**** BEGIN LOGGING AT Mon Sep 06 02:59:57 2010 Sep 06 06:56:51 build #74 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/74 Sep 06 07:05:00 acoul * r22952 /packages/net/quagga-unstable/ (4 files in 2 dirs): net/quagga-unstable: Update to v0.99.17, refresh patches Sep 06 07:25:40 acoul * r22953 /trunk/package/busybox/ (9 files in 2 dirs): package/busybox: update to busybox-1.7.2, include upstream patches, refresh patches Sep 06 08:47:24 hi. I'm running a mips malta vmlinux with qemu and it freezes after finding the netcard. I'd like to find out why so I enabled debugging symbols in the kernel and ran qemu with gdb. When I connect to the gdb server, it can't resolve any symbols... Am I doing something wrong? Sep 06 10:44:44 blogic: ping Sep 06 10:47:00 KanjiMonster: hi Sep 06 10:48:53 blogic: I wondered wether the arcor/vodaphone easybox 802 is the same platform as the 801. Sep 06 10:50:24 do you have one ? Sep 06 10:50:33 i have been trying to get hold of a sample for ages Sep 06 10:50:43 i can swap it against a unit that is already supported :) Sep 06 10:50:51 blogic: http://cgi.ebay.de/ws/eBayISAPI.dll?ViewItem&item=200514373669 Sep 06 10:52:37 I just saw the offer and wondered if its still openwrt-capable (in theory) Sep 06 10:53:18 it is Sep 06 10:56:21 KanjiMonster: 30 euro is a lot Sep 06 10:56:27 i will wait till i get a donation Sep 06 10:56:39 if it were 10 euro or so i would get it Sep 06 10:57:08 but i also need a speedport w303V w503V, vodafone a802, KPN units, 2 netgear units Sep 06 10:57:22 all of them can run openwrt once i get them and added mach support Sep 06 10:57:40 after that i will only keep them in the shelf to test incase support gets broken Sep 06 10:57:54 so 30 euro is a lto specially as i need 6 different units :) Sep 06 10:58:13 just spent 95 euro last week to get more units to support, so the budget for this month is gone really Sep 06 10:59:35 ah, I see Sep 06 11:00:09 i normally wait till the units are 10 euro or less on ebay until i order lots of them Sep 06 11:00:21 the speedport w502V is a nice and cheap unit Sep 06 11:00:32 has a atheros legacy wifi that runs witrh madwifi Sep 06 11:00:42 so if oyu need a supported unit that might be good Sep 06 11:06:25 I was more looking at potential new victims for playing around (and saw the offer today - 30 euros for the hardware seems quite ok, especially since it seems to come pre-equipped with a serial header) Sep 06 11:08:13 [florian], ping Sep 06 11:08:29 <[florian]> Wipster: pong Sep 06 11:13:44 [florian], hi, the problem narge has with the ethernet is a bit odd. From what I understand the cpmac is detecting the phy and bringing it up but the packets are scrambled with or without the switch driver. If its seeing the PHY and its sending, how is that different from when it was a fixed phy that might cause the header problem? Sep 06 11:16:51 [florian], does detecting the phy require the switch driver to be there or can it just send? Sep 06 11:27:29 <[florian]> Wipster: this is expected with a marvell switch, those beasts require a hack for the packets not to be mangled Sep 06 11:27:50 <[florian]> Wipster: and it requires a hack in the ethernet driver to account for these mangled packets Sep 06 11:31:50 [florian], ahha I see, so how did it work for him when fixed phy or the init_hack was used? Sep 06 11:35:00 <[florian]> Wipster: I guess the switch was simply put in its default mode where it does not mangle the skb's Sep 06 11:39:34 [florian], hmmmm interesting so when the driver is disabled why does it still mangle, shouldn't it be back to that default? Only difference is the ethernet can 'see' it Sep 06 11:40:20 <[florian]> Wipster: I do not have the switch datasheet, but it may happen that if it "finds" a host being alive behind the MDIO/MDC lines, it may assume that the host knows how the switch shall be configured Sep 06 11:40:36 <[florian]> Wipster: there is an example about how to deal with switch switches in the ar231x ethernet driver Sep 06 11:42:19 [florian], ahhhh yes... learning all the time. I shall have a look at that when I have a moment. Thanks for explaining that one Sep 06 11:43:34 <[florian]> Wipster: about the reason this used to work with fixed phy is still pure speculation, but that's the only likely solution that I see Sep 06 11:44:15 acoul * r22954 /trunk/package/kernel/modules/crypto.mk: package/kernel: update crypto.mk for kernel 2.6.36 Sep 06 11:44:37 hi. I'm running a mips malta vmlinux with qemu and it freezes after finding the netcard. I'd like to find out why so I enabled debugging symbols in the kernel and ran qemu with gdb. When I connect to the gdb server, it can't resolve any symbols... Am I doing something wrong? Sep 06 11:45:04 [florian], yeh sounds like a good conclusion Sep 06 12:13:06 hi. I'm running a mips malta vmlinux with qemu and it freezes after finding the netcard. I'd like to find out why so I enabled debugging symbols in the kernel and ran qemu with gdb. When I connect to the gdb server, it can't resolve any symbols... Am I doing something wrong? Sep 06 12:13:34 oops, sorry about "double posting" Sep 06 12:19:25 xMff: thanks for committing the kvm stuff ;-) and greetings from Croatia :P Sep 06 12:52:21 stintel: u live there or on holiday ? Sep 06 13:51:38 hcg * r22955 /packages/lang/python-cjson/ (4 files in 2 dirs): [packages] add package python-cjson Sep 06 13:56:31 what is included in an openwrt SDK? Sep 06 13:59:49 hcg * r22956 /trunk/target/linux/omap35xx/gumstix/profiles/gumstix.mk: [omap35xx] add package python-cjson to gumstix profile Sep 06 16:37:55 blogic * r22957 /trunk/ (package/ifx-vmmc/Makefile target/linux/ifxmips/extract.sh): Sep 06 16:37:55 [ifxmips] Sep 06 16:37:55 extracted firmware files were generated with an incorrect umask Sep 06 16:42:12 blogic * r22958 /branches/backfire/ (package/ifx-vmmc/Makefile target/linux/ifxmips/extract.sh): Sep 06 16:42:12 [ifxmips] Sep 06 16:42:12 merge [22957] in backfire Sep 06 16:42:12 extracted firmware files were generated with an incorrect umask Sep 06 17:19:09 kristianpaul: toolchain and firmware utils Sep 06 17:29:08 xMff: okay, something related to do auto packaging too maybe? Sep 06 17:29:31 not quite Sep 06 17:29:44 its mainly intended for cross compiling single packages Sep 06 17:30:08 but you can use the toolchain manually as well, e.g. to compile sources without packaging them first Sep 06 17:32:00 you me "inside" the SDK as i compile apps on my computer,or i will end doing exports anyway? Sep 06 17:32:29 expots? Sep 06 17:34:31 exports* Sep 06 17:34:43 i meant enviroment variables Sep 06 17:34:58 export foo_path ..... Sep 06 17:35:01 ah Sep 06 17:35:05 you'll need exports Sep 06 17:35:12 it is not a chroot environment or something Sep 06 17:35:43 i see Sep 06 17:39:37 obsy * r22959 /packages/net/vsftpd/Makefile: [packages] vsftpd: update to 2.3.2 Sep 06 19:07:35 build #75 of ps3 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/75 Sep 06 20:01:13 jow * r22960 /trunk/ (3 files in 3 dirs): [package] opkg: add --force-run-hooks option to run postinstall scripts in offline root mode, utilize it for host package install Sep 06 20:28:50 acoul * r22961 /packages/libs/libpam/ (5 files in 2 dirs): libs/libpam: update libpam to version 1.1.2, refresh patches Sep 06 20:44:38 someone here who builds openwrt on snow leopard (especially target dockstar ...) Sep 06 20:45:57 get very strange errors when buildroot compiles host/u-boot-2009.11 (don't even know why he is compiling it ...) Sep 06 20:46:20 iirc the dockstar target includes a second stage loader now Sep 06 20:46:26 thats the u-boot it tries to compile Sep 06 20:46:53 <_trine> it compiles on ubuntu Sep 06 20:49:09 but why does it compile HOST/u-boot ... Sep 06 20:49:16 it is the wrong architecture ... isn't it? Sep 06 20:49:24 should it build target/u-boot than? Sep 06 20:49:29 probably for image preparation Sep 06 20:49:51 put some V=99 log into pastebin Sep 06 20:49:56 _trine: yes ... id works under ubuntu ... but i'm most of the time under macosx though ... and it would be good to compile here aswell (bigger hdd *g*) Sep 06 20:50:06 xMff: k one moment Sep 06 20:50:46 there is some struct type missmatch Sep 06 20:50:47 http://pastebin.com/WEfd4YK8 Sep 06 20:52:49 jow * r22962 /branches/backfire/package/uhttpd/ (11 files in 3 dirs): [backfire] merge r22630, r22692, r22805 Sep 06 20:53:13 seems it has some problems with the ulonglong type ... mhhh is it a stdint type? Sep 06 20:53:22 probably not Sep 06 20:53:30 I recall fixing a very similar issue for freebsd Sep 06 20:53:34 let me search it Sep 06 20:56:03 heh well... Sep 06 20:56:07 https://dev.openwrt.org/changeset/22344/trunk/tools/mkimage Sep 06 20:56:47 mhh Sep 06 20:56:51 #ifndef ulonglong #define ulonglong unsigned long long #endif Sep 06 20:57:07 chances are high that you just need to add a || defined(__APPLE__) somewhere Sep 06 20:57:08 but the sha1.h is from macosx somehow ... i can't understand why they are using undefined types ... Sep 06 20:57:11 will test it Sep 06 20:57:20 yeah could be Sep 06 20:57:24 will grep for ulonglong Sep 06 20:57:57 config-win.h:typedef unsigned __int64 ulonglong; /* Microsofts 64 bit types */ <- PFUI! *G* Sep 06 20:59:13 <_trine> by the way am I right in thinking we now have mysql Sep 06 20:59:28 Having some issues with 10.03 mtd utils :root@OpenWrt:/# nandwrite -m -j /dev/mtdblock0 = MEMGETINFO: Inappropriate ioctl for device Sep 06 20:59:55 nandwrite needs the non-block devices Sep 06 20:59:56 _trine: think i've seen a commit latly which would indicate that - yeah ;) Sep 06 21:00:16 a.k.a. /dev/mtd0 Sep 06 21:00:53 <_trine> Memphiz, I just update svn and the feeds when I saw it Sep 06 21:01:42 ist cool ... Sep 06 21:01:45 xMff: sorry, root@OpenWrt:/# nandwrite -m -j /dev/mtd0 = Unknown flash (not normal NAND) Sep 06 21:01:52 i think i have some applications for mysql on dockstar :) Sep 06 21:02:53 <_trine> I have cherokee and wordpress running on a dockstar but with debian at the moment Sep 06 21:04:02 xMff: I think this is why I cant flash the device. -> nandtest /dev/mtd0 = Bad block at 0x00000000 Sep 06 21:04:46 there are several other bad blocks as well, I think the bbt got erased somehow Sep 06 21:06:04 <_trine> Memphiz, have you noticed Atelco have increased the price of dockstars now Sep 06 21:06:43 ohh no ... i did only notice that the store where i bought mine has increased the price to 50 bucks Sep 06 21:09:57 <_trine> still only 24.90 though Sep 06 21:10:12 <_trine> they increased the price by 5 euro Sep 06 21:12:05 <_trine> but the price is most definitely rising Sep 06 21:12:43 <_trine> 2 days ago there were for suppliers selling for 24.90 and today only 2 Sep 06 21:12:54 <_trine> /for/four Sep 06 21:16:58 xMff i think the problem is, that he has /usr/local/include as include path instead of /usr/include ... any idea where this path is set in buildroot? Sep 06 21:17:11 nowhere Sep 06 21:17:23 how is it choosed Sep 06 21:17:28 it must come from the package itself iirc# Sep 06 21:17:33 kk Sep 06 21:26:02 mb * r22963 /trunk/include/host.mk: host.mk: Don't use bash-ism Sep 06 21:27:57 damn why is this fscking sha1.h using stdints without including the needed header ... i don't get this ... thats so ugly *g* Sep 06 21:28:22 think i have to go to bed ... hate when this happens *G* Sep 06 21:28:29 probably because it worked on Linux :P Sep 06 21:28:47 mhh don't know why it should ... but you're right ... it works on linux ... Sep 06 21:29:18 i've adapted compiler.h so that it should include stdint.h before including sha1.h ... but it still spits on uint32 ... don't know how this could happen Sep 06 21:30:01 If uboot gets corrupted or overwritten, is there a way to recover, short of replacing the chip? Sep 06 21:30:08 no Sep 06 21:30:12 well jtag Sep 06 21:32:07 what utility is used to replace uboot with jtag? Sep 06 21:32:38 openocd probably Sep 07 00:33:57 blogic: can you fix the stray f insertion in r22958? see #7887 **** ENDING LOGGING AT Tue Sep 07 02:59:57 2010