LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Zhenyu Wang <zhenyuw@linux.intel.com>
To: Alex Williamson <alex.williamson@redhat.com>
Cc: Cornelia Huck <cohuck@redhat.com>,
	kwankhede@nvidia.com, kvm@vger.kernel.org,
	linux-kernel@vger.kernel.org, pasic@linux.ibm.com,
	zhenyuw@linux.intel.com, intel-gvt-dev@lists.freedesktop.org,
	intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH v4 1/2] vfio/mdev: Check globally for duplicate devices
Date: Wed, 23 May 2018 12:53:28 +0800	[thread overview]
Message-ID: <20180523045328.ooyr2dxrd7hqxj6e@zhen-hp.sh.intel.com> (raw)
In-Reply-To: <20180522095337.1a3043e6@w520.home>

[-- Attachment #1: Type: text/plain, Size: 4306 bytes --]

On 2018.05.22 09:53:37 -0600, Alex Williamson wrote:
> [Cc +GVT-g maintainers/lists]
> 
> On Tue, 22 May 2018 10:13:46 +0200
> Cornelia Huck <cohuck@redhat.com> wrote:
> 
> > On Fri, 18 May 2018 13:10:25 -0600
> > Alex Williamson <alex.williamson@redhat.com> wrote:
> > 
> > > When we create an mdev device, we check for duplicates against the
> > > parent device and return -EEXIST if found, but the mdev device
> > > namespace is global since we'll link all devices from the bus.  We do
> > > catch this later in sysfs_do_create_link_sd() to return -EEXIST, but
> > > with it comes a kernel warning and stack trace for trying to create
> > > duplicate sysfs links, which makes it an undesirable response.
> > > 
> > > Therefore we should really be looking for duplicates across all mdev
> > > parent devices, or as implemented here, against our mdev device list.
> > > Using mdev_list to prevent duplicates means that we can remove
> > > mdev_parent.lock, but in order not to serialize mdev device creation
> > > and removal globally, we add mdev_device.active which allows UUIDs to
> > > be reserved such that we can drop the mdev_list_lock before the mdev
> > > device is fully in place.
> > > 
> > > Two behavioral notes; first, mdev_parent.lock had the side-effect of
> > > serializing mdev create and remove ops per parent device.  This was
> > > an implementation detail, not an intentional guarantee provided to
> > > the mdev vendor drivers.  Vendor drivers can trivially provide this
> > > serialization internally if necessary.  Second, review comments note
> > > the new -EAGAIN behavior when the device, and in particular the remove
> > > attribute, becomes visible in sysfs.  If a remove is triggered prior
> > > to completion of mdev_device_create() the user will see a -EAGAIN
> > > error.  While the errno is different, receiving an error during this
> > > period is not, the previous implementation returned -ENODEV for the
> > > same condition.  Furthermore, the consistency to the user is improved
> > > in the case where mdev_device_remove_ops() returns error.  Previously
> > > concurrent calls to mdev_device_remove() could see the device
> > > disappear with -ENODEV and return in the case of error.  Now a user
> > > would see -EAGAIN while the device is in this transitory state.
> > > 
> > > Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> > > ---
> > >  Documentation/vfio-mediated-device.txt |    5 ++
> > >  drivers/vfio/mdev/mdev_core.c          |  102 +++++++++++---------------------
> > >  drivers/vfio/mdev/mdev_private.h       |    2 -
> > >  3 files changed, 42 insertions(+), 67 deletions(-)  
> > 
> > Reviewed-by: Cornelia Huck <cohuck@redhat.com>
> > 
> > I think it is better to deal with any possible vendor driver
> > implications on top of this (I still believe that vfio-ccw is fine).
> 
> Thanks Cornelia.  So if vfio-ccw is fine, presumably NVIDIA is fine,
> then this leaves GVT-g to see if there's any fallout.  Zhenyu & Zhi,
> I've linked the series under discussion here below[1].  The question to
> you is the first of the two behavioral notes listed above, does GVT-g
> have any dependency on the mdev core providing serialization per mdev
> parent device for the create and remove callbacks within the
> mdev_parent_ops?  This was never an intended feature of the
> implementation and as noted it should be trivial for for an mdev vendor
> driver to provide equivalent course grained serialization if
> necessary.  Of course it would be better to implement that sooner
> rather than later if required.
> 
> I see that __intel_gvt_create_vgpu() makes use of gvt->lock, which
> would seem to already provide this level of per-parent locking. The
> remove path makes use of this same lock, so I think we're ok, but
> looking for an explicit ack so there are no surprises.  I'd like
> to queue this series for v4.18.  Thanks,
> 

yeah, we don't expect mdev core for parent serialization for create and
remove of mdev device. Series look good to me.

Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>


> Alex
> 
> [1] https://lkml.org/lkml/2018/5/18/1035

-- 
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

  reply	other threads:[~2018-05-23  5:03 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-18 19:10 [PATCH v4 0/2] vfio/mdev: Device namespace protection Alex Williamson
2018-05-18 19:10 ` [PATCH v4 1/2] vfio/mdev: Check globally for duplicate devices Alex Williamson
2018-05-18 19:37   ` Kirti Wankhede
2018-05-22  8:13   ` Cornelia Huck
2018-05-22 15:53     ` Alex Williamson
2018-05-23  4:53       ` Zhenyu Wang [this message]
2018-05-18 19:10 ` [PATCH v4 2/2] vfio/mdev: Re-order sysfs attribute creation Alex Williamson
2018-05-18 19:38   ` Kirti Wankhede
2018-05-22  8:14   ` Cornelia Huck
2018-05-18 19:37 ` [PATCH v4 0/2] vfio/mdev: Device namespace protection Kirti Wankhede
2018-05-22 17:17 ` Halil Pasic
2018-05-22 18:38   ` Alex Williamson
2018-05-23  8:56     ` Cornelia Huck
2018-05-23 12:29       ` Halil Pasic
2018-05-23 13:34         ` Cornelia Huck

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180523045328.ooyr2dxrd7hqxj6e@zhen-hp.sh.intel.com \
    --to=zhenyuw@linux.intel.com \
    --cc=alex.williamson@redhat.com \
    --cc=cohuck@redhat.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-gvt-dev@lists.freedesktop.org \
    --cc=kvm@vger.kernel.org \
    --cc=kwankhede@nvidia.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pasic@linux.ibm.com \
    --subject='Re: [PATCH v4 1/2] vfio/mdev: Check globally for duplicate devices' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).