Re: [PULL 00/32] VFIO updates 2020-10-26 (for QEMU 5.2 soft-freeze)

From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
To: Alex Williamson <alex.williamson@redhat.com>, mszeredi@redhat.com
Cc: "Artem Polyakov" <artemp@nvidia.com>,
	"Peter Maydell" <peter.maydell@linaro.org>,
	"Zhengui li" <lizhengui@huawei.com>,
	"Yan Zhao" <yan.y.zhao@intel.com>,
	"Zhi Wang" <zhi.wang.linux@gmail.com>,
	"Pierre Morel" <pmorel@linux.ibm.com>,
	"Cornelia Huck" <cohuck@redhat.com>,
	"QEMU Developers" <qemu-devel@nongnu.org>,
	"Shameer Kolothum" <shameerali.kolothum.thodi@huawei.com>,
	"Max Reitz" <mreitz@redhat.com>,
	"Eric Auger" <eric.auger@redhat.com>,
	"Kirti Wankhede" <kwankhede@nvidia.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"Neo Jia" <cjia@nvidia.com>,
	"Amey Narkhede" <ameynarkhede03@gmail.com>,
	"Philippe Mathieu-Daudé" <philmd@redhat.com>,
	"Matthew Rosato" <mjrosato@linux.ibm.com>
Subject: Re: [PULL 00/32] VFIO updates 2020-10-26 (for QEMU 5.2 soft-freeze)
Date: Wed, 28 Oct 2020 09:18:59 +0000	[thread overview]
Message-ID: <20201028091859.GA3701@work-vm> (raw)
In-Reply-To: <20201027200021.00fac851@x1.home>

* Alex Williamson (alex.williamson@redhat.com) wrote:
> On Tue, 27 Oct 2020 23:42:57 +0000
> Peter Maydell <peter.maydell@linaro.org> wrote:
> 
> > On Mon, 26 Oct 2020 at 19:39, Alex Williamson
> > <alex.williamson@redhat.com> wrote:
> > > ----------------------------------------------------------------
> > > VFIO update 2020-10-26
> > >
> > >  * Migration support (Kirti Wankhede)
> > >  * s390 DMA limiting (Matthew Rosato)
> > >  * zPCI hardware info (Matthew Rosato)
> > >  * Lock guard (Amey Narkhede)
> > >  * Print fixes (Zhengui li)  
> > 
> > I get a conflict here in
> > include/standard-headers/linux/fuse.h:
> > 
> > ++<<<<<<< HEAD
> >  +#define FUSE_ATTR_FLAGS               (1 << 27)
> > ++=======
> > + #define FUSE_SUBMOUNTS                (1 << 27)
> > ++>>>>>>> remotes/awilliam/tags/vfio-update-20201026.0  
> > 
> > I assume these should not both be trying to use the same value,
> > so something has gone wrong somewhere. The conflicting commit
> > now in master is Max's 97d741cc96dd08 ("linux/fuse.h: Pull in from Linux").
> > 
> > Can you sort out the correct resolution between you, please?
> > (My guess is that Max's commit is the erroneous one because
> > it doesn't look like it was created via a standard update
> > from the kernel headers.)

Eww that's messy; copying in Miklos to see what's going on.

> So as near as I can tell, QEMU commit 97d741cc96dd ("linux/fuse.h: Pull
> in from Linux") is fantasy land.  The only thing I can find of this
> FUSE_ATTR_FLAGS outside Max's QEMU series is this[1] posting where the
> fuse maintainer announces that he's replaced FUSE_ATTR_FLAGS with
> FUSE_SUBMOUNTS, but the usage is "slightly different".  Reading that
> thread, it seems that virtiofsd probably needed an update but I can't
> see that it ever happened.
> 
> I'm not comfortable trying to update Max's series to try to determine
> if FUSE_SUBMOUNTS can be interchanged with FUSE_ATTR_FLAGS, where the
> latter appears to be used to express the new field in struct fuse_attr
> exists, but the former appears to be a feature.  My guess would be that
> maybe FUSE_KERNEL_MINOR_VERSION needs to be tested instead for this new
> field??

They're part of the init flags sent in the negotiation; so they should
be fine.

> Anyway, I hate to pull the big hammer, but I think Max's series is
> bogus.  The only thing I can propose is to revert it in its entirety,
> after which this series applies cleanly.  I'll post a patch to do that
> as I think the code currently in master is on pretty shaky ground with
> respect to interpreting flag bits differently from those the kernel
> defines.  Thanks,

Yeh the kernel header is king in the definition of those flags.
It maybe bet, but I'd like to see what Miklos says.

Dave

> Alex
> 
> [1] https://marc.info/?l=fuse-devel&m=160069539811247
-- 
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK