tdb_seqnum trick for brlocks?

[Jeremy Allison]

On Sun, Feb 18, 2007 at 12:15:58AM +0100, Volker Lendecke wrote:
> On Sat, Feb 17, 2007 at 06:13:22PM +0100, Volker Lendecke wrote:
> > Attached find a patch that applies the tdb_seqnum trick
> 
> The attached patch is broken, I did not expect that
> brl_get_locks_internal() would actually return a
> write-locked record when called with read_only==True. Fixed
> that bug locally, but for the discussion it's not relevant I
> think.

Can I see the updated patch ?

In this :

" I did not expect that brl_get_locks_internal() would actually return a
write-locked record when called with read_only==True."

are you talking about the !fsp->lockdb_clean case where
we explicitly change the read_only flag ?

More info please (I'm trying to catch up after NY this
week :-).

Jeremy.