[PATCH] Fix for bug #9706 - Parameter is incorrect on Android

[Jeremy Allison]

On Tue, Mar 19, 2013 at 05:15:34PM +0100, Stefan (metze) Metzmacher wrote:
> Am 19.03.2013 00:22, schrieb Jeremy Allison:
> > On Mon, Mar 18, 2013 at 02:44:26PM -0700, Jeremy Allison wrote:
> >>
> >>> I've attached my current patches, but they're not complete
> >>> and fail make test. I'll try to fix this tomorrow.
> >>
> >> I'll also update your patches and re-run tests.
> > 
> > Ok, here is a set for master that passes make test,
> > and includes the fixes for the test suite you mentioned
> > (correct negotiation and checking of unix extensions,
> > etc.).
> > 
> > It also includes the replacement of the announcing of
> > CAP_LARGE_READX | CAP_LARGE_WRITEX in the sessionsetupX
> > as I still maintain that's needed :-).
> > 
> > Please review. At the very least this should help you
> > finish your modified patches also.
> > 
> > Once we've finalized these changes I'll back-port to
> > 4.0.x and upload to the bug report.
> 
> here's my current patchset it didn't pass a full make test yet.

Ok, this looks like we're converging on a common patchset.

I noticed you added a different constant name when re-adding the
CAP_LARGE_READX | CAP_LARGE_WRITEX but that's ok by me,
so long as the functionality gets re-added :-).

Remember, it's not just for Samba servers, there are other
servers out there (the MacOSX one, EMC, NetApp etc.) and
unless we test against *all* of them then it's much safer
to keep doing what we were doing which is sending the
CAP_LARGE_READX | CAP_LARGE_WRITEX bits on the sessionsetupX.

I also noticed you changed my fixes to create_outbuf() to
a more invasive change than I had, but as it achieves the
same effect I'm not going to moan too much if it makes it
clearer for you :-).

I've been testing with make test TESTS=samba3.smbtorture_s3
so not doing a "full" make test. I'll do so today with your
changes and see if I get faults.

Cheers,

	Jeremy.