[CentOS] smbmount: mount() syscall gone bad ?

Mon Oct 9 15:37:15 UTC 2006
Rodrigo Barbosa <rodrigob at darkover.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I have seen a weird problem in one of my clients. The reported that
when issuing a smbmount command (both client and server are CentOS 4.4),
it would take up to 30 second for the mount to complete.

To make a long history short, by using strace and the source, I noticed
smbmnt is first calling mount() using ascii options for *data. Then
that would timeout, and it would try again using a struct, finally working.

Looking at the source code, it will first try ascii if kernel >= 2.4,
and fallback to struct if it doesn't work. So changing smbmnt.c to
always using the struct first solved my problem. I also noticed that
debian will always go first for the struct aproach.

Anyone seen this before ?

- -- 
Rodrigo Barbosa
"Quid quid Latine dictum sit, altum viditur"
"Be excellent to each other ..." - Bill & Ted (Wyld Stallyns)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFFKmyrpdyWzQ5b5ckRAnq4AJ9Z2ZCe3denIoWFK5ePWF7INHqMjwCgs+9c
bNCRI/9fJ9B1Cg2gGD4hPHM=
=Qa4R
-----END PGP SIGNATURE-----