| 00:07.28 | *** join/#uclibc hiyuh (n=hiyuh@ZQ177022.ppp.dion.ne.jp) |
| 00:23.34 | landley | tharvey: so you can only run pthread_create from init? |
| 00:23.53 | landley | Last I checked, other processes were created by fork()... |
| 00:32.23 | tharvey | landley, if a child of a fork calls pthread_create it never returns (the thread is created but it never returns) |
| 00:32.45 | landley | Threading in uClibc is a bit screwed up. |
| 00:33.01 | landley | They keep meaning to do NPTL threads, but nobody's merging the three different implementations together, last I checked... |
| 00:33.03 | tharvey | I ran into this with stunnel - which by default forks to the background - when a connection is received it creates a thread which would hang. Run stunnel in the background and its fine |
| 00:33.30 | tharvey | thats what I was wondering... if pthreads in uclibc is ok - this is the first issue I've run into but I'm new to uclibc |
| 00:33.40 | landley | I don't use threads in my own programs. |
| 00:33.43 | landley | I really haven't tested them. |
| 00:33.54 | landley | I did a threaded "hello world" to check and it worked, that's as far as I got. |
| 00:33.55 | tharvey | there are several options in buildroot... should I be trying the others? (like NPTL?) |
| 00:34.03 | landley | I don't use buildroot. |
| 00:34.09 | landley | buildroot != uClibc. |
| 00:34.23 | tharvey | ya... do a fork in your hello world and and create the thread from the child and it will hang |
| 00:34.38 | tharvey | ya I know, but this is the closest thing to a channel buildroot has |
| 00:36.44 | tharvey | the toolchain configuration of buildroot allows selection of thread library: libpthreads, linuxthreads (default), NPTL - is linuxthreads not recommended? |
| 00:37.41 | tharvey | sorry meant to say 'linuxthreads' vs 'linuxthreads (stable/old)' vs 'nptl' (not libpthreads) |
| 00:37.47 | tharvey | which do you use with uclibc? |
| 00:54.41 | *** join/#uclibc nabcore (n=nabcore@85-211-244-118.dsl.pipex.com) |
| 00:54.52 | nabcore | Can anyone see why busybox's /sbin/init is taking so long to start here: https://dev.openwrt.org/ticket/2400#comment:4 |
| 01:02.20 | nabcore | I'm assuming the first thing done by /sbin/init is to read /etc/inittab and then (given sysinit points to rcS), execute rcS? |
| 01:16.31 | *** part/#uclibc nabcore (n=nabcore@85-211-244-118.dsl.pipex.com) |
| 02:05.20 | *** join/#uclibc wrobbie (n=rob@cm22.kappa84.maxonline.com.sg) |
| 02:24.30 | *** join/#uclibc blindvt_ (n=bf@M985P001.adsl.highway.telekom.at) |
| 04:27.22 | Xride | blindvt_: s/look-device/loop-device/ |
| 04:29.27 | Xride | blindvt_: but thanks (8 |
| 05:41.41 | *** join/#uclibc kernelzitowz (i=kernelpa@189.2.18.91) |
| 05:48.39 | neli | blindvt_: good to know, I thought it perhaps was policy or so |
| 06:15.17 | *** join/#uclibc Xride (n=xride@0x5550f21e.adsl.cybercity.dk) |
| 06:17.03 | Xride | . |
| 06:17.20 | Xride | blindvt_: how do i get it to build the bzImage? |
| 06:21.22 | *** join/#uclibc doc2 (n=doc@p57B85725.dip.t-dialin.net) |
| 06:24.09 | *** join/#uclibc blindvt__ (n=bf@M922P004.adsl.highway.telekom.at) |
| 06:32.54 | *** join/#uclibc Sgt-Donan (n=Loutre@mto30-1-82-242-53-8.fbx.proxad.net) |
| 06:36.20 | *** join/#uclibc Gerrath (n=Gerrath_@unaffiliated/gerrath) |
| 06:56.26 | *** join/#uclibc wrobbie (n=rob@cm22.kappa84.maxonline.com.sg) |
| 07:02.45 | *** join/#uclibc Redhatter-LM2E (n=stuartl@gentoo/developer/redhatter) |
| 07:54.22 | *** join/#uclibc hw (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 09:01.20 | *** join/#uclibc CIA-10 (n=CIA@208.69.182.149) |
| 10:14.23 | blindvt` | Xride, "Kernel"->select linux and specify bzImage there (along your config if you don't want to use my i386 stub config |
| 10:25.34 | CIA-10 | 03vda * r20078 10busybox/miscutils/hdparm.c: (log message trimmed) |
| 10:25.34 | CIA-10 | hdparm: get rid of static variables |
| 10:25.34 | CIA-10 | process_dev 5309 5311 +2 |
| 10:28.04 | CIA-10 | 03vda * r20079 10busybox/miscutils/hdparm.c: (log message trimmed) |
| 10:28.04 | CIA-10 | hdparm: simplify timing measurement (it wa the last last user |
| 10:28.04 | CIA-10 | of setitimer in the tree). |
| 10:29.18 | CIA-10 | 03vda * r20080 10busybox/ (5 files in 4 dirs): (log message trimmed) |
| 10:29.18 | CIA-10 | assorted static vars removal |
| 10:29.18 | CIA-10 | function old new delta |
| 10:35.02 | *** join/#uclibc Jacmet (n=peko@vps.korsgaard.com) |
| 11:07.21 | Xride | is there a make *-menuconfig for the kernel in buildroot? |
| 11:21.49 | CIA-10 | 03vda * r20081 10busybox/miscutils/hdparm.c: (log message trimmed) |
| 11:21.50 | CIA-10 | hdparm: reduce possibility of numeric overflow in -T. |
| 11:21.50 | CIA-10 | do_time 386 410 +24 |
| 11:37.39 | CIA-10 | 03vda * r20082 10busybox/selinux/setfiles.c: trivial fix in setfiles. spotted by Yuichi Nakamura <ynakam@hitachisoft.jp> |
| 11:37.58 | *** join/#uclibc Baby (n=miry@195.37.62.208) |
| 11:40.58 | *** join/#uclibc antab (n=antab@postur.valka.is) |
| 12:41.35 | *** join/#uclibc whirm (n=whirm@173.pool85-58-82.dynamic.orange.es) |
| 12:55.27 | whirm | hi all! |
| 13:06.27 | *** join/#uclibc prpplague (n=DaveAnde@mail.americanmicrosystems.com) |
| 13:16.53 | *** join/#uclibc emil|k (n=emil|k@e179237108.adsl.alicedsl.de) |
| 13:29.53 | *** join/#uclibc hiyuh (n=hiyuh@KD125054017176.ppp-bb.dion.ne.jp) |
| 13:33.41 | *** join/#uclibc howtoo (n=howtoo@195.159.62.66) |
| 13:38.10 | CIA-10 | 03vda * r20083 10busybox/miscutils/hdparm.c: hdparm: more robust (re overflows) code. +38 bytes. |
| 13:42.40 | CIA-10 | 03vda * r20084 10busybox/miscutils/hdparm.c: hdparm: paranoia-induced fix against theoretical div-by-0 case |
| 14:25.07 | *** join/#uclibc blindvt_ (n=bf@M925P029.adsl.highway.telekom.at) |
| 14:33.21 | *** join/#uclibc tchan (n=tchan@lunar-linux/developer/tchan) |
| 15:52.20 | *** join/#uclibc landley (n=landley@cpe-70-112-19-53.austin.res.rr.com) |
| 15:53.15 | *** join/#uclibc svolpe_gerrath (n=Gerrath_@unaffiliated/gerrath) |
| 16:18.17 | CIA-10 | 03ulf * r20085 10buildroot/ (Makefile target/linux/Makefile.in toolchain/uClibc/uclibc.mk): Add administration for linux,uclibc and main |
| 16:26.42 | CIA-10 | 03aldot * r20086 10buildroot/toolchain/binutils/ (2 files in 2 dirs): - fix texinfo version check. The next release will most likely supercede these |
| 16:27.33 | *** join/#uclibc doc2 (i=doc@p54AFAF16.dip0.t-ipconnect.de) |
| 16:58.27 | *** join/#uclibc PiterPunk (i=3354@slackware.com) |
| 16:59.32 | PiterPunk | Hi, i am using mdev to dynamic device creation in initrd; it works fine with hda/sda devices, but i have problems with cciss devices. |
| 17:00.04 | PiterPunk | They are expected to be in /dev/ida/cNdNpN (N is a number) but mdev is creating /dev/ida!cNdNpN. |
| 17:00.14 | PiterPunk | Any idea about how to fix that? |
| 17:00.29 | PiterPunk | Using busybox-1.4.2 |
| 17:03.32 | solar | sure have mdev make symlinks |
| 17:05.00 | PiterPunk | solar: need a newer busybox to do that? |
| 17:05.27 | PiterPunk | solar: I am puzzled because that "!" in device name. |
| 17:06.08 | PiterPunk | mdev canīt handle the ida!cNdNpN and make ida/cNdNpN? |
| 17:08.13 | solar | the driver is telling /sys/block/* that what it's real name is. |
| 17:08.35 | solar | newer busybox. Not really. it works with pretty much every mdev release |
| 17:08.44 | solar | see the mdev.wtf |
| 17:09.05 | PiterPunk | solar: But udev with the same information makes ida/cNdNpN. |
| 17:09.11 | PiterPunk | Let's see |
| 17:17.13 | CIA-10 | 03aldot * r20087 10buildroot/toolchain/kernel-headers/ (2 files in 2 dirs): - bump version |
| 17:17.18 | PiterPunk | solar: Hmmm looking in mdev.c it doesn't seems to support subdirectory creation. |
| 17:25.52 | ncopa | mdev does not create dirs |
| 17:25.54 | ncopa | true |
| 17:25.57 | ncopa | i have a wrapperscript |
| 17:26.42 | PiterPunk | ncopa: Can you show that script? |
| 17:27.43 | ncopa | http://rafb.net/p/PgmClF77.html |
| 17:28.00 | ncopa | for /dev/hd* |
| 17:28.39 | ncopa | sorry that was not what you wanted |
| 17:28.52 | PiterPunk | -:( |
| 17:29.27 | ncopa | http://rafb.net/p/juDidv82.html |
| 17:29.29 | ncopa | try that |
| 17:29.57 | ncopa | creates sudirs for alsa and zap devices |
| 17:31.14 | ncopa | wanna see how I load usb devices? |
| 17:31.20 | ncopa | autoload drivers to webcams etc |
| 17:31.27 | ncopa | with mdev wrapper script |
| 17:31.40 | PiterPunk | Yes, of course |
| 17:32.12 | ncopa | http://rafb.net/p/rxhwsB86.html |
| 17:34.15 | PiterPunk | Thanks! This helps a lot -:) |
| 17:35.03 | *** join/#uclibc Gerrath (n=Gerrath_@unaffiliated/gerrath) |
| 17:38.16 | ncopa | in case you wanna try on a livecd |
| 17:38.42 | ncopa | http://wiki.alpinelinux.org/ |
| 17:39.12 | ncopa | busybox/uclibc power in real life :) |
| 17:39.26 | ncopa | actually, the wiki itself runs uclibc/busybox |
| 17:48.36 | *** join/#uclibc whirm (n=whirm@173.pool85-58-82.dynamic.orange.es) |
| 18:07.07 | *** join/#uclibc whirm (n=whirm@173.pool85-58-82.dynamic.orange.es) |
| 18:11.17 | *** join/#uclibc ashes (n=ashes@2001:5c0:8fff:ffff:0:0:0:1) |
| 18:21.26 | *** join/#uclibc svolpe_gerrath (n=Gerrath_@unaffiliated/gerrath) |
| 18:21.47 | *** part/#uclibc ncopa (n=ncopa@213.234.126.134) |
| 18:21.54 | *** join/#uclibc JockeHome (n=JockeHom@84-217-2-81.tn.glocalnet.net) |
| 18:30.04 | *** join/#uclibc hiyuh (n=hiyuh@KD125054017176.ppp-bb.dion.ne.jp) |
| 18:39.07 | *** join/#uclibc Chase (n=cndougla@cpe-75-185-70-183.columbus.res.rr.com) |
| 18:39.34 | Chase | I'm getting a SIGKILL on a pthreads program, and I'm wondering how I can go about debugging it? |
| 18:39.51 | Chase | the program is a test which just creates a bunch of threads and then attempts to join them |
| 18:40.04 | *** join/#uclibc likewhoa (n=y0@ool-18b9285a.dyn.optonline.net) |
| 18:40.36 | Chase | when it joins, I've printed out steps through pthread_join in uClibc, but that doesn't tell me much because the SIGKILL is printed out at different times during different runs |
| 18:41.02 | Chase | is there any way to debug a SIGKILL? |
| 19:01.28 | *** join/#uclibc antab (n=antab@buttbuddies.cannabis.is) |
| 19:19.59 | blindvt_ | Chase, i'd start with building a threadlib with debugging support to see if you can reproduce it under a debugger, perhaps that would help. And you can trap signals fine, yes |
| 19:25.12 | Chase | I've tried gdb'ing it, but I get |
| 19:25.12 | Chase | dlopen failed on 'libthread_db.so.1' - (null) |
| 19:25.12 | Chase | GDB will not be able to debug pthreads. |
| 19:30.54 | *** join/#uclibc Gerrath (n=Gerrath_@unaffiliated/gerrath) |
| 19:43.28 | CIA-10 | 03ulf * r20088 10buildroot/ (Makefile target/linux/Makefile.in toolchain/uClibc/uclibc.mk): Cleanup cross toolchain admin, split uclibc-menuconfig into same + uclibc-update to avoid overwriting board support file |
| 19:46.59 | CIA-10 | 03ulf * r20089 10buildroot/Makefile: Add some more host tools |
| 19:49.15 | CIA-10 | 03ulf * r20090 10buildroot/package/Makefile.in: Sort HOST/TARGET_CONFIGURE_OPTS, add some more, needed for gcc-4.1.2 |
| 19:51.08 | CIA-10 | 03ulf * r20091 10buildroot/target/device/ (Atmel/toolchain/avr32/gcc-4.1.2-avr32-2.2/ Toolchain.in): Update AVR external gcc tarball to 2.2 |
| 19:53.14 | CIA-10 | 03ulf * r20092 10buildroot/toolchain/gcc/gcc-uclibc-3.x.mk: Add bells and whistles to allow gcc-4.1.2-AVR32 buildprocess to progress further, not home yet though |
| 19:56.19 | CIA-10 | 03ulf * r20093 10buildroot/toolchain/br2_autoreconf: add autoreconf script that can handle gcc (aclocal needs to handle -I switch) |
| 20:10.25 | CIA-10 | 03aldot * r20094 10buildroot/ (Makefile package/Makefile.in toolchain/Makefile.in): |
| 20:10.25 | CIA-10 | - add FC, remove some duplicate entries, cleanup |
| 20:10.25 | CIA-10 | - fixup after Ulf: CXX doesn't take CFLAGS but CXXFLAGS |
| 20:14.38 | CIA-10 | 03vapier * r20095 10uClibc/ldso/ldso/dl-hash.c: (log message trimmed) |
| 20:14.38 | CIA-10 | Make sure that the DSO has an hash table into its elf, otherwise skip |
| 20:14.38 | CIA-10 | the lookup over next DSO. |
| 20:39.21 | CIA-10 | 03vapier * r20096 10uClibc/libc/inet/resolv.c: cleanup code style |
| 20:42.27 | CIA-10 | 03vapier * r20097 10uClibc/libc/inet/resolv.c: do not check the pointer before freeing it since free(NULL) works just fine |
| 20:52.11 | CIA-10 | 03aldot * r20098 10buildroot/TODO: |
| 20:52.11 | CIA-10 | - mark the autotools.in part as taken |
| 20:52.11 | CIA-10 | - fix a typo |
| 20:52.31 | *** join/#uclibc emil|k`work_ (n=emil@pD950F2F1.dip0.t-ipconnect.de) |
| 21:17.58 | CIA-10 | 03vapier * r20099 10uClibc/libc/inet/resolv.c: (log message trimmed) |
| 21:17.58 | CIA-10 | This patch will fix a problem when the same host is defined with both IPv4 and |
| 21:17.58 | CIA-10 | IPv6 entries in /etc/hosts. Previous only the first of these host would work, |
| 21:44.36 | CIA-10 | 03ulf * r20100 10buildroot/toolchain/uClibc/uclibc.mk: No reliance on dirs in uClibc |
| 22:13.11 | *** join/#uclibc CIA-10 (n=CIA@208.69.182.149) |
| 22:14.01 | CIA-10 | 03vda * r20104 10busybox/libbb/ (getopt32.c procps.c): more "if (p) free(p)" fixes; fix a typo in comment |
| 22:15.20 | CIA-10 | 03aldot * r20105 10buildroot/package/Makefile.in: - cleanup configure options |
| 22:16.48 | CIA-10 | 03ulf * r20106 10buildroot/toolchain/uClibc/uclibc.mk: Removed uclibc source dependency on host-sed |
| 22:20.33 | CIA-10 | 03aldot * r20107 10buildroot/target/device/Atmel/ (5 files): - rename to match Config.in.foo naming convention |
| 22:26.01 | *** join/#uclibc blindvt__ (n=bf@M937P031.adsl.highway.telekom.at) |
| 22:36.26 | CIA-10 | 03vda * r20108 10busybox/networking/httpd.c: httpd: trivial simplification |
| 23:46.00 | CIA-10 | 03vda * r20109 10busybox/archival/libunarchive/decompress_bunzip2.c: (log message trimmed) |
| 23:46.00 | CIA-10 | bunzip2: trim verbose messages |
| 23:46.00 | CIA-10 | <PROTECTED> |