qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Liu, Yi L" <yi.l.liu@linux.intel.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-devel@nongnu.org, mst@redhat.com,
	david@gibson.dropbear.id.au, alex.williamson@redhat.com,
	eric.auger.pro@gmail.com, yi.l.liu@intel.com, peterx@redhat.com,
	kevin.tian@intel.com, jasowang@redhat.com
Subject: Re: [Qemu-devel] [PATCH v3 08/12] hw/pci: introduce pci_device_notify_iommu()
Date: Mon, 5 Mar 2018 16:42:07 +0800	[thread overview]
Message-ID: <20180305084207.GE2482@sky-dev> (raw)
In-Reply-To: <f9ee648a-ff93-b57a-4b05-6bd0f40c15bf@redhat.com>

On Fri, Mar 02, 2018 at 04:12:01PM +0100, Paolo Bonzini wrote:
> On 01/03/2018 11:33, Liu, Yi L wrote:
> > This patch adds pci_device_notify_iommu() for notify virtual IOMMU
> > emulator when assigned device is added. And adds a new notify_func
> > in PCIBus. vIOMMU emulator provides the instance of this notify_func.
> > 
> > Reason:
> > When virtual IOMMU is exposed to guest, vIOMMU emulator needs to
> > programm host IOMMU to setup DMA mapping for assigned devices. This
> > is a per-device operation, to be efficient, vIOMMU emulator needs
> > to record the assigned devices.
> > 
> > Example: devices assigned thru vfio, vfio_realize would call
> > pci_device_notify_iommu() to notify vIOMMU emulator to record necessary
> > info for assigned device.
> 
> I think the notification should not be left to the individual device.
> Instead, the add notification should be done in pci_setup_sva_ops, and
> the delete notification should be done in pci_qdev_unrealize, before
> calling pc->exit, and only if dev->sva_ops is not NULL.

Agreed. I think it works together with your comments against
"[PATCH v3 05/12] hw/pci: introduce PCISVAOps to PCIDevice". Would apply
it in next version.
 
> In general I think it's better to change your names from "assigned_dev"
> to "sva_dev", because the point of the list is to only iterate over
> devices that might be interested in using SVA.

For "assigned_dev", my purpose is to distinguish "assigned devices" from
emulated devices. Only the SVA usage on "assigned devices" is cared here.
But it is true only SVA capable device is interested. So I may need to
rename it as "assigned_sva_dev". How about your opinion?

Thanks,
Yi Liu

  reply	other threads:[~2018-03-05  8:59 UTC|newest]

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-01 10:33 [Qemu-devel] [PATCH v3 00/12] Introduce new iommu notifier framework for virt-SVA Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 01/12] memory: rename existing iommu notifier to be iommu mr notifier Liu, Yi L
2018-03-02 15:01   ` Paolo Bonzini
2018-03-05 10:09     ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 02/12] vfio: rename GuestIOMMU to be GuestIOMMUMR Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 03/12] hw/core: introduce IOMMUSVAContext for virt-SVA Liu, Yi L
2018-03-02 15:13   ` Paolo Bonzini
2018-03-05  8:10     ` Liu, Yi L
2018-03-06  8:51   ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 04/12] vfio/pci: add notify framework based on IOMMUSVAContext Liu, Yi L
2018-03-05  7:45   ` Peter Xu
2018-03-05  8:05     ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 05/12] hw/pci: introduce PCISVAOps to PCIDevice Liu, Yi L
2018-03-02 15:10   ` Paolo Bonzini
2018-03-05  8:11     ` Liu, Yi L
2018-03-06 10:33   ` Liu, Yi L
2018-04-12  2:36     ` David Gibson
2018-04-12 11:06       ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 06/12] vfio/pci: provide vfio_pci_sva_ops instance Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 07/12] vfio/pci: register sva notifier Liu, Yi L
2018-03-06  6:44   ` Peter Xu
2018-03-06  8:00     ` Liu, Yi L
2018-03-06 12:09       ` Peter Xu
2018-03-08 11:22         ` Liu, Yi L
2018-03-09  7:05           ` Peter Xu
2018-03-09 10:25             ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 08/12] hw/pci: introduce pci_device_notify_iommu() Liu, Yi L
2018-03-02 15:12   ` Paolo Bonzini
2018-03-05  8:42     ` Liu, Yi L [this message]
2018-03-06 10:18       ` Paolo Bonzini
2018-03-06 11:03         ` Liu, Yi L
2018-03-06 11:22           ` Paolo Bonzini
2018-03-06 11:27             ` Liu, Yi L
2018-03-02 16:06   ` Paolo Bonzini
2018-03-05  8:43     ` Liu, Yi L
2018-03-05 10:43       ` Peter Xu
2018-03-06 10:19         ` Paolo Bonzini
2018-03-06 10:47           ` Peter Xu
2018-03-06 11:06             ` Liu, Yi L
2018-03-05  8:27   ` Peter Xu
2018-03-05  8:46     ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 09/12] intel_iommu: record assigned devices in a list Liu, Yi L
2018-03-02 15:08   ` Paolo Bonzini
2018-03-05  9:39     ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 10/12] intel_iommu: bind guest pasid table to host Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 11/12] intel_iommu: add framework for PASID AddressSpace management Liu, Yi L
2018-03-02 14:52   ` Paolo Bonzini
2018-03-05  9:12     ` Liu, Yi L
2018-03-02 15:00   ` Paolo Bonzini
2018-03-05  9:11     ` Liu, Yi L
2018-03-06 10:26       ` Paolo Bonzini
2018-03-08 10:42         ` Liu, Yi L
2018-03-01 10:33 ` [Qemu-devel] [PATCH v3 12/12] intel_iommu: bind device to PASID tagged AddressSpace Liu, Yi L
2018-03-02 14:51   ` Paolo Bonzini
2018-03-05  9:56     ` Liu, Yi L
2018-03-06 11:43   ` Peter Xu
2018-03-08  9:39     ` Liu, Yi L
2018-03-09  7:59       ` Peter Xu
2018-03-09  8:09         ` Tian, Kevin
2018-03-09 11:05         ` Liu, Yi L
2018-03-06  6:55 ` [Qemu-devel] [PATCH v3 00/12] Introduce new iommu notifier framework for virt-SVA Peter Xu
2018-03-06  7:45   ` Liu, Yi L
2018-03-07  5:38     ` Peter Xu
2018-03-08  9:10       ` Liu, Yi L
  -- strict thread matches above, loose matches on Subject: below --
2018-03-01 10:31 Liu, Yi L
2018-03-01 10:31 ` [Qemu-devel] [PATCH v3 08/12] hw/pci: introduce pci_device_notify_iommu() Liu, Yi L

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=20180305084207.GE2482@sky-dev \
    --to=yi.l.liu@linux.intel.com \
    --cc=alex.williamson@redhat.com \
    --cc=david@gibson.dropbear.id.au \
    --cc=eric.auger.pro@gmail.com \
    --cc=jasowang@redhat.com \
    --cc=kevin.tian@intel.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=peterx@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=yi.l.liu@intel.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).