global_oplock_break: comment and possible bug

[jonathan bright]

possible bug:

the call chain:

  reply_mknew() -> open_file_shared() -> file_new() ->
  attempt_close_oplock_file() -> oplock_break()

results in an abort() if the smbd recently sent
the windows client an exclusive oplock break.


comment:  

the malloc() in oplock_break() doesn't seem necessary,
as global_oplock_break should protect a static variable.
this wouldn't be so bad, but if true the comment is 
misleading.


------
jonathan bright
jonathan.bright at onebox.com
415.820.7322 - voicemail/fax
www.brightconsulting.com