**** BEGIN LOGGING AT Wed Oct 24 02:59:57 2007 Oct 24 04:35:06 nbd * r9422 /trunk/package/madwifi/ (29 files in 2 dirs): upgrade madwifi to r2756 Oct 24 08:19:20 juhosg * r9423 /trunk/target/linux/adm5120/files/ (8 files in 3 dirs): [adm5120] minor cleanups Oct 24 13:38:51 hi, all! Oct 24 13:39:28 [florian], hi! a miracle happened - I've found 7zip binary source in one of tarballs :) Oct 24 13:39:54 [florian], so, one less binary-only executable in archive. Oct 24 13:40:08 <[florian]> slapin: hey, this is good news Oct 24 13:42:18 florian * r9424 /trunk/package/base-files/files/etc/init.d/6tunnel: Mark 6tunnel as executable, avoids ": /etc/init.d/rcS: line 8: /etc/rc.d/S466tunnel: Permission denied" Oct 24 13:43:16 florian * r9425 /trunk/target/linux/au1000/patches/016-watchdog_pdev_gpio.patch: Oct 24 13:43:16 The watchdog driver is now a platform device driver and reads its gpio line Oct 24 13:43:16 while registering Oct 24 15:18:25 [florian], I've just pushed all changes to that git. Oct 24 15:21:20 matteo * r9426 /trunk/package/ar7-atm/Makefile: ar7-atm: updated the firmware image (#2532) Oct 24 15:27:26 ejka, hi, could you please check email? Oct 24 15:29:52 matteo: Thanks for closing #2532. Oct 24 15:30:19 farnz: i wish i did before, but i noticed just now Oct 24 15:39:12 Closing of that ticket resolves the DSP version issue; but still does not fix the avsar_modem_stats file containing nonsense Oct 24 15:39:49 I tried an old image, r7119, and still see the avsar_modem_stats nonsense Oct 24 15:41:29 nabcore: Then it's the TI DSP firmware or the DSLAM. Oct 24 15:42:18 nabcore: If it's *just* the upstream SNR that goes completely haywire, it could well be Pipex's DSLAM. Oct 24 15:43:35 nabcore: And even if it's the DSP firmware (the ar0700xx.bin file), it's not going to be easily fixed. Oct 24 15:44:15 farnz; in the ticket, you say that using the updated firmware solves the avsar_modem_stats nonsense, have you tested these values at different points in time? Oct 24 15:45:09 nabcore: Yes, and I've just rechecked (uptime 26 hours). Oct 24 15:46:08 ok; fair enough Oct 24 15:46:14 nabcore: What's the ATUC Vendor Code in your stats? Oct 24 15:46:49 ATUC Vendor Code: 50000000 Oct 24 15:48:47 brb; 2 mins Oct 24 15:53:02 nabcore: You're on a different make of DSLAM to me; if it's *just* upstream SNR, I'd suspect Pipex's DSLAM (and I'd like to know what manufacturer firmware gives you). Oct 24 15:54:01 yes; I was thinking of testing that: http://kbserver.netgear.com/release_notes/D103044.asp Oct 24 15:56:00 actually farnz; I'm going to test that now Oct 24 15:58:21 nbd * r9427 /packages/ (11 files in 11 dirs): fix up downloading in python packages Oct 24 16:05:04 farnz; with netgear latest firmware: http://pastebin.ca/748087 Oct 24 16:06:39 do anybody knows of some mips disassembler for .o files or at least objdump command line to do so? I need to check wheather .c file I found produces the same functionality as bundled binary-only .o file... Oct 24 16:07:51 slapin: hi Oct 24 16:08:04 nabcore: Much older firmware though. Oct 24 16:08:32 yeah.. Oct 24 16:08:35 slapin: objdump -d not good enough? Oct 24 16:08:57 aett, hi! Oct 24 16:09:48 slapin: i bricked my di524-up :/ Oct 24 16:10:37 do you have access to taht rome loader? Oct 24 16:10:46 nabcore: If it's stable over time on the old driver, I'd suspect the new DSP firmware. Oct 24 16:11:01 aett, how did you manage to do so? Oct 24 16:11:31 aett, I just have stuff everybody have. Oct 24 16:11:39 aett, do you see something on serial? Oct 24 16:12:14 nope Oct 24 16:12:40 aett, all I know is on that page. so I think for your case you need jtag to recover. Oct 24 16:12:59 I picked the wrong flashing option in rome Oct 24 16:13:00 farnz; you could be right Oct 24 16:13:14 see http://forum.openwrt.org/viewtopic.php?pid=57997#p57997 Oct 24 16:13:16 aett, what option? bootloader one? Oct 24 16:13:34 nabcore: In which case, keep an eye out for an AR7 router getting a new release of GPL source. Oct 24 16:13:36 yep Oct 24 16:14:00 aett, I could dump loader for you Oct 24 16:14:20 slapin: i got little frustrated with it... didnt want to boot anything Oct 24 16:14:49 flashing doesnt work via jtag afaik Oct 24 16:14:52 aett, do you have jtag soldered? Oct 24 16:15:05 aett, why? Oct 24 16:15:47 i just punched resistor legs through the holes Oct 24 16:15:47 aett, at least you should be able to put bootloader to RAM Oct 24 16:16:44 i dont think theres a boot vector Oct 24 16:17:01 aett, mail me your email, so I could dump b/l for you at home, and mail it. Oct 24 16:17:15 aett, jtag should control CPU at least. Oct 24 16:17:35 aett, but you will need to configure RAM first. Oct 24 16:17:39 i dont know really Oct 24 16:18:26 its supports 32 megs of ram btw Oct 24 16:18:34 aett, I'd try jtag approach properly, because similar machines are flashed through it as I see. Oct 24 16:18:44 aett, by what means? Oct 24 16:19:06 it has A12 line Oct 24 16:19:15 aett, I can't buy RAM chips here anyway Oct 24 16:20:28 Well, I'm very annoyed by the whole jtag thing Oct 24 16:20:55 it only supports ejtag 2.0 Oct 24 16:24:18 aett, I hope you will share your experience Oct 24 16:25:56 i do have a copy of rome loader but i think its corrupted Oct 24 16:26:31 my email is aet@rasterburn.or Oct 24 16:34:18 http://www.rasterburn.org/~aet/di524-up/bootloader Oct 24 16:39:04 farnz; using the latest offical netgear firmware (ar0700xx.bin: 6.01.00.12) with the latest trunk also causes the issue Oct 24 16:40:00 http://pastebin.ca/748137 Oct 24 16:40:03 nabcore: You'd need to change the DSL driver, too (sangam-atm). Oct 24 16:40:47 nabcore: TI have changed the format of the stats message between the firmware and the driver, so pre 7.02 firmware will have issues with a 7.02 driver. Oct 24 16:40:58 ah ok Oct 24 16:41:18 as an aside, it seems to work in the sense that I can online Oct 24 16:42:34 nabcore: Quick background for you; there's two processors in the AR7. The MIPS core, which runs OpenWRT, and the DSP core, which runs the ar0700xx.bin file. The two can communicate, but if TI change the comms format, things go wrong. Oct 24 16:42:46 Then TI haven't changed the data path comms format. Oct 24 16:43:33 nbd * r9428 /trunk/package/busybox/patches/470-insmod_search.patch: fix absolute path to .ko files in insmod Oct 24 16:44:30 farnz; I assume they communicate across the vlynq bus? Oct 24 16:45:37 or are they "closer" ? Oct 24 16:45:48 nabcore: IIRC, they're in the same IC. Oct 24 16:47:13 ok Oct 24 16:47:47 Has anyone attempted writing their own firmware code for the DSP core? Oct 24 16:48:08 (I'm assuming this would be quite hard to do) Oct 24 16:48:33 Not that I'm aware of; bear in mind that the "documentation" you have is the existing firmware file, and the knowledge that TI make DSP chips. Oct 24 16:48:56 :) Oct 24 16:50:37 http://mcmcc.bat.ru/dlinkt/acorp-russia/ar7_dsp72_feature_control.pdf Oct 24 16:54:25 AndyI: Thanks for that; it explains some of the weirder constants in the source. Oct 24 16:56:42 AndyI: Interesting that they've got a 7.4 DSP firmware, but not source. Oct 24 16:57:06 nabcore: And AndyI's document gives one possible explanation for your weird SNR; it's the minimum SNR on any bin in use for bit carrying. Oct 24 16:59:34 http://mcmcc.bat.ru/dlinkt/acorp-russia/ar7_dsp62_feature_control.pdf Oct 24 17:00:03 It's a shame that mcmcc.bat.ru don't have more of these TI docs :( Oct 24 17:00:40 AndyI: That's an older version of the same documentation. Oct 24 17:00:43 There is a scheme for 7300. It is necessary for 7100 Oct 24 17:01:35 AndyI: Both nabcore and I have 7300 chips. Oct 24 17:01:55 AndyI: I note that he's got what claims to be 7.04 firmware, but without new driver source. Oct 24 17:02:14 Interest jtag pin for tnetd7100 Oct 24 17:08:25 AndyI: Not sure what you mean? Oct 24 17:10:24 Share the scheme for TNETD7100 Oct 24 17:11:24 AndyI: I have no knowledge of the pinouts of any TNETD chip :( Oct 24 17:11:32 Weedy: ping Oct 24 17:12:31 Where to put? Oct 24 17:13:18 somebody knows something about a TNETV chip? Oct 24 17:14:11 nbd * r9429 /trunk/target/linux/x86/base-files/etc/preinit.arch: silence a bogus preinit message Oct 24 17:14:38 AndyI: http://wiki.openwrt.org/AR7Port would be a helpful place Oct 24 17:17:06 florian * r9430 /trunk/target/linux/adm5120/router_le/base-files/sbin/wget2nand: The build system actually changed the routerboard kernel (#2584) Oct 24 17:17:58 Has put look Oct 24 17:18:00 <[florian]> crazy_imp: tnetv are wireless chips from texas Oct 24 17:19:06 [florian]: interesting, the ti hotline told me it's for voip (and both, wireless and voip are not on the box i have...) Oct 24 17:20:10 AndyI: Those are really useful. Thanks. Oct 24 17:21:32 help my with firmware for http://wiki.openwrt.org/OpenWrtDocs/Hardware/ADLS2110HR and http://wiki.openwrt.org/OpenWrtDocs/Hardware/ADLS2110HR_7 Oct 24 17:23:11 AndyI: It looks like OpenWRT SVN would work, if you can fit the image in 2MB. Oct 24 17:25:50 nbd * r9431 /packages/net/olsrd/ (Makefile files/olsr.config files/olsrd.init): fix olsrd cflags handling, add new init script and config by Alina Friedrichsen Oct 24 17:26:07 Unfortunately I cannot collect. Oct 24 17:30:27 Probably to collect an image for 2mb and 7100 for check? Oct 24 17:42:09 nbd * r9432 /trunk/package/dnsmasq/files/dnsmasq.conf: disable filterwin2k in dnsmasq (see #2566) Oct 24 17:44:23 farnz: i'm about to apply your patch from #2525 Oct 24 17:44:39 farnz: has the +ipv6 option been tested without active ipv6 support? Oct 24 17:45:51 Is 7.10 branched yet (or is it cancelled ?) Oct 24 17:46:07 i don't think there will be a 7.10 Oct 24 17:46:10 7.11 is more likely Oct 24 17:56:22 nbd * r9433 /trunk/package/madwifi/patches/128-revert_ioaccess.patch: Oct 24 17:56:22 revert changes to io access functions done in madwifi trunk Oct 24 17:56:22 (they seem wrong to me, and they break xscale) Oct 24 17:56:22 fixes #2591 Oct 24 17:57:10 h3sp4wn: trunk still needs a while to stabilize Oct 24 17:58:29 nbd: any chance of a uci firewall config in 7.11? Oct 24 17:59:21 if i get it done in time Oct 24 18:01:36 nbd: yeah the madwifi on the last trunk build I did was not working (haven't investigated it as of yet - don't really need it here right now) Oct 24 18:01:52 h3sp4wn: what platform? Oct 24 18:02:04 nbd: wgt634u Oct 24 18:02:16 hm. maybe it's fixed with 9433 as well Oct 24 18:02:22 (xscale is 200 miles away - should get another one) Oct 24 18:06:26 nbd * r9434 /trunk/tools/firmware-utils/src/motorola-bin.c: add updated motorola-bin.c file from #2483 (by D. Hugh Redelmeier) Oct 24 18:44:08 florian * r9435 /trunk/package/base-files/files/etc/sysctl.conf: Enable IPv6 forwarding by default (#2527) Oct 24 19:23:49 florian * r9436 /trunk/ (2 files in 2 dirs): Fix the r8169 kernel module (#2554) Oct 24 19:49:40 florian * r9437 /trunk/package/wpa_supplicant/files/config.i686: Provide an i686 wpa_supplicant configuration file Oct 24 19:50:07 florian * r9438 /trunk/package/madwifi/Makefile: i686 should use the i386 HAL Oct 24 20:03:15 florian * r9439 /trunk/rules.mk: Treat i686 as an i386 if present, fixes build errors with wireless-tools on i686 (x86 mediacenter) Oct 24 22:21:27 florian * r9441 /packages/utils/digitemp/ (. Makefile): Add digitemp (#2546) Oct 24 22:59:28 nbd, HIGH LANTENCY PONG Oct 25 00:16:24 [florian], ping Oct 25 02:58:37 thepeople * r9442 /trunk/package/dnsmasq/Makefile: remove unneeded config file section fixes #2594 **** ENDING LOGGING AT Thu Oct 25 02:59:57 2007