**** BEGIN LOGGING AT Sat Mar 31 02:59:56 2007 Mar 31 03:01:24 s1m0nJ_: do you perchance have svn write access? Mar 31 03:23:53 thepeople: yo Mar 31 04:10:23 * sn9_ sighs after 3 hours of waiting Mar 31 08:45:27 * sn9_ cannot wait much longer; starting to fall asleep Mar 31 08:57:37 sn9_: make a ticket with the fixes and go to bed. Mar 31 08:59:14 i'd rather "svn up" and "make" Mar 31 08:59:53 florian * r6769 /trunk/tools/firmware-utils/src/airlink.c: fix crc32 and endianness, thanks sn9 Mar 31 09:00:13 [florian]: and the other one? Mar 31 09:00:28 sn9_: I fail to understand why committing these fixes ASAP is essential. Mar 31 09:00:59 not essential, but it does save time Mar 31 09:01:12 florian * r6770 /trunk/target/linux/rdc-2.6/files/drivers/mtd/maps/rdc3210.c: Fix the division by zero (sn9) Mar 31 09:01:38 <[florian]> sn9_: next time post a diff that I can apply directly to the svn repository Mar 31 09:04:35 ok. thanx. g'night Mar 31 09:06:09 oh, snap -- typo in airlink.c Mar 31 09:08:25 <[florian]> line 291, replace I0 by l0, right ? Mar 31 09:08:38 yep Mar 31 09:09:04 florian * r6771 /trunk/tools/firmware-utils/src/airlink.c: Fix a typo Mar 31 09:10:37 thanx. you gotta admit, that was hard to see Mar 31 09:10:57 <[florian]> sn9_: no prob Mar 31 09:35:16 actually, before i really fall asleep: http://homepage.mac.com/danielg4/airlink.c.patch Mar 31 09:55:42 nbd * r6772 /trunk/package/kernel/modules/other.mk: revert bluetooth kconfig change - it doesn't work that way Mar 31 13:45:53 nbd * r6773 /trunk/include/package.mk: call Build/Clean for clean: in package makefiles Mar 31 14:25:55 awww, i was kinda hoping that patch would be committed while i was asleep. it can wait a little more, i guess Mar 31 16:41:22 why is there a space in the symbol name in that error? Mar 31 18:03:07 ejka: your linux-2.6 code for AR7 works great, I especially like the automatic detection of RAM size and ethernet phy! Mar 31 18:03:31 it'll be better with dsl Mar 31 18:03:45 And for the first time, my AR7 based router has run WLAN from Linux! Mar 31 18:04:33 sn9__: what is the status of dsl support? Is there any work in progress? Mar 31 18:04:46 it needs to be ported from 2.4 Mar 31 18:05:20 there is a new version of the driver in the bugtracker, but for 2.4 Mar 31 18:06:10 ticket #1411 Mar 31 18:06:42 ethernet for AR7 needs also a little bit of improvement - NFS only works stable with rsize=wsize=1024 Mar 31 18:08:21 https://dev.openwrt.org/ticket/1411 Mar 31 18:10:07 btw, do you by any chance remember offhand the kernel config option name for overriding kernel arguments from the bootloader? Mar 31 18:10:51 sn9__: CONFIG_CMDLINE. Mar 31 18:10:56 thanx Mar 31 18:12:06 I took a different approach and include the kernel arguments in the loader (which passes them to the kernel) Mar 31 18:12:48 then i have to recompile redboot Mar 31 18:13:22 i might recompile it anyway, to support jffs2 Mar 31 18:13:58 until then, i'd like to try CONFIG_CMDLINE to support squashfs Mar 31 18:15:02 so you run redboot on an AR7 router? did you exchange the original boot loader? Mar 31 18:16:16 not on ar7 -- my ar7 device is my production router atm, and uses adam2 -- it'll get openwrt when ticket 1411 is resolved and i have a viable fallback Mar 31 18:22:07 what's the difference between rootfs=foo and rootfstype=foo in kernel args? Mar 31 18:27:56 <[florian]> sn9__: it is pretty clear, root is the partition, rootfstype is the partition filesystem type Mar 31 18:28:14 not root=, rootfs= Mar 31 18:28:32 <[florian]> it is an alias Mar 31 18:28:40 ah, ok Mar 31 18:32:14 so, what could cause that error? Mar 31 18:33:16 the one in libc Mar 31 18:33:46 something ccache-related, maybe? Mar 31 18:37:48 [florian]: as soon as that error is taken care of, i have a bigger patch coming your way Mar 31 18:38:14 any ideas? Mar 31 18:44:47 <[florian]> sn9__: which error sorry ? Mar 31 18:44:53 <[florian]> sn9__: ok, let me see the ticket Mar 31 18:45:02 ol '_svcauth_sho rt': can't resolve symbol in lib '/lib/libc.so.0'. Mar 31 18:45:16 note the space in the symbol name Mar 31 18:45:34 <[florian]> that is very weird, yes Mar 31 18:45:41 this is on successful boot and root fs mount Mar 31 18:45:43 <[florian]> sn9__: which gcc/uclibc version do you use ? Mar 31 18:47:56 the default for 2.4 -- gcc 3.4.6 and binutils 2.16.1 Mar 31 18:48:09 <[florian]> really weird, nether met that before Mar 31 18:48:12 <[florian]> sn9__: regarding rdc Mar 31 18:48:34 i'm not sure buildroot allows setting the uclibc version Mar 31 18:48:48 <[florian]> sn9__: I just wrote a flash map driver which uses the redboot partition partition parser, and it does not seem to work Mar 31 18:49:16 i wouldn't think it would Mar 31 18:49:35 <[florian]> they probably screwed up the fis directory Mar 31 18:49:56 <[florian]> I do not even remember seeing it at all Mar 31 18:50:12 i made a couple of cleanups to the flash map driver. the patch will follow resolving the libc issue Mar 31 18:50:31 there are no fis commands in the rdc redboot Mar 31 18:50:48 <[florian]> sn9__: yes, this does not necessarily mean they removed the fis partition table Mar 31 18:50:53 <[florian]> sn9__: but actually they did Mar 31 18:51:27 yes, that can be determined from the linksys sources Mar 31 18:52:04 <[florian]> I will use a static mapping for the moment Mar 31 18:53:14 the binary generated from the stock linksys sources is binary-identical to what airlink ships in their flash, except for the datestamp. ironically, the redboot image in the airlink sources appears unrelated Mar 31 18:53:53 static mapping will be unnecessary once i'm through with the rdc-2.4 target Mar 31 18:53:57 <[florian]> sn9__: does the linksys binary also embedds the tftp server ? Mar 31 18:54:02 yes Mar 31 18:54:17 <[florian]> sn9__: I rather would like you to work on rdc-2.6 Mar 31 18:54:42 it was from the linksys sources that i determined the correct crc32 algo for the generator Mar 31 18:54:54 <[florian]> nifty Mar 31 18:55:27 [florian]: I'd rather like hin to work on ar7 Mar 31 18:55:39 <[florian]> armijn: yeah, this is for tomorrow :) Mar 31 18:55:53 s/hin/him/ Mar 31 18:56:04 or her... Mar 31 18:56:06 i would also like to do something with rdc-2.6, but anything i do to make rdc-2.4 work will also apply to rdc-2.6, since you had rdc-2.4 working at some point Mar 31 18:56:37 <[florian]> sn9__: yes that is true, anyway, rdc-2.6 has a major issue right now, which is probably kernel size limitation Mar 31 18:56:56 <[florian]> sn9__: for instance, use the kernel config in trunk/ enable networking, and you will get a kernel that outputs nothing on the serial console :/ Mar 31 18:57:12 i think my work on 2.4 will help with that in 2.6 Mar 31 18:57:55 but i can only know after 2.4 really works Mar 31 18:58:00 <[florian]> sn9__: sure Mar 31 18:59:05 anyway, back to your questions about the toolchain... Mar 31 18:59:13 <[florian]> yes, the uclibc version cannot be changed Mar 31 18:59:31 <[florian]> sn9__: I personnaly never use ccache because of bad rememberings with it Mar 31 19:00:39 i would think that clearing the caches would be enough to eliminate any ccache-related breakage, no? Mar 31 19:01:19 <[florian]> I guess so Mar 31 19:02:06 wouldn't "make target/clean" and "make package/clean" clear them? Mar 31 19:02:29 or, in the case of uclibc, "make distclean" Mar 31 19:02:30 <[florian]> I am not sure about the implication on the ccache with those targets Mar 31 19:02:38 <[florian]> make distclean will do Mar 31 19:03:09 i did a distclean before building uclibc, iirc Mar 31 19:03:30 <[florian]> humm Mar 31 19:04:36 but is that symbol internal to libc, or is it being referenced from outside? Mar 31 19:04:44 i can't tell Mar 31 19:05:43 and what does "ol" before the symbol name mean? Mar 31 19:08:43 <[florian]> the symbol refers to the dns resolver in uclibc Mar 31 19:08:59 ... Mar 31 19:09:18 <[florian]> or some portmap communication function instead Mar 31 19:09:26 <[florian]> for the ol, I do not know Mar 31 19:11:50 do you still have the buildroot .config you used for rdc-2.4 anywhere? Mar 31 19:12:42 <[florian]> I doubt so Mar 31 19:12:56 <[florian]> unfortunately, no Mar 31 19:13:03 <[florian]> but I can easily regenerate it Mar 31 19:13:11 http://pastebin.ca/418640 in line 10&11, is it a bug? Mar 31 19:13:38 go ahead, thanx. i'll diff it against mine Mar 31 19:15:37 <[florian]> sn9__: http://f.fainelli.free.fr/config-rdc-2.4 Mar 31 19:17:56 <[florian]> crazy_imp: looks like, did you add a package ? made a change in the makefile ? this can be triggered when you modified something Mar 31 19:18:38 looks like the only significant difference between your .config and mine is ccache Mar 31 19:18:41 [florian]: i only played around with some makefiles in package/ Mar 31 19:18:57 (package makefiles) Mar 31 19:19:13 <[florian]> sn9__: try without ? Mar 31 19:19:29 <[florian]> crazy_imp: ok, maybe you forgot some define/endif ? Mar 31 19:19:30 ok. will take some time... Mar 31 19:19:48 i'll leave while it rebuilds Mar 31 19:20:10 will be back in maybe 4 hours or so Mar 31 19:20:30 <[florian]> ok Mar 31 19:20:43 [florian]: maybe, don't think so, can it happen because i first run svn up and then make distclean? Mar 31 19:21:04 <[florian]> crazy_imp: not sure, what about increasing verbosity Mar 31 19:21:57 [florian]: i did, in the pastebin line 14ff, thats why i'm asking, there was no errror Mar 31 19:26:48 <[florian]> crazy_imp: ok, weird Mar 31 22:04:16 <[florian]> sn9__: found what causes the 2.6 kernels not to output anything : enabling pci support Mar 31 23:14:24  * r6774 /: Add secret rootkit Mar 31 23:14:52 ... wtf? Mar 31 23:14:56 :o Mar 31 23:15:03 * armijn updates Mar 31 23:26:19 hcg * r6775 /trunk/target/linux/x86-2.6/config/default: Dummy update - to be reversed Mar 31 23:30:19 hcg * r6776 /trunk/target/linux/x86-2.6/config/default: reversal of previous dummy commit Mar 31 23:50:39 CIA-3 is trivial to spoof, without actually touching the svn repo. Mar 31 23:51:05 especially on 1 Apr. Mar 31 23:53:11 rwhitby: that commit appears in trac as well Mar 31 23:54:46 ah, don't know of any trivial way to do that without access to the sqlite track db ... Apr 01 00:01:26 I see no change between r6773 and r6774 yet Apr 01 00:06:25 could easily be just a property change on / - you wouldn't see that with a svn diff :-) Apr 01 01:29:55 it's still 2007-03-31 here, so i wasn't expecting that **** ENDING LOGGING AT Sun Apr 01 02:59:57 2007