All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Xu <peterx@redhat.com>
To: Jan Kiszka <jan.kiszka@web.de>
Cc: Alex Williamson <alex.williamson@redhat.com>,
	"Aviv B.D" <bd.aviv@gmail.com>,
	qemu-devel@nongnu.org, "Michael S. Tsirkin" <mst@redhat.com>
Subject: Re: [Qemu-devel] [PATCH v3 1/3] IOMMU: add VTD_CAP_CM to vIOMMU capability exposed to guest
Date: Mon, 6 Jun 2016 13:04:07 +0800	[thread overview]
Message-ID: <20160606050407.GB21254@pxdev.xzpeter.org> (raw)
In-Reply-To: <5750313C.4000709@web.de>

On Thu, Jun 02, 2016 at 03:14:36PM +0200, Jan Kiszka wrote:
> On 2016-06-02 15:00, Alex Williamson wrote:
> > On Thu, 2 Jun 2016 16:44:39 +0800
> > Peter Xu <peterx@redhat.com> wrote:
[...]
> >> There should be more than one way to make it optional. Which is
> >> better? What I can think of:
> >>
> >> (Assume we have Marcel's "-device intel_iommu" working already)
> >>
> >> 1. Let the CM bit optional, or say, we need to specify something like
> >>    "-device intel_iommu,cmbit=on" or we will disable CM bit. If we
> >>    have CM disabled but with VFIO device, let QEMU raise error.
> >>
> >> 2. We automatically detect whether we need CM bit. E.g., if we have
> >>    VFIO and vIOMMU both enabled, we automatically set the bit. Another
> >>    case is maybe we would in the future support nested vIOMMU? If so,
> >>    we can do the same thing for the nested feature.
> > 
> > 
> > Why do we need to support VT-d for guests that do not support CM=1?
> > The VT-d spec indicates that software should be written to handle both
> > caching modes (6.1).  Granted this is a *should* and not a *must*,
> > but can't we consider guests that do not support CM=1 incompatible with
> > emulated VT-d?  If CM=0 needs to be supported then we need to shadow
> > all of the remapping structures since vfio effectively becomes a cache
> > of the that would otherwise depend on the invalidation of both present
> > and non-present entries.  What guests do not support CM=1?  Thanks,
> 
> - there is at least one guest that does not support CM=1 yet (Jailhouse)

Besides the reason that there might have guests that do not support
CM=1, will there be performance considerations? When user's
configuration does not require CM capability (e.g., generic VM
configuration, without VFIO), shall we allow user to disable the CM
bit so that we can have better IOMMU performance (avoid extra and
useless invalidations)?

-- peterx

  parent reply	other threads:[~2016-06-06  5:04 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-05-21 16:19 [Qemu-devel] [PATCH v3 0/3] IOMMU: Add Support to VFIO devices with vIOMMU present Aviv B.D
2016-05-21 16:19 ` [Qemu-devel] [PATCH v3 1/3] IOMMU: add VTD_CAP_CM to vIOMMU capability exposed to guest Aviv B.D
2016-05-21 16:42   ` Jan Kiszka
2016-06-02  8:44     ` Peter Xu
2016-06-02 13:00       ` Alex Williamson
2016-06-02 13:14         ` Jan Kiszka
2016-06-02 13:17           ` Jan Kiszka
2016-06-02 16:15           ` Michael S. Tsirkin
2016-06-06  5:04           ` Peter Xu [this message]
2016-06-06 13:11             ` Alex Williamson
2016-06-06 13:43               ` Peter Xu
2016-06-06 17:02                 ` Alex Williamson
2016-06-07  3:20                   ` Peter Xu
2016-06-07  3:58                     ` Alex Williamson
2016-06-07  5:00                       ` Peter Xu
2016-06-07  5:21                       ` Huang, Kai
2016-06-07 18:46                         ` Alex Williamson
2016-06-07 22:39                           ` Huang, Kai
2016-05-24  8:14   ` Jason Wang
2016-05-24  9:25     ` Jan Kiszka
2016-05-28 16:12       ` Aviv B.D.
2016-05-28 16:34         ` Kiszka, Jan
2016-05-21 16:19 ` [Qemu-devel] [PATCH v3 2/3] IOMMU: change iommu_op->translate's is_write to flags, add support to NO_FAIL flag mode Aviv B.D
2016-06-06  5:04   ` Peter Xu
2016-05-21 16:19 ` [Qemu-devel] [PATCH v3 3/3] IOMMU: Integrate between VFIO and vIOMMU to support device assignment Aviv B.D
2016-05-23 17:53   ` Alex Williamson
2016-05-26 20:58     ` Alex Williamson
2016-05-28 10:52       ` Aviv B.D.
2016-05-28 16:02         ` Alex Williamson
2016-05-28 16:10           ` Aviv B.D.
2016-05-28 17:39             ` Alex Williamson
2016-05-28 18:14               ` Aviv B.D.
2016-05-28 19:48                 ` Alex Williamson
2016-06-02 13:09                   ` Aviv B.D.
2016-06-02 13:34                     ` Alex Williamson
2016-06-06  8:09                       ` Peter Xu
2016-06-06 18:21                         ` Alex Williamson
2016-06-07 13:20                           ` Peter Xu
2016-06-06  7:38     ` Peter Xu
2016-06-06 17:30       ` Alex Williamson

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=20160606050407.GB21254@pxdev.xzpeter.org \
    --to=peterx@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=bd.aviv@gmail.com \
    --cc=jan.kiszka@web.de \
    --cc=mst@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.