root@kavics:~# /usr/sbin/nmbd -D Aborted root@kavics:~# /usr/sbin/smbd -D Aborted Here you can find the strace output: http://frugalware.org/~voroskoi/strace_nmbd.out
You compiled glibc 2.4 yourself didn't you?
No, i'm using frugalware-current package. ftp://ftp.frugalware.org/pub/frugalware/frugalware-current/source/base/glibc There you can find the FrugalBuild. I'm using the i686 version, so no patch have been used. CFLAGS are -march=i686 -O2 -pipe Let me know if you need any additional information!
SuSE 10.1 uses glibc 2.4 without any problems. Can you get a gdb back trace of smbd ? This appears to be the same thing as BUG 3655.
severity should be determined by the developers and not the reporter.
(In reply to comment #3) > SuSE 10.1 uses glibc 2.4 without any problems. Can you get a gdb > back trace of smbd ? This appears to be the same thing as BUG 3655. > (gdb) run Starting program: /usr/sbin/smbd (no debugging symbols found) Program received signal SIGABRT, Aborted. 0xffffe410 in __kernel_vsyscall () (gdb) bt #0 0xffffe410 in __kernel_vsyscall () #1 0xb7b35581 in ?? () #2 0xb7c27ff4 in ?? () #3 0xb7a0c6b0 in ?? () #4 0xbf857af0 in ?? () #5 0xb7b36be9 in ?? () #6 0x00000006 in ?? () #7 0xbf857a64 in ?? () #8 0x00000000 in ?? () So you think that it's a bug in our glibc?
I haven't looked into it very deeply. But that has been the conclusion of others. I've copied Guenther on this. Perhaps he can offer some advice as to why this is not seen on the SuSE glibc 2.4 builds.
Now i've recompiled glibc without -march=i686 -O2 -pipe, but the problem is still there. I've recompile samba as well, but didn't help. I have no idea why does it work on suse, but BUG 3655 shows that there are problems on mandriva too. Have you tried to use samba on suse with a clean (not full of patches) glibc build? Maybe you would be able to reproduce this problem.
I am encountering the same problem. My system has glibc-2.4, gcc-4.1.0 and Linux kernel 2.6.16, all built from source. Here's the output I receive when trying to start the daemons: bash-3.1$ /sbin/nmbd -i -S Netbios nameserver version 3.0.22 started. Copyright Andrew Tridgell and the Samba Team 1992-2006 nmbd: gconv_db.c:232: __gconv_release_step: Assertion `step->__end_fct == ((void *)0)' failed. =============================================================== INTERNAL ERROR: Signal 6 in pid 29554 (3.0.22) Please read the Trouble-Shooting section of the Samba3-HOWTO From: http://www.samba.org/samba/docs/Samba3-HOWTO.pdf =============================================================== PANIC: internal error BACKTRACE: 1 stack frames: #0 /sbin/nmbd(smb_panic2+0x8a) [0x800a446a] Aborted bash-3.1$ /sbin/smbd -i -S smbd version 3.0.22 started. Copyright Andrew Tridgell and the Samba Team 1992-2006 smbd: gconv_db.c:232: __gconv_release_step: Assertion `step->__end_fct == ((void *)0)' failed. =============================================================== INTERNAL ERROR: Signal 6 in pid 29555 (3.0.22) Please read the Trouble-Shooting section of the Samba3-HOWTO From: http://www.samba.org/samba/docs/Samba3-HOWTO.pdf =============================================================== PANIC: internal error BACKTRACE: 1 stack frames: #0 /sbin/smbd(smb_panic2+0x8a) [0x8021451a] Aborted
Same problems here using glibc-2.4, kernel 2.6.16.11 and gcc-4.1. All compiled from sources. The weird thing is if I start smbd and nmbd while running in X, it starts up fine
And the same here on a new, clean LFS with glibc-2.4, Kernel 2.6.16-ck9 and gcc-4.1.0. I`ve tried out samba-3.0.22 and 3.0.23-pre1.
Created attachment 1909 [details] fix against glibc2.4
I've added a patch that works for me. IMO you can close this bug now.
confirmed glibc bug.