From: Peter Xu <peterx@redhat.com>
To: "Liu, Yi L" <yi.l.liu@intel.com>
Cc: "Tian, Kevin" <kevin.tian@intel.com>,
"jacob.jun.pan@linux.intel.com" <jacob.jun.pan@linux.intel.com>,
Yi Sun <yi.y.sun@linux.intel.com>,
"kvm@vger.kernel.org" <kvm@vger.kernel.org>,
"mst@redhat.com" <mst@redhat.com>,
"Tian, Jun J" <jun.j.tian@intel.com>,
"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
"eric.auger@redhat.com" <eric.auger@redhat.com>,
"alex.williamson@redhat.com" <alex.williamson@redhat.com>,
"pbonzini@redhat.com" <pbonzini@redhat.com>,
"Sun, Yi Y" <yi.y.sun@intel.com>,
David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [RFC v2 14/22] vfio/pci: add iommu_context notifier for pasid bind/unbind
Date: Wed, 6 Nov 2019 09:25:30 -0500 [thread overview]
Message-ID: <20191106142530.GB29717@xz-x1> (raw)
In-Reply-To: <A2975661238FB949B60364EF0F2C25743A0EF2F1@SHSMSX104.ccr.corp.intel.com>
On Wed, Nov 06, 2019 at 12:22:46PM +0000, Liu, Yi L wrote:
> > From: David Gibson
> > Sent: Tuesday, November 5, 2019 12:02 AM
> > To: Liu, Yi L <yi.l.liu@intel.com>
> > Subject: Re: [RFC v2 14/22] vfio/pci: add iommu_context notifier for pasid
> > bind/unbind
> >
> > On Thu, Oct 24, 2019 at 08:34:35AM -0400, Liu Yi L wrote:
> > > This patch adds notifier for pasid bind/unbind. VFIO registers this
> > > notifier to listen to the dual-stage translation (a.k.a. nested
> > > translation) configuration changes and propagate to host. Thus vIOMMU
> > > is able to set its translation structures to host.
> > >
> > > Cc: Kevin Tian <kevin.tian@intel.com>
> > > Cc: Jacob Pan <jacob.jun.pan@linux.intel.com>
> > > Cc: Peter Xu <peterx@redhat.com>
> > > Cc: Eric Auger <eric.auger@redhat.com>
> > > Cc: Yi Sun <yi.y.sun@linux.intel.com>
> > > Cc: David Gibson <david@gibson.dropbear.id.au>
> > > Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
> > > ---
> > > hw/vfio/pci.c | 39 +++++++++++++++++++++++++++++++++++++++
> > > include/hw/iommu/iommu.h | 11 +++++++++++
> > > 2 files changed, 50 insertions(+)
> > >
> > > diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
> > > index 8721ff6..012b8ed 100644
> > > --- a/hw/vfio/pci.c
> > > +++ b/hw/vfio/pci.c
> > > @@ -2767,6 +2767,41 @@ static void
> > vfio_iommu_pasid_free_notify(IOMMUCTXNotifier *n,
> > > pasid_req->free_result = ret;
> > > }
> > >
> > > +static void vfio_iommu_pasid_bind_notify(IOMMUCTXNotifier *n,
> > > + IOMMUCTXEventData *event_data)
> > > +{
> > > +#ifdef __linux__
> >
> > Is hw/vfio/pci.c even built on non-linux hosts?
>
> I'm not quite sure. It's based a comment from RFC v1. I think it could somehow
> prevent compiling issue when doing code porting. So I added it. If it's impossible
> to build on non-linux hosts per your experience, I can remove it to make things
> simple.
To my understanding this should not be needed because VFIO doesn't
work with non-linux after all (as said)... while...
>
> > > + VFIOIOMMUContext *giommu_ctx = container_of(n, VFIOIOMMUContext, n);
> > > + VFIOContainer *container = giommu_ctx->container;
> > > + IOMMUCTXPASIDBindData *pasid_bind =
> > > + (IOMMUCTXPASIDBindData *) event_data->data;
> > > + struct vfio_iommu_type1_bind *bind;
> > > + struct iommu_gpasid_bind_data *bind_data;
> > > + unsigned long argsz;
> > > +
> > > + argsz = sizeof(*bind) + sizeof(*bind_data);
> > > + bind = g_malloc0(argsz);
> > > + bind->argsz = argsz;
> > > + bind->bind_type = VFIO_IOMMU_BIND_GUEST_PASID;
> > > + bind_data = (struct iommu_gpasid_bind_data *) &bind->data;
> > > + *bind_data = *pasid_bind->data;
> > > +
> > > + if (pasid_bind->flag & IOMMU_CTX_BIND_PASID) {
> > > + if (ioctl(container->fd, VFIO_IOMMU_BIND, bind) != 0) {
> > > + error_report("%s: pasid (%llu:%llu) bind failed: %d", __func__,
> > > + bind_data->gpasid, bind_data->hpasid, -errno);
> > > + }
> > > + } else if (pasid_bind->flag & IOMMU_CTX_UNBIND_PASID) {
> > > + if (ioctl(container->fd, VFIO_IOMMU_UNBIND, bind) != 0) {
> > > + error_report("%s: pasid (%llu:%llu) unbind failed: %d", __func__,
> > > + bind_data->gpasid, bind_data->hpasid, -errno);
> > > + }
> > > + }
> > > +
> > > + g_free(bind);
> > > +#endif
> > > +}
> > > +
> > > static void vfio_realize(PCIDevice *pdev, Error **errp)
> > > {
> > > VFIOPCIDevice *vdev = PCI_VFIO(pdev);
> > > @@ -3079,6 +3114,10 @@ static void vfio_realize(PCIDevice *pdev, Error **errp)
> > > iommu_context,
> > > vfio_iommu_pasid_free_notify,
> > > IOMMU_CTX_EVENT_PASID_FREE);
> > > + vfio_register_iommu_ctx_notifier(vdev,
> > > + iommu_context,
> > > + vfio_iommu_pasid_bind_notify,
> > > + IOMMU_CTX_EVENT_PASID_BIND);
> > > }
> > >
> > > return;
> > > diff --git a/include/hw/iommu/iommu.h b/include/hw/iommu/iommu.h
> > > index 4352afd..4f21aa1 100644
> > > --- a/include/hw/iommu/iommu.h
> > > +++ b/include/hw/iommu/iommu.h
> > > @@ -33,6 +33,7 @@ typedef struct IOMMUContext IOMMUContext;
> > > enum IOMMUCTXEvent {
> > > IOMMU_CTX_EVENT_PASID_ALLOC,
> > > IOMMU_CTX_EVENT_PASID_FREE,
> > > + IOMMU_CTX_EVENT_PASID_BIND,
> > > IOMMU_CTX_EVENT_NUM,
> > > };
> > > typedef enum IOMMUCTXEvent IOMMUCTXEvent;
> > > @@ -50,6 +51,16 @@ union IOMMUCTXPASIDReqDesc {
> > > };
> > > typedef union IOMMUCTXPASIDReqDesc IOMMUCTXPASIDReqDesc;
> > >
> > > +struct IOMMUCTXPASIDBindData {
> > > +#define IOMMU_CTX_BIND_PASID (1 << 0)
> > > +#define IOMMU_CTX_UNBIND_PASID (1 << 1)
> > > + uint32_t flag;
> > > +#ifdef __linux__
> > > + struct iommu_gpasid_bind_data *data;
> >
> > Embedding a linux specific structure in the notification message seems
> > dubious to me.
>
> Just similar as your above comment in this thread. If we don't want to add
> it there, then here it is also unnecessary.
... I'm not sure but maybe we need this here because I _think_ vt-d
should even work on Windows. However instead of __linux__ over *data,
should you cover the whole IOMMUCTXPASIDBindData?
--
Peter Xu
next prev parent reply other threads:[~2019-11-06 14:26 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-24 12:34 [RFC v2 00/22] intel_iommu: expose Shared Virtual Addressing to VM Liu Yi L
2019-10-24 12:34 ` [RFC v2 01/22] update-linux-headers: Import iommu.h Liu Yi L
2019-10-24 12:34 ` [RFC v2 02/22] header update VFIO/IOMMU vSVA APIs against 5.4.0-rc3+ Liu Yi L
2019-10-24 12:34 ` [RFC v2 03/22] intel_iommu: modify x-scalable-mode to be string option Liu Yi L
2019-11-01 14:57 ` Peter Xu
2019-11-05 9:14 ` Liu, Yi L
2019-11-05 12:50 ` Peter Xu
2019-11-06 9:50 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 04/22] hw/iommu: introduce IOMMUContext Liu Yi L
2019-10-27 17:39 ` David Gibson
2019-11-06 11:18 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 05/22] vfio/common: add iommu_ctx_notifier in container Liu Yi L
2019-11-01 14:58 ` Peter Xu
2019-11-06 11:08 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 06/22] hw/pci: modify pci_setup_iommu() to set PCIIOMMUOps Liu Yi L
2019-10-27 17:43 ` David Gibson
2019-11-06 8:18 ` Liu, Yi L
2019-11-01 18:09 ` Peter Xu
2019-11-06 8:15 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 07/22] hw/pci: introduce pci_device_iommu_context() Liu Yi L
2019-10-29 11:50 ` David Gibson
2019-11-06 8:20 ` Liu, Yi L
2019-11-01 18:09 ` Peter Xu
2019-11-06 8:14 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 08/22] intel_iommu: provide get_iommu_context() callback Liu Yi L
2019-11-01 14:55 ` Peter Xu
2019-11-06 11:07 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 09/22] vfio/pci: add iommu_context notifier for pasid alloc/free Liu Yi L
2019-10-29 12:15 ` David Gibson
2019-11-01 17:26 ` Peter Xu
2019-11-06 12:46 ` Liu, Yi L
2019-11-06 12:14 ` Liu, Yi L
2019-11-20 4:27 ` David Gibson
2019-11-26 7:07 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 10/22] intel_iommu: add virtual command capability support Liu Yi L
2019-11-01 18:05 ` Peter Xu
2019-11-06 12:40 ` Liu, Yi L
2019-11-06 14:00 ` Peter Xu
2019-11-12 6:27 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 11/22] intel_iommu: process pasid cache invalidation Liu Yi L
2019-11-02 16:05 ` Peter Xu
2019-11-06 5:55 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 12/22] intel_iommu: add present bit check for pasid table entries Liu Yi L
2019-11-02 16:20 ` Peter Xu
2019-11-06 8:14 ` Liu, Yi L
2019-10-24 12:34 ` [RFC v2 13/22] intel_iommu: add PASID cache management infrastructure Liu Yi L
2019-11-04 17:08 ` Peter Xu
2019-11-04 20:06 ` Peter Xu
2019-11-06 7:56 ` Liu, Yi L
2019-11-07 15:46 ` Peter Xu
2019-10-24 12:34 ` [RFC v2 14/22] vfio/pci: add iommu_context notifier for pasid bind/unbind Liu Yi L
2019-11-04 16:02 ` David Gibson
2019-11-06 12:22 ` Liu, Yi L
2019-11-06 14:25 ` Peter Xu [this message]
2019-10-24 12:34 ` [RFC v2 15/22] intel_iommu: bind/unbind guest page table to host Liu Yi L
2019-11-04 20:25 ` Peter Xu
2019-11-06 8:10 ` Liu, Yi L
2019-11-06 14:27 ` Peter Xu
2019-10-24 12:34 ` [RFC v2 16/22] intel_iommu: replay guest pasid bindings " Liu Yi L
2019-10-24 12:34 ` [RFC v2 17/22] intel_iommu: replay pasid binds after context cache invalidation Liu Yi L
2019-10-24 12:34 ` [RFC v2 18/22] intel_iommu: do not passdown pasid bind for PASID #0 Liu Yi L
2019-10-24 12:34 ` [RFC v2 19/22] vfio/pci: add iommu_context notifier for PASID-based iotlb flush Liu Yi L
2019-10-24 12:34 ` [RFC v2 20/22] intel_iommu: process PASID-based iotlb invalidation Liu Yi L
2019-10-24 12:34 ` [RFC v2 21/22] intel_iommu: propagate PASID-based iotlb invalidation to host Liu Yi L
2019-10-24 12:34 ` [RFC v2 22/22] intel_iommu: process PASID-based Device-TLB invalidation Liu Yi L
2019-10-25 6:21 ` [RFC v2 00/22] intel_iommu: expose Shared Virtual Addressing to VM no-reply
2019-10-25 6:30 ` no-reply
2019-10-25 9:49 ` Jason Wang
2019-10-25 10:12 ` Tian, Kevin
2019-10-31 4:33 ` Jason Wang
2019-10-31 5:39 ` Tian, Kevin
2019-10-31 14:07 ` Liu, Yi L
2019-11-01 7:29 ` Jason Wang
2019-11-01 7:46 ` Tian, Kevin
2019-11-01 8:04 ` Jason Wang
2019-11-01 8:09 ` Jason Wang
2019-11-02 7:35 ` Tian, Kevin
2019-11-04 17:22 ` Peter Xu
2019-11-05 9:09 ` 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=20191106142530.GB29717@xz-x1 \
--to=peterx@redhat.com \
--cc=alex.williamson@redhat.com \
--cc=david@gibson.dropbear.id.au \
--cc=eric.auger@redhat.com \
--cc=jacob.jun.pan@linux.intel.com \
--cc=jun.j.tian@intel.com \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=yi.l.liu@intel.com \
--cc=yi.y.sun@intel.com \
--cc=yi.y.sun@linux.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).