:20:46 [florian]: ok, i have the dump of the ar360w3g flash contents Oct 08 07:32:54 <[Fate]> is there a way on atheros/redboot to just load the kernel/rootfs and execute it without flashing it? Oct 08 07:34:33 [Fate]: fis load, then exec? Oct 08 07:35:48 <[Fate]> to which load addresses is irrelevant? Oct 08 07:35:54 <[Fate]> or platform-specific? Oct 08 07:36:24 oops, not fis load, just load Oct 08 07:38:43 <[Fate]> I guess my problem is how to have kernel and rootfs placed at correct addresses to that the kernel finds the latter without actually changing the flash Oct 08 07:41:05 IIRC the openwrt atheros kernel is hardcoded to be loaded rom 0x80041000, might want to confirm that though Oct 08 07:41:22 s/rom/from/ Oct 08 07:44:36 night [Fate] Oct 08 07:44:38 * Bartman007 crashes into bed Oct 08 07:46:00 <[Fate]> n8 Oct 08 08:12:42 Weedy: y0 Oct 08 08:33:54 [florian]: around? Oct 08 08:36:25 <[florian]> sn9: yep Oct 08 08:37:26 <[florian]> sn9: read you got a dump of the ar360w3g, which bootloader does it use ? Oct 08 08:38:01 it's RDC mnBIOS, like in the anas350 that [mbm] examined Oct 08 08:38:21 <[florian]> what is it capable of ? Oct 08 08:38:35 the entire bootloader fits into 8KB, with room to spare Oct 08 08:39:29 <[florian]> so it basically drivers the uart and does a jump to the kernel ? Oct 08 08:39:59 it doesn't use the uart apparently Oct 08 08:40:29 it does have some text in it, apparently for a nonexistent vga Oct 08 08:43:11 however, the erase block containing the bootloader also appears to contain a recovery webserver Oct 08 08:43:40 <[florian]> weird Oct 08 08:47:42 mnBIOS appears to have its own flash partitioning scheme Oct 08 08:48:02 <[florian]> it is shared with the OS ? Oct 08 08:48:04 if memory serves, [mbm] noticed it on the anas350 Oct 08 08:48:27 no, the kernel just sees one big fat mtd0 Oct 08 08:48:52 uses initrd instead of rootfs Oct 08 08:51:40 each "partition" begins with a sixteen-byte header Oct 08 08:52:19 at 0x0 is a 32-bit value, possibly a crc of some sort Oct 08 08:52:33 at 0x4 is always 0xffffffff Oct 08 08:53:29 at 0x8 is the length of the data in the partition, not including the header nor the padding Oct 08 08:54:10 at 0xc is a 16-bit unknown field, possibly a header checksum of some sort Oct 08 08:54:37 at 0xe is magic 4c 69 Oct 08 08:56:28 there does not appear to be any rhyme or reason to where each partition begins Oct 08 08:56:45 some are not even on erase-block boundaries Oct 08 08:57:30 the partition layout used is: Oct 08 08:57:44 0x000000 kernel Oct 08 08:58:19 0x100000 initrd Oct 08 08:58:59 0x3a0000 config Oct 08 08:59:45 0x3b0000 webif pages Oct 08 09:01:23 0x3f0000 recovery webif (?) Oct 08 09:02:11 0x3f8000 serial num info (?) Oct 08 09:02:35 0x3fa000 bootloader Oct 08 09:10:31 if the initrd *must* begin at 0x100000, this device might not be supportable at all Oct 08 09:11:13 [florian], nbd: what do you think? Oct 08 09:38:48 eek! http://www.macsat.com/macsat/component/option,com_openwiki/Itemid,66/id,braindead_by_design_-_the_gui-fw_update_and_recoveryloader/ Oct 08 09:42:27 hehe, at least they implemented a checksum ;) Oct 08 09:51:49 <[florian]> sn9: we might want to replace the bootloader on this device to get into it properly ) Oct 08 09:52:40 without working jtag? Oct 08 10:36:06 lars * r12894 /packages/libs/net-snmp/Makefile: Oct 08 10:36:06 Install net-snmp-config into $(STAGING_DIR)/host/bin rather then Oct 08 10:36:06 $(STAGING_DIR)/usr/bin. Fixes build of packages which depend on libnet-snmp like Oct 08 10:36:06 collectd. Oct 08 12:08:34 nbd * r12895 /trunk/package/mac80211/patches/299-compat_minstrel.patch: mac80211/compat-wireless: don't patch a generated file Oct 08 12:09:00 nbd * r12896 /trunk/include/ (package.mk site/mipsel-linux2.4-uclibc): fix inotify detection for linux 2.4 (fixes glib2 and packages that depend on it on brcm-2.4) Oct 08 12:09:51 nbd * r12897 /branches/8.09/include/ (package.mk site/mipsel-linux2.4-uclibc): merge r12896 to 8.09 Oct 08 14:05:03 nbd: so, seriously - I just deleted 40 spam pages thanks to the bug in the wiki Oct 08 14:05:44 come on, guys - it's not that freaking hard. Do it or let someone do it for you Oct 08 14:06:16 every page I edit is irrevocably (i.e. cannot be re-created) deleted Oct 08 14:06:21 we're flooded with spam daily Oct 08 14:08:01 I'm going to start editing the landing pages soon Oct 08 14:12:05 glp [florian] [mbm] Kaloz: I'm going to be the squeaky wheel here. It's trivial to fix and needs fixing Oct 08 14:35:19 +1 Oct 08 15:05:18 lars * r12898 /packages/libs/libxml2/Makefile: Update libxml2 to 2.7.2. Oct 08 15:16:45 lars * r12899 /packages/multimedia/gstreamer/Makefile: gstreamer does not depend on check. Oct 08 15:17:21 lars * r12900 /packages/libs/libxml2/Makefile: Set md5sum for libxml2. Oct 08 15:20:15 lars * r12901 /packages/libs/check/Makefile: check does not depend on libxml2 and set md5sum for check. Oct 08 15:34:48 lars * r12902 /packages/libs/taglib/Makefile: Oct 08 15:34:48 Install taglib-config to $(STAGING_DIR)/host/bin rather then Oct 08 15:34:48 $(STAGING_DIR)/usr/bin. Oct 08 15:46:03 aoz: kaloz is responsible for the server Oct 08 15:47:52 nbd: can you gently persuade him? ;) Oct 08 15:48:34 i'll talk to him Oct 08 15:48:35 at the risk of being harsh, I wouldn't call it particularly responsible Oct 08 15:50:26 nbd: thanks Oct 08 15:50:29 lars * r12903 /packages/XOrg/lib/imlib2/Makefile: Update imlib2 to 1.4.1. Oct 08 15:51:19 aoz: without wanting to stir things up -- i have spoken to Kaloz about 3 weeks ago. he said he had some private things that took a lot of his time but, during our conversation, said he would most likely be able to set something up within a week. that would have been 2 weeks ago Oct 08 15:54:09 danage: indeed, I got the same story a week before, and research seems to show others got the same answer most of the past year Oct 08 15:55:50 i'm sure he means to take care of it Oct 08 15:56:32 some things just take time Oct 08 15:56:42 the new server will be there Oct 08 15:57:20 yes, things take time; most of us intend well. The fact remains that a year ago when I started working w/OWRT, it was precisely the same situation and story. Oct 08 15:58:55 hmmz Oct 08 15:58:58 hell, I've volunteered a server and my time to convert & maintain Oct 08 15:59:58 blogic: i think the current (and past) state of the issue hinders both development and user adoption of openwrt Oct 08 16:00:16 plus, fixing the issue would take care of a lot of recurring questions in the channel Oct 08 16:00:25 I could do it now - wrote a bot that guts a Moin wiki, and have snagged most of the content already, I'd just rather have an official blessing. Oct 08 16:01:00 aoz: i think what Kaloz is planning is a second wiki, allowing manual migration of contents to a new one Oct 08 16:03:26 lars * r12904 /packages/libs/libart/Makefile: Oct 08 16:03:26 Install libart2-config to $(STAGING_DIR)/host/bin rather then Oct 08 16:03:26 $(STAGING_DIR)/usr/bin. Oct 08 16:04:05 lars * r12905 /packages/libs/libaudiofile/Makefile: Oct 08 16:04:05 Install audiofile-config to $(STAGING_DIR)/host/bin rather then Oct 08 16:04:05 $(STAGING_DIR)/usr/bin. Oct 08 16:04:27 that's what I have been told as well; it's still usually pretty easy to port one wiki's data to another Oct 08 16:09:26 nbd: what do you think? Oct 08 16:09:34 blogic: same Oct 08 16:10:12 danage: what can i say Oct 08 16:13:21 my driving force: Oct 08 16:13:38 reduce support needs by being able to say, "it's in the wiki, RTFW" Oct 08 16:13:59 * aoz likes to document once, be an ass many times Oct 08 16:20:49 lars * r12906 /packages/ (libs/libusb/Makefile utils/sispmctl/Makefile): Oct 08 16:20:49 Install libusb-config to $(STAGING_DIR)/host/bin rather then Oct 08 16:20:49 $(STAGING_DIR)/usr/bin. Oct 08 16:20:49 sispmctl doesn't need help finding libusb-config anymore. Oct 08 16:31:03 lars * r12907 /packages/libs/ossp-js/Makefile: Oct 08 16:31:03 Install js-config to $(STAGING_DIR)/host/bin rather then Oct 08 16:31:03 $(STAGING_DIR)/usr/bin. Oct 08 16:32:58 nbd * r12908 /packages/net/arpwatch/Makefile: work around wrong permissions in the arpwatch tarball - fixes avr32 compile issues Oct 08 16:33:03 nbd * r12909 /packages/utils/rcs/patches/100-cross_compile.patch: fix spurious rcs build errors Oct 08 16:33:05 nbd * r12910 /packages/libs/file/Makefile: work around permission problems in the file package - fixes avr32 compile error Oct 08 16:33:10 nbd * r12911 /packages/libs/fftw3/Makefile: fftw3 is broken on avr32 Oct 08 16:33:14 nbd * r12912 /packages/utils/avrdude/Makefile: work around permission problems in the avrdude tarball - fixes avr32 compile error Oct 08 16:33:22 nbd * r12913 /packages/net/ipsec-tools/Makefile: work around permission problems in the ipsec-tools tarball - fixes avr32 compile error Oct 08 16:42:36 lars * r12914 /packages/libs/libol/Makefile: Let libol-config return paths relative to $(STAGING_DIR). Oct 08 16:55:23 lars * r12915 /packages/libs/libgpg-error/Makefile: Oct 08 16:55:23 Install gpg-error-config to $(STAGING_DIR)/host/bin rather then Oct 08 16:55:23 $(STAGING_DIR)/usr/bin. Oct 08 16:59:23 blogic: hi Oct 08 17:02:44 What's the process for adding a new port to OpenWrt? I've posted my patches to the mailing list but there's no feedback. Contacted Kaloz; he provided very useful feedback which I incorporated into a new patch set, but I've gotten no further response. Oct 08 17:03:28 mailing list Oct 08 17:03:42 keep posting and prodding; took me months to get my patches for sierra_wireless in Oct 08 17:07:03 MateIn4: if this is about the mi424wr, I was hoping to get a jtag dump of my device before I flashed it, but I should forget about that, the units are cheap enough on ebay that I can always pickup a second. Oct 08 17:15:11 Bartman007: yes, I'm referring to the mi424wr. Note that it's now possible to get a complete flash dump and install redboot without JTAG. Oct 08 17:26:10 I guess I'm a little confused about the mailing list. Should I keep sending the same patches once a month? Oct 08 17:27:46 depends on how insistent you want to be Oct 08 17:29:50 lars * r12916 /packages/libs/libdnet/Makefile: Oct 08 17:29:50 Install dnet-config to $(STAGING_DIR)/host/bin rather then Oct 08 17:29:50 $(STAGING_DIR)/usr/bin. Oct 08 17:31:26 well, I'm just trying to find out if there's a step I've missed. I've created a wiki page and sent out patches. Of course, this may not be a device people are interested in. I think it's pretty good for the price. Oct 08 17:32:28 no steps missing, IMHO - you're doing the right thing. Oct 08 17:32:50 thanks, and I'll leave it at that. Oct 08 17:33:54 MateIn4: one problem Oct 08 17:34:39 your page is under OpenWrtDocs/Hardware /Actiontec/MI424-WR Oct 08 17:34:44 ^^ notice the space Oct 08 17:35:33 I don't see the space: http://wiki.openwrt.org/OpenWrtDocs/Hardware/Actiontec/MI424-WR Oct 08 17:36:09 feh, I took an old copy/paste - MoinMoin-- again! Oct 08 17:36:52 Funny, I haven't had any problems with the wiki (other than deleting a page) Oct 08 17:37:58 nah, it's the DWIM magic again - malform your URL and it'll point you *somewhere* Oct 08 17:43:44 lars * r12917 /packages/libs/opencdk/Makefile: Oct 08 17:43:44 Install opencdk-config to $(STAGING_DIR)/host/bin rather then Oct 08 17:43:44 $(STAGING_DIR)/usr/bin. Oct 08 17:48:51 Weedy: ho Oct 08 17:49:03 aoz: patches that add DWIM functionality to openwrt are welcomed :-) Oct 08 17:50:17 :D Oct 08 17:50:27 blogic: mah firewall is borked again, what happened too pre/postrouting Oct 08 17:50:34 ? Oct 08 17:50:39 any detail ? Oct 08 17:50:46 yeah Oct 08 17:50:50 i have all of tomorrow set aside for the fw Oct 08 17:50:55 can you email the info ? Oct 08 17:50:56 theres no pre/postrouting Oct 08 17:51:01 blogic@openwrt.org Oct 08 17:51:05 ok Oct 08 17:51:10 will look into it Oct 08 17:52:00 http://openwrt.pastebin.ca/1222916 Oct 08 17:52:13 blogic: we need redirect support :) Oct 08 17:52:20 blogic: port -> port Oct 08 17:52:30 for prxy foo and such stuff Oct 08 17:52:54 *hehe* only if you have time to look Oct 08 17:58:26 juhosg * r12918 / (2 files in 2 dirs): [atheros] fix section mismatch warning Oct 08 18:02:20 lars * r12919 /packages/libs/dbus-glib/Makefile: Add libtool fixup to dbus-glib. Oct 08 18:02:39 lars * r12920 /packages/libs/libaudiofile/Makefile: Add libtool fixup to libaudiofile. Oct 08 18:02:56 lars * r12921 /packages/libs/libdnet/Makefile: Add libtool fixup to libdnet. Oct 08 18:03:11 lars * r12922 /packages/libs/libusb/Makefile: Add libtool fixup to libusb. Oct 08 18:03:30 lars * r12923 /packages/libs/taglib/Makefile: Add libtool fixup to taglib. Oct 08 18:03:48 lars * r12924 /packages/libs/opencdk/Makefile: Add libtool fixup to opencdk. Oct 08 18:10:04 * mib_xtin8i looks at topic and says: nice!!!!! didn't know there was a schedule :D Oct 08 18:10:15 well I do know there was Oct 08 18:10:50 I just didn't know everything else after the delays Oct 08 18:13:06 mib_xtin8i: neither did the devs Oct 08 18:15:58 s/everything/anything//g Oct 08 18:16:29 I want a refund of 3 times the €0 I paid for openwrt :D Oct 08 18:18:14 hum, btw, can i ask a question here or just in #openwrt? noone there was able to help Oct 08 18:18:29 useful help -- priceless Oct 08 18:19:07 hi sn9 :) Oct 08 18:19:13 hello Oct 08 18:19:25 time is priceless :) Oct 08 18:20:43 they say that time is money; unfortunately the inverse is not also true Oct 08 18:21:08 ok, found out the answer I was looking for :D make target_install was "renamed" to make target/install Oct 08 18:21:21 lol, I'd swear I'd tried that.... Oct 08 18:21:34 had to "make" again to see the output of all phases Oct 08 18:21:54 i don't think it ever had a _ Oct 08 18:21:58 it was - Oct 08 18:22:41 in 2006 it was, at least that's what it says here: http://downloads.openwrt.org/docs/buildroot-documentation.html :) Oct 08 18:23:17 may be a typo Oct 08 18:25:38 target_clean was a valid make target in whiterussian... Oct 08 18:28:09 WR was already a has-been in 2006 Oct 08 18:30:07 sn9: on the active development side, but 0.9 was tagged on 2007-02-05 Oct 08 18:30:59 active development is all we have left now Oct 08 18:31:44 latest SVN revisions (a few days ago) complain a patch is already is already applied when patching bridge-utils Oct 08 18:31:53 however hadn't the time to properly verify the issue Oct 08 18:31:58 so didn't fill a ticket Oct 08 18:32:11 I'm using r12006 anyway Oct 08 19:08:01 sn9: how is your shell? Oct 08 19:08:26 cause im about to launch my router across the room Oct 08 19:10:55 my shell is still running Oct 08 19:11:02 and yours? Oct 08 19:11:16 complaing about newlines Oct 08 19:11:26 bah cant spell Oct 08 19:11:44 anyway this works on the cmdline Oct 08 19:11:47 MODEM=$(for i in /sys/bus/usb/devices/*/uevent; do if cat "$i"|grep airprime 1>&-; then echo "$i"; fi ; done|head -n1) Oct 08 19:12:06 you are trying to use comgt? Oct 08 19:12:12 and dies in the script Oct 08 19:12:14 line 3: syntax error: newline unexpected (expecting ")") Oct 08 19:12:18 yar Oct 08 19:12:45 what editor are you using? Oct 08 19:12:50 vi Oct 08 19:13:05 weedy: use vi -b just in case Oct 08 19:13:34 in vi, you can arrow up and down to see which lines are separate and which are wrapped Oct 08 19:14:01 my console is wide enough Oct 08 19:15:14 why don't you use grep -q, anyway? Oct 08 19:15:49 and nuke the "cat" Oct 08 19:16:55 shorter version: Oct 08 19:17:15 nbd * r12925 /trunk/scripts/env: scripts/env: fix file syncing on env switch Oct 08 19:18:02 MODEM=$(for i in /sys/bus/usb/devices/*/uevent; do grep -l airprime "$i"; done|head -n 1) Oct 08 19:18:24 nbd * r12926 /branches/8.09/scripts/env: merge r12925 to 8.09 Oct 08 19:18:30 even shorter version: Oct 08 19:19:00 MODEM=$(grep -l airprime /sys/bus/usb/devices/*/uevent|head -n 1) Oct 08 19:20:52 its still whining Oct 08 19:21:09 what is the error now? Oct 08 19:21:15 i scp'ed it too my desktop Oct 08 19:21:18 cat -A Oct 08 19:21:25 there is no new line Oct 08 19:21:51 idea: Oct 08 19:21:58 MODEM="$(grep -l airprime /sys/bus/usb/devices/*/uevent|head -n 1)" Oct 08 19:22:35 .................... if thats it i need to be shot Oct 08 19:23:03 nope Oct 08 19:23:10 same error? Oct 08 19:23:17 well at least i did fail bash-101 Oct 08 19:23:59 same error? Oct 08 19:24:06 yeah Oct 08 19:24:31 are you sure that's the line it complains about? Oct 08 19:24:50 /etc/hotplug.d/usb/15-Sprint: line 3: syntax error: newline unexpected (expecting ")") Oct 08 19:25:23 root@OpenWrt:~# head -n3 /etc/hotplug.d/usb/15-Sprint Oct 08 19:25:23 #!/bin/sh -x Oct 08 19:25:23 case "$ACTION" in Oct 08 19:25:23 MODEM="$(grep -l airprime /sys/bus/usb/devices/*/uevent|head -n1)" Oct 08 19:25:43 shouldn't it be -n 1? Oct 08 19:26:22 no difference Oct 08 19:26:24 wait, this is in a case? Oct 08 19:26:35 that makes zero sense Oct 08 19:27:04 -ENOPARSE Oct 08 19:27:34 your newline is in line 2 Oct 08 19:27:51 woot Oct 08 19:27:55 i fail bash Oct 08 19:28:05 in ... Oct 08 19:28:08 in what? Oct 08 19:28:42 i moved it out of the case Oct 08 19:29:38 ? Oct 08 19:33:00 Weedy: what was that supposed to do? Oct 08 19:34:05 # grep -l airprime /sys/bus/usb/devices/*/uevent|head -n 1 Oct 08 19:34:05 /sys/bus/usb/devices/1-1:1.0/uevent Oct 08 19:34:23 my modem fires hotplug like 6 times Oct 08 19:35:16 i needed a way to figure out which is the first iteration Oct 08 19:35:38 define "first" Oct 08 19:37:09 im switching the wan over to the 3g modem Oct 08 19:37:33 my first version did it 3 or more times Oct 08 19:37:52 cause the modem makes an ass load of add/remove events Oct 08 19:38:06 so yeah i only need it running once Oct 08 19:38:31 is it usb, or expresscard/cardbus? Oct 08 19:38:49 usb Oct 08 19:38:58 anybody care to svn delete the empty package/{arptables,ebtables}/patches dirs? Oct 08 19:39:10 * Weedy points to usb in above paths Oct 08 19:39:52 Weedy: use locking of some sort Oct 08 19:40:23 Weedy: most of the pcmcia and implementations I've seen are the USB version w/ a usb-to-pcmcia chip, and expresscard has usb2.0 as part of the formfactor :-P Oct 08 19:43:06 on add, check for the existence of /tmp/sprint_is_added, and create it if absent Oct 08 19:43:16 do the reverse on remove Oct 08 19:43:23 [ test ] && { } continues on true right? Oct 08 19:44:33 on add: Oct 08 19:45:05 [ -e /tmp/sprint_is_added ] || exit Oct 08 19:45:14 oops Oct 08 19:45:18 [ -e /tmp/sprint_is_added ] && exit Oct 08 19:45:34 touch /tmp/sprint_is_added Oct 08 19:45:47 on remove: Oct 08 19:45:54 [ -e /tmp/sprint_is_added ] || exit Oct 08 19:46:01 rm /tmp/sprint_is_added Oct 08 19:46:24 actually, replace "touch" with ">" Oct 08 20:07:01 sn9: Have you got tips on dumping flash? I've got a USR8200, which I'm about to start work on. dd appears not to be on the default install, and I'm still building the JTAG. There is a 'flash_dump' command in the bootloader... Oct 08 20:10:34 is that a brcm63xx device? Oct 08 20:12:27 No, IXP425 Oct 08 20:12:52 so, it uses redboot? Oct 08 20:13:11 No, "OpenRG" Oct 08 20:13:26 but there are instructions for putting redboot on Oct 08 20:13:28 that's a bootloader? Oct 08 20:14:25 well, it boots, and has flash load and tftp stuff Oct 08 20:14:34 it is linux based Oct 08 20:15:25 lars * r12927 /packages/multimedia/gst-plugins-base/patches/ (. 010-dont-build-tests.patch): Disable building of tests. Oct 08 20:17:30 do you have linux shell access? Oct 08 20:18:24 Yes, but there are very (very) few commands installed Oct 08 20:18:36 is dmesg there? Oct 08 20:18:39 juhosg * r12928 / (2 files in 2 dirs): [atheros] fix 2.6.26.x boot failure on ar2313 (closes #4082) Oct 08 20:20:17 sn9: No, dmesg isn't there Oct 08 20:20:49 no serial bootlog, either? Oct 08 20:20:51 (But I have serial port access, and can see the boot log - there is apparently only 1 mtd device created) Oct 08 20:21:28 what are the addresses for the mtd device listed in the bootlog? Oct 08 20:21:31 Using predefined MTD partitions. Creating 1 MTD partitions on "IXP425 Flash Bank 0": 0x00000000-0x01000000 : "openrg" Oct 08 20:21:54 ok, and in /proc/mtd ? Oct 08 20:22:18 mtd0: 01000000 00020000 "openrg" Oct 08 20:22:44 16M flash, 128K blocks Oct 08 20:22:55 is there an httpd? Oct 08 20:23:07 flash size is right Oct 08 20:23:13 I'll just look... Oct 08 20:25:51 Not obviously an httpd, there's an 'l2tpd' (not running) and a running 'pluto' Oct 08 20:27:21 isn't openrg done by jungo? Oct 08 20:27:26 pluto is openswan Oct 08 20:27:31 Thanks Oct 08 20:27:39 well now, google answered that question :-) Oct 08 20:28:11 Bartman007: Yes, which is why I'm trying to replace it ASAP - the purist in me is going to keep a backup though (I'm pretty sure it'll never go back on) Oct 08 20:28:13 they did the stock firmware on the verizon/actiontec mi424wr also Oct 08 20:29:54 noz: check ps Oct 08 20:31:57 yes, did - there's very little running: openrg (their shell app), smbd/nmbd, pluto and pluto_adns, sh, and that's it Oct 08 20:32:34 what does their shell app do? Oct 08 20:35:02 configuration stuff, mostly setting single things, and saving conf Oct 08 20:35:34 routing, ifconfig, ping - that sort of stuff as well, but tailored with no extra Oct 08 20:36:20 so, it handles its webif? Oct 08 20:40:32 is file creation time stored with seconds? Oct 08 20:41:23 many filesystems do not store creation at all Oct 08 20:42:15 sn9: Don't know yet, can't get any web interface yet Oct 08 20:42:23 crw-r--r-- 1 root root 21, 0 Oct 8 20:37 sg0 Oct 08 20:42:32 i need to know seconds Oct 08 20:42:51 noz: the original firmware has no webif? Oct 08 20:43:17 Don't know yet - looks like it Oct 08 20:43:30 noz: I've been working on another ixp425 based router mi424-wr that is also using Jungo's OpenRG. I've written a small python program that can make a copy of the existing flash image and replace the firmware. I've also tested it on the WRV54G. They key is to have telnet access to the device. Oct 08 20:45:22 MateIn4: That sounds promising - I'll try to get telnet working. Oct 08 20:46:24 noz: see if there's a "Local administration" option in the menus. Take a look at the mi424-wr wiki page to download the python script. Oct 08 20:48:11 doh! Let me actually plug a network cable in before saying there's no webif! Oct 08 20:52:34 noz: one other thing, the jungo software has it's own shell commands to do basic things (tried on mi424-wr and wrv54g). see if they enabled the 'shell' command to get to a busybox shell. Oct 08 20:53:27 Yes, I can get to a busybox shell, but it's very cut down. Oct 08 20:54:21 that's also my experience. there's very few commands and very little that can be done there. poking around /proc was useful. Oct 08 20:55:36 btw, the python script uses jungo's openrg commands to make a copy of the image. it is kind of slow though. Oct 08 20:56:29 https://dev.openwrt.org/browser/trunk/package/comgt/files/3g.sh Oct 08 20:56:45 lines 44-47 that doesn't work Oct 08 20:56:55 MateIn4: Which commands? image_dump ? Oct 08 20:57:57 using flash_dump Oct 08 20:58:34 OK. Oct 08 21:34:17 MateIn4: Do you think the MI424-WR redboot image will work on the USR8200? It is an IXP425, but I don't know what customisations were done for the MI424-WR... Oct 08 21:34:45 (i.e. How specific is a redboot build, beyond the processor type?) Oct 08 21:35:52 noz: the biggest issue was finding out where the phys are located. this may not match with your board and that could be a problem. Does your unit also have a ks8995 4 port switch? Oct 08 21:36:49 No - it's a Marvell 88E6060. The stock firmware lists the PHYs as it boots Oct 08 21:37:49 take a look at the phy numbers and compare those with the values I've put up in mi424-wr wiki. also check which npe is connected to the switch. Oct 08 21:38:39 but I think you'll need to make a custom version of redboot just to handle these things. Oct 08 21:39:55 noz: you may want to put up the output of the boot process on the wiki. Oct 08 21:39:57 comes up as ixp1 in ifconfig - the PHYs are 0 at 9, then 1@16 .. 5@20, 6@24..11@29 Why there are 12 PHYs, I don't know Oct 08 21:41:10 Already done (indirectly) - it's at http://home.comcast.net/~johnrw/image/USR8200_notes.txt Oct 08 21:41:39 your setup is different. it's not too hard to change. also, the openwrt port is not difficult either. the biggest issue is gathering all the details about the HW Oct 08 21:42:51 Sure - I'm part way through that, and have (almost) all the chips id'ed. Some even have data sheets! Oct 08 21:44:20 I'm looking at your notes. try a 'help all' to get a list of all possible commands. Oct 08 21:44:23 I think I might wait until all the bits arrive for my JTAG, if I have to build a custom RedBoot. Oct 08 21:45:08 good idea. I used openocd to debug. It was a nightmare to make it work. Oct 08 21:45:16 :) Oct 08 21:45:46 I have a USBprog, which will do openocd (among many other things) Oct 08 21:46:37 Long term, we're aiming to solder a miniPCI connector on, and add wireless to the board. It's a while away yet, though... Oct 08 21:48:01 looks very similar to the mi424-wr Oct 08 21:48:21 of course, I've got wireless and coax and you have usb & firewire Oct 08 21:48:50 But you can buy wireless cards ;) Oct 08 21:49:37 I think you should be able to use the python script to do a flash dump without any worries. Oct 08 21:50:37 Yep - many thanks for the pointer. Oct 08 21:50:49 good luck Oct 08 21:51:53 lars * r12929 /trunk/tools/automake/ (Makefile files/ files/aclocal): Oct 08 21:51:53 Add wrapper around aclocal which also search in the targets staging dir aclocal Oct 08 21:51:53 folder for m4 files. Oct 08 21:52:05 BTW, how much is the usr8200? are they on sale? Oct 08 21:52:51 They retail at an alarming $350, I believe, but these cost much less on eBay. Oct 08 21:53:27 (so I believe - they are a donation to OpenWrt) Oct 08 21:54:26 noz: The gateway 7001 a few of us have had an RRP of $400 Oct 08 21:55:48 That's too much for me to hack on. the mi424-wr were dumped for around $25 Oct 08 21:56:29 MateIn4: we grabbed them off ebay for $50 :-) Oct 08 21:56:38 That's almost cheaper than the dev tools you need. Oct 08 21:56:48 <[Fate]> Add wrapper around aclocal which also search in the targets staging dir aclocal Oct 08 21:56:49 <[Fate]> doh Oct 08 21:56:56 <[Fate]> I just read "around alcohol" :( Oct 08 21:56:59 yeah, that's what got me into openwrt! Oct 08 21:58:01 noz: I dunno, I got my Flyswatter for free from the TinCan Tools guy :-) Oct 08 21:58:56 really nice guy, we were in a booth next to him at LinuxWorld. Oct 08 22:00:22 I've been buying mine, but I'm still down at EUR32 for a USBprog, EUR13 for the level convertor and ~£15 for 2 USB-serial cables (one 3.3v, one full) Oct 08 22:01:23 I was able to do the entire port without resorting to the serial port, but did rely on JTAG quite a bit. Oct 08 22:02:09 The custom PSU I'm building to replace a long line of wall-warts is probably going to be the most expensive bit, but only because I'm cheating and buying complete regulators, rather than designing it Oct 08 22:04:55 * aoz wants a nice 12v desk-mount supply with variable rails Oct 08 22:06:50 There are some nice Maxim or LinearTech chips, but it's a fair amount of design work, and then you have to make the board... Oct 08 22:12:22 * Bartman007 has a small ATX powersupply turned bench board Oct 08 22:13:36 works well enough when I need it Oct 08 22:13:56 Trouble with ATX powersupplies is that they typically won't start unless you load them Oct 08 22:15:27 you mean real load, or PS-ON? Oct 08 22:15:35 real load Oct 08 22:15:36 something more like this: http://www.powersupplydepot.com/productview.asp?product=14602+PS Oct 08 22:16:30 Indeed. But I can't justify that, nice though it would be Oct 08 22:16:53 http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply Oct 08 22:17:08 ^^ with a little more electronics & work, could be just as good Oct 08 22:18:06 noz: I've dealt with enough that work just fine Oct 08 22:18:35 they even solve the switching problem (load resistor) Oct 08 22:19:10 aoz: Good link - the power resistor is the load that you need to ensure it's on. Oct 08 22:20:33 I'm aiming for low power and high efficiency. The switched-mode regulators I'm using quote 95-96% efficiency for 12V (running off an 18V laptop PSU) Oct 08 22:20:59 My primary aim is reduction of clutter and small size Oct 08 22:21:25 Secondary is high efficiency Oct 08 22:21:40 noz: have a link to the actual units you're purchasing? Oct 08 22:23:45 163-394 part number at uk.rs-online.com Oct 08 22:24:49 hmm, seems the anas350 and the mgb111 are the same board Oct 08 22:25:12 noz: ah, that's roughly the price I've seen stateside for equivalent parts. Oct 08 22:25:24 i just updated the wiki to reflect that Oct 08 22:25:30 Idea is one single mains block (I'm not into designing a PSU to go all the way down from 240V - it's too much hassle), and then high-efficiency SM regulators to go from the 18V or so to 12V, 5V, 9V, etc. Oct 08 22:27:05 noz: yeah. I wish >2A switching regulators were cheaper, that's why I decided against a similar setup a couple years ago Oct 08 22:27:48 fortunately I only need 12V @ 1A (so far...!) Oct 08 22:28:02 this device has a 12V 4A (!) brick Oct 08 22:28:07 ouch Oct 08 22:28:28 well, it's a nas Oct 08 22:32:09 MateIn4: Think I'm only going to get 1/2 the flash (8MB) without modifying the script - however, it seems to be working - thanks Oct 08 22:33:43 it also has a 3V Li battery Oct 08 22:34:18 noz: yes, the flash size is hardcoded to 8MB Oct 08 22:37:20 lars * r12930 /packages/libs/libxslt/Makefile: Use xml2-config. Oct 08 22:39:41 MateIn4: not any more ;) I'll send you the updated script if you like (defaults to 8MB) Oct 08 22:40:02 noz: perhaps the script should cat /proc/mtd to find out the flash size Oct 08 22:40:09 Could do Oct 08 22:41:58 I would like to see this script added to openwrt somewhere to help other developers Oct 08 22:42:29 It's on the wiki (now linked in two places) Oct 08 22:42:49 ... just needs more links Oct 08 23:12:17 noz: I've updated the script to query /proc/mtd for the flash size Oct 08 23:38:48 lars * r12931 /packages/libs/libxslt/Makefile: Cleanup libxslt makefile. Oct 09 02:08:20 thepeople: as you know that you are knowledgeable you might know how to handle this... :) Oct 09 02:08:44 it's a openwrt/webif question about setting up bridging Oct 09 02:09:35 i just build the latest revision from trunk but the default network settings doesn't let me reach the internet from my lan Oct 09 02:10:24 i can connect to my rooter with ssh and wrt and when i'm logged in on the router i can reach the internet from it Oct 09 02:10:50 i can't even ping or anything from lan to internet Oct 09 02:11:24 my guess is that i can change the settings for the bridge somehow Oct 09 02:11:36 is this a know problem or is it just me? Oct 09 02:13:25 do you have forwarding set up in your /etc/config/firewall? Oct 09 02:14:10 i haven't changed anything from the default settings Oct 09 02:14:27 it's a brand new flash and they usually work out of the box for me Oct 09 02:15:42 and i thought openwrt handled bridging between wlan and lan in some other way than port forwarding Oct 09 02:16:47 right now i have flashed back to an older revision that works but i want a newer one :) Oct 09 02:18:11 so i was thinking that there might have been changes in how the default settings for bridging and the network is handled in recent revisions **** ENDING LOGGING AT Thu Oct 09 02:59:57 2008