**** BEGIN LOGGING AT Thu Dec 30 02:59:58 2010
Dec 30 05:41:49 <russell--> uh-oh, more problems on wgt634u?
Dec 30 05:42:41 <russell--> hmm.
Dec 30 05:44:02 <russell--> no ethernet
Dec 30 05:44:16 <russell--> r24843
Dec 30 05:45:03 <russell--> b44: Invalid MAC address found in EEPROM
Dec 30 05:46:13 <russell--> b44 ssb0:0: Problem fetching invariants of chip, aborting
Dec 30 05:46:13 <russell--> b44: probe of ssb0:0 failed with error -22
Dec 30 05:48:04 <russell--> ifconfig eth0 -auto
Dec 30 05:48:04 <russell--> ifconfig eth0 -auto
Dec 30 05:48:22 <russell--> that gets me an address in CFE
Dec 30 06:02:59 <sn9> oh geez, not the weird nvram mac address issue again...
Dec 30 06:06:22 <russell--> r24748 has same problem
Dec 30 06:06:51 <russell--> i notice r24749 was supposed to address this
Dec 30 06:06:58 <russell--> i wonder if it was ever tested
Dec 30 06:18:07 <russell--> r24196 is okay
Dec 30 06:21:50 <russell--> r24513 is bad
Dec 30 07:40:53 * russell-- wants to throttle Hauke
Dec 30 07:52:28 <sn9> at least he didn't make the retarded commits acoul made
Dec 30 08:11:17 <stintel> hey, errare humanum est, no need to call people names
Dec 30 08:13:52 <sn9> i called the code names, not any people
Dec 30 08:15:23 <russell--> people make mistakes, i just wish people took a little more care in not breaking working devices when trying to support new ones
Dec 30 08:16:04 <russell--> wgt634u has been busted for 5 of the last 6 months in trunk
Dec 30 08:21:49 <stintel> true, but it's trunk after all and it's called kamikaze for a reason
Dec 30 08:22:43 <sn9> kamikaze is a vodka-based drink, as are backfire and white russian
Dec 30 08:22:53 <russell--> my mistake was in assuming that when it got fixed, that it was because the problem was understood.
Dec 30 08:25:06 <russell--> pretty sure i know which commit broke it, just building to find out now.
Dec 30 08:25:30 <builder> build #44 of au1000 is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/44
Dec 30 08:26:10 <sn9> knowing which commit broke something does not really illuminate a fix
Dec 30 08:27:03 <russell--> it provides important clues
Dec 30 08:27:19 <sn9> on occasion
Dec 30 08:27:40 <russell--> one fix is just to revert it
Dec 30 08:28:57 <sn9> not always feasible
Dec 30 08:29:36 <sn9> especially if a kernel upgrade is what breaks stuff
Dec 30 08:31:47 <russell--> i can say that's not the case here, and anyway i don't agree you can't just roll back to what works while the problems are ironed out
Dec 30 08:32:13 <russell--> you don't have to commit brokenness to make progress
Dec 30 08:38:37 <sn9> progress is fixing it
Dec 30 08:40:03 <russell--> my point is, it was fixed before it was broken.  by testing it before commiting, it stays fixed.
Dec 30 08:40:47 <russell--> trying r24264
Dec 30 08:47:53 <russell--> ethernet works
Dec 30 08:48:10 <russell--> building r24266
Dec 30 09:18:53 <builder> build #43 of mpc52xx is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/43
Dec 30 09:27:30 <russell--> Hauke: /me is looking for you! ;-)
Dec 30 09:28:31 <russell--> https://dev.openwrt.org/ticket/8549
Dec 30 09:29:17 <russell--> yeah, r24266 is the broken commit
Dec 30 09:29:27 <russell--> *confirmed*
Dec 30 09:32:05 <Hauke> russell_: here is a fix for some problem on the device: https://dev.openwrt.org/changeset/24749
Dec 30 09:32:16 <russell--> no, that isn't the fix
Dec 30 09:32:39 <Hauke> I think netgear does things different that the offer just of fun ;-)
Dec 30 09:33:00 <russell--> did you read the ticket?
Dec 30 09:37:31 <Hauke> russell_: what are the values of et1macaddr, et0macaddr and il0macaddr?
Dec 30 09:37:39 <Hauke> in the nvram?
Dec 30 09:37:46 <russell--> they aren't in nvram!
Dec 30 09:37:58 <russell--> at least, not in the section you read
Dec 30 09:38:53 <Hauke> yes but you can read them out with the nvram program
Dec 30 09:39:25 <russell--> hang on, i need to reflash a working version
Dec 30 09:40:09 <russell--> it'll be a few minutes
Dec 30 09:42:13 <russell--> i think i explained before, the FLSH section starts at offset 0x18000 in the erase block, all the CFE variables start at 0x1e000.
Dec 30 09:45:48 <russell--> on r22295:
Dec 30 09:45:50 <russell--> root@OpenWrt:/# nvram
Dec 30 09:45:53 <russell--> Could not open nvram! Possible reasons are: - No device found (/proc not mounted or no nvram present) ...
Dec 30 10:05:45 <russell--> Hauke: i updated the ticket with a hexdump of /dev/mtdblock4
Dec 30 10:07:24 <russell--> r24266 removes the calls that look at the section where et0macaddr etc live
Dec 30 10:17:31 <russell--> Hauke: did anyone actually test r24749 on a wgt634u?
Dec 30 10:18:13 <phaidros> wiki is back
Dec 30 10:18:20 <Hauke> russell_: yes I got the same bug report from mirko and this fixed the problem for him
Dec 30 10:18:31 <phaidros> sry 4 inconvenience, we had a burned power supply :)
Dec 30 10:20:46 <russell--> Hauke: i build r24843 and still saw the rpoblem
Dec 30 10:21:20 <Hauke> russell_: have you chacked that the new patch was really applied?
Dec 30 10:21:31 <russell--> i am doing that now
Dec 30 10:21:41 <russell--> (not sure why it wouldn't be)
Dec 30 10:22:44 <Hauke> sometimes the build system does notice changes in the patches
Dec 30 10:23:33 <builder> build #57 of at91 is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/57
Dec 30 10:27:31 <russell--> hmm.  looks like it didn't get applied.
Dec 30 10:27:51 <russell--> nvram_parse_macaddr body is only two lines long
Dec 30 10:28:06 * russell-- rebuilding the kernel
Dec 30 10:30:26 <Hauke> russell_: ok that would explain this
Dec 30 10:30:35 <russell--> i'll confirm
Dec 30 10:30:52 <Hauke> we should have a look at the build system so that it find new patches
Dec 30 10:31:08 <Hauke> *finds
Dec 30 10:31:25 <russell--> i can also report that reverting r24749 and r24266 works too ;-)
Dec 30 10:32:24 <Hauke> hmm I should get this netgear device for testing as it is different from the others ;-)
Dec 30 10:34:43 <russell--> yes, you should! :)
Dec 30 10:35:57 <russell--> if you can't find someone i germany to give you one, let me know.  i have a pile (though shipping to europe is expensive and slow)
Dec 30 10:40:00 <russell--> btw, on one of my test compiles, i erroneously saved a .config with the default target profile.  i later went back and switched to Netgear WGT634U, but it didn't update my wireless driver configuration.
Dec 30 10:41:36 <russell--> Hauke: i can confirm r24843 with a recompiled kernel is working now.  thanks for your help!
Dec 30 10:42:49 <russell--> oh, and maybe you could apply the patch I posted yesterday for the WGT634U
Dec 30 10:43:20 <russell--> http://www.pastebin.ca/2032998
Dec 30 10:44:57 <russell--> kmod-usb-ohci is not needed and throws error messages during boot
Dec 30 10:52:36 <KanjiMonster> russell--: changes in target/linux/*/files or /patches don't trigger a rebuild of the kernel unless the kernel version changes
Dec 30 10:53:15 * russell-- steps on landmine of faulty expectations
Dec 30 10:54:09 <KanjiMonster> yeah, it's really not that intuitive since packagages get rebuilt on changes
Dec 30 10:54:53 <KanjiMonster> or better said it triggers a clean for packages but not for the kernel
Dec 30 11:04:07 <builder> build #54 of ubicom32 is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/54
Dec 30 11:23:06 <KanjiMonster> I didn't know ecos/redboot exists for bcm47xx ...
Dec 30 11:29:54 <russell--> link?
Dec 30 11:49:25 <builder> build #57 of brcm47xx is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/57
Dec 30 11:54:14 <KanjiMonster> russell--: ftp://downloads.netgear.com/files/GPL/FS728TPv2_GPL_code.zip - contains ecos/redboot for bcm47xx (or better for bcm53xxx)
Dec 30 12:14:52 <russell--> nice
Dec 30 12:39:08 <swalker> ^ most of the recent bustage is more missing (mostly keyboard) symbols
Dec 30 12:43:30 <KanjiMonster> russell--: it seems like it's a bit of a redboot/cfe crossbreed, with a cfe functions wrapper to get the normal broadcom code to work on redboot
Dec 30 12:45:50 <russell--> interesting
Dec 30 12:47:43 <builder> build #43 of xburst is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/xburst/builds/43
Dec 30 14:59:12 <mazilo> Any perl Embed experts here for OpenWRT who can help me out?
Dec 30 15:15:38 <builder> build #45 of kirkwood is complete: Failure [failed shell_9]  Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/45
Dec 30 15:20:24 <builder> build #45 of ixp4xx is complete: Failure [failed shell_9]  Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/45
Dec 30 15:51:14 <CIA-77> nbd * r24855 /trunk/package/mac80211/patches/560-ath9k_rx_error_handling.patch: ath9k: fix rx dma descriptor error flag handling
Dec 30 15:51:18 <CIA-77> nbd * r24856 /trunk/package/relayd/ (. Makefile): add relayd, can be used as a replacement for client bridge, but does not have any uci integration yet
Dec 30 15:51:28 <CIA-77> nbd * r24857 /trunk/tools/mkimage/patches/: tools/mkimage: remove the broken portability patch, it is no longer necessary with the new version
Dec 30 15:51:40 <CIA-77> nbd * r24858 /trunk/package/mac80211/patches/ (3 files): ath9k: add more pending fixes from linux-wireless@
Dec 30 15:57:30 <KanjiMonster> grmpf
Dec 30 15:57:34 <KanjiMonster> cshore_: ping
Dec 30 16:28:11 <builder> build #54 of orion is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/54
Dec 30 16:48:14 <mazilo> On my Seagate DockStar running on OpenWRT, if I executed "perl -V', it returns the environment paths points to the full staging directory as shown in http://pastebin.com/ptTMpHLn
Dec 30 16:48:31 <mazilo> This doesn't seem to be correct.
Dec 30 16:49:46 <mazilo> On my Linux desktop computer, if I do a 'staging_dir/target-arm_v5te_uClibc-0.9.31_eabi/usr/bin/perl -V', it bumps as shown in http://pastebin.com/MyTbwnwS
Dec 30 16:50:09 <mazilo> Does this mean the OpenWRT Makefile for perl package is in question?
Dec 30 17:01:10 <builder> build #50 of ps3 is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/50
Dec 30 17:10:28 <h3sp4wn> mazilo: Do you have an actual problem with it (perl is a bitch to cross compile)
Dec 30 17:10:49 <h3sp4wn> mazilo: If just the output of that is broken then I think its a minor issue]
Dec 30 17:11:11 <mazilo> h3sp4wn: yes. I run into perl problem.
Dec 30 17:12:18 <mazilo> h3sp4wn: Can you please give me about 10 minutes. I am trying to summarize and post in OpenWRT forum.
Dec 30 17:13:08 <h3sp4wn> mazilo: ok
Dec 30 17:22:15 <CIA-77> juhosg * r24859 /trunk/ (65 files in 16 dirs): cns21xx: add initial support for the Cavium CNS21xx SoC
Dec 30 17:23:37 <mazilo> h3sp4wn: Can you please take a look at my post (https://forum.openwrt.org/viewtopic.php?id=27911)?
Dec 30 17:27:15 <h3sp4wn> mazilo: What happens if you manually do PERL5LIB=/path/to/Config.pm
Dec 30 17:28:27 <mazilo> h3sp4wn: Can you please tell me how can I do that on my DockStar running OpenWRT?
Dec 30 17:29:13 <h3sp4wn> mazilo: You can do it in sh or from the perl script itself
Dec 30 17:29:47 <h3sp4wn> PERL5LIB=/whatever:/whatever ; export PERL5LIB
Dec 30 17:30:09 <h3sp4wn> or you can do use - use lib "/whatever"
Dec 30 17:30:24 <h3sp4wn> use lib "/what/ever/";
Dec 30 17:30:27 <h3sp4wn> (sorry)
Dec 30 17:30:55 <mazilo> h3sp4wn: This is what I got http://pastebin.com/vbpEUeTF
Dec 30 18:03:00 <KanjiMonster> hey Hauke
Dec 30 18:11:46 <builder> build #51 of ramips is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/ramips/builds/51
Dec 30 18:40:34 <KanjiMonster> Hauke: I successfully ported the serial flash driver to 2.6. Writing seems to work also since I successfully broke the stock image with the trx update thingy (something I thought I'd avoid by booting from ram/tftp ;)
Dec 30 19:59:58 <mazilo> h3sp4wn: R U there?
Dec 30 20:13:19 <CIA-77> juhosg * r24860 /trunk/target/linux/generic/ (7 files): generic: more CONFIG_KEYBOARD_* symbols
Dec 30 20:27:07 <builder> build #43 of octeon is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/43
Dec 30 22:03:24 <builder> build #44 of etrax is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/44
Dec 30 22:51:23 <builder> build #48 of ar7 is complete: Exception [exception failed slave lost shell_13 compile_12]  Build details are at http://tksite.gotdns.org:8010/builders/ar7/builds/48
Dec 30 23:42:04 <sn9> feeds/packages/net/xtables-addons no longer compiles, even after a distclean
**** ENDING LOGGING AT Fri Dec 31 02:59:58 2010