Been trying to run this on FC3 pulling stuff off a NT4 PDC - it just segfaults on 3.0.21 & 3.0.21a I'm building the RPMS from the tar ball on the host using the makerpms.sh script Reverted to the 3.0.10 issued by Fedora and no segfault. From what I can see (with strace) it segfaults while reading from the socket connecting the PDC - not the first read, but after quite a few gettimeofday()/read() operations. Will post more info if anyone is interested.
On Sun, 2006-01-08 at 22:53 +1100, taso wrote:> Been trying to run this on FC3 pulling stuff off a NT4 PDC - it just segfaults on 3.0.21 & 3.0.21a > I'm building the RPMS from the tar ball on the host using the makerpms.sh script > > Reverted to the 3.0.10 issued by Fedora and no segfault. > > From what I can see (with strace) it segfaults while reading from the socket connecting the PDC - > not the first read, but after quite a few gettimeofday()/read() operations. > > Will post more info if anyone is interested.It shouldn't ever segfault. But what we really need is (at least) a gdb backtrace, and more usefully a valgrind trace. Probably best to file the result in bugzilla. Oh, and watch out that you don't include production passwords in the trace :-) Andrew Bartlett -- Andrew Bartlett samba.org/~abartlet Authentication Developer, Samba Team samba.org Student Network Administrator, Hawker College hawkerc.net -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part Url : lists.samba.org/archive/samba/attachments/20060109/60daf411/attachment.bin
On Sun, Jan 08, 2006 at 10:53:14PM +1100, taso wrote:> Been trying to run this on FC3 pulling stuff off a NT4 PDC - it just > segfaults on 3.0.21 & 3.0.21a > I'm building the RPMS from the tar ball on the host using the makerpms.sh > script > > Reverted to the 3.0.10 issued by Fedora and no segfault. > > From what I can see (with strace) it segfaults while reading from the > socket connecting the PDC - > not the first read, but after quite a few gettimeofday()/read() operations. > > Will post more info if anyone is interested.Can you send in a gdb backtrace - run it under gdb and when it segfaults type "bt" and post the output. Thanks, Jeremy.