Data Corruption bug with Samba's vfs_iouring and Linux 5.6.7/5.7rc3

[Jeremy Allison]

On Sat, May 09, 2020 at 12:04:55AM +0200, Stefan Metzmacher via samba-technical wrote:
> 
> Do you have some comments on the offset < 0 on write case?

I think the offset < 0 case in MS-FSA 2.1.5.3
is talking about the behavior of the interface
into the NtDLL layer in the Windows kernel,
not the SMB2 layer.

I think at the SMB2 layer all offsets for
read and write are implicitly unsigned.

I'll write some test code though, which
sets the high bit for reads and writes
on a zero length file and see what Windows10
returns.