From: Kirti Wankhede <kwankhede@nvidia.com>
To: Christoph Hellwig <hch@lst.de>,
Alex Williamson <alex.williamson@redhat.com>
Cc: Diana Craciun <diana.craciun@oss.nxp.com>,
Cornelia Huck <cohuck@redhat.com>,
Eric Auger <eric.auger@redhat.com>,
Jason Gunthorpe <jgg@ziepe.ca>,
Terrence Xu <terrence.xu@intel.com>, <kvm@vger.kernel.org>,
Jason Gunthorpe <jgg@nvidia.com>,
Kevin Tian <kevin.tian@intel.com>
Subject: Re: [PATCH 11/14] vfio: clean up the check for mediated device in vfio_iommu_type1
Date: Fri, 17 Sep 2021 01:25:59 +0530 [thread overview]
Message-ID: <c4ef0d8a-39f4-4834-f8f2-beffd2f2f8ae@nvidia.com> (raw)
In-Reply-To: <20210913071606.2966-12-hch@lst.de>
On 9/13/2021 12:46 PM, Christoph Hellwig wrote:
> Pass the group flags to ->attach_group and remove the messy check for
> the bus type.
>
I like the way vfio_group_type is used in this patch, that removes messy
way to call symbol_get(mdev_bus_type).
Any thoughts on how VFIO_IOMMU, i.e. IOMMU backed mdev, can be implemented?
For IOMMU backed mdev, mdev->dev->iommu_group should be same as
mdev->type->parent->dev->iommu_group or in other words, parent device
would be DMA alias for the mdev device with the restriction - single
mdev device can be created for the physical device. Is it possible to
link iommu_group of these two devices some way?
Thanks,
Kirti
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
> Reviewed-by: Kevin Tian <kevin.tian@intel.com>
> ---
> drivers/vfio/vfio.c | 32 +++++------------------------
> drivers/vfio/vfio.h | 27 +++++++++++++++++++++++-
> drivers/vfio/vfio_iommu_spapr_tce.c | 2 +-
> drivers/vfio/vfio_iommu_type1.c | 19 ++---------------
> 4 files changed, 34 insertions(+), 46 deletions(-)
>
> diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
> index 6589e296ef348c..08b27b64f0f935 100644
> --- a/drivers/vfio/vfio.c
> +++ b/drivers/vfio/vfio.c
> @@ -68,30 +68,6 @@ struct vfio_unbound_dev {
> struct list_head unbound_next;
> };
>
> -enum vfio_group_type {
> - /*
> - * Physical device with IOMMU backing.
> - */
> - VFIO_IOMMU,
> -
> - /*
> - * Virtual device without IOMMU backing. The VFIO core fakes up an
> - * iommu_group as the iommu_group sysfs interface is part of the
> - * userspace ABI. The user of these devices must not be able to
> - * directly trigger unmediated DMA.
> - */
> - VFIO_EMULATED_IOMMU,
> -
> - /*
> - * Physical device without IOMMU backing. The VFIO core fakes up an
> - * iommu_group as the iommu_group sysfs interface is part of the
> - * userspace ABI. Users can trigger unmediated DMA by the device,
> - * usage is highly dangerous, requires an explicit opt-in and will
> - * taint the kernel.
> - */
> - VFIO_NO_IOMMU,
> -};
> -
> struct vfio_group {
> struct kref kref;
> int minor;
> @@ -219,7 +195,7 @@ static long vfio_noiommu_ioctl(void *iommu_data,
> }
>
> static int vfio_noiommu_attach_group(void *iommu_data,
> - struct iommu_group *iommu_group)
> + struct iommu_group *iommu_group, enum vfio_group_type type)
> {
> return 0;
> }
> @@ -1129,7 +1105,8 @@ static int __vfio_container_attach_groups(struct vfio_container *container,
> int ret = -ENODEV;
>
> list_for_each_entry(group, &container->group_list, container_next) {
> - ret = driver->ops->attach_group(data, group->iommu_group);
> + ret = driver->ops->attach_group(data, group->iommu_group,
> + group->type);
> if (ret)
> goto unwind;
> }
> @@ -1387,7 +1364,8 @@ static int vfio_group_set_container(struct vfio_group *group, int container_fd)
> driver = container->iommu_driver;
> if (driver) {
> ret = driver->ops->attach_group(container->iommu_data,
> - group->iommu_group);
> + group->iommu_group,
> + group->type);
> if (ret)
> goto unlock_out;
> }
> diff --git a/drivers/vfio/vfio.h b/drivers/vfio/vfio.h
> index a78de649eb2f16..a6713022115155 100644
> --- a/drivers/vfio/vfio.h
> +++ b/drivers/vfio/vfio.h
> @@ -4,6 +4,30 @@
> * Author: Alex Williamson <alex.williamson@redhat.com>
> */
>
> +enum vfio_group_type {
> + /*
> + * Physical device with IOMMU backing.
> + */
> + VFIO_IOMMU,
> +
> + /*
> + * Virtual device without IOMMU backing. The VFIO core fakes up an
> + * iommu_group as the iommu_group sysfs interface is part of the
> + * userspace ABI. The user of these devices must not be able to
> + * directly trigger unmediated DMA.
> + */
> + VFIO_EMULATED_IOMMU,
> +
> + /*
> + * Physical device without IOMMU backing. The VFIO core fakes up an
> + * iommu_group as the iommu_group sysfs interface is part of the
> + * userspace ABI. Users can trigger unmediated DMA by the device,
> + * usage is highly dangerous, requires an explicit opt-in and will
> + * taint the kernel.
> + */
> + VFIO_NO_IOMMU,
> +};
> +
> /* events for the backend driver notify callback */
> enum vfio_iommu_notify_type {
> VFIO_IOMMU_CONTAINER_CLOSE = 0,
> @@ -20,7 +44,8 @@ struct vfio_iommu_driver_ops {
> long (*ioctl)(void *iommu_data, unsigned int cmd,
> unsigned long arg);
> int (*attach_group)(void *iommu_data,
> - struct iommu_group *group);
> + struct iommu_group *group,
> + enum vfio_group_type);
> void (*detach_group)(void *iommu_data,
> struct iommu_group *group);
> int (*pin_pages)(void *iommu_data,
> diff --git a/drivers/vfio/vfio_iommu_spapr_tce.c b/drivers/vfio/vfio_iommu_spapr_tce.c
> index 3efd09faeca4a8..936a26b13c0b01 100644
> --- a/drivers/vfio/vfio_iommu_spapr_tce.c
> +++ b/drivers/vfio/vfio_iommu_spapr_tce.c
> @@ -1239,7 +1239,7 @@ static long tce_iommu_take_ownership_ddw(struct tce_container *container,
> }
>
> static int tce_iommu_attach_group(void *iommu_data,
> - struct iommu_group *iommu_group)
> + struct iommu_group *iommu_group, enum vfio_group_type type)
> {
> int ret = 0;
> struct tce_container *container = iommu_data;
> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
> index 42a6be1fb7265e..a48e9f597cb213 100644
> --- a/drivers/vfio/vfio_iommu_type1.c
> +++ b/drivers/vfio/vfio_iommu_type1.c
> @@ -36,7 +36,6 @@
> #include <linux/uaccess.h>
> #include <linux/vfio.h>
> #include <linux/workqueue.h>
> -#include <linux/mdev.h>
> #include <linux/notifier.h>
> #include <linux/dma-iommu.h>
> #include <linux/irqdomain.h>
> @@ -1934,20 +1933,6 @@ static bool vfio_iommu_has_sw_msi(struct list_head *group_resv_regions,
> return ret;
> }
>
> -static bool vfio_bus_is_mdev(struct bus_type *bus)
> -{
> - struct bus_type *mdev_bus;
> - bool ret = false;
> -
> - mdev_bus = symbol_get(mdev_bus_type);
> - if (mdev_bus) {
> - ret = (bus == mdev_bus);
> - symbol_put(mdev_bus_type);
> - }
> -
> - return ret;
> -}
> -
> /*
> * This is a helper function to insert an address range to iova list.
> * The list is initially created with a single entry corresponding to
> @@ -2172,7 +2157,7 @@ static void vfio_iommu_iova_insert_copy(struct vfio_iommu *iommu,
> }
>
> static int vfio_iommu_type1_attach_group(void *iommu_data,
> - struct iommu_group *iommu_group)
> + struct iommu_group *iommu_group, enum vfio_group_type type)
> {
> struct vfio_iommu *iommu = iommu_data;
> struct vfio_iommu_group *group;
> @@ -2207,7 +2192,7 @@ static int vfio_iommu_type1_attach_group(void *iommu_data,
> if (ret)
> goto out_free;
>
> - if (vfio_bus_is_mdev(bus)) {
> + if (type == VFIO_EMULATED_IOMMU) {
> if (!iommu->external_domain) {
> INIT_LIST_HEAD(&domain->group_list);
> iommu->external_domain = domain;
>
next prev parent reply other threads:[~2021-09-16 19:56 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-13 7:15 cleanup vfio iommu_group creation v5 Christoph Hellwig
2021-09-13 7:15 ` [PATCH 01/14] vfio: Move vfio_iommu_group_get() to vfio_register_group_dev() Christoph Hellwig
2021-09-14 1:57 ` Tian, Kevin
2021-09-14 5:46 ` Christoph Hellwig
2021-09-14 6:11 ` Tian, Kevin
2021-09-13 7:15 ` [PATCH 02/14] vfio: factor out a vfio_iommu_driver_allowed helper Christoph Hellwig
2021-09-23 22:59 ` Alex Williamson
2021-09-13 7:15 ` [PATCH 03/14] vfio: remove the iommudata check in vfio_noiommu_attach_group Christoph Hellwig
2021-09-13 7:15 ` [PATCH 04/14] vfio: factor out a vfio_group_find_or_alloc helper Christoph Hellwig
2021-09-14 2:00 ` Tian, Kevin
2021-09-13 7:15 ` [PATCH 05/14] vfio: refactor noiommu group creation Christoph Hellwig
2021-09-14 2:07 ` Tian, Kevin
2021-09-13 7:15 ` [PATCH 06/14] vfio: remove the iommudata hack for noiommu groups Christoph Hellwig
2021-09-15 16:44 ` Jason Gunthorpe
2021-09-13 7:15 ` [PATCH 07/14] vfio: simplify iommu group allocation for mediated devices Christoph Hellwig
2021-09-14 2:23 ` Tian, Kevin
2021-09-13 7:16 ` [PATCH 08/14] vfio: remove unused method from vfio_iommu_driver_ops Christoph Hellwig
2021-09-13 7:16 ` [PATCH 09/14] vfio: move the vfio_iommu_driver_ops interface out of <linux/vfio.h> Christoph Hellwig
2021-09-13 7:16 ` [PATCH 10/14] vfio: remove the unused mdev iommu hook Christoph Hellwig
2021-09-13 7:16 ` [PATCH 11/14] vfio: clean up the check for mediated device in vfio_iommu_type1 Christoph Hellwig
2021-09-16 19:55 ` Kirti Wankhede [this message]
2021-09-16 22:18 ` Jason Gunthorpe
2021-09-17 4:49 ` Tian, Kevin
2021-09-17 5:05 ` Christoph Hellwig
2021-09-17 6:51 ` Kirti Wankhede
2021-09-17 12:53 ` Jason Gunthorpe
2021-09-30 16:46 ` Alex Williamson
2021-09-30 16:57 ` Jason Gunthorpe
2021-10-01 3:20 ` Christoph Hellwig
2021-09-13 7:16 ` [PATCH 12/14] vfio/spapr_tce: reject mediated devices Christoph Hellwig
2021-09-13 7:16 ` [PATCH 13/14] vfio/iommu_type1: remove the "external" domain Christoph Hellwig
2021-09-23 22:59 ` Alex Williamson
2021-09-23 23:06 ` Jason Gunthorpe
2021-09-13 7:16 ` [PATCH 14/14] vfio/iommu_type1: remove IS_IOMMU_CAP_DOMAIN_IN_CONTAINER Christoph Hellwig
2021-09-15 18:07 ` cleanup vfio iommu_group creation v5 Jason Gunthorpe
-- strict thread matches above, loose matches on Subject: below --
2021-08-26 13:34 cleanup vfio iommu_group creation v4 Christoph Hellwig
2021-08-26 13:34 ` [PATCH 11/14] vfio: clean up the check for mediated device in vfio_iommu_type1 Christoph Hellwig
2021-08-25 16:19 cleanup vfio iommu_group creation v3 Christoph Hellwig
2021-08-25 16:19 ` [PATCH 11/14] vfio: clean up the check for mediated device in vfio_iommu_type1 Christoph Hellwig
2021-08-26 3:51 ` Tian, Kevin
2021-08-24 14:46 cleanup vfio iommu_group creation v2 Christoph Hellwig
2021-08-24 14:46 ` [PATCH 11/14] vfio: clean up the check for mediated device in vfio_iommu_type1 Christoph Hellwig
2021-08-25 0:28 ` Jason Gunthorpe
2021-08-25 5:34 ` Christoph Hellwig
2021-08-25 12:35 ` Jason Gunthorpe
2021-08-11 15:14 cleanup vfio iommu_group creation Christoph Hellwig
2021-08-11 15:14 ` [PATCH 11/14] vfio: clean up the check for mediated device in vfio_iommu_type1 Christoph Hellwig
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=c4ef0d8a-39f4-4834-f8f2-beffd2f2f8ae@nvidia.com \
--to=kwankhede@nvidia.com \
--cc=alex.williamson@redhat.com \
--cc=cohuck@redhat.com \
--cc=diana.craciun@oss.nxp.com \
--cc=eric.auger@redhat.com \
--cc=hch@lst.de \
--cc=jgg@nvidia.com \
--cc=jgg@ziepe.ca \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=terrence.xu@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).