From: "Liu, Yi L" <yi.l.liu@linux.intel.com>
To: Auger Eric <eric.auger@redhat.com>
Cc: qemu-devel@nongnu.org, mst@redhat.com,
david@gibson.dropbear.id.au, pbonzini@redhat.com,
alex.williamson@redhat.com, tianyu.lan@intel.com,
kevin.tian@intel.com, yi.l.liu@intel.com, jasowang@redhat.com,
peterx@redhat.com
Subject: Re: [Qemu-devel] [RESEND PATCH 5/6] vfio/pci: add notify framework based on IOMMUObject
Date: Tue, 14 Nov 2017 22:24:07 +0800 [thread overview]
Message-ID: <20171114142407.GC3895@sky-dev> (raw)
In-Reply-To: <7b16c99e-faad-9369-c165-c0dde588021b@redhat.com>
Hi Eric,
On Tue, Nov 14, 2017 at 11:23:03AM +0100, Auger Eric wrote:
> Hi Yi L,
>
> On 03/11/2017 13:01, Liu, Yi L wrote:
> > This patch introduce a notify framework for IOMMUObject.iommu_notifiers.
> > Introduce VFIOGuestIOMMUObject is to link VFIO Container and the new
> > IOMMUObject notififiers.
> notifiers
> >
> > VFIOGuestIOMMUObject instance is allocated when device is assigned and
> > meanwhile vIOMMU is exposed to guest.
> >
> > If there is IOMMUObject behind the device AddressSpace(a.ka vIOMMU exposed).
> > The VFIOGuestIOMMUObject instance would be allocated and inserted to the
> > VFIOContainer.giommu_object_list.
> >
> > Signed-off-by: Liu, Yi L <yi.l.liu@linux.intel.com>
> > ---
> > hw/vfio/pci.c | 39 ++++++++++++++++++++++++++++++++++++++-
> > include/hw/vfio/vfio-common.h | 8 ++++++++
> > 2 files changed, 46 insertions(+), 1 deletion(-)
> >
> > diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
> > index c977ee3..5b77c7e 100644
> > --- a/hw/vfio/pci.c
> > +++ b/hw/vfio/pci.c
> > @@ -2642,6 +2642,8 @@ static void vfio_realize(PCIDevice *pdev, Error **errp)
> > VFIOPCIDevice *vdev = DO_UPCAST(VFIOPCIDevice, pdev, pdev);
> > VFIODevice *vbasedev_iter;
> > VFIOGroup *group;
> > + AddressSpace *as;
> > + IOMMUObject *iommu;
> > char *tmp, group_path[PATH_MAX], *group_name;
> > Error *err = NULL;
> > ssize_t len;
> > @@ -2694,7 +2696,8 @@ static void vfio_realize(PCIDevice *pdev, Error **errp)
> >
> > trace_vfio_realize(vdev->vbasedev.name, groupid);
> >
> > - group = vfio_get_group(groupid, pci_device_iommu_address_space(pdev), errp);
> > + as = pci_device_iommu_address_space(pdev);
> > + group = vfio_get_group(groupid, as, errp);
> > if (!group) {
> > goto error;
> > }
> > @@ -2877,6 +2880,17 @@ static void vfio_realize(PCIDevice *pdev, Error **errp)
> > vfio_register_req_notifier(vdev);
> > vfio_setup_resetfn_quirk(vdev);
> >
> > + iommu = address_space_iommu_get(as);
> > + if (iommu != NULL) {
> > + VFIOGuestIOMMUObject *giommu;
> nit: void line needed
accepted.
> > + giommu = g_malloc0(sizeof(*giommu));
> > + giommu->iommu = iommu;
> > + giommu->container = group->container;
> > + QLIST_INSERT_HEAD(&group->container->giommu_object_list,
> There is no QLIST_INIT anywhere for container's giommu_object_list?
yes, need to add. thx for the catching.
Thanks,
Yi L
>
> Thanks
>
> Eric
> > + giommu,
> > + giommu_next);
> > + }
> > +
> > return;
> >
> > out_teardown:
> > @@ -2907,6 +2921,28 @@ static void vfio_instance_finalize(Object *obj)
> > vfio_put_group(group);
> > }
> >
> > +static void vfio_release_iommu_object(PCIDevice *pdev)
> > +{
> > + VFIOPCIDevice *vdev = DO_UPCAST(VFIOPCIDevice, pdev, pdev);
> > + AddressSpace *as;
> > + IOMMUObject *iommu;
> > +
> > + as = pci_device_iommu_address_space(pdev);
> > + iommu = address_space_iommu_get(as);
> > + if (iommu != NULL) {
> > + VFIOGuestIOMMUObject *giommu, *tmp;
> > + VFIOGroup *group;
> > + group = vdev->vbasedev.group;
> > +
> > + QLIST_FOREACH_SAFE(giommu,
> > + &group->container->giommu_object_list,
> > + giommu_next, tmp) {
> > + QLIST_REMOVE(giommu, giommu_next);
> > + g_free(giommu);
> > + }
> > + }
> > + return;
> > +}
> > static void vfio_exitfn(PCIDevice *pdev)
> > {
> > VFIOPCIDevice *vdev = DO_UPCAST(VFIOPCIDevice, pdev, pdev);
> > @@ -2915,6 +2951,7 @@ static void vfio_exitfn(PCIDevice *pdev)
> > vfio_unregister_err_notifier(vdev);
> > pci_device_set_intx_routing_notifier(&vdev->pdev, NULL);
> > vfio_disable_interrupts(vdev);
> > + vfio_release_iommu_object(pdev);
> > if (vdev->intx.mmap_timer) {
> > timer_free(vdev->intx.mmap_timer);
> > }
> > diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h
> > index 702a085..e4963cc 100644
> > --- a/include/hw/vfio/vfio-common.h
> > +++ b/include/hw/vfio/vfio-common.h
> > @@ -88,6 +88,7 @@ typedef struct VFIOContainer {
> > * future
> > */
> > QLIST_HEAD(, VFIOGuestIOMMUMR) giommu_mr_list;
> > + QLIST_HEAD(, VFIOGuestIOMMUObject) giommu_object_list;
> > QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list;
> > QLIST_HEAD(, VFIOGroup) group_list;
> > QLIST_ENTRY(VFIOContainer) next;
> > @@ -101,6 +102,13 @@ typedef struct VFIOGuestIOMMUMR {
> > QLIST_ENTRY(VFIOGuestIOMMUMR) giommu_next;
> > } VFIOGuestIOMMUMR;
> >
> > +typedef struct VFIOGuestIOMMUObject {
> > + VFIOContainer *container;
> > + IOMMUObject *iommu;
> > + IOMMUNotifier n;
> > + QLIST_ENTRY(VFIOGuestIOMMUObject) giommu_next;
> > +} VFIOGuestIOMMUObject;
> > +
> > typedef struct VFIOHostDMAWindow {
> > hwaddr min_iova;
> > hwaddr max_iova;
> >
next prev parent reply other threads:[~2017-11-14 14:40 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-03 12:01 [Qemu-devel] [RESEND PATCH 0/6] Introduce new iommu notifier framework Liu, Yi L
2017-11-03 12:01 ` [Qemu-devel] [RESEND PATCH 1/6] memory: rename existing iommu notifier to be iommu mr notifier Liu, Yi L
2017-11-03 12:01 ` [Qemu-devel] [RESEND PATCH 2/6] memory: introduce AddressSpaceOps and IOMMUObject Liu, Yi L
2017-11-13 5:56 ` David Gibson
2017-11-13 8:28 ` Peter Xu
2017-11-14 0:59 ` David Gibson
2017-11-14 3:31 ` Peter Xu
2017-12-18 5:41 ` David Gibson
2017-11-16 8:57 ` Liu, Yi L
2017-12-18 6:14 ` David Gibson
2017-12-18 9:17 ` Liu, Yi L
2017-12-18 11:22 ` David Gibson
2017-12-20 6:32 ` Liu, Yi L
2017-12-20 11:01 ` David Gibson
2017-12-22 6:47 ` Liu, Yi L
2017-11-13 9:58 ` Liu, Yi L
2017-11-14 8:53 ` Auger Eric
2017-11-14 13:59 ` Liu, Yi L
2017-11-14 21:52 ` Auger Eric
2017-11-15 2:36 ` Liu, Yi L
2017-11-15 7:16 ` Peter Xu
2017-12-18 11:35 ` David Gibson
2017-12-20 6:47 ` Liu, Yi L
2017-12-20 11:18 ` David Gibson
2017-12-21 8:40 ` Liu, Yi L
2018-01-03 0:28 ` David Gibson
2018-01-04 9:40 ` Liu, Yi L
2018-01-12 10:25 ` Liu, Yi L
2018-01-16 6:04 ` David Gibson
2017-12-18 6:30 ` David Gibson
2017-11-14 10:21 ` Auger Eric
2017-11-14 14:20 ` Liu, Yi L
2017-12-18 11:38 ` David Gibson
2017-11-03 12:01 ` [Qemu-devel] [RESEND PATCH 3/6] intel_iommu: provide AddressSpaceOps.iommu_get instance Liu, Yi L
2017-11-03 12:01 ` [Qemu-devel] [RESEND PATCH 4/6] vfio: rename GuestIOMMU to be GuestIOMMUMR Liu, Yi L
2017-11-03 12:01 ` [Qemu-devel] [RESEND PATCH 5/6] vfio/pci: add notify framework based on IOMMUObject Liu, Yi L
2017-11-14 10:23 ` Auger Eric
2017-11-14 14:24 ` Liu, Yi L [this message]
2017-11-03 12:01 ` [Qemu-devel] [RESEND PATCH 6/6] vfio/pci: register vfio_iommu_bind_pasidtbl_notify notifier 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=20171114142407.GC3895@sky-dev \
--to=yi.l.liu@linux.intel.com \
--cc=alex.williamson@redhat.com \
--cc=david@gibson.dropbear.id.au \
--cc=eric.auger@redhat.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=tianyu.lan@intel.com \
--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).