2.0.6: smbmount still loses connection after awhile to a NT4 server

[not]

hello,

I know this topic has already been discused, and I remember that the final
word was the this issue would be fixed in 2.0.6.

I'm running a fresh install of Mandrake 7.0  with kernel 
2.2.14-15mdksecure
with samba 2.0.6 installed.

I can smbmount a share on my NT4 server without problem.
the connection is establised and everything works fine:

Proto Recv-Q Send-Q Local Address           Foreign Address         State
tcp        0      0 box:1031             123.345.67.:netbios-ssn 
ESTABLISHED

I go away for a while. When I come back I get this:
tcp        0      0 box:1031             123.345.67.:netbios-ssn   CLOSE

If I then try to do a df smbmount crashes and I get the following output:
df: .: Input/output error

here's an output of strace:

...
7614  close(1020)                       = -1 EBADF (Bad file descriptor)
7614  close(1021)                       = -1 EBADF (Bad file descriptor)
7614  close(1022)                       = -1 EBADF (Bad file descriptor)
7614  close(1023)                       = -1 EBADF (Bad file descriptor)
7614  rt_sigaction(SIGUSR1, {0x804bd74, [USR1], SA_RESTART|0x4000000}, 
NULL, 8) = 0
7614  pause()                           = ? ERESTARTNOHAND (To be 
restarted)
7614  --- SIGUSR1 (User defined signal 1) ---
7614  sigreturn()                       = ? (mask now [])
7614  getpid()                          = 7614
7614  brk(0x80c9000)                    = 0x80c9000
7614  brk(0x80da000)                    = 0x80da000
7614  socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 0
7614  fcntl(0, F_GETFL)                 = 0x2 (flags O_RDWR)
7614  fcntl(0, F_SETFL, O_RDWR|O_NONBLOCK) = 0
7614  --- SIGSEGV (Segmentation fault) ---
7612  <... wait4 resumed> 0xbfffdec8, 0, NULL) = ? ERESTARTSYS (To be 
restarted)
7612  --- SIGTERM (Terminated) ---
7612  munmap(0x124000, 4096)            = 0
7612  _exit(0)                          = ?

So has it been fixed in 2.0.6 or not? Is there another problem?
Any ideas welcome,

Eric