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

[Jeremy Allison]

On Tue, May 05, 2020 at 11:39:14AM -0600, Jens Axboe wrote:
> > 
> > I'll try and see if I can get an arch binary build that has the
> > vfs_io_uring module and reproduce.
> 
> Got that done, and I can now mount it on Linux. Been trying pretty
> hard to trigger any corruptions on reads, but it works for me. Checked
> that we see short reads, and we do, and that it handles it just fine.
> So pretty blank right now on what this could be.
> 
> FWIW, I'm mounting on Linux as:
> 
> # mount -t cifs -o ro,guest //arch/data /smb

The reporter claims this only happens with
a Windows client, as the Linux and Samba libsmb
clients don't pipeline the reads in the same
way for multiple files the same way that Windows
does.

Here is the bug report with all the details I've
managed to get out of the original reporter:

https://bugzilla.samba.org/show_bug.cgi?id=14361

I also can't reproduce the data corruption, even
from a Windows client although I've only tried
so far on my home Ubuntu 19.04 kernel 5.3.0-51-generic #44-Ubuntu SMP
box.

I tried running 2 powershell processes on the
Windows client and doing a copy in both of them
to ensure we are opening multiple files simultaneously
and going through the io_uring module (which I
saw happening by adding debug statements) but
I never see corruption.

I'm planning to try running against a Fedora32
VM next, as that's case the reporter claims will
reproduce the issue.

Thanks for helping me look at this !

Jeremy.