RE: [PATCH v4 2/3] scsi: storvsc: Use vmbus_requestor to generate transaction IDs for VMBus hardening

From: Michael Kelley <mikelley@microsoft.com>
To: Wei Liu <wei.liu@kernel.org>,
	Nathan Chancellor <natechancellor@gmail.com>
Cc: Andres Beltran <t-mabelt@microsoft.com>,
	KY Srinivasan <kys@microsoft.com>,
	Haiyang Zhang <haiyangz@microsoft.com>,
	Stephen Hemminger <sthemmin@microsoft.com>,
	"linux-hyperv@vger.kernel.org" <linux-hyperv@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"parri.andrea@gmail.com" <parri.andrea@gmail.com>,
	Andres Beltran <lkmlabelt@gmail.com>,
	"James E.J. Bottomley" <jejb@linux.ibm.com>,
	"Martin K. Petersen" <martin.petersen@oracle.com>,
	"linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>
Subject: RE: [PATCH v4 2/3] scsi: storvsc: Use vmbus_requestor to generate transaction IDs for VMBus hardening
Date: Fri, 17 Jul 2020 13:54:07 +0000	[thread overview]
Message-ID: <SN6PR2101MB1056F67B43877AC7298D2748D77C0@SN6PR2101MB1056.namprd21.prod.outlook.com> (raw)
In-Reply-To: <20200717104556.ul5s6hmtlerjpi3g@liuwe-devbox-debian-v2>

From: Wei Liu <wei.liu@kernel.org>  Sent: Friday, July 17, 2020 3:46 AM
> On Wed, Jul 08, 2020 at 09:25:12AM +0000, Wei Liu wrote:
> > On Wed, Jul 08, 2020 at 09:21:05AM +0000, Wei Liu wrote:
> > [...]
> > > > If I revert this commit, everything works fine:
> > > >
> > > > PS C:\Users\natec> wsl --shutdown
> > > > PS C:\Users\natec> wsl -d ubuntu -- /bin/bash
> > > > nathan@Ryzen-9-3900X:/mnt/c/Users/natec$ cat /proc/version
> > > > Linux version 5.8.0-rc4-next-20200707-microsoft-standard+ (nathan@Ryzen-9-3900X)
> (gcc (Ubuntu 9.3.0-10ubuntu2) 9.3.0, GNU ld (GNU Binutils for Ubuntu) 2.34) #1 SMP Tue Jul
> 7 16:35:06 MST 2020
> > > > nathan@Ryzen-9-3900X:/mnt/c/Users/natec$ git -C ~/src/linux-next lo -2
> > > > 0ff017dff922 (HEAD -> master) Revert "scsi: storvsc: Use vmbus_requestor to
> generate transaction IDs for VMBus hardening"
> > > > 5b2a702f85b3 (tag: next-20200707, origin/master, origin/HEAD) Add linux-next specific
> files for 20200707
> > > > nathan@Ryzen-9-3900X:/mnt/c/Users/natec$
> > > >
> > > > The kernel was built using the following commands:
> > > >
> > > > $ mkdir -p out/x86_64
> > > >
> > > > $ curl -LSso out/x86_64/.config
> https://raw.githubusercontent.com/microsoft/WSL2-Linux-Kernel/linux-msft-wsl-4.19.y/Microsoft/config-wsl
> > > >
> > > > $ scripts/config --file out/x86_64/.config -d RAID6_PQ_BENCHMARK -e
> NET_9P_VIRTIO
> > > >
> > > > $ make -skj"$(nproc)" O=out/x86_64 olddefconfig bzImage
> > > >
> > > > I don't really know how to get more information than this as WSL seems
> > > > rather opaque but I am happy to provide any information.
> > >
> > > Linux kernel uses Hyper-V's crash reporting facility to spit out
> > > information when it dies. It is said that you can see that information
> > > in the "Event Viewer" program.
> > >
> > > (I've never tried this though -- not using WSL2)
> > >
> >
> > If this doesn't work, another idea is to install a traditional VM on
> > Hyper-V and replace the kernel with your own.
> >
> > With such setup, you should be able to add an emulated serial port to
> > the VM and grab more information.
> 
> Hi Nathan, do you need more help on this?
> 
> MSFT is also working on reproducing this internally.
> 
> We're ~2 weeks away from the next merge window so it would be good if we
> can get to the bottom of this as quickly as possible.
> 

On the Microsoft side we now have a repro of the problem when running
in WSLv2.  The symptoms match exactly what Nathan has reported.  We
will debug it from here.  Thanks for reporting this!

Michael