From: Alex Williamson <alex.williamson@redhat.com>
To: Jason Gunthorpe <jgg@nvidia.com>
Cc: Yishai Hadas <yishaih@nvidia.com>,
bhelgaas@google.com, saeedm@nvidia.com,
linux-pci@vger.kernel.org, kvm@vger.kernel.org,
netdev@vger.kernel.org, kuba@kernel.org, leonro@nvidia.com,
kwankhede@nvidia.com, mgurtovoy@nvidia.com, maorg@nvidia.com
Subject: Re: [PATCH V4 mlx5-next 13/13] vfio/mlx5: Use its own PCI reset_done error handler
Date: Wed, 27 Oct 2021 10:48:55 -0600 [thread overview]
Message-ID: <20211027104855.7d35be05.alex.williamson@redhat.com> (raw)
In-Reply-To: <20211027155339.GE2744544@nvidia.com>
On Wed, 27 Oct 2021 12:53:39 -0300
Jason Gunthorpe <jgg@nvidia.com> wrote:
> On Wed, Oct 27, 2021 at 09:29:43AM -0600, Alex Williamson wrote:
>
> > The reset_done handler sets deferred_reset = true and if it's possible
> > to get the state_mutex, will reset migration data and device_state as
> > part of releasing that mutex. If there's contention on state_mutex,
> > the deferred_reset field flags that this migration state is still stale.
> >
> > So, I assume that it's possible that a user resets the device via ioctl
> > or config space, there was contention and the migration state is still
> > stale, right?
>
> If this occurs it is a userspace bug and the goal here is to maintain
> kernel integrity.
>
> > The user then goes to read device_state, but the staleness of the
> > migration state is not resolved until *after* the stale device state is
> > copied to the user buffer.
>
> This is not preventable in the general case. Assume we have sane
> locking and it looks like this:
>
> CPU0 CPU1
> ioctl state change
> mutex_lock
> copy_to_user(state == !RUNNING)
> mutex_unlock
> ioctl reset
> mutex_lock
> state = RUNNING
> mutex_unlock
> return to userspace
> return to userspace
> Userspace sees state != RUNNING
>
> Same issue. Userspace cannot race state manipulating ioctls and expect
> things to make any sense.
>
> In all cases contention on the mutex during reset causes the reset to
> order after the mutex is released. This is true with this approach and
> it is true with a simple direct use of mutex.
>
> In either case userspace will see incoherent results, and it is
> userspace error to try and run the kernel ioctls this way.
>
> > What did the user do wrong to see stale data? Thanks,
>
> Userspace allowed two state effecting IOCTLs to run concurrently.
>
> Userspace must block reset while it is manipulating migration states.
Ok, I see. I didn't digest that contention on state_mutex can only
occur from a concurrent migration region access and the stale state is
resolved at the end of that concurrent access, not some subsequent
access. I agree we have no obligation to resolve anything about the
state that concurrent access would see. Thanks,
Alex
next prev parent reply other threads:[~2021-10-27 16:49 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-26 9:05 [PATCH V4 mlx5-next 00/13] Add mlx5 live migration driver Yishai Hadas
2021-10-26 9:05 ` [PATCH V4 mlx5-next 01/13] PCI/IOV: Add pci_iov_vf_id() to get VF index Yishai Hadas
2021-10-26 9:05 ` [PATCH V4 mlx5-next 02/13] net/mlx5: Reuse exported virtfn index function call Yishai Hadas
2021-10-26 9:05 ` [PATCH V4 mlx5-next 03/13] net/mlx5: Disable SRIOV before PF removal Yishai Hadas
2021-10-26 9:05 ` [PATCH V4 mlx5-next 04/13] PCI/IOV: Add pci_iov_get_pf_drvdata() to allow VF reaching the drvdata of a PF Yishai Hadas
2021-10-26 9:05 ` [PATCH V4 mlx5-next 05/13] net/mlx5: Expose APIs to get/put the mlx5 core device Yishai Hadas
2021-10-26 9:05 ` [PATCH V4 mlx5-next 06/13] vfio: Fix VFIO_DEVICE_STATE_SET_ERROR macro Yishai Hadas
2021-10-26 15:32 ` Cornelia Huck
2021-10-26 15:50 ` Alex Williamson
2021-10-26 15:56 ` Cornelia Huck
2021-10-26 16:18 ` Leon Romanovsky
2021-10-26 16:32 ` Cornelia Huck
2021-10-26 16:42 ` Leon Romanovsky
2021-10-26 16:57 ` Cornelia Huck
2021-10-26 9:05 ` [PATCH V4 mlx5-next 07/13] vfio: Add a macro for VFIO_DEVICE_STATE_ERROR Yishai Hadas
2021-10-26 15:37 ` Cornelia Huck
2021-10-26 9:06 ` [PATCH V4 mlx5-next 08/13] vfio/pci_core: Make the region->release() function optional Yishai Hadas
2021-10-26 9:06 ` [PATCH V4 mlx5-next 09/13] net/mlx5: Introduce migration bits and structures Yishai Hadas
2021-10-26 9:06 ` [PATCH V4 mlx5-next 10/13] vfio/mlx5: Expose migration commands over mlx5 device Yishai Hadas
2021-10-26 9:06 ` [PATCH V4 mlx5-next 11/13] vfio/mlx5: Implement vfio_pci driver for mlx5 devices Yishai Hadas
2021-10-26 22:42 ` Alex Williamson
2021-10-26 23:46 ` Jason Gunthorpe
2021-10-26 9:06 ` [PATCH V4 mlx5-next 12/13] vfio/pci: Expose vfio_pci_aer_err_detected() Yishai Hadas
2021-10-26 22:45 ` Alex Williamson
2021-10-26 9:06 ` [PATCH V4 mlx5-next 13/13] vfio/mlx5: Use its own PCI reset_done error handler Yishai Hadas
2021-10-26 23:16 ` Alex Williamson
2021-10-26 23:50 ` Jason Gunthorpe
2021-10-27 15:29 ` Alex Williamson
2021-10-27 15:53 ` Jason Gunthorpe
2021-10-27 16:48 ` Alex Williamson [this message]
2021-10-27 16:53 ` Jason Gunthorpe
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=20211027104855.7d35be05.alex.williamson@redhat.com \
--to=alex.williamson@redhat.com \
--cc=bhelgaas@google.com \
--cc=jgg@nvidia.com \
--cc=kuba@kernel.org \
--cc=kvm@vger.kernel.org \
--cc=kwankhede@nvidia.com \
--cc=leonro@nvidia.com \
--cc=linux-pci@vger.kernel.org \
--cc=maorg@nvidia.com \
--cc=mgurtovoy@nvidia.com \
--cc=netdev@vger.kernel.org \
--cc=saeedm@nvidia.com \
--cc=yishaih@nvidia.com \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).