**** BEGIN LOGGING AT Fri Jul 30 02:59:58 2010 Jul 30 06:12:47 _trine : did you test my rtl8366s patch for reactivating the switch already? Jul 30 06:19:48 <_trine> Memphiz, morning Jul 30 06:21:54 yeah its a good one i think ;) Jul 30 06:22:07 like all friday mornings *hehe* Jul 30 06:22:59 <_trine> yes I tested your code Jul 30 06:23:13 <_trine> however it appears it is not complete Jul 30 06:23:24 whats missing? Jul 30 06:23:26 <_trine> for my switch on the wndr3700 Jul 30 06:23:58 <_trine> it would be better to confer with nbd on that matter Jul 30 06:24:05 <_trine> as I am not a coder Jul 30 06:25:23 ok i'll do that ... Jul 30 06:25:34 <_trine> thanks Jul 30 06:26:21 thanks for testing anyway Jul 30 06:26:43 <_trine> thats ok I tested it the following day Jul 30 06:26:48 <_trine> it compiled ok Jul 30 06:27:36 <_trine> but it didnt seem to do anything and you could not issue the swconfig command to enable the new param Jul 30 06:28:05 ohh Jul 30 06:28:19 did you make a "make target/clean-linux" after applying the patch? Jul 30 06:28:27 <_trine> no Jul 30 06:28:35 otherwise the new code didn't get into build_dir ... Jul 30 06:28:49 mhh sorry ... maybe i've had mentioned that Jul 30 06:29:16 i thinkg that caused that swconfig didn't want the new attribute Jul 30 06:29:46 swconfig dev rtl8366s help should print a description of a new attribute called "enable_learning". if this isn't there, my code didn't get compiled in Jul 30 06:30:02 <_trine> when nbd looked at it provisionally he felt there maybe some of the code missing Jul 30 06:31:31 https://forum.openwrt.org/viewtopic.php?id=25873 <- this guy also did the patch ... at least he didn't say anything about failing with the swconfig command Jul 30 06:31:42 <_trine> however I'm not a coder so I don't know Jul 30 06:31:56 i'll talk with nbd when he gets up :) Jul 30 06:32:10 <_trine> I will read this Jul 30 06:33:53 <_trine> swconfig dev rtl8366s enable_learning 1 Jul 30 06:34:11 <_trine> apparently there is a 'set' command missing here Jul 30 06:34:20 ahh Jul 30 06:34:25 damnit ... Jul 30 06:34:57 your right ... i failed on that one ;) Jul 30 06:35:57 <_trine> I am just at the moment messing with a dockstar and am installing openwrt onto a drive to run it Jul 30 06:36:30 <_trine> when I have finished this and nbd is around later I will give it another try Jul 30 06:36:38 great thx Jul 30 06:37:10 <_trine> these little dockstar's look very interesting Jul 30 06:37:48 <_trine> 128 flash 256 ram 1.2 GHz processor 1 gig lan ports and 4 usb ports all for less than 25 euro Jul 30 06:37:52 yes i also read about them Jul 30 06:38:12 how many 1gig lan ports do they have? Jul 30 06:38:19 <_trine> only 1 Jul 30 06:38:40 <_trine> so you would need to plug it into the back of a router Jul 30 06:38:50 <_trine> but it would be a workhorse Jul 30 06:39:14 thank god ... otherwise i would regret now that i've bought 2 1043nd ;) Jul 30 06:39:15 <_trine> very nice for freeswitch etc Jul 30 06:39:29 yeah the power is great Jul 30 06:39:48 and by usb one could plug an ath9k usbstick or something like that *G* Jul 30 06:39:54 mhh Jul 30 06:40:01 *ebaying for dockstar* Jul 30 06:40:55 <_trine> http://www.zentralverkauf.de/a/33461945/src/goo/Computer/Geh%C3%A4use/Seagate-FreeAgent-DockStar-LAN-+-USB2.0.html Jul 30 06:40:56 _trine do you know about the current draw Jul 30 06:41:09 <_trine> on this unit? Jul 30 06:41:14 yes Jul 30 06:41:38 <_trine> its less than 8 watts including a free agent hard drive Jul 30 06:41:48 <_trine> and a lot less without a drive Jul 30 06:42:15 <_trine> if you just ran it on a flash drive Jul 30 06:42:43 mhh ... feeling a prickle in my fingers *G* Jul 30 06:43:10 i think i read the dockstar thread first *G* Jul 30 06:43:19 <_trine> make sure you do not plug it into a computer that is online without first disabling 'phone home' Jul 30 06:43:45 <_trine> http://ahsoftware.de/dockstar/#motivation Jul 30 06:43:56 <_trine> http://www.dslreports.com/forum/r24479163-Re-Other-Own-a-FreeSwitch-on-a-Seagate-DockStar-for-2499 Jul 30 06:44:08 <_trine> http://www.dslreports.com/forum/r24479163-Other-Own-a-FreeSwitch-on-a-Seagate-DockStar-for-2499 Jul 30 06:44:21 <_trine> http://www.cs.cmu.edu/~ecc/dockstar-howto.html Jul 30 06:44:55 <_trine> the route I chose to go down was not to flash the nand directly on the dockstar with openwrt Jul 30 06:45:53 <_trine> Instead I installed a new uboot in an unused partition which then allows the dockstar to boot off a pen drive if it sees a bootable OS Jul 30 06:46:31 <_trine> if it does not see another OS on the pendrive it defaults back to its own OS Jul 30 06:46:53 <_trine> I think that is a much better approach Jul 30 06:48:38 uboot can boot off an usb stick? Jul 30 06:48:45 <_trine> yes Jul 30 06:48:53 impressive Jul 30 06:48:54 <_trine> but not the original one Jul 30 06:49:08 and how do you jump into your patched uboot? Jul 30 06:49:55 <_trine> all the instructions are in those links I gave you Jul 30 06:50:32 <_trine> here Jul 30 06:50:36 <_trine> http://www.plugapps.com/index.php5?title=PlugApps:Pogoplug_Setboot Jul 30 06:53:55 not bad ... Jul 30 06:54:03 how far did you come so far? Jul 30 06:54:03 craigc * r22415 /packages/net/multiwan/ (Makefile files/usr/bin/multiwan): [packages] multiwan: Changed route acquistion to grab everything except default routes. Jul 30 06:54:35 <_trine> I have it booting off a pendrive into pogo linux Jul 30 06:56:00 well then i live you alone for now ... go go go ;) Jul 30 06:57:14 build #78 of ep93xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/78 Jul 30 07:07:51 mhh on this hardware it would be doable to run an ubuntu rootfs *just thinking* Jul 30 07:09:50 <_trine> Memphiz, [root@plugbox ~]# date Jul 30 07:09:50 <_trine> Fri Jul 30 08:08:20 PDT 2010 Jul 30 07:09:50 <_trine> [root@plugbox ~]# Jul 30 07:13:20 so you did it? Jul 30 07:13:56 but it looks like the wrong timezone ;op Jul 30 07:15:41 _trine does zentralverkauf.de ship with dhl so that i could use a packetstation for delivery? ^^ Jul 30 07:16:27 i'll by one ... i'll find an application for this nice little bastard when its here ;) Jul 30 07:25:21 <_trine> Memphiz, I live in the UK so I dont know that Jul 30 07:26:00 <_trine> Memphiz, there are quite a few different places that sell those units Jul 30 07:30:26 i've ordered one already ... ohh i thought you are from germany too Jul 30 07:32:00 <_trine> you're lucky those units are twice as expensive here in the UK Jul 30 07:33:23 yeah i'm very amazed about the price :) Jul 30 07:34:01 did the "date" command above indicate that you've got openwrt running from usb stick? Jul 30 07:34:17 <_trine> no Jul 30 07:34:30 <_trine> it a linux called pogo linux Jul 30 07:34:55 ahh ok ... that one from the howto ... i see Jul 30 07:35:01 <_trine> yes Jul 30 07:35:29 what will you use your unit for, when its working? Jul 30 07:35:58 <_trine> well I am intending running freeswitch in it etc Jul 30 07:36:32 <_trine> but with a 1.2 GHz processor and 256 ram it could do much more than that Jul 30 07:36:48 256mb ram? not 128mb? Jul 30 07:36:58 <_trine> you could also install gentoo or ubuntu if you wanted to Jul 30 07:37:13 <_trine> 128 flash 256 ram Jul 30 07:37:29 i think you have to build a openwrt kernel with ramdisk and usb drivers in it ... otherwise the kernel can't find the rootfs on the usb stick ... Jul 30 07:37:56 yeah ... i think about hanging an usb to hdmi adapter at this box and connect it to my flat tv :) ... as little surfstation or something like that Jul 30 07:38:35 <_trine> I already use the usb drivers in my wndr3700 to run everything off a usb pendrive Jul 30 07:38:54 including the kernel? Jul 30 07:39:17 <_trine> I use extroot Jul 30 07:39:21 ahh ok Jul 30 07:40:37 so the usb drivers are on flash ... but if you want to boot completly from usb on your dockstar ... you need to have the usbdrivers on ramdisk i think ... Jul 30 07:54:49 hello Jul 30 08:53:13 nbd ... when you booted up today you can tell me whats missing on my swconfig patch for rtl8366x switches ;) Jul 30 09:05:38 Memphiz: while I'm not nbd, I remember he mentioned that the learning should be enabled by default, and you can manually disable it Jul 30 09:05:55 so the logic should be reversed Jul 30 09:07:49 KanjiMonster yes ... that would be the final patch ... this was just for testing purposes ... if this is everything he claimed, i'll put a "real" patch together on the weekend and send it in for comittment :) Jul 30 09:08:41 i did it that way to be sure not to brick my device (serial doesn't working so i had to be carefull here) Jul 30 09:09:36 oh, sweet! Jul 30 09:09:54 looks like circa r22409 is bricking wgt634u's Jul 30 09:10:01 oh Jul 30 09:10:45 looks like it's overwriting nvram or something again Jul 30 09:10:54 Memphiz: oh, working network stuff without serial is fun Jul 30 09:11:03 russell_: would be nice to fix that before 10.03.1 Jul 30 09:11:14 indeed Jul 30 09:11:51 KanjiMonster but it is doable ;) Jul 30 09:12:33 Memphiz: as long as you still can access the failsafe mode/flash from the bootloader ;) Jul 30 09:13:07 I don't know how often I locked myself out trying to get the ar8316 vlan support to work Jul 30 09:13:35 russell_: https://dev.openwrt.org/search?q=nvram&noquickjump=1&changeset=on Jul 30 09:13:58 my serial is working only in receiving data ... i have to fix this somehow ... so that i can write to the device and unbrick in case i have to ... its a strange problem - but i think i'll play with some pullups ;) Jul 30 09:14:42 Memphiz: ah, but thats a lot more than no serial, this at least allows you to do prinktk-debugging Jul 30 09:15:31 yes it is not so bad ... but when i introduce a kernel panic by patching shit in the kernel - its a brick after that ;) - but i can look how it bricks over serial *hrhr* Jul 30 09:15:48 heh Jul 30 09:16:42 hmm, weird Jul 30 09:17:04 its really odd ... i've never had any issues with serial ttl shifter ... but at my 1043nd it just doesn't do Jul 30 09:17:48 mhh i think i should try to borrow me an oscilloscope from work to see what is going on there Jul 30 09:20:35 so, i flash an older image (i use this so i can easily mtd in a second step) from CFE, reboot, get to a console with jffs all built and mounted, then scp my new image, and mtd -r write foo.trx linux, and when it reboots it hangs at booting "Starting program at 0x80001000", when i power cycle printenv shows a bunch of stuff missing, like STARTUP, etc. I setenv them back, i reboot, it hangs in the same place, this time the printenv says i'm Jul 30 09:20:35 apart from that, the patch looks (imo) good; it follows the code style, is well documented, I don't see anything missing for a new option, and it seems to work - what more do you want ;) Jul 30 09:20:36 still good. wtf? Jul 30 09:21:37 KanjiMonster - find a bug in chrome to get 3113,70$ and make holiday on hawaii ... but that is completly unrelated :) Jul 30 09:21:47 i just reflashed my starting image from cfe and it's booting apparently normally Jul 30 09:21:58 russell_: perhaps the image is too big and overwrites the nvram? Jul 30 09:22:04 indeed weird Jul 30 09:22:09 nope Jul 30 09:22:16 -rw-r--r-- 1 russell staff 2494464 Jul 29 04:01 openwrt-brcm47xx-squashfs.trx Jul 30 09:23:42 i'm rusty, how do i rebuild everything. maybe the six months of collecting dust caused my build tree to go haywire Jul 30 09:24:11 that is, what do i nuke to ensure everything rebuilds from clean bits? Jul 30 09:25:02 make distclean Jul 30 09:27:55 i'm trying an image i built in april now, not sure if i ever got around to trying it Jul 30 09:28:58 hmm. that one boots okay, circa r21175 Jul 30 09:29:22 is anything using ath5k out of the box these days? Jul 30 09:31:57 it even reboots, so it didn't seem to munch nvram, which is nice Jul 30 09:33:16 russell_: nbd mentioned there's still a lot to do until ath5k is ready Jul 30 09:35:06 i recall aaron kaplan was having a lot of trouble with madwifi and olsrd-needed-vap combinations or something a year or two ago, i'm wondering if ath5k might have fixed that. Jul 30 09:36:31 just for kicks, i'm going to try my death-image again Jul 30 09:37:07 russell_: iirc the whole vpa stuff does not work as well as with madwifi yet Jul 30 09:37:12 *vap Jul 30 09:37:54 what is *with* those people! /me shakes an angry impatience stick at the linux-wireless yahoos Jul 30 09:39:50 hmm. boot hangs again. i suppose it might just be the console Jul 30 09:41:03 I would guess ath5k looses to ath9k in attention since it is probably for obsolete hardware ... Jul 30 09:41:13 *loses Jul 30 09:43:29 i went to one of the linux-wireless summits in 2006 or 2007, when they were starting down the mac80211 road. speed it up, dudes! ;-) Jul 30 09:46:58 KanjiMonster: if you want ath5k to work, how about making it work yourself? scratch your own itch and stuff ;) *ducks* Jul 30 09:47:02 err Jul 30 09:47:09 @russell_ Jul 30 09:47:13 * russell_ assumes that by taking long naps, life will be better when i wake up. Jul 30 09:48:57 nvram was fine after the second flash of the death-image Jul 30 09:49:07 might have been a fluke Jul 30 09:50:11 serial console hangs at "Starting program at 0x80001000" is the symptom Jul 30 09:55:09 _trine mhh are you sure the dockstar has 256mb ram ... in your links they say it has 256mb flash and 128mb ram ... Jul 30 09:59:15 acoul * r22416 /trunk/target/linux/atheros/ (4 files in 4 dirs): [atheros] add missing bits & bytes from r22412 Jul 30 10:19:57 et0macaddr value :flag =2 value=00-0f-b5 ... Jul 30 10:20:06 that =2 is different Jul 30 10:25:42 hmm. also, stock bootloader seems to give me this: Jul 30 10:25:43 Loading: 0x80500000/1151874 0x80619382/35022 0x80621c50/24 Entry at 0x80500000 Jul 30 10:25:43 but i'm seeing this now: Jul 30 10:25:43 Loading: 0x80000000/7864 Entry at 0x80001000 Jul 30 10:37:33 Memphiz: ah, I just remembered, enabling the learning /might/ trigger the same thing that is prevented on atheros switches by regularly flushing the learned addresses: here were some problems when moving ports/going from wireless to wired or vice versa) - perhaps nbd can shed some light on this whether this is an atheros specific problem Jul 30 10:38:59 yeah ... i thought that it would have a reason why juhosg explicitly turned off the learning ... Jul 30 10:39:44 you mean that with learning there are problems when changeing the wired machine on one port to another? Jul 30 10:39:54 because it doesn't relearn the mac on the port? Jul 30 10:41:33 well, at least on the ar8316 I only saw the "obvious" problem - as long as the moved device didn't send a frame thus triggering the relearning, it wasn't reachable Jul 30 10:42:31 ahh ... that is configurable Jul 30 10:42:49 i have activated the "drop packets whose destination adress is not learned" Jul 30 10:43:00 maybe we have to deactivate this for exactly this special case Jul 30 10:43:15 the problem is that the device is learned, but still has the wrong old port[ Jul 30 10:43:18 -[ Jul 30 10:43:25 so enable learning and don't dropping unlearned DAs could be the final stadium Jul 30 10:43:37 yeah Jul 30 10:43:53 and probably setting the ageing time to a minimum Jul 30 10:44:00 then wie should reduce the aging ... by default the mac is valid for 300 secs Jul 30 10:44:04 yea Jul 30 10:44:48 how is this on other switch drivers? Jul 30 10:44:58 the bigger problem are clients moving from wired to wireless - unless they try to access other lan devices their traffic never touches the switch so the switch never learns the device is now at the cpu port Jul 30 10:45:03 we should try to get an equal configuration for switches here Jul 30 10:45:27 for atheros devices its the same, 300 secs Jul 30 10:45:52 but shouldn't devices have different macs for wired and wireless? Jul 30 10:46:00 ah, right Jul 30 10:46:17 this should prevent issues when going from wired to wireless Jul 30 10:46:59 but you're right ... my 1043nd has one mac for wired and wireless *disgusting* Jul 30 10:50:21 strange ... i would think, that it should be a hardware feature when the linkstate is changing it should flush this port ... Jul 30 10:50:33 or trigger relearning Jul 30 10:52:52 nbd * r22417 /trunk/package/mac80211/patches/524-mac80211_offchannel_fix.patch: mac80211: minor fix for the offchannel status change notification Jul 30 10:55:24 Memphiz: well, forwarding frames with unknown DAs to all other ports is the default behaviour on atheros switches, so it's probably best to not enable dropping them Jul 30 10:57:40 yes Jul 30 10:57:44 i also think so now Jul 30 10:59:43 but the problem when moving one device from one port to another and not sending anything from this device is a regular problem i think ... ageing is implemented for this (after the old entry is aged out, the traffic is forwarded to all ports and when the moved device responds to it, the switch can learn the new port of the mac) Jul 30 11:00:17 or are there other approaches for this on other switches? Jul 30 11:05:23 not that I know of. But I only know atheros switches, and have taken a look at some realtek datasheets, but not really looked for anything about this Jul 30 11:08:39 Memphiz: do you have the datasheets for the realtek switches? Jul 30 11:09:16 if not: http://realtek.info/pdf/rtl8366s_8366sr_datasheet_vpre-1.4_20071022.pdf Jul 30 11:09:51 you could probaly emulate this behaviour (regularly checking the port states, and if one goes down, flush all entries for this port), but I don't know how much of a performance inpact the continous reading of switch registers is for the cpu. Iirc reading registers wasn't that fast Jul 30 11:12:46 nbd yes i have Jul 30 11:13:01 there is no possibility to alter the aging time ... it is between 200 and 400 sec. typicall 300s Jul 30 11:13:41 nbd: that datasheet is incomplete, it only describes the basic phy registers, but not the extended ones (though you find some values in the functional descriptions) Jul 30 11:13:51 KanjiMonster polling the link status is bad ... if there was en interrupt on linkchange this could be implemented easaly i think ... Jul 30 11:14:32 http://realtek.info/pdf/rtl8366_8369_datasheet_1-1.pdf <- i use this one Jul 30 11:15:13 as i read the datasheet i think we just should stick with fixed ageing (as it is not configurable) and don't drop unlearned DAs ... so in worstcase the lan device is available after 400 secs after moved from port to port Jul 30 11:15:47 i don't think this is a so common wish to switch devices between port ... or what do you think @ nbd and KanjiMonster Jul 30 11:18:48 i think that's the way to go Jul 30 11:20:10 well, if we ever implement link status polling, we can issue a table flush whenever we see a link status change Jul 30 11:20:27 then we have at least a part of the problem covered Jul 30 11:20:35 yes ... we should have this in mind Jul 30 11:21:00 and we should implement a flush op as an swconfig attribute Jul 30 11:21:05 so that user space can trigger this Jul 30 11:21:17 why only a part? i thought this was the only "problem" (not so nice behaviour) Jul 30 11:21:30 think multiple APs and a client roaming between them Jul 30 11:22:01 (APs connected to each other via cable) Jul 30 11:22:02 ohh yes ... i missed the wireless problems Jul 30 11:22:21 isn't there a "link" state change aswell? Jul 30 11:22:45 forget my last question ... ;) Jul 30 11:22:47 link status polling should probably implemented anyways, to make the behaviour between switches consistent (afaik broadcom switches change their link state depending on whether a device is connected or not) Jul 30 11:23:13 well, the switch driver does ;) Jul 30 11:24:18 well i can make a patch for enable learning in hw init (and allow non learned DAs to pass) and add an swconfig attribute to flush the tables (toggleing learning off and on again should be enough) ... Jul 30 11:25:51 KanjiMonster you mean the broadcom flushes the port when link state changes? because link state changes if a device is connected or not should be the case aswell on rtl switches (otherwise the leds couldn't indicate the link state ;o) ) Jul 30 11:27:36 Memphiz: I mean the linux-visible interface link state - the leds are normally controlled by the hardware, not software Jul 30 11:28:08 ahh ... now i understand *G* Jul 30 11:29:44 the aging time on atheros switches is configurable, minimum is 17s for the ar8216 and 7s for ar8316 Jul 30 11:30:05 seems atheros has made a better work then realtek Jul 30 11:31:52 realtek has always been infamous for their hardware designs ;) Jul 30 11:32:24 and for their cheap prices ;) Jul 30 11:39:09 yeah. also the rtl8139 was /the/ ethernet chip almost everyone had - slow, but cheap, and drivers available for everything - I have several older pci ethernet cards, and none has a win7 driver, exept the realtek one Jul 30 12:05:40 awesome, rebuild nukes my nvram too Jul 30 12:24:53 russell_: hmm, there should be a check thaqt prevents it Jul 30 12:26:07 i don't see console messages as the newly flashed r22409 boots, and when i reboot, the nvram is reset to non-functionally values Jul 30 12:30:35 that and i don't see it with r21175 are all i really know at this point Jul 30 12:39:22 * russell_ trying again, really completely from scratch this time Jul 30 13:02:41 yeah, completely from scratch one is hoarked as well Jul 30 13:03:26 * russell_ is going to sleep on it and maybe try to bisect when i wake up Jul 30 16:50:29 build #82 of sibyte is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/82 Jul 30 17:49:26 * russell_ starts bisecting Jul 30 18:26:56 circa r21945 looks suspicious Jul 30 18:27:18 r21740 is good Jul 30 18:57:07 cshore, hows it goin? Jul 30 18:57:23 ok, still working on FreeSWITCH Jul 30 18:57:37 and you? Jul 30 18:57:37 nice. nice. Jul 30 18:57:42 how's the little one? Jul 30 18:58:04 im just configuring my new linux install. got a new HDD just for linux :) 1TB Jul 30 18:58:10 r22055 is good too Jul 30 18:58:11 Also, I see you've been getting into devel stuff, that's cool Jul 30 18:58:25 shes doing great. doc just called and said that she passed her hearing test :) Jul 30 18:59:06 yeah. im gonna try to see if i can get this package added to the repository when i get it finished. Jul 30 18:59:33 what package is that? Jul 30 19:03:55 mjproxy Jul 30 19:04:20 its a proxy for making the magic jack service work for linux Jul 30 19:04:27 cool Jul 30 19:05:14 i didnt write the code though Jul 30 19:05:57 and who ever did didnt put a GPL or anything :/ Jul 30 19:06:18 ah, I see, so you're trying to get a hold of them? Jul 30 19:06:58 *ahold Jul 30 19:07:11 yep Jul 30 19:13:30 finally got firefox to work with multiple x screens :) Jul 30 19:13:52 dual head...I like that myself too Jul 30 19:14:52 for sure. now i can watchtv or play CoD while having firefox open on the other screen Jul 30 19:17:51 ugh. so tired. and I just got up a few hours ago. Jul 30 19:24:06 build #79 of octeon is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/79 Jul 30 19:35:00 build #74 of gemini is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/74 Jul 30 20:01:43 r22248 good; r22338 bad Jul 30 20:26:44 build #74 of etrax is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/74 Jul 30 20:30:49 ping nbd Jul 30 20:37:17 r22293 good; r22338 bad Jul 30 20:39:25 russell_: what problem do you have? Jul 30 20:39:40 cshore: pong Jul 30 20:40:04 wgt634u, a) doesn't show any boot up messages; b) if i wait too long, nvram gets corrupted Jul 30 20:40:40 nbd: how did you do the serial cable so that you didn't have to figure out which was tx/rx/gnd/vcc? Jul 30 20:40:44 Hauke looks like the responsible party ;-) Jul 30 20:40:54 a) is caused by 22398 Jul 30 20:41:33 i think i saw it on 22338 too Jul 30 20:42:31 this commit produces a bad image checksum, so the CFE does not boot the image Jul 30 20:42:35 cshore: i use usb stuff. usb stuff typically doesn't require vcc Jul 30 20:42:45 so i look for gnd on the board, as it's easy to recognize Jul 30 20:42:59 then figure out tx by trial and error (since i know where rx is on my cable) Jul 30 20:43:03 and then i figure out rx Jul 30 20:43:09 nbd: yes, exactly! Jul 30 20:43:14 nbd: right, ok, I though you had a no guessing cable Jul 30 20:43:23 nope Jul 30 20:43:32 that would probably cost more and the complexity wouldn't be worth it Jul 30 20:43:33 ;) Jul 30 20:43:38 nbd: true Jul 30 20:43:44 rx vs tx, only two possibities Jul 30 20:43:54 nbd: I'm just looking at my cable options Jul 30 20:44:10 nbd: I need a new usb to serial cable and don't want to have to hack up another one Jul 30 20:44:18 i have this one: http://www.avrfreaks.net/index.php?func=viewItem&item_id=745&module=Freaks%20Tools Jul 30 20:44:21 the $20 ftdi one is pretty easy route Jul 30 20:44:23 it works well Jul 30 20:44:30 <_trine> make one Jul 30 20:45:08 Hauke: is it fixed somewhere? Jul 30 20:45:14 _trine: it's harder to get parts to make one than to buy a bought kit (and more expensive too) Jul 30 20:45:21 no I am just doing the commit Jul 30 20:45:22 <_trine> http://homepage.ntlworld.com/tenerifeworld/max3232.jpg Jul 30 20:45:26 <_trine> no it isnt Jul 30 20:45:44 sweet, i'll test when it's available Jul 30 20:45:50 <_trine> you can get free samples if you ask Jul 30 20:45:54 _trine: $6.00 on ebay for a bought kit Jul 30 20:46:01 _trine: ? where? Jul 30 20:46:53 <_trine> http://www.maxim-ic.com/ Jul 30 20:47:14 <_trine> http://www.electro-tech-online.com/general-electronics-chat/9566-free-sample-chips.html Jul 30 20:47:41 Hauke: can you ping me when the commit is in, please? Jul 30 20:48:40 yes will do Jul 30 20:48:55 <_trine> most of the large electronic companies will supply free samples of almost anything if you ask especially if you say you are a student or are associated with an educational establishment Jul 30 20:49:21 <_trine> or in this case a developer Jul 30 20:49:39 <_trine> ask and thou shalt receive Jul 30 20:49:48 <_trine> :) Jul 30 21:04:52 Hauke: r22311 appears to have a console problem (that is, i see nothing after "Starting program at 0x80001000") but it doesn't seem to do nasty things to nvram. Jul 30 21:08:50 * russell_ suspects the "Swapping serial ports" stuff Jul 30 21:19:52 hauke * r22418 /trunk/ (2 files in 2 dirs): (log message trimmed) Jul 30 21:19:53 mostly revert commit r22396 Jul 30 21:19:53 CFE does not boot images generated with these checksums because of Jul 30 21:19:53 wrong checksum. Jul 30 21:19:53 After flashing then with tftp to my Asus wl500-GPv1 the following messages Jul 30 21:19:53 are show: Jul 30 21:19:53 Null Rescue Flag. Jul 30 21:20:20 russell_: commited Jul 30 21:20:25 i see Jul 30 21:20:28 thanks Jul 30 21:20:37 but ... i fear that's not what i'm seeing Jul 30 21:20:56 yes the nvram showul not de affected by this Jul 30 21:21:03 or the serial Jul 30 21:21:51 something between 22293 and 22311 Jul 30 22:18:59 nbd * r22419 /trunk/package/mac80211/ (7 files in 2 dirs): mac80211: update to 2010-07-29, add pending patches to fix further issues with calibration Jul 30 22:42:38 cshore, ping Jul 30 22:47:07 nbd, ping. Jul 30 22:47:47 a few days ago Hauke said that newer releases of trunk was unable to boot. is this still true? Jul 30 22:49:42 hmm Jul 30 22:49:55 after looking at the trunk log, looks like he just fixed it ;p Jul 30 22:56:49 florian * r22420 /packages/net/squid/ (6 files in 2 dirs): [package] update squid to 2.7STABLE9, thanks to Tomasz Mon Jul 30 23:00:59 ping RealOpty Jul 30 23:01:58 rather pong RealOpty Jul 30 23:02:04 hehe Jul 30 23:02:07 i was gonna say Jul 30 23:02:17 if u need help finding the Tx and Rx Jul 30 23:02:30 you can use a voltage meter. Jul 30 23:08:04 nah, I just wanted to be lazy Jul 30 23:08:57 I found the problem any (I think). The board is 3.3V not 5V and my USB-Serial is 5V Jul 30 23:09:27 (Well 6V is what I measured actually) Jul 30 23:10:11 So it's a voltage level issue, I think Jul 30 23:14:54 ahh Jul 30 23:15:16 i have a link on the wiki for a 3.3 ttl cable thats $0.53 usd ;p Jul 30 23:15:49 doh! now you tell me Jul 30 23:16:04 Where's the link? Jul 30 23:16:19 http://nuwiki.openwrt.org/toh/simpletech/simpleshare/serialport?s[]=simpleshare Jul 30 23:17:46 cut off the end and you have 3 wires. Jul 30 23:17:52 grnd, rx, tx Jul 30 23:18:12 yeah, I already ordered a cable, but doh! Jul 30 23:18:23 damn Jul 30 23:18:57 the shipping on that cable cost 5X the cost of the cable lol Jul 30 23:23:43 jow * r22421 /trunk/target/linux/generic/ (config-2.6.34 config-2.6.35): [kernel] fix a typo in config-2.6.34 and config-2.6.35, thanks frogonwheels Jul 30 23:27:37 ah, I hate days when I can't focus on work....need to get my job done! Jul 30 23:30:07 heh I slacked the whole week at the office Jul 30 23:30:41 and here i spent all day with my nose to the grindstone, just to be thwarted at the end! Jul 30 23:31:10 this wgt634u board suddenly doesn't want to boot the known-good image i've been using all day Jul 30 23:31:44 heh, I have a promise to keep....trying to get some coding done for Monday Jul 30 23:32:17 ooh, this is cool! Jul 30 23:32:21 Starting program at 0x80001000 Jul 30 23:32:21 **Exception 8: EPC=801EA5FC, Cause=00000008, VAddr=448302C8 Jul 30 23:32:21 RA=80001434, PRID=00029007 Jul 30 23:32:33 that was with a random older image Jul 30 23:32:40 still, an improvement Jul 30 23:32:42 xMff: sounds like fun :) I wondered how you manage to spend so much time on channel ! Jul 30 23:33:24 well I must do iPad coding at work but most of the staff is on vacation, so we play openarena Jul 30 23:34:22 need to speed things up a bit next week ;) Jul 30 23:36:02 there's always a price Jul 30 23:38:37 cshore: are you aware of any bcm63xx boards that do vlan stuff? Jul 30 23:39:03 xMff: yes, the comtrend routers, and the gw6200, at the least Jul 30 23:39:29 have user in #openwrt why got some device from ebay Jul 30 23:39:35 and the network config is unusual Jul 30 23:39:53 http://openwrt.pastebin.com/7gT5tBUi Jul 30 23:40:01 he claimed that was the default Jul 30 23:40:33 he bricked it when trying to assign port 4 to vlan 0 Jul 30 23:40:44 oops Jul 30 23:40:48 s/why/who/ Jul 30 23:41:18 hmmm....did it come with OpenWRT on it, or what? Jul 30 23:41:29 yes Jul 30 23:41:35 http://pastebin.com/L0S3qkVw Jul 30 23:41:44 96358VW2 Jul 30 23:42:45 I've seen code for those in the Makefile and board kernel stuff Jul 30 23:43:03 but I don't remember what they are Jul 30 23:43:16 I can look, if you need Jul 30 23:43:24 no need to Jul 30 23:43:43 whoever flashed the board took a jffs2 image Jul 30 23:43:53 so resetting is impossible Jul 30 23:44:05 a, needs serial so he can do CFE Jul 30 23:44:10 *ah Jul 30 23:44:11 he has serial Jul 30 23:44:22 but I'm not sure whether the running image is vanilla Jul 30 23:44:33 and he rather unexperienced Jul 30 23:44:41 so I will not tell him to reflash right now ^^ Jul 30 23:44:47 ok Jul 31 01:38:53 i think i found the wgt634u breaking commit, it's r22296 Jul 31 02:13:49 russell_: what arch is that btw? Jul 31 02:28:53 brcm47xx **** ENDING LOGGING AT Sat Jul 31 02:59:57 2010