From: Nicolin Chen <nicolinc@nvidia.com>
To: <jgg@nvidia.com>, <kevin.tian@intel.com>, <joro@8bytes.org>,
<will@kernel.org>, <robin.murphy@arm.com>,
<alex.williamson@redhat.com>, <shuah@kernel.org>
Cc: <yi.l.liu@intel.com>, <linux-kernel@vger.kernel.org>,
<iommu@lists.linux.dev>, <kvm@vger.kernel.org>,
<linux-kselftest@vger.kernel.org>, <baolu.lu@linux.intel.com>
Subject: [PATCH v1 5/8] iommufd: Add replace support in iommufd_access_set_ioas()
Date: Wed, 1 Feb 2023 23:05:09 -0800 [thread overview]
Message-ID: <bfe5aed6d354ef547979f0b256c8a3f9bd5b223b.1675320212.git.nicolinc@nvidia.com> (raw)
In-Reply-To: <cover.1675320212.git.nicolinc@nvidia.com>
Support an access->ioas replacement in iommufd_access_set_ioas(), which
sets the access->ioas to NULL provisionally so that any further incoming
iommufd_access_pin_pages() callback can be blocked.
Then, call access->ops->unmap() to clean up the entire iopt. To allow an
iommufd_access_unpin_pages() callback to happen via this unmap() call,
add an ioas_unpin pointer so the unpin routine won't be affected by the
"access->ioas = NULL" trick above.
Also, a vdev without an ops->dma_unmap implementation cannot replace its
access->ioas pointer. So add an iommufd_access_ioas_is_attached() helper
to sanity that.
Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
---
drivers/iommu/iommufd/device.c | 28 +++++++++++++++++++++++--
drivers/iommu/iommufd/iommufd_private.h | 1 +
drivers/vfio/iommufd.c | 4 ++++
include/linux/iommufd.h | 1 +
4 files changed, 32 insertions(+), 2 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c
index f4bd6f532a90..8118d5262800 100644
--- a/drivers/iommu/iommufd/device.c
+++ b/drivers/iommu/iommufd/device.c
@@ -509,11 +509,23 @@ int iommufd_access_set_ioas(struct iommufd_access *access, u32 ioas_id)
iommufd_ref_to_users(obj);
}
+ /*
+ * Set ioas to NULL to block any further iommufd_access_pin_pages().
+ * iommufd_access_unpin_pages() can continue using access->ioas_unpin.
+ */
+ access->ioas = NULL;
+
if (cur_ioas) {
+ if (new_ioas) {
+ mutex_unlock(&access->ioas_lock);
+ access->ops->unmap(access->data, 0, ULONG_MAX);
+ mutex_lock(&access->ioas_lock);
+ }
iopt_remove_access(&cur_ioas->iopt, access);
refcount_dec(&cur_ioas->obj.users);
}
+ access->ioas_unpin = new_ioas;
access->ioas = new_ioas;
mutex_unlock(&access->ioas_lock);
@@ -527,6 +539,18 @@ int iommufd_access_set_ioas(struct iommufd_access *access, u32 ioas_id)
}
EXPORT_SYMBOL_NS_GPL(iommufd_access_set_ioas, IOMMUFD);
+bool iommufd_access_ioas_is_attached(struct iommufd_access *access)
+{
+ bool attached;
+
+ mutex_lock(&access->ioas_lock);
+ attached = !!access->ioas;
+ mutex_unlock(&access->ioas_lock);
+
+ return attached;
+}
+EXPORT_SYMBOL_NS_GPL(iommufd_access_ioas_is_attached, IOMMUFD);
+
/**
* iommufd_access_notify_unmap - Notify users of an iopt to stop using it
* @iopt: iopt to work on
@@ -587,11 +611,11 @@ void iommufd_access_unpin_pages(struct iommufd_access *access,
return;
mutex_lock(&access->ioas_lock);
- if (!access->ioas) {
+ if (!access->ioas_unpin) {
mutex_unlock(&access->ioas_lock);
return;
}
- iopt = &access->ioas->iopt;
+ iopt = &access->ioas_unpin->iopt;
down_read(&iopt->iova_rwsem);
iopt_for_each_contig_area(&iter, area, iopt, iova, last_iova)
diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h
index 2f4bb106bac6..593138bb37b8 100644
--- a/drivers/iommu/iommufd/iommufd_private.h
+++ b/drivers/iommu/iommufd/iommufd_private.h
@@ -261,6 +261,7 @@ struct iommufd_access {
struct iommufd_object obj;
struct iommufd_ctx *ictx;
struct iommufd_ioas *ioas;
+ struct iommufd_ioas *ioas_unpin;
struct mutex ioas_lock;
const struct iommufd_access_ops *ops;
void *data;
diff --git a/drivers/vfio/iommufd.c b/drivers/vfio/iommufd.c
index 78a8e4641cbf..7e09defbcffe 100644
--- a/drivers/vfio/iommufd.c
+++ b/drivers/vfio/iommufd.c
@@ -187,6 +187,10 @@ int vfio_iommufd_emulated_attach_ioas(struct vfio_device *vdev, u32 *pt_id)
if (!vdev->iommufd_access)
return -ENOENT;
+ if (!vdev->ops->dma_unmap && pt_id &&
+ iommufd_access_ioas_is_attached(vdev->iommufd_access))
+ return -EBUSY;
+
iommufd_access_set_ioas(vdev->iommufd_access, *pt_id);
return 0;
}
diff --git a/include/linux/iommufd.h b/include/linux/iommufd.h
index 0e30f2ce27cd..e8d764b2c14c 100644
--- a/include/linux/iommufd.h
+++ b/include/linux/iommufd.h
@@ -44,6 +44,7 @@ iommufd_access_create(struct iommufd_ctx *ictx,
const struct iommufd_access_ops *ops, void *data);
void iommufd_access_destroy(struct iommufd_access *access);
int iommufd_access_set_ioas(struct iommufd_access *access, u32 ioas_id);
+bool iommufd_access_ioas_is_attached(struct iommufd_access *access);
void iommufd_ctx_get(struct iommufd_ctx *ictx);
--
2.39.1
next prev parent reply other threads:[~2023-02-02 7:06 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-02 7:05 [PATCH v1 0/8] Add IO page table replacement support Nicolin Chen
2023-02-02 7:05 ` [PATCH v1 1/8] iommu: Move dev_iommu_ops() to private header Nicolin Chen
2023-02-02 7:05 ` [PATCH v1 2/8] iommu: Introduce a new iommu_group_replace_domain() API Nicolin Chen
2023-02-02 10:21 ` Baolu Lu
2023-02-02 19:14 ` Nicolin Chen
2023-02-03 1:33 ` Baolu Lu
2023-02-03 1:41 ` Nicolin Chen
2023-02-03 2:35 ` Baolu Lu
2023-02-03 8:26 ` Tian, Kevin
2023-02-03 15:03 ` Jason Gunthorpe
2023-02-03 17:53 ` Nicolin Chen
2023-02-06 6:57 ` Tian, Kevin
2023-02-06 13:25 ` Jason Gunthorpe
2023-02-07 0:32 ` Tian, Kevin
2023-02-07 12:22 ` Jason Gunthorpe
2023-02-08 4:25 ` Tian, Kevin
2023-02-08 12:42 ` Jason Gunthorpe
2023-02-07 19:16 ` Nicolin Chen
2023-02-03 17:45 ` Nicolin Chen
2023-02-06 6:40 ` Tian, Kevin
2023-02-02 7:05 ` [PATCH v1 3/8] iommufd: Create access in vfio_iommufd_emulated_bind() Nicolin Chen
2023-02-03 9:23 ` Tian, Kevin
2023-02-03 9:28 ` Tian, Kevin
2023-02-02 7:05 ` [PATCH v1 4/8] iommufd/selftest: Add IOMMU_TEST_OP_ACCESS_SET_IOAS coverage Nicolin Chen
2023-02-02 7:05 ` Nicolin Chen [this message]
2023-02-03 10:10 ` [PATCH v1 5/8] iommufd: Add replace support in iommufd_access_set_ioas() Tian, Kevin
2023-02-03 12:31 ` Jason Gunthorpe
2023-02-03 22:25 ` Nicolin Chen
2023-02-02 7:05 ` [PATCH v1 6/8] iommufd/selftest: Add coverage for access->ioas replacement Nicolin Chen
2023-02-02 7:05 ` [PATCH v1 7/8] iommufd/device: Use iommu_group_replace_domain() Nicolin Chen
2023-02-06 8:46 ` Tian, Kevin
2023-02-06 19:17 ` Nicolin Chen
2023-02-07 0:37 ` Tian, Kevin
2023-02-02 7:05 ` [PATCH v1 8/8] vfio-iommufd: Support IO page table replacement Nicolin Chen
2023-02-06 8:49 ` Tian, Kevin
2023-02-06 18:54 ` Nicolin Chen
2023-02-03 8:09 ` [PATCH v1 0/8] Add IO page table replacement support Tian, Kevin
2023-02-03 15:04 ` Jason Gunthorpe
2023-02-06 6:39 ` Tian, Kevin
2023-02-06 13:26 ` Jason Gunthorpe
2023-02-07 0:34 ` Tian, Kevin
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=bfe5aed6d354ef547979f0b256c8a3f9bd5b223b.1675320212.git.nicolinc@nvidia.com \
--to=nicolinc@nvidia.com \
--cc=alex.williamson@redhat.com \
--cc=baolu.lu@linux.intel.com \
--cc=iommu@lists.linux.dev \
--cc=jgg@nvidia.com \
--cc=joro@8bytes.org \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=robin.murphy@arm.com \
--cc=shuah@kernel.org \
--cc=will@kernel.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
all inboxes | Powered by JetHome®