Problems with profiling

[Richard Sharpe]

Hi,

Apart from the fact that profiling is not as cleanly implemented as it 
could be, there are problems of the following sort:

1. Some SMB ops call VFS ops. The VFS ops were not using the _NESTED 
versions of the profiling calls.

2. Some trans2 calls do other functions, that use the _NESTED ops, which 
then call VFS ops that were not using their own nested profiling macros.

Should yet another hack be added to this festering sore, of should a more 
generalized structure be added?

Regards
-----
Richard Sharpe, rsharpe at ns.aus.com, rsharpe at samba.org, 
sharpe at ethereal.com