Design change for oplock/open code?

Jeremy Allison jra at samba.org
Fri Apr 26 15:46:30 MDT 2013


On Fri, Apr 26, 2013 at 05:38:50PM +0200, Volker Lendecke wrote:
> 
> Under
> 
> https://git.samba.org/?p=vl/samba.git/.git;a=shortlog;h=refs/heads/defer_open
> 
> find a patchset that I see as complete for upstream. The
> nice thing is the overall diff:
> 
>  17 files changed, 76 insertions(+), 434 deletions(-)
> 
> Please review & push if okay.
> 
> Thanks,

Great work ! Pushed, thanks !

I made 2 changes. One was to add a comment in
include/vfs.h saying the VFS interface number
wasn't yet being changed as we're not releasing
and noting the fields within the files_struct
that got deleted. The second was to up the debug
level on an error return from dbwrap_record_watch_recv()
from 1 to 5 as it can commonly get NT_STATUS_IO_TIMEOUT if
there really is a sharing violation.

Maybe you can add a followup patch that removes
the code under the '#if 0's you added, as it will
tidy up the files.

On a different note, if you could take a look
at the change :

[PATCH] Fix bug #9794 - opening/editing or copying MS files causes a core dump with invalid lock order

I posted yesterday I'd appreciate it. In my
testing it fixes the test case you used to
reproduce the issue (setting create mask to
000) and hopefully will be good enough for
4.0.next, although I'm still working on
my patch to remove the possibility of calling
create_file_fchmod() inside of a file create.

Should have that ready for master next week.

The originator of the bug #9794 complained
it didn't fix things for him, but I think that's
because he has a custom VFS.

Cheers,

	Jeremy.


More information about the samba-technical mailing list