From: Jason Gunthorpe <jgg@nvidia.com>
To: iommu@lists.linux.dev, Kevin Tian <kevin.tian@intel.com>,
linux-kselftest@vger.kernel.org
Cc: kvm@vger.kernel.org, Nicolin Chen <nicolinc@nvidia.com>,
Yi Liu <yi.l.liu@intel.com>
Subject: [PATCH v2 09/17] iommufd: Add enforced_cache_coherency to iommufd_hw_pagetable_alloc()
Date: Tue, 7 Mar 2023 20:35:48 -0400 [thread overview]
Message-ID: <9-v2-51b9896e7862+8a8c-iommufd_alloc_jgg@nvidia.com> (raw)
In-Reply-To: <0-v2-51b9896e7862+8a8c-iommufd_alloc_jgg@nvidia.com>
Logically the HWPT should have the coherency set properly for the device
that it is being created for when it is created.
This was happening implicitly if the immediate_attach was set because
iommufd_hw_pagetable_attach() does it as the first thing.
Do it unconditionally so !immediate_attach works properly.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
---
drivers/iommu/iommufd/device.c | 20 +++++-------------
drivers/iommu/iommufd/hw_pagetable.c | 27 +++++++++++++++++++++++++
drivers/iommu/iommufd/iommufd_private.h | 1 +
3 files changed, 33 insertions(+), 15 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c
index d6d5c2fcc33889..ddde14d6d1352c 100644
--- a/drivers/iommu/iommufd/device.c
+++ b/drivers/iommu/iommufd/device.c
@@ -293,21 +293,11 @@ int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt,
if (idev->igroup->hwpt != NULL && idev->igroup->hwpt != hwpt)
return -EINVAL;
- /*
- * Try to upgrade the domain we have, it is an iommu driver bug to
- * report IOMMU_CAP_ENFORCE_CACHE_COHERENCY but fail
- * enforce_cache_coherency when there are no devices attached to the
- * domain.
- */
- if (idev->enforce_cache_coherency && !hwpt->enforce_cache_coherency) {
- if (hwpt->domain->ops->enforce_cache_coherency)
- hwpt->enforce_cache_coherency =
- hwpt->domain->ops->enforce_cache_coherency(
- hwpt->domain);
- if (!hwpt->enforce_cache_coherency) {
- WARN_ON(list_empty(&idev->igroup->device_list));
- return -EINVAL;
- }
+ /* Try to upgrade the domain we have */
+ if (idev->enforce_cache_coherency) {
+ rc = iommufd_hw_pagetable_enforce_cc(hwpt);
+ if (rc)
+ return rc;
}
rc = iopt_table_enforce_dev_resv_regions(&hwpt->ioas->iopt, idev->dev,
diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c
index 566eba0cd9b917..2584f9038b29a2 100644
--- a/drivers/iommu/iommufd/hw_pagetable.c
+++ b/drivers/iommu/iommufd/hw_pagetable.c
@@ -25,6 +25,20 @@ void iommufd_hw_pagetable_destroy(struct iommufd_object *obj)
refcount_dec(&hwpt->ioas->obj.users);
}
+int iommufd_hw_pagetable_enforce_cc(struct iommufd_hw_pagetable *hwpt)
+{
+ if (hwpt->enforce_cache_coherency)
+ return 0;
+
+ if (hwpt->domain->ops->enforce_cache_coherency)
+ hwpt->enforce_cache_coherency =
+ hwpt->domain->ops->enforce_cache_coherency(
+ hwpt->domain);
+ if (!hwpt->enforce_cache_coherency)
+ return -EINVAL;
+ return 0;
+}
+
/**
* iommufd_hw_pagetable_alloc() - Get an iommu_domain for a device
* @ictx: iommufd context
@@ -60,6 +74,19 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas,
goto out_abort;
}
+ /*
+ * Set the coherency mode before we do iopt_table_add_domain() as some
+ * iommus have a per-PTE bit that controls it and need to decide before
+ * doing any maps. It is an iommu driver bug to report
+ * IOMMU_CAP_ENFORCE_CACHE_COHERENCY but fail enforce_cache_coherency on
+ * a new domain.
+ */
+ if (idev->enforce_cache_coherency) {
+ rc = iommufd_hw_pagetable_enforce_cc(hwpt);
+ if (WARN_ON(rc))
+ goto out_abort;
+ }
+
mutex_lock(&idev->igroup->lock);
/*
diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h
index 8bf053f4d4a9ce..471a3fdff1e0b6 100644
--- a/drivers/iommu/iommufd/iommufd_private.h
+++ b/drivers/iommu/iommufd/iommufd_private.h
@@ -254,6 +254,7 @@ struct iommufd_hw_pagetable {
struct iommufd_hw_pagetable *
iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas,
struct iommufd_device *idev, bool immediate_attach);
+int iommufd_hw_pagetable_enforce_cc(struct iommufd_hw_pagetable *hwpt);
int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt,
struct iommufd_device *idev);
struct iommufd_hw_pagetable *
--
2.39.2
next prev parent reply other threads:[~2023-03-08 0:36 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-08 0:35 [PATCH v2 00/17] Add iommufd physical device operations for replace and alloc hwpt Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 01/17] iommufd: Move isolated msi enforcement to iommufd_device_bind() Jason Gunthorpe
2023-03-08 2:55 ` Baolu Lu
2023-03-09 12:53 ` Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 02/17] iommufd: Add iommufd_group Jason Gunthorpe
2023-03-08 3:18 ` Baolu Lu
2023-03-09 15:51 ` Jason Gunthorpe
2023-03-10 10:30 ` Tian, Kevin
2023-03-08 0:35 ` [PATCH v2 03/17] iommufd: Replace the hwpt->devices list with iommufd_group Jason Gunthorpe
2023-03-08 11:56 ` Baolu Lu
2023-03-10 10:36 ` Tian, Kevin
2023-03-10 14:16 ` Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 04/17] iommu: Export iommu_get_resv_regions() Jason Gunthorpe
2023-03-08 12:06 ` Baolu Lu
2023-03-08 14:54 ` Jason Gunthorpe
2023-03-10 10:40 ` Tian, Kevin
2023-03-08 0:35 ` [PATCH v2 05/17] iommufd: Keep track of each device's reserved regions instead of groups Jason Gunthorpe
2023-03-08 12:32 ` Baolu Lu
2023-03-08 15:02 ` Jason Gunthorpe
2023-03-09 1:43 ` Baolu Lu
2023-03-10 10:42 ` Tian, Kevin
2023-03-08 0:35 ` [PATCH v2 06/17] iommufd: Use the iommufd_group to avoid duplicate MSI setup Jason Gunthorpe
2023-03-10 10:42 ` Tian, Kevin
2023-03-08 0:35 ` [PATCH v2 07/17] iommufd: Make sw_msi_start a group global Jason Gunthorpe
2023-03-10 10:45 ` Tian, Kevin
2023-03-08 0:35 ` [PATCH v2 08/17] iommufd: Move putting a hwpt to a helper function Jason Gunthorpe
2023-03-08 12:42 ` Baolu Lu
2023-03-10 10:45 ` Tian, Kevin
2023-03-08 0:35 ` Jason Gunthorpe [this message]
2023-03-08 13:04 ` [PATCH v2 09/17] iommufd: Add enforced_cache_coherency to iommufd_hw_pagetable_alloc() Baolu Lu
2023-03-08 15:06 ` Jason Gunthorpe
2023-03-09 2:03 ` Baolu Lu
2023-03-09 2:05 ` Baolu Lu
2023-03-08 0:35 ` [PATCH v2 10/17] iommufd: Reorganize iommufd_device_attach into iommufd_device_change_pt Jason Gunthorpe
2023-03-08 13:46 ` Baolu Lu
2023-03-10 11:26 ` Tian, Kevin
2023-03-10 15:01 ` Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 11/17] iommu: Introduce a new iommu_group_replace_domain() API Jason Gunthorpe
2023-03-08 13:48 ` Baolu Lu
2023-03-08 0:35 ` [PATCH v2 12/17] iommufd: Add iommufd_device_replace() Jason Gunthorpe
2023-03-09 1:14 ` Baolu Lu
2023-03-10 11:48 ` Tian, Kevin
2023-03-10 15:04 ` Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 13/17] iommufd: Make destroy_rwsem use a lock class per object type Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 14/17] iommufd/selftest: Test iommufd_device_replace() Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 15/17] iommufd: Add IOMMU_HWPT_ALLOC Jason Gunthorpe
2023-03-09 1:21 ` Baolu Lu
2023-03-08 0:35 ` [PATCH v2 16/17] iommufd/selftest: Return the real idev id from selftest mock_domain Jason Gunthorpe
2023-03-08 0:35 ` [PATCH v2 17/17] iommufd/selftest: Add a selftest for IOMMU_HWPT_ALLOC Jason Gunthorpe
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=9-v2-51b9896e7862+8a8c-iommufd_alloc_jgg@nvidia.com \
--to=jgg@nvidia.com \
--cc=iommu@lists.linux.dev \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=nicolinc@nvidia.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