01:42.51 | *** join/#uclibc ambroseL (~bjb@67.71.251.210) |
03:40.13 | *** join/#uclibc robert_ (~chatzilla@220.248.128.10) |
03:40.34 | *** part/#uclibc robert_ (~chatzilla@220.248.128.10) |
03:45.53 | *** join/#uclibc sjhill (~sjhill@spvl-02-110.dsl.netins.net) |
04:16.14 | *** join/#uclibc divan (~divan@83.170.242.206) |
04:45.55 | *** join/#uclibc _divan (~divan@83.170.242.206) |
05:28.08 | *** join/#uclibc jenna (~cherryRed@209.8.233.240) |
05:28.37 | *** part/#uclibc jenna (~cherryRed@209.8.233.240) |
05:37.26 | *** join/#uclibc ashes (~ashes@modemcable202.48-130-66.mc.videotron.ca) |
06:13.42 | *** join/#uclibc provolone (~root@70-32-14-201.frdrmd.adelphia.net) |
06:18.40 | provolone | how do I configure busybox to find uclibc ? how do I build mipsel-uclibc-gcc ? I have been using mipsel-linux-gcc. |
06:21.58 | ashes | make CC=mipsel-uclibc-gcc menuconfig ? |
06:22.03 | ashes | i guessing |
06:27.20 | provolone | I was just asking if it makes a difference and how everyone else is getting that file. I am having another problem related to "mipsel-linux/bin/ld: cannot find -lc" and I was trying to back track and see where I have done things differently than everyone else |
06:46.06 | provolone | what if I want to supply the kernel headers directly with buildroot instead of downloading them |
06:46.50 | *** join/#uclibc simmo (~simmo@adsl-18-27.swiftdsl.com.au) |
07:13.23 | *** join/#uclibc psaksa (~pate@ip212-226-134-137.adsl.kpnqwest.fi) |
08:33.35 | ashes | um |
08:33.51 | ashes | the kernel doesn't seem to get stripped very well |
08:33.52 | *** join/#uclibc jenna (~cherryRed@209.8.233.234) |
08:34.06 | ashes | strip -R .comment -R .note vmlinux |
08:34.10 | *** part/#uclibc jenna (~cherryRed@209.8.233.234) |
08:34.22 | ashes | that removed like 700KB |
08:41.00 | *** join/#uclibc psaksa (~pate@ip212-226-134-137.adsl.kpnqwest.fi) |
08:52.01 | *** join/#uclibc Jena (~cherryRed@209.8.233.234) |
08:52.41 | *** part/#uclibc Jena (~cherryRed@209.8.233.234) |
09:13.47 | *** join/#uclibc GyrosGei1r (~geier@honey.kvedulv.de) |
09:29.55 | *** join/#uclibc Jenna (~cherryRed@209.8.233.237) |
09:30.07 | *** part/#uclibc Jenna (~cherryRed@209.8.233.237) |
10:06.37 | *** join/#uclibc Md (md@md.staff.freenode) |
10:07.57 | Md | I need to compile an uclibc (cross)toolchain but I cannot use buildroot (because I must use the patched to death tarball provided by AXis), can somebody point me to some instructions? |
10:24.28 | *** join/#uclibc ralf (~matteo@openjlab.org) |
11:20.35 | *** join/#uclibc vrm (~vrm@9.55.101-84.rev.gaoland.net) |
11:21.15 | *** join/#uclibc Jenna (~cherryRed@209.8.233.227) |
11:21.31 | *** part/#uclibc Jenna (~cherryRed@209.8.233.227) |
12:14.44 | Md | I finally managed to create an almost-working toolchain... the last issue is that binaries use /lib/ld.so.1 instead of the proper /lib/ld-uClibc-0.9.27.so. how can I fix this? |
12:35.57 | GyrosGeier | Md, -dynamic-linker |
12:36.14 | GyrosGeier | Md, your specs file can set that |
12:38.35 | Md | but how can I build the toolchain correctly in the first place? |
12:43.12 | GyrosGeier | Md, patch the file that generates the specs file |
12:43.41 | GyrosGeier | Md, at least that's what I did to make it all work on the picotux |
12:43.50 | GyrosGeier | Md, and what I will submit to CVS as soon as it has been forward-patched to CVS |
12:44.35 | GyrosGeier | Md, (and which is what led me to accept armeb-unknown-uclinux-uclibc as a sane host triplet) |
13:02.24 | Md | GyrosGeier: it did not work... I added %{!dynamic-linker:-dynamic-linker /lib/ld-uClibc.so.0}} to the "*link:" section of the specs file but looks like it's being ignored. do I need to put it somewhere else? |
13:02.58 | GyrosGeier | Md, no, that should work |
13:03.44 | Md | I checked with strace, and the file is being read. but cris-axis-linux-gnu-gcc -dumpspecs does not show my change, so I'm a bit puzzled... |
13:04.19 | GyrosGeier | Md, dumpspecs dumps the default specs, not the current ones AFAIK |
13:04.50 | Md | yes, I see that it's not reading any file |
13:05.47 | GyrosGeier | Md, do you really close the single opening brace with two closing braces? |
13:06.25 | Md | oops... yes, I pasted it from another specs file :-) |
13:06.45 | Md | this time it worked, thank you |
13:09.38 | *** join/#uclibc vrm_ (~vrm@204.55.101-84.rev.gaoland.net) |
13:19.55 | *** join/#uclibc vrm__ (~vrm@81.55.101-84.rev.gaoland.net) |
13:36.47 | *** join/#uclibc mr_claus (random@p5499006A.dip0.t-ipconnect.de) |
13:52.35 | GyrosGeier | Hrm |
13:52.47 | GyrosGeier | Can I always expect a PHDR at offset 0x34? |
13:53.01 | GyrosGeier | in a loaded ELF image, that is. |
15:01.48 | *** join/#uclibc ambroseL (~bjb@67.71.251.210) |
16:04.21 | *** join/#uclibc rfs (~rsiemsen@gw-rossvideo.storm.ca) |
16:44.41 | solar | dumpelf {/usr,}/{s,}bin/* | grep "Program Header #0" | sort -u |
16:44.59 | solar | and did not find a single case where the initial phdr was not at 0x34 |
17:10.59 | *** join/#uclibc JockeHome (~JockeHome@84.217.10.128) |
17:22.57 | *** join/#uclibc SpanKYx (~UserBah@pool-151-203-10-222.bos.east.verizon.net) |
17:32.42 | provolone | in buildroot how can I specify a different kernel header ? |
17:35.01 | *** join/#uclibc JockeHome (~JockeHome@84.217.89.96) |
17:43.58 | *** join/#uclibc ChanServ (ChanServ@services.) |
17:43.58 | *** mode/#uclibc [+o ChanServ] by irc.freenode.net |
17:45.45 | provolone | anyone ? I have a different kernel for my target than what is in the buildroot menconfig. How can I change to the correct one like in the uclibc config |
17:55.44 | rfs | provolone: see toolchain/kernel-headers directory. |
17:55.57 | rfs | in particular the Config.in and kernel-headers.mk files. |
17:57.34 | solar | GyrosGeie: still here? |
17:57.36 | solar | nope |
18:12.25 | provolone | cant I just use uclibc without buildroot ? |
18:12.48 | provolone | this buildroot is overbloated |
18:14.21 | provolone | it has a menuconfig but then it drops me to the annoying prompt system for uclibc config. uclibc menuconfig accually asks for the kernel header location |
18:16.12 | provolone | this is a poor ui |
18:18.36 | provolone | and it makes me frustrated |
18:20.17 | rfs | provolone: go into uclibc build directory and run make menuconfig there... |
18:20.30 | provolone | right and copy the config |
18:22.34 | provolone | I think im going for a bike ride first. |
18:28.29 | rfs | ah, priorities, good ;) |
19:06.55 | *** join/#uclibc ashes (~ashes@modemcable202.48-130-66.mc.videotron.ca) |
19:24.41 | provolone | I wonder why on earth I should copmpile yet another gcc cross compilier when I have one that already works. I wonder why I should download one automatically with buildroot. I wonder why I would want to waste disk space. I wonder why busybox didnt find all of the uclibc libraries. I wonder why I am forced to use this 'all inclusive' trying to be everything to everyone' bloat package of buildroot just so I can use uclibc |
19:26.12 | Md | talking about this... I know /why/ I should build a new cross compiler, but I can't (because I need to use the hacked gcc tree from AXis). how can I use the system toolchain instead of building a new one? |
19:38.22 | *** join/#uclibc mallok (~mallok20@85.137.30.130) |
19:59.06 | solar | symlinks guys :) |
20:03.43 | Md | solar: ok. where? |
20:19.50 | *** join/#uclibc GyrosGeier (~geier@port-212-202-79-27.dynamic.qsc.de) |
20:41.59 | solar | well assuming your attempting to make use of buildroot |
20:42.22 | solar | you would symlink build_$ARCH etc. |
20:42.37 | solar | but if you already have a toolchain built I'm confused why one would need buildroot |
20:42.40 | ashes | i combined the disc encryption howto, the encrypted root filesystem howto, the loop-aes readme, and made 8 encrypted partitions, including a randomly keyed swap, encrypted partition table, and gnupg encrypted disc key, in a 400 lines, 2300 words, of notes |
20:43.16 | solar | thats alot of encryption. |
20:43.17 | ashes | and im adding squashfs and lzma floppies |
20:43.33 | solar | if your rootfs is encrypted does that mean you must enter a password in the bootup process? |
20:43.45 | ashes | yes |
20:44.03 | solar | ahh so not really ideal for embedded things then |
20:44.47 | ashes | depends how often you reboot |
20:44.47 | Md | solar: I need it at at least to build uclibc |
20:44.51 | Md | but it's useful for busybox and other stuff too |
20:45.03 | solar | what arch are you building for? |
20:45.06 | Md | CRIS |
20:45.17 | solar | oh thats right you said axis already |
20:45.47 | ashes | and then |
20:46.04 | ashes | im running grsecurity on it with all the options enabled |
20:46.19 | solar | Md: I thought axis pushed most/all of the patches into uclibc/buildroot |
20:47.15 | Md | solar: I don't know about uclibc, but the official binutils and gcc are patched to death and I do not want to risk trying the official releases |
20:50.22 | solar | axis sent erik the cris development board as far as I know. I also see an axis dev active on the uclibc ml from time to. But being that Erik has had one for a while should mean that everything should already work |
20:52.16 | solar | but if you really want to trick buildroot into it. Your going to have to hack on the build system itself |
20:52.22 | solar | or make alot of symlinks |
20:52.40 | solar | and touch a few files (.compiled|.patched|etc..) |
20:52.57 | solar | in the end I can't really see that being faster than using buildroot itself. |
20:53.43 | solar | and if patches are indeed needed to resupport cris then probably best imo todo the trial and error thing with building so it gets done right |
20:54.10 | Md | me neither... I'd like to play with my axis board but the official toolchain is very depressing |
20:55.10 | solar | yes. |
20:55.11 | Md | I manually built my own to have decent uclibc support (the official toolchain only provides a gcc wrapper), but I'd rather work improving some already established project than reinvent my "build world" scripts from scratch |
20:55.30 | solar | a while ago zynot forked off from gentoo and added support for cris |
20:56.00 | Md | zynot? |
20:56.49 | solar | yeah.. |
20:56.50 | Md | found it... but I'm not really a gentoo person |
20:57.01 | solar | I know your deb |
22:09.35 | Md | solar: FYI, I see that a recent uclibc snapshot has all the patches in the cris tarball (which is based on 0.9.27), so I would "only" need to make the toolchain fit |
22:18.19 | solar | thats pretty good news |
23:18.19 | ashes | i got a usb drive :) |
23:18.42 | ashes | i can boot this thing |
23:18.47 | ashes | i hope |
23:19.00 | ashes | there's a penguin on the box |
23:48.42 | *** join/#uclibc ralf`away (~matteo@openjlab.org) |