Nested translation is a hardware feature that is supported by many modern IOMMU hardwares. It has two stages of address translations to get access to the physical address. A stage-1 translation table is owned by userspace (e.g. by a guest OS), while a stage-2 is owned by kernel. Any change to a stage-1 translation table should be followed by an IOTLB invalidation.
Take Intel VT-d as an example, the stage-1 translation table is guest I/O page table. As the below diagram shows, the guest I/O page table pointer in GPA (guest physical address) is passed to host and be used to perform a stage-1 translation. Along with it, a modification to present mappings in the guest I/O page table should be followed by an IOTLB invalidation.
.-------------. .---------------------------. | vIOMMU | | Guest I/O page table | | | '---------------------------' .----------------/ | PASID Entry |--- PASID cache flush --+ '-------------' | | | V | | I/O page table pointer in GPA '-------------' Guest ------| Shadow |---------------------------|-------- v v v Host .-------------. .------------------------. | pIOMMU | | FS for GIOVA->GPA | | | '------------------------' .----------------/ | | PASID Entry | V (Nested xlate) '----------------.----------------------------------. | | | SS for GPA->HPA, unmanaged domain| | | '----------------------------------' '-------------' Where: - FS = First stage page tables - SS = Second stage page tables <Intel VT-d Nested translation>
In IOMMUFD, all the translation tables are tracked by hw_pagetable (hwpt) and each hwpt is backed by an iommu_domain allocated from an iommu driver. So in this series hw_pagetable and iommu_domain means the same thing if no special note. IOMMUFD has already supported allocating hw_pagetable linked with an IOAS. However, a nesting case requires IOMMUFD to allow allocating hw_pagetable with driver specific parameters and interface to sync stage-1 IOTLB as user owns the stage-1 translation table.
This series is based on the iommu hw info reporting series [1] and nested parent domain allocation [2]. It first extends domain_alloc_user to allocate hwpt with user data by allowing the IOMMUFD internal infrastructure to accept user_data and parent hwpt, relaying the user_data/parent to the iommu core to allocate IOMMU_DOMAIN_NESTED. And it then extends the IOMMU_HWPT_ALLOC ioctl to accept user data and a parent hwpt ID.
Note that this series is the part-1 set of a two-part nesting series. It does not include the cache invalidation interface, which will be added in the part 2.
Complete code can be found in [3], QEMU could can be found in [4].
At last, this is a team work together with Nicolin Chen, Lu Baolu. Thanks them for the help. ^_^. Look forward to your feedbacks.
[1] https://lore.kernel.org/linux-iommu/20230818101033.4100-1-yi.l.liu@intel.com... - merged [2] https://lore.kernel.org/linux-iommu/20230928071528.26258-1-yi.l.liu@intel.co... - merged [3] https://github.com/yiliu1765/iommufd/tree/iommufd_nesting [4] https://github.com/yiliu1765/qemu/tree/zhenzhong/wip/iommufd_nesting_rfcv1
Change log:
v5: - Split the iommufd nesting series into two parts of alloc_user and invalidation (Jason) - Split IOMMUFD_OBJ_HW_PAGETABLE to IOMMUFD_OBJ_HWPT_PAGING/_NESTED, and do the same with the structures/alloc()/abort()/destroy(). Reworked the selftest accordingly too. (Jason) - Move hwpt/data_type into struct iommu_user_data from standalone op arguments. (Jason) - Rename hwpt_type to be data_type, the HWPT_TYPE to be HWPT_ALLOC_DATA, _TYPE_DEFAULT to be _ALLOC_DATA_NONE (Jason, Kevin) - Rename iommu_copy_user_data() to iommu_copy_struct_from_user() (Kevin) - Add macro to the iommu_copy_struct_from_user() to calculate min_size (Jason) - Fix two bugs spotted by ZhaoYan
v4: https://lore.kernel.org/linux-iommu/20230921075138.124099-1-yi.l.liu@intel.c... - Separate HWPT alloc/destroy/abort functions between user-managed HWPTs and kernel-managed HWPTs - Rework invalidate uAPI to be a multi-request array-based design - Add a struct iommu_user_data_array and a helper for driver to sanitize and copy the entry data from user space invalidation array - Add a patch fixing TEST_LENGTH() in selftest program - Drop IOMMU_RESV_IOVA_RANGES patches - Update kdoc and inline comments - Drop the code to add IOMMU_RESV_SW_MSI to kernel-managed HWPT in nested translation, this does not change the rule that resv regions should only be added to the kernel-managed HWPT. The IOMMU_RESV_SW_MSI stuff will be added in later series as it is needed only by SMMU so far.
v3: https://lore.kernel.org/linux-iommu/20230724110406.107212-1-yi.l.liu@intel.c... - Add new uAPI things in alphabetical order - Pass in "enum iommu_hwpt_type hwpt_type" to op->domain_alloc_user for sanity, replacing the previous op->domain_alloc_user_data_len solution - Return ERR_PTR from domain_alloc_user instead of NULL - Only add IOMMU_RESV_SW_MSI to kernel-managed HWPT in nested translation (Kevin) - Add IOMMU_RESV_IOVA_RANGES to report resv iova ranges to userspace hence userspace is able to exclude the ranges in the stage-1 HWPT (e.g. guest I/O page table). (Kevin) - Add selftest coverage for the new IOMMU_RESV_IOVA_RANGES ioctl - Minor changes per Kevin's inputs
v2: https://lore.kernel.org/linux-iommu/20230511143844.22693-1-yi.l.liu@intel.co... - Add union iommu_domain_user_data to include all user data structures to avoid passing void * in kernel APIs. - Add iommu op to return user data length for user domain allocation - Rename struct iommu_hwpt_alloc::data_type to be hwpt_type - Store the invalidation data length in iommu_domain_ops::cache_invalidate_user_data_len - Convert cache_invalidate_user op to be int instead of void - Remove @data_type in struct iommu_hwpt_invalidate - Remove out_hwpt_type_bitmap in struct iommu_hw_info hence drop patch 08 of v1
v1: https://lore.kernel.org/linux-iommu/20230309080910.607396-1-yi.l.liu@intel.c...
Thanks, Yi Liu
Jason Gunthorpe (2): iommufd: Rename IOMMUFD_OBJ_HW_PAGETABLE to IOMMUFD_OBJ_HWPT_PAGING iommufd/device: Wrap IOMMUFD_OBJ_HWPT_PAGING-only configurations
Lu Baolu (1): iommu: Add IOMMU_DOMAIN_NESTED
Nicolin Chen (6): iommufd: Derive iommufd_hwpt_paging from iommufd_hw_pagetable iommufd: Share iommufd_hwpt_alloc with IOMMUFD_OBJ_HWPT_NESTED iommufd: Add a nested HW pagetable object iommu: Add iommu_copy_struct_from_user helper iommufd/selftest: Add nested domain allocation for mock domain iommufd/selftest: Add coverage for IOMMU_HWPT_ALLOC with nested HWPTs
Yi Liu (1): iommu: Pass in parent domain with user_data to domain_alloc_user op
drivers/iommu/intel/iommu.c | 4 +- drivers/iommu/iommufd/device.c | 184 +++++++++----- drivers/iommu/iommufd/hw_pagetable.c | 237 ++++++++++++++---- drivers/iommu/iommufd/iommufd_private.h | 67 +++-- drivers/iommu/iommufd/iommufd_test.h | 18 ++ drivers/iommu/iommufd/main.c | 10 +- drivers/iommu/iommufd/selftest.c | 137 ++++++++-- drivers/iommu/iommufd/vfio_compat.c | 6 +- include/linux/iommu.h | 72 +++++- include/uapi/linux/iommufd.h | 31 ++- tools/testing/selftests/iommu/iommufd.c | 120 +++++++++ .../selftests/iommu/iommufd_fail_nth.c | 3 +- tools/testing/selftests/iommu/iommufd_utils.h | 31 ++- 13 files changed, 752 insertions(+), 168 deletions(-)
From: Lu Baolu baolu.lu@linux.intel.com
Introduce a new domain type for a user I/O page table, which is nested on top of another user space address represented by a PAGING domain. This new domain can be allocated by the domain_alloc_user op, and attached to a device through the existing iommu_attach_device/group() interfaces.
The mappings of a nested domain are managed by user space software, so it is not necessary to have map/unmap callbacks.
Signed-off-by: Lu Baolu baolu.lu@linux.intel.com Reviewed-by: Kevin Tian kevin.tian@intel.com Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- include/linux/iommu.h | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 3861d66b65c1..ce08a3c41446 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -65,6 +65,9 @@ struct iommu_domain_geometry {
#define __IOMMU_DOMAIN_SVA (1U << 4) /* Shared process address space */
+#define __IOMMU_DOMAIN_NESTED (1U << 5) /* User-managed address space nested + on a stage-2 translation */ + #define IOMMU_DOMAIN_ALLOC_FLAGS ~__IOMMU_DOMAIN_DMA_FQ /* * This are the possible domain-types @@ -91,6 +94,7 @@ struct iommu_domain_geometry { __IOMMU_DOMAIN_DMA_API | \ __IOMMU_DOMAIN_DMA_FQ) #define IOMMU_DOMAIN_SVA (__IOMMU_DOMAIN_SVA) +#define IOMMU_DOMAIN_NESTED (__IOMMU_DOMAIN_NESTED)
struct iommu_domain { unsigned type;
domain_alloc_user op already accepts user flags for domain allocation, add a parent domain pointer and a driver specific user data support as well.
Add a struct iommu_user_data as a bundle of data_ptr/data_len/type from an iommufd core uAPI structure. Make the user data opaque to the core, since a userspace driver must match the kernel driver. In the future, if drivers share some common parameter, there would be a generic parameter as well.
Define an enum iommu_hwpt_data_type (with IOMMU_HWPT_DATA_NONE type) for iommu drivers to add their own driver specific user data per hw_pagetable.
Signed-off-by: Lu Baolu baolu.lu@linux.intel.com Co-developed-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/intel/iommu.c | 4 +++- drivers/iommu/iommufd/hw_pagetable.c | 3 ++- drivers/iommu/iommufd/selftest.c | 4 +++- include/linux/iommu.h | 27 ++++++++++++++++++++++++--- include/uapi/linux/iommufd.h | 8 ++++++++ 5 files changed, 40 insertions(+), 6 deletions(-)
diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 017aed5813d8..8d78149e793e 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -4075,7 +4075,9 @@ static struct iommu_domain *intel_iommu_domain_alloc(unsigned type) }
static struct iommu_domain * -intel_iommu_domain_alloc_user(struct device *dev, u32 flags) +intel_iommu_domain_alloc_user(struct device *dev, u32 flags, + struct iommu_domain *parent, + const struct iommu_user_data *user_data) { struct iommu_domain *domain; struct intel_iommu *iommu; diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c index 8b3d2875d642..548079c8ca0a 100644 --- a/drivers/iommu/iommufd/hw_pagetable.c +++ b/drivers/iommu/iommufd/hw_pagetable.c @@ -96,7 +96,8 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, hwpt->ioas = ioas;
if (ops->domain_alloc_user) { - hwpt->domain = ops->domain_alloc_user(idev->dev, flags); + hwpt->domain = ops->domain_alloc_user(idev->dev, flags, + NULL, NULL); if (IS_ERR(hwpt->domain)) { rc = PTR_ERR(hwpt->domain); hwpt->domain = NULL; diff --git a/drivers/iommu/iommufd/selftest.c b/drivers/iommu/iommufd/selftest.c index fe7e3c7d933a..2bfe8dc9adc3 100644 --- a/drivers/iommu/iommufd/selftest.c +++ b/drivers/iommu/iommufd/selftest.c @@ -171,7 +171,9 @@ static struct iommu_domain *mock_domain_alloc(unsigned int iommu_domain_type) }
static struct iommu_domain * -mock_domain_alloc_user(struct device *dev, u32 flags) +mock_domain_alloc_user(struct device *dev, u32 flags, + struct iommu_domain *parent, + const struct iommu_user_data *user_data) { struct iommu_domain *domain;
diff --git a/include/linux/iommu.h b/include/linux/iommu.h index ce08a3c41446..cdb989b853a5 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -231,6 +231,21 @@ struct iommu_iotlb_gather { bool queued; };
+/** + * struct iommu_user_data - iommu driver specific user space data info + * @type: The data type of the user buffer + * @uptr: Pointer to the user buffer for copy_from_user() + * @len: The length of the user buffer in bytes + * + * A user space data is an uAPI that is defined in include/uapi/linux/iommufd.h + * @type, @uptr and @len should be just copied from an iommufd core uAPI struct. + */ +struct iommu_user_data { + unsigned int type; + void __user *uptr; + size_t len; +}; + /** * struct iommu_ops - iommu ops and capabilities * @capable: check capability @@ -245,8 +260,12 @@ struct iommu_iotlb_gather { * parameters as defined in include/uapi/linux/iommufd.h. * Unlike @domain_alloc, it is called only by IOMMUFD and * must fully initialize the new domain before return. - * Upon success, a domain is returned. Upon failure, - * ERR_PTR must be returned. + * Upon success, if the @user_data is valid and the @parent + * points to a kernel-managed domain, the new domain must be + * IOMMU_DOMAIN_NESTED type; otherwise, the @parent must be + * NULL while the @user_data can be optionally provided, the + * new domain must support __IOMMU_DOMAIN_PAGING. + * Upon failure, ERR_PTR must be returned. * @probe_device: Add device to iommu driver handling * @release_device: Remove device from iommu driver handling * @probe_finalize: Do final setup work after the device is added to an IOMMU @@ -279,7 +298,9 @@ struct iommu_ops {
/* Domain allocation and freeing by the iommu driver */ struct iommu_domain *(*domain_alloc)(unsigned iommu_domain_type); - struct iommu_domain *(*domain_alloc_user)(struct device *dev, u32 flags); + struct iommu_domain *(*domain_alloc_user)(struct device *dev, u32 flags, + struct iommu_domain *parent, + const struct iommu_user_data *user_data);
struct iommu_device *(*probe_device)(struct device *dev); void (*release_device)(struct device *dev); diff --git a/include/uapi/linux/iommufd.h b/include/uapi/linux/iommufd.h index be7a95042677..b459bb7ac579 100644 --- a/include/uapi/linux/iommufd.h +++ b/include/uapi/linux/iommufd.h @@ -356,6 +356,14 @@ enum iommufd_hwpt_alloc_flags { IOMMU_HWPT_ALLOC_NEST_PARENT = 1 << 0, };
+/** + * enum iommu_hwpt_data_type - IOMMU HWPT Data Type + * @IOMMU_HWPT_DATA_NONE: no data + */ +enum iommu_hwpt_data_type { + IOMMU_HWPT_DATA_NONE, +}; + /** * struct iommu_hwpt_alloc - ioctl(IOMMU_HWPT_ALLOC) * @size: sizeof(struct iommu_hwpt_alloc)
From: Jason Gunthorpe jgg@nvidia.com
To add a new IOMMUFD_OBJ_HWPT_NESTED, rename the HWPT object to confine it to PAGING hwpts/domains. The following patch will separate the hwpt structure as well.
Signed-off-by: Jason Gunthorpe jgg@nvidia.com Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/iommufd/device.c | 10 +++++----- drivers/iommu/iommufd/hw_pagetable.c | 2 +- drivers/iommu/iommufd/iommufd_private.h | 2 +- drivers/iommu/iommufd/main.c | 2 +- drivers/iommu/iommufd/selftest.c | 2 +- 5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c index e88fa73a45e6..a5f9f20b2a9b 100644 --- a/drivers/iommu/iommufd/device.c +++ b/drivers/iommu/iommufd/device.c @@ -579,7 +579,7 @@ static int iommufd_device_change_pt(struct iommufd_device *idev, u32 *pt_id, return PTR_ERR(pt_obj);
switch (pt_obj->type) { - case IOMMUFD_OBJ_HW_PAGETABLE: { + case IOMMUFD_OBJ_HWPT_PAGING: { struct iommufd_hw_pagetable *hwpt = container_of(pt_obj, struct iommufd_hw_pagetable, obj);
@@ -617,8 +617,8 @@ static int iommufd_device_change_pt(struct iommufd_device *idev, u32 *pt_id, /** * iommufd_device_attach - Connect a device to an iommu_domain * @idev: device to attach - * @pt_id: Input a IOMMUFD_OBJ_IOAS, or IOMMUFD_OBJ_HW_PAGETABLE - * Output the IOMMUFD_OBJ_HW_PAGETABLE ID + * @pt_id: Input a IOMMUFD_OBJ_IOAS, or IOMMUFD_OBJ_HWPT_PAGING + * Output the IOMMUFD_OBJ_HWPT_PAGING ID * * This connects the device to an iommu_domain, either automatically or manually * selected. Once this completes the device could do DMA. @@ -646,8 +646,8 @@ EXPORT_SYMBOL_NS_GPL(iommufd_device_attach, IOMMUFD); /** * iommufd_device_replace - Change the device's iommu_domain * @idev: device to change - * @pt_id: Input a IOMMUFD_OBJ_IOAS, or IOMMUFD_OBJ_HW_PAGETABLE - * Output the IOMMUFD_OBJ_HW_PAGETABLE ID + * @pt_id: Input a IOMMUFD_OBJ_IOAS, or IOMMUFD_OBJ_HWPT_PAGING + * Output the IOMMUFD_OBJ_HWPT_PAGING ID * * This is the same as:: * diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c index 548079c8ca0a..e100caa502ed 100644 --- a/drivers/iommu/iommufd/hw_pagetable.c +++ b/drivers/iommu/iommufd/hw_pagetable.c @@ -86,7 +86,7 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, if (flags && !ops->domain_alloc_user) return ERR_PTR(-EOPNOTSUPP);
- hwpt = iommufd_object_alloc(ictx, hwpt, IOMMUFD_OBJ_HW_PAGETABLE); + hwpt = iommufd_object_alloc(ictx, hwpt, IOMMUFD_OBJ_HWPT_PAGING); if (IS_ERR(hwpt)) return hwpt;
diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h index 3064997a0181..07ce35f09599 100644 --- a/drivers/iommu/iommufd/iommufd_private.h +++ b/drivers/iommu/iommufd/iommufd_private.h @@ -113,7 +113,7 @@ enum iommufd_object_type { IOMMUFD_OBJ_NONE, IOMMUFD_OBJ_ANY = IOMMUFD_OBJ_NONE, IOMMUFD_OBJ_DEVICE, - IOMMUFD_OBJ_HW_PAGETABLE, + IOMMUFD_OBJ_HWPT_PAGING, IOMMUFD_OBJ_IOAS, IOMMUFD_OBJ_ACCESS, #ifdef CONFIG_IOMMUFD_TEST diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c index e71523cbd0de..ad7c8c9ea46a 100644 --- a/drivers/iommu/iommufd/main.c +++ b/drivers/iommu/iommufd/main.c @@ -482,7 +482,7 @@ static const struct iommufd_object_ops iommufd_object_ops[] = { [IOMMUFD_OBJ_IOAS] = { .destroy = iommufd_ioas_destroy, }, - [IOMMUFD_OBJ_HW_PAGETABLE] = { + [IOMMUFD_OBJ_HWPT_PAGING] = { .destroy = iommufd_hw_pagetable_destroy, .abort = iommufd_hw_pagetable_abort, }, diff --git a/drivers/iommu/iommufd/selftest.c b/drivers/iommu/iommufd/selftest.c index 2bfe8dc9adc3..eca5b1d0919d 100644 --- a/drivers/iommu/iommufd/selftest.c +++ b/drivers/iommu/iommufd/selftest.c @@ -350,7 +350,7 @@ get_md_pagetable(struct iommufd_ucmd *ucmd, u32 mockpt_id, struct iommufd_object *obj;
obj = iommufd_get_object(ucmd->ictx, mockpt_id, - IOMMUFD_OBJ_HW_PAGETABLE); + IOMMUFD_OBJ_HWPT_PAGING); if (IS_ERR(obj)) return ERR_CAST(obj); hwpt = container_of(obj, struct iommufd_hw_pagetable, obj);
From: Jason Gunthorpe jgg@nvidia.com
Some of the configurations during the attach/replace() should only apply to IOMMUFD_OBJ_HWPT_PAGING. Once IOMMUFD_OBJ_HWPT_NESTED gets introduced in a following patch, keeping them unconditionally in the common routine will not work.
Wrap all of those PAGING-only configurations together into helpers. Do a hwpt_is_paging check whenever calling them or their fallback routines.
Also, move "num_devices++" closer to the place using it.
Signed-off-by: Jason Gunthorpe jgg@nvidia.com Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/iommufd/device.c | 137 ++++++++++++++++-------- drivers/iommu/iommufd/iommufd_private.h | 5 + 2 files changed, 99 insertions(+), 43 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c index a5f9f20b2a9b..3414276bbd15 100644 --- a/drivers/iommu/iommufd/device.c +++ b/drivers/iommu/iommufd/device.c @@ -325,6 +325,35 @@ static int iommufd_group_setup_msi(struct iommufd_group *igroup, return 0; }
+static int iommufd_hwpt_paging_attach(struct iommufd_hw_pagetable *hwpt, + struct iommufd_device *idev) +{ + int rc; + + lockdep_assert_held(&idev->igroup->lock); + + /* 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, + &idev->igroup->sw_msi_start); + if (rc) + return rc; + + if (list_empty(&idev->igroup->device_list)) { + rc = iommufd_group_setup_msi(idev->igroup, hwpt); + if (rc) { + iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); + return rc; + } + } + return 0; +} + int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, struct iommufd_device *idev) { @@ -337,18 +366,12 @@ int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, goto err_unlock; }
- /* Try to upgrade the domain we have */ - if (idev->enforce_cache_coherency) { - rc = iommufd_hw_pagetable_enforce_cc(hwpt); + if (hwpt_is_paging(hwpt)) { + rc = iommufd_hwpt_paging_attach(hwpt, idev); if (rc) goto err_unlock; }
- rc = iopt_table_enforce_dev_resv_regions(&hwpt->ioas->iopt, idev->dev, - &idev->igroup->sw_msi_start); - if (rc) - goto err_unlock; - /* * Only attach to the group once for the first device that is in the * group. All the other devices will follow this attachment. The user @@ -357,10 +380,6 @@ int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, * attachment. */ if (list_empty(&idev->igroup->device_list)) { - rc = iommufd_group_setup_msi(idev->igroup, hwpt); - if (rc) - goto err_unresv; - rc = iommu_attach_group(hwpt->domain, idev->igroup->group); if (rc) goto err_unresv; @@ -371,7 +390,8 @@ int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, mutex_unlock(&idev->igroup->lock); return 0; err_unresv: - iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); + if (hwpt_is_paging(hwpt)) + iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); err_unlock: mutex_unlock(&idev->igroup->lock); return rc; @@ -388,7 +408,8 @@ iommufd_hw_pagetable_detach(struct iommufd_device *idev) iommu_detach_group(hwpt->domain, idev->igroup->group); idev->igroup->hwpt = NULL; } - iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); + if (hwpt_is_paging(hwpt)) + iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); mutex_unlock(&idev->igroup->lock);
/* Caller must destroy hwpt */ @@ -407,40 +428,36 @@ iommufd_device_do_attach(struct iommufd_device *idev, return NULL; }
-static struct iommufd_hw_pagetable * -iommufd_device_do_replace(struct iommufd_device *idev, - struct iommufd_hw_pagetable *hwpt) +static void iommufd_group_remove_reserved_iova(struct iommufd_group *igroup, + struct iommufd_hw_pagetable *hwpt) { - struct iommufd_group *igroup = idev->igroup; - struct iommufd_hw_pagetable *old_hwpt; - unsigned int num_devices = 0; struct iommufd_device *cur; - int rc;
- mutex_lock(&idev->igroup->lock); + lockdep_assert_held(&igroup->lock);
- if (igroup->hwpt == NULL) { - rc = -EINVAL; - goto err_unlock; - } + list_for_each_entry(cur, &igroup->device_list, group_item) + iopt_remove_reserved_iova(&hwpt->ioas->iopt, cur->dev); +}
- if (hwpt == igroup->hwpt) { - mutex_unlock(&idev->igroup->lock); - return NULL; - } +static int iommufd_group_do_replace_paging(struct iommufd_group *igroup, + struct iommufd_hw_pagetable *hwpt) +{ + struct iommufd_hw_pagetable *old_hwpt = igroup->hwpt; + struct iommufd_device *cur; + int rc; + + lockdep_assert_held(&igroup->lock);
/* Try to upgrade the domain we have */ list_for_each_entry(cur, &igroup->device_list, group_item) { - num_devices++; if (cur->enforce_cache_coherency) { rc = iommufd_hw_pagetable_enforce_cc(hwpt); if (rc) - goto err_unlock; + return rc; } }
- old_hwpt = igroup->hwpt; - if (hwpt->ioas != old_hwpt->ioas) { + if (hwpt_is_paging(old_hwpt) && hwpt->ioas != old_hwpt->ioas) { list_for_each_entry(cur, &igroup->device_list, group_item) { rc = iopt_table_enforce_dev_resv_regions( &hwpt->ioas->iopt, cur->dev, NULL); @@ -448,23 +465,57 @@ iommufd_device_do_replace(struct iommufd_device *idev, goto err_unresv; } } - - rc = iommufd_group_setup_msi(idev->igroup, hwpt); + rc = iommufd_group_setup_msi(igroup, hwpt); if (rc) goto err_unresv; + return 0; + +err_unresv: + iommufd_group_remove_reserved_iova(igroup, hwpt); + return rc; +} + +static struct iommufd_hw_pagetable * +iommufd_device_do_replace(struct iommufd_device *idev, + struct iommufd_hw_pagetable *hwpt) +{ + struct iommufd_group *igroup = idev->igroup; + struct iommufd_hw_pagetable *old_hwpt; + unsigned int num_devices = 0; + struct iommufd_device *cur; + int rc; + + mutex_lock(&idev->igroup->lock); + + if (igroup->hwpt == NULL) { + rc = -EINVAL; + goto err_unlock; + } + + if (hwpt == igroup->hwpt) { + mutex_unlock(&idev->igroup->lock); + return NULL; + } + + if (hwpt_is_paging(hwpt)) { + rc = iommufd_group_do_replace_paging(igroup, hwpt); + if (rc) + goto err_unlock; + }
rc = iommu_group_replace_domain(igroup->group, hwpt->domain); if (rc) goto err_unresv;
- if (hwpt->ioas != old_hwpt->ioas) { - list_for_each_entry(cur, &igroup->device_list, group_item) - iopt_remove_reserved_iova(&old_hwpt->ioas->iopt, - cur->dev); - } + old_hwpt = igroup->hwpt; + if (hwpt_is_paging(old_hwpt) && + (!hwpt_is_paging(hwpt) || hwpt->ioas != old_hwpt->ioas)) + iommufd_group_remove_reserved_iova(igroup, old_hwpt);
igroup->hwpt = hwpt;
+ list_for_each_entry(cur, &igroup->device_list, group_item) + num_devices++; /* * Move the refcounts held by the device_list to the new hwpt. Retain a * refcount for this thread as the caller will free it. @@ -478,8 +529,8 @@ iommufd_device_do_replace(struct iommufd_device *idev, /* Caller must destroy old_hwpt */ return old_hwpt; err_unresv: - list_for_each_entry(cur, &igroup->device_list, group_item) - iopt_remove_reserved_iova(&hwpt->ioas->iopt, cur->dev); + if (hwpt_is_paging(hwpt)) + iommufd_group_remove_reserved_iova(igroup, hwpt); err_unlock: mutex_unlock(&idev->igroup->lock); return ERR_PTR(rc); diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h index 07ce35f09599..6244ffddee6e 100644 --- a/drivers/iommu/iommufd/iommufd_private.h +++ b/drivers/iommu/iommufd/iommufd_private.h @@ -240,6 +240,11 @@ struct iommufd_hw_pagetable { struct list_head hwpt_item; };
+static inline bool hwpt_is_paging(struct iommufd_hw_pagetable *hwpt) +{ + return hwpt->obj.type == IOMMUFD_OBJ_HWPT_PAGING; +} + struct iommufd_hw_pagetable * iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, struct iommufd_device *idev, u32 flags,
From: Nicolin Chen nicolinc@nvidia.com
To prepare for IOMMUFD_OBJ_HWPT_NESTED, derive struct iommufd_hwpt_paging from struct iommufd_hw_pagetable, by leaving the common members in struct iommufd_hw_pagetable. Add a __iommufd_object_alloc and to_hwpt_paging() helpers for the new structure.
Then, update "hwpt" to "hwpt_paging" throughout the files, accordingly.
Suggested-by: Jason Gunthorpe jgg@nvidia.com Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/iommufd/device.c | 78 ++++++++++++-------- drivers/iommu/iommufd/hw_pagetable.c | 98 ++++++++++++++----------- drivers/iommu/iommufd/iommufd_private.h | 37 +++++++--- drivers/iommu/iommufd/main.c | 4 +- drivers/iommu/iommufd/vfio_compat.c | 6 +- 5 files changed, 131 insertions(+), 92 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c index 3414276bbd15..cc9f5d0e5b7e 100644 --- a/drivers/iommu/iommufd/device.c +++ b/drivers/iommu/iommufd/device.c @@ -293,7 +293,7 @@ u32 iommufd_device_to_id(struct iommufd_device *idev) EXPORT_SYMBOL_NS_GPL(iommufd_device_to_id, IOMMUFD);
static int iommufd_group_setup_msi(struct iommufd_group *igroup, - struct iommufd_hw_pagetable *hwpt) + struct iommufd_hwpt_paging *hwpt_paging) { phys_addr_t sw_msi_start = igroup->sw_msi_start; int rc; @@ -311,8 +311,9 @@ static int iommufd_group_setup_msi(struct iommufd_group *igroup, * matches what the IRQ layer actually expects in a newly created * domain. */ - if (sw_msi_start != PHYS_ADDR_MAX && !hwpt->msi_cookie) { - rc = iommu_get_msi_cookie(hwpt->domain, sw_msi_start); + if (sw_msi_start != PHYS_ADDR_MAX && !hwpt_paging->msi_cookie) { + rc = iommu_get_msi_cookie(hwpt_paging->common.domain, + sw_msi_start); if (rc) return rc;
@@ -320,12 +321,12 @@ static int iommufd_group_setup_msi(struct iommufd_group *igroup, * iommu_get_msi_cookie() can only be called once per domain, * it returns -EBUSY on later calls. */ - hwpt->msi_cookie = true; + hwpt_paging->msi_cookie = true; } return 0; }
-static int iommufd_hwpt_paging_attach(struct iommufd_hw_pagetable *hwpt, +static int iommufd_hwpt_paging_attach(struct iommufd_hwpt_paging *hwpt_paging, struct iommufd_device *idev) { int rc; @@ -334,20 +335,22 @@ static int iommufd_hwpt_paging_attach(struct iommufd_hw_pagetable *hwpt,
/* Try to upgrade the domain we have */ if (idev->enforce_cache_coherency) { - rc = iommufd_hw_pagetable_enforce_cc(hwpt); + rc = iommufd_hwpt_paging_enforce_cc(hwpt_paging); if (rc) return rc; }
- rc = iopt_table_enforce_dev_resv_regions(&hwpt->ioas->iopt, idev->dev, + rc = iopt_table_enforce_dev_resv_regions(&hwpt_paging->ioas->iopt, + idev->dev, &idev->igroup->sw_msi_start); if (rc) return rc;
if (list_empty(&idev->igroup->device_list)) { - rc = iommufd_group_setup_msi(idev->igroup, hwpt); + rc = iommufd_group_setup_msi(idev->igroup, hwpt_paging); if (rc) { - iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); + iopt_remove_reserved_iova(&hwpt_paging->ioas->iopt, + idev->dev); return rc; } } @@ -367,7 +370,7 @@ int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, }
if (hwpt_is_paging(hwpt)) { - rc = iommufd_hwpt_paging_attach(hwpt, idev); + rc = iommufd_hwpt_paging_attach(to_hwpt_paging(hwpt), idev); if (rc) goto err_unlock; } @@ -391,7 +394,8 @@ int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, return 0; err_unresv: if (hwpt_is_paging(hwpt)) - iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); + iopt_remove_reserved_iova(&to_hwpt_paging(hwpt)->ioas->iopt, + idev->dev); err_unlock: mutex_unlock(&idev->igroup->lock); return rc; @@ -409,7 +413,8 @@ iommufd_hw_pagetable_detach(struct iommufd_device *idev) idev->igroup->hwpt = NULL; } if (hwpt_is_paging(hwpt)) - iopt_remove_reserved_iova(&hwpt->ioas->iopt, idev->dev); + iopt_remove_reserved_iova(&to_hwpt_paging(hwpt)->ioas->iopt, + idev->dev); mutex_unlock(&idev->igroup->lock);
/* Caller must destroy hwpt */ @@ -428,19 +433,20 @@ iommufd_device_do_attach(struct iommufd_device *idev, return NULL; }
-static void iommufd_group_remove_reserved_iova(struct iommufd_group *igroup, - struct iommufd_hw_pagetable *hwpt) +static void +iommufd_group_remove_reserved_iova(struct iommufd_group *igroup, + struct iommufd_hwpt_paging *hwpt_paging) { struct iommufd_device *cur;
lockdep_assert_held(&igroup->lock);
list_for_each_entry(cur, &igroup->device_list, group_item) - iopt_remove_reserved_iova(&hwpt->ioas->iopt, cur->dev); + iopt_remove_reserved_iova(&hwpt_paging->ioas->iopt, cur->dev); }
static int iommufd_group_do_replace_paging(struct iommufd_group *igroup, - struct iommufd_hw_pagetable *hwpt) + struct iommufd_hwpt_paging *hwpt_paging) { struct iommufd_hw_pagetable *old_hwpt = igroup->hwpt; struct iommufd_device *cur; @@ -451,27 +457,28 @@ static int iommufd_group_do_replace_paging(struct iommufd_group *igroup, /* Try to upgrade the domain we have */ list_for_each_entry(cur, &igroup->device_list, group_item) { if (cur->enforce_cache_coherency) { - rc = iommufd_hw_pagetable_enforce_cc(hwpt); + rc = iommufd_hwpt_paging_enforce_cc(hwpt_paging); if (rc) return rc; } }
- if (hwpt_is_paging(old_hwpt) && hwpt->ioas != old_hwpt->ioas) { + if (hwpt_is_paging(old_hwpt) && + hwpt_paging->ioas != to_hwpt_paging(old_hwpt)->ioas) { list_for_each_entry(cur, &igroup->device_list, group_item) { rc = iopt_table_enforce_dev_resv_regions( - &hwpt->ioas->iopt, cur->dev, NULL); + &hwpt_paging->ioas->iopt, cur->dev, NULL); if (rc) goto err_unresv; } } - rc = iommufd_group_setup_msi(igroup, hwpt); + rc = iommufd_group_setup_msi(igroup, hwpt_paging); if (rc) goto err_unresv; return 0;
err_unresv: - iommufd_group_remove_reserved_iova(igroup, hwpt); + iommufd_group_remove_reserved_iova(igroup, hwpt_paging); return rc; }
@@ -498,7 +505,8 @@ iommufd_device_do_replace(struct iommufd_device *idev, }
if (hwpt_is_paging(hwpt)) { - rc = iommufd_group_do_replace_paging(igroup, hwpt); + rc = iommufd_group_do_replace_paging(igroup, + to_hwpt_paging(hwpt)); if (rc) goto err_unlock; } @@ -509,8 +517,10 @@ iommufd_device_do_replace(struct iommufd_device *idev,
old_hwpt = igroup->hwpt; if (hwpt_is_paging(old_hwpt) && - (!hwpt_is_paging(hwpt) || hwpt->ioas != old_hwpt->ioas)) - iommufd_group_remove_reserved_iova(igroup, old_hwpt); + (!hwpt_is_paging(hwpt) || + to_hwpt_paging(hwpt)->ioas != to_hwpt_paging(old_hwpt)->ioas)) + iommufd_group_remove_reserved_iova(igroup, + to_hwpt_paging(old_hwpt));
igroup->hwpt = hwpt;
@@ -530,7 +540,8 @@ iommufd_device_do_replace(struct iommufd_device *idev, return old_hwpt; err_unresv: if (hwpt_is_paging(hwpt)) - iommufd_group_remove_reserved_iova(igroup, hwpt); + iommufd_group_remove_reserved_iova(igroup, + to_hwpt_paging(old_hwpt)); err_unlock: mutex_unlock(&idev->igroup->lock); return ERR_PTR(rc); @@ -558,6 +569,7 @@ iommufd_device_auto_get_domain(struct iommufd_device *idev, */ bool immediate_attach = do_attach == iommufd_device_do_attach; struct iommufd_hw_pagetable *destroy_hwpt; + struct iommufd_hwpt_paging *hwpt_paging; struct iommufd_hw_pagetable *hwpt;
/* @@ -566,10 +578,11 @@ iommufd_device_auto_get_domain(struct iommufd_device *idev, * other. */ mutex_lock(&ioas->mutex); - list_for_each_entry(hwpt, &ioas->hwpt_list, hwpt_item) { - if (!hwpt->auto_domain) + list_for_each_entry(hwpt_paging, &ioas->hwpt_list, hwpt_item) { + if (!hwpt_paging->auto_domain) continue;
+ hwpt = &hwpt_paging->common; if (!iommufd_lock_obj(&hwpt->obj)) continue; destroy_hwpt = (*do_attach)(idev, hwpt); @@ -590,12 +603,13 @@ iommufd_device_auto_get_domain(struct iommufd_device *idev, goto out_unlock; }
- hwpt = iommufd_hw_pagetable_alloc(idev->ictx, ioas, idev, - 0, immediate_attach); - if (IS_ERR(hwpt)) { - destroy_hwpt = ERR_CAST(hwpt); + hwpt_paging = iommufd_hwpt_paging_alloc(idev->ictx, ioas, idev, + 0, immediate_attach); + if (IS_ERR(hwpt_paging)) { + destroy_hwpt = ERR_CAST(hwpt_paging); goto out_unlock; } + hwpt = &hwpt_paging->common;
if (!immediate_attach) { destroy_hwpt = (*do_attach)(idev, hwpt); @@ -605,7 +619,7 @@ iommufd_device_auto_get_domain(struct iommufd_device *idev, destroy_hwpt = NULL; }
- hwpt->auto_domain = true; + hwpt_paging->auto_domain = true; *pt_id = hwpt->obj.id;
iommufd_object_finalize(idev->ictx, &hwpt->obj); diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c index e100caa502ed..52df5b2bb329 100644 --- a/drivers/iommu/iommufd/hw_pagetable.c +++ b/drivers/iommu/iommufd/hw_pagetable.c @@ -8,56 +8,60 @@ #include "../iommu-priv.h" #include "iommufd_private.h"
-void iommufd_hw_pagetable_destroy(struct iommufd_object *obj) +void iommufd_hwpt_paging_destroy(struct iommufd_object *obj) { - struct iommufd_hw_pagetable *hwpt = - container_of(obj, struct iommufd_hw_pagetable, obj); + struct iommufd_hwpt_paging *hwpt_paging = + container_of(obj, struct iommufd_hwpt_paging, common.obj);
- if (!list_empty(&hwpt->hwpt_item)) { - mutex_lock(&hwpt->ioas->mutex); - list_del(&hwpt->hwpt_item); - mutex_unlock(&hwpt->ioas->mutex); + if (!list_empty(&hwpt_paging->hwpt_item)) { + mutex_lock(&hwpt_paging->ioas->mutex); + list_del(&hwpt_paging->hwpt_item); + mutex_unlock(&hwpt_paging->ioas->mutex);
- iopt_table_remove_domain(&hwpt->ioas->iopt, hwpt->domain); + iopt_table_remove_domain(&hwpt_paging->ioas->iopt, + hwpt_paging->common.domain); }
- if (hwpt->domain) - iommu_domain_free(hwpt->domain); + if (hwpt_paging->common.domain) + iommu_domain_free(hwpt_paging->common.domain);
- refcount_dec(&hwpt->ioas->obj.users); + refcount_dec(&hwpt_paging->ioas->obj.users); }
-void iommufd_hw_pagetable_abort(struct iommufd_object *obj) +void iommufd_hwpt_paging_abort(struct iommufd_object *obj) { - struct iommufd_hw_pagetable *hwpt = - container_of(obj, struct iommufd_hw_pagetable, obj); + struct iommufd_hwpt_paging *hwpt_paging = + container_of(obj, struct iommufd_hwpt_paging, common.obj);
/* The ioas->mutex must be held until finalize is called. */ - lockdep_assert_held(&hwpt->ioas->mutex); + lockdep_assert_held(&hwpt_paging->ioas->mutex);
- if (!list_empty(&hwpt->hwpt_item)) { - list_del_init(&hwpt->hwpt_item); - iopt_table_remove_domain(&hwpt->ioas->iopt, hwpt->domain); + if (!list_empty(&hwpt_paging->hwpt_item)) { + list_del_init(&hwpt_paging->hwpt_item); + iopt_table_remove_domain(&hwpt_paging->ioas->iopt, + hwpt_paging->common.domain); } - iommufd_hw_pagetable_destroy(obj); + iommufd_hwpt_paging_destroy(obj); }
-int iommufd_hw_pagetable_enforce_cc(struct iommufd_hw_pagetable *hwpt) +int iommufd_hwpt_paging_enforce_cc(struct iommufd_hwpt_paging *hwpt_paging) { - if (hwpt->enforce_cache_coherency) + struct iommu_domain *paging_domain = hwpt_paging->common.domain; + + if (hwpt_paging->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) + if (paging_domain->ops->enforce_cache_coherency) + hwpt_paging->enforce_cache_coherency = + paging_domain->ops->enforce_cache_coherency( + paging_domain); + if (!hwpt_paging->enforce_cache_coherency) return -EINVAL; return 0; }
/** - * iommufd_hw_pagetable_alloc() - Get an iommu_domain for a device + * iommufd_hwpt_paging_alloc() - Get a PAGING iommu_domain for a device * @ictx: iommufd context * @ioas: IOAS to associate the domain with * @idev: Device to get an iommu_domain for @@ -72,12 +76,13 @@ int iommufd_hw_pagetable_enforce_cc(struct iommufd_hw_pagetable *hwpt) * iommufd_object_abort_and_destroy() or iommufd_object_finalize() is called on * the returned hwpt. */ -struct iommufd_hw_pagetable * -iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, - struct iommufd_device *idev, u32 flags, - bool immediate_attach) +struct iommufd_hwpt_paging * +iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, + struct iommufd_device *idev, u32 flags, + bool immediate_attach) { const struct iommu_ops *ops = dev_iommu_ops(idev->dev); + struct iommufd_hwpt_paging *hwpt_paging; struct iommufd_hw_pagetable *hwpt; int rc;
@@ -86,14 +91,17 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, if (flags && !ops->domain_alloc_user) return ERR_PTR(-EOPNOTSUPP);
- hwpt = iommufd_object_alloc(ictx, hwpt, IOMMUFD_OBJ_HWPT_PAGING); - if (IS_ERR(hwpt)) - return hwpt; + hwpt_paging = __iommufd_object_alloc(ictx, hwpt_paging, + IOMMUFD_OBJ_HWPT_PAGING, + common.obj); + if (IS_ERR(hwpt_paging)) + return ERR_CAST(hwpt_paging); + hwpt = &hwpt_paging->common;
- INIT_LIST_HEAD(&hwpt->hwpt_item); + INIT_LIST_HEAD(&hwpt_paging->hwpt_item); /* Pairs with iommufd_hw_pagetable_destroy() */ refcount_inc(&ioas->obj.users); - hwpt->ioas = ioas; + hwpt_paging->ioas = ioas;
if (ops->domain_alloc_user) { hwpt->domain = ops->domain_alloc_user(idev->dev, flags, @@ -119,7 +127,7 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, * a new domain. */ if (idev->enforce_cache_coherency) { - rc = iommufd_hw_pagetable_enforce_cc(hwpt); + rc = iommufd_hwpt_paging_enforce_cc(hwpt_paging); if (WARN_ON(rc)) goto out_abort; } @@ -136,11 +144,11 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, goto out_abort; }
- rc = iopt_table_add_domain(&hwpt->ioas->iopt, hwpt->domain); + rc = iopt_table_add_domain(&ioas->iopt, hwpt->domain); if (rc) goto out_detach; - list_add_tail(&hwpt->hwpt_item, &hwpt->ioas->hwpt_list); - return hwpt; + list_add_tail(&hwpt_paging->hwpt_item, &ioas->hwpt_list); + return hwpt_paging;
out_detach: if (immediate_attach) @@ -153,6 +161,7 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) { struct iommu_hwpt_alloc *cmd = ucmd->cmd; + struct iommufd_hwpt_paging *hwpt_paging; struct iommufd_hw_pagetable *hwpt; struct iommufd_device *idev; struct iommufd_ioas *ioas; @@ -172,12 +181,13 @@ int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) }
mutex_lock(&ioas->mutex); - hwpt = iommufd_hw_pagetable_alloc(ucmd->ictx, ioas, - idev, cmd->flags, false); - if (IS_ERR(hwpt)) { - rc = PTR_ERR(hwpt); + hwpt_paging = iommufd_hwpt_paging_alloc(ucmd->ictx, ioas, idev, + cmd->flags, false); + if (IS_ERR(hwpt_paging)) { + rc = PTR_ERR(hwpt_paging); goto out_unlock; } + hwpt = &hwpt_paging->common;
cmd->out_hwpt_id = hwpt->obj.id; rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd)); diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h index 6244ffddee6e..a7ed384717a5 100644 --- a/drivers/iommu/iommufd/iommufd_private.h +++ b/drivers/iommu/iommufd/iommufd_private.h @@ -171,7 +171,7 @@ struct iommufd_object *_iommufd_object_alloc(struct iommufd_ctx *ictx, size_t size, enum iommufd_object_type type);
-#define iommufd_object_alloc(ictx, ptr, type) \ +#define __iommufd_object_alloc(ictx, ptr, type, obj) \ container_of(_iommufd_object_alloc( \ ictx, \ sizeof(*(ptr)) + BUILD_BUG_ON_ZERO( \ @@ -180,6 +180,9 @@ struct iommufd_object *_iommufd_object_alloc(struct iommufd_ctx *ictx, type), \ typeof(*(ptr)), obj)
+#define iommufd_object_alloc(ictx, ptr, type) \ + __iommufd_object_alloc(ictx, ptr, type, obj) + /* * The IO Address Space (IOAS) pagetable is a virtual page table backed by the * io_pagetable object. It is a user controlled mapping of IOVA -> PFNs. The @@ -231,8 +234,12 @@ int iommufd_vfio_ioas(struct iommufd_ucmd *ucmd); */ struct iommufd_hw_pagetable { struct iommufd_object obj; - struct iommufd_ioas *ioas; struct iommu_domain *domain; +}; + +struct iommufd_hwpt_paging { + struct iommufd_hw_pagetable common; + struct iommufd_ioas *ioas; bool auto_domain : 1; bool enforce_cache_coherency : 1; bool msi_cookie : 1; @@ -245,24 +252,32 @@ static inline bool hwpt_is_paging(struct iommufd_hw_pagetable *hwpt) return hwpt->obj.type == IOMMUFD_OBJ_HWPT_PAGING; }
-struct iommufd_hw_pagetable * -iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, - struct iommufd_device *idev, u32 flags, - bool immediate_attach); -int iommufd_hw_pagetable_enforce_cc(struct iommufd_hw_pagetable *hwpt); +static inline struct iommufd_hwpt_paging * +to_hwpt_paging(struct iommufd_hw_pagetable *hwpt) +{ + return container_of(hwpt, struct iommufd_hwpt_paging, common); +} + +struct iommufd_hwpt_paging * +iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, + struct iommufd_device *idev, u32 flags, + bool immediate_attach); +int iommufd_hwpt_paging_enforce_cc(struct iommufd_hwpt_paging *hwpt_paging); int iommufd_hw_pagetable_attach(struct iommufd_hw_pagetable *hwpt, struct iommufd_device *idev); struct iommufd_hw_pagetable * iommufd_hw_pagetable_detach(struct iommufd_device *idev); -void iommufd_hw_pagetable_destroy(struct iommufd_object *obj); -void iommufd_hw_pagetable_abort(struct iommufd_object *obj); +void iommufd_hwpt_paging_destroy(struct iommufd_object *obj); +void iommufd_hwpt_paging_abort(struct iommufd_object *obj); int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd);
static inline void iommufd_hw_pagetable_put(struct iommufd_ctx *ictx, struct iommufd_hw_pagetable *hwpt) { - lockdep_assert_not_held(&hwpt->ioas->mutex); - if (hwpt->auto_domain) + struct iommufd_hwpt_paging *hwpt_paging = to_hwpt_paging(hwpt); + + lockdep_assert_not_held(&hwpt_paging->ioas->mutex); + if (hwpt_paging->auto_domain) iommufd_object_deref_user(ictx, &hwpt->obj); else refcount_dec(&hwpt->obj.users); diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c index ad7c8c9ea46a..b4a24ef041e6 100644 --- a/drivers/iommu/iommufd/main.c +++ b/drivers/iommu/iommufd/main.c @@ -483,8 +483,8 @@ static const struct iommufd_object_ops iommufd_object_ops[] = { .destroy = iommufd_ioas_destroy, }, [IOMMUFD_OBJ_HWPT_PAGING] = { - .destroy = iommufd_hw_pagetable_destroy, - .abort = iommufd_hw_pagetable_abort, + .destroy = iommufd_hwpt_paging_destroy, + .abort = iommufd_hwpt_paging_abort, }, #ifdef CONFIG_IOMMUFD_TEST [IOMMUFD_OBJ_SELFTEST] = { diff --git a/drivers/iommu/iommufd/vfio_compat.c b/drivers/iommu/iommufd/vfio_compat.c index 6c810bf80f99..538fbf76354d 100644 --- a/drivers/iommu/iommufd/vfio_compat.c +++ b/drivers/iommu/iommufd/vfio_compat.c @@ -255,7 +255,7 @@ static int iommufd_vfio_unmap_dma(struct iommufd_ctx *ictx, unsigned int cmd,
static int iommufd_vfio_cc_iommu(struct iommufd_ctx *ictx) { - struct iommufd_hw_pagetable *hwpt; + struct iommufd_hwpt_paging *hwpt_paging; struct iommufd_ioas *ioas; int rc = 1;
@@ -264,8 +264,8 @@ static int iommufd_vfio_cc_iommu(struct iommufd_ctx *ictx) return PTR_ERR(ioas);
mutex_lock(&ioas->mutex); - list_for_each_entry(hwpt, &ioas->hwpt_list, hwpt_item) { - if (!hwpt->enforce_cache_coherency) { + list_for_each_entry(hwpt_paging, &ioas->hwpt_list, hwpt_item) { + if (!hwpt_paging->enforce_cache_coherency) { rc = 0; break; }
From: Nicolin Chen nicolinc@nvidia.com
Allow iommufd_hwpt_alloc() to have a common routine but jump to different allocators corresponding to different user input pt_obj types, either an IOMMUFD_OBJ_IOAS for a PAGING hwpt or an IOMMUFD_OBJ_HWPT_PAGING as the parent for a NESTED hwpt.
Also, move the "flags" validation to the hwpt allocator (paging), so that later the hwpt_nested allocator can do its own separate flags validation.
Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/iommufd/hw_pagetable.c | 43 ++++++++++++++++++---------- 1 file changed, 28 insertions(+), 15 deletions(-)
diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c index 52df5b2bb329..24cd84a5fe8f 100644 --- a/drivers/iommu/iommufd/hw_pagetable.c +++ b/drivers/iommu/iommufd/hw_pagetable.c @@ -81,6 +81,7 @@ iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, struct iommufd_device *idev, u32 flags, bool immediate_attach) { + const u32 valid_flags = IOMMU_HWPT_ALLOC_NEST_PARENT; const struct iommu_ops *ops = dev_iommu_ops(idev->dev); struct iommufd_hwpt_paging *hwpt_paging; struct iommufd_hw_pagetable *hwpt; @@ -90,6 +91,8 @@ iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas,
if (flags && !ops->domain_alloc_user) return ERR_PTR(-EOPNOTSUPP); + if (flags & ~valid_flags) + return ERR_PTR(-EOPNOTSUPP);
hwpt_paging = __iommufd_object_alloc(ictx, hwpt_paging, IOMMUFD_OBJ_HWPT_PAGING, @@ -161,33 +164,41 @@ iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) { struct iommu_hwpt_alloc *cmd = ucmd->cmd; - struct iommufd_hwpt_paging *hwpt_paging; struct iommufd_hw_pagetable *hwpt; + struct iommufd_ioas *ioas = NULL; + struct iommufd_object *pt_obj; struct iommufd_device *idev; - struct iommufd_ioas *ioas; int rc;
- if ((cmd->flags & (~IOMMU_HWPT_ALLOC_NEST_PARENT)) || cmd->__reserved) + if (cmd->__reserved) return -EOPNOTSUPP;
idev = iommufd_get_device(ucmd, cmd->dev_id); if (IS_ERR(idev)) return PTR_ERR(idev);
- ioas = iommufd_get_ioas(ucmd->ictx, cmd->pt_id); - if (IS_ERR(ioas)) { - rc = PTR_ERR(ioas); + pt_obj = iommufd_get_object(ucmd->ictx, cmd->pt_id, IOMMUFD_OBJ_ANY); + if (IS_ERR(pt_obj)) { + rc = -EINVAL; goto out_put_idev; }
- mutex_lock(&ioas->mutex); - hwpt_paging = iommufd_hwpt_paging_alloc(ucmd->ictx, ioas, idev, - cmd->flags, false); - if (IS_ERR(hwpt_paging)) { - rc = PTR_ERR(hwpt_paging); - goto out_unlock; + if (pt_obj->type == IOMMUFD_OBJ_IOAS) { + struct iommufd_hwpt_paging *hwpt_paging; + + ioas = container_of(pt_obj, struct iommufd_ioas, obj); + mutex_lock(&ioas->mutex); + hwpt_paging = iommufd_hwpt_paging_alloc(ucmd->ictx, ioas, idev, + cmd->flags, false); + if (IS_ERR(hwpt_paging)) { + rc = PTR_ERR(hwpt_paging); + goto out_unlock; + } + hwpt = &hwpt_paging->common; + } else { + rc = -EINVAL; + goto out_put_pt; } - hwpt = &hwpt_paging->common;
cmd->out_hwpt_id = hwpt->obj.id; rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd)); @@ -199,8 +210,10 @@ int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) out_hwpt: iommufd_object_abort_and_destroy(ucmd->ictx, &hwpt->obj); out_unlock: - mutex_unlock(&ioas->mutex); - iommufd_put_object(&ioas->obj); + if (ioas) + mutex_unlock(&ioas->mutex); +out_put_pt: + iommufd_put_object(pt_obj); out_put_idev: iommufd_put_object(&idev->obj); return rc;
From: Nicolin Chen nicolinc@nvidia.com
IOMMU_HWPT_ALLOC already supports iommu_domain allocation for usersapce. But it can only allocate a hw_pagetable that associates to a given IOAS, i.e. only a kernel-managed hw_pagetable of IOMMUFD_OBJ_HWPT_PAGING type.
IOMMU drivers can now support user-managed hw_pagetables, for two-stage translation use cases that require user data input from the user space.
Add a new IOMMUFD_OBJ_HWPT_NESTED type with its abort/destroy(). Pair it with a new iommufd_hwpt_nested structure and its to_hwpt_nested() helper. Update the to_hwpt_paging() helper, so a NESTED-type hw_pagetable can be handled in the callers, for example iommufd_hw_pagetable_enforce_rr().
Screen the inputs including the parent PAGING-type hw_pagetable that has a need of a new nest_parent flag in the iommufd_hwpt_paging structure.
Extend the IOMMU_HWPT_ALLOC ioctl to accept an IOMMU driver specific data input. Also, update the @pt_id to accept hwpt_id too besides an ioas_id. Then, use them to allocate a hw_pagetable of IOMMUFD_OBJ_HWPT_NESTED type using the iommufd_hw_pagetable_alloc_nested() allocator.
Signed-off-by: Nicolin Chen nicolinc@nvidia.com Co-developed-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/iommufd/device.c | 1 + drivers/iommu/iommufd/hw_pagetable.c | 105 ++++++++++++++++++++++++ drivers/iommu/iommufd/iommufd_private.h | 29 +++++-- drivers/iommu/iommufd/main.c | 4 + include/uapi/linux/iommufd.h | 23 +++++- 5 files changed, 154 insertions(+), 8 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c index cc9f5d0e5b7e..d687f8bb5c44 100644 --- a/drivers/iommu/iommufd/device.c +++ b/drivers/iommu/iommufd/device.c @@ -644,6 +644,7 @@ static int iommufd_device_change_pt(struct iommufd_device *idev, u32 *pt_id, return PTR_ERR(pt_obj);
switch (pt_obj->type) { + case IOMMUFD_OBJ_HWPT_NESTED: case IOMMUFD_OBJ_HWPT_PAGING: { struct iommufd_hw_pagetable *hwpt = container_of(pt_obj, struct iommufd_hw_pagetable, obj); diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c index 24cd84a5fe8f..362c694aac9b 100644 --- a/drivers/iommu/iommufd/hw_pagetable.c +++ b/drivers/iommu/iommufd/hw_pagetable.c @@ -44,6 +44,22 @@ void iommufd_hwpt_paging_abort(struct iommufd_object *obj) iommufd_hwpt_paging_destroy(obj); }
+void iommufd_hwpt_nested_destroy(struct iommufd_object *obj) +{ + struct iommufd_hwpt_nested *hwpt_nested = + container_of(obj, struct iommufd_hwpt_nested, common.obj); + + if (hwpt_nested->common.domain) + iommu_domain_free(hwpt_nested->common.domain); + + refcount_dec(&hwpt_nested->parent->common.obj.users); +} + +void iommufd_hwpt_nested_abort(struct iommufd_object *obj) +{ + iommufd_hwpt_nested_destroy(obj); +} + int iommufd_hwpt_paging_enforce_cc(struct iommufd_hwpt_paging *hwpt_paging) { struct iommu_domain *paging_domain = hwpt_paging->common.domain; @@ -105,6 +121,7 @@ iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, /* Pairs with iommufd_hw_pagetable_destroy() */ refcount_inc(&ioas->obj.users); hwpt_paging->ioas = ioas; + hwpt_paging->nest_parent = flags & IOMMU_HWPT_ALLOC_NEST_PARENT;
if (ops->domain_alloc_user) { hwpt->domain = ops->domain_alloc_user(idev->dev, flags, @@ -161,6 +178,73 @@ iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas, return ERR_PTR(rc); }
+/** + * iommufd_hwpt_nested_alloc() - Get a NESTED iommu_domain for a device + * @ictx: iommufd context + * @parent: Parent PAGING-type hwpt to associate the domain with + * @idev: Device to get an iommu_domain for + * @flags: Flags from userspace + * @user_data: user_data pointer. Must be valid + * + * Allocate a new iommu_domain (must be IOMMU_DOMAIN_NESTED) and return it as + * a NESTED hw_pagetable. The given parent PAGING-type hwpt must be capable of + * being a parent. + */ +static struct iommufd_hwpt_nested * +iommufd_hwpt_nested_alloc(struct iommufd_ctx *ictx, + struct iommufd_hwpt_paging *parent, + struct iommufd_device *idev, u32 flags, + const struct iommu_user_data *user_data) +{ + const struct iommu_ops *ops = dev_iommu_ops(idev->dev); + struct iommufd_hwpt_nested *hwpt_nested; + struct iommufd_hw_pagetable *hwpt; + int rc; + + if (flags != 0) + return ERR_PTR(-EOPNOTSUPP); + if (!user_data) + return ERR_PTR(-EINVAL); + if (user_data->type == IOMMU_HWPT_DATA_NONE) + return ERR_PTR(-EINVAL); + if (parent->auto_domain) + return ERR_PTR(-EINVAL); + if (!parent->nest_parent) + return ERR_PTR(-EINVAL); + + if (!ops->domain_alloc_user) + return ERR_PTR(-EOPNOTSUPP); + + hwpt_nested = __iommufd_object_alloc(ictx, hwpt_nested, + IOMMUFD_OBJ_HWPT_NESTED, + common.obj); + if (IS_ERR(hwpt_nested)) + return ERR_CAST(hwpt_nested); + hwpt = &hwpt_nested->common; + + refcount_inc(&parent->common.obj.users); + hwpt_nested->parent = parent; + + hwpt->domain = ops->domain_alloc_user(idev->dev, 0, + parent->common.domain, + user_data); + if (IS_ERR(hwpt->domain)) { + rc = PTR_ERR(hwpt->domain); + hwpt->domain = NULL; + goto out_abort; + } + + if (WARN_ON_ONCE(hwpt->domain->type != IOMMU_DOMAIN_NESTED)) { + rc = -EINVAL; + goto out_abort; + } + return hwpt_nested; + +out_abort: + iommufd_object_abort_and_destroy(ictx, &hwpt->obj); + return ERR_PTR(rc); +} + int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) { struct iommu_hwpt_alloc *cmd = ucmd->cmd; @@ -186,6 +270,10 @@ int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) if (pt_obj->type == IOMMUFD_OBJ_IOAS) { struct iommufd_hwpt_paging *hwpt_paging;
+ if (cmd->data_type != IOMMU_HWPT_DATA_NONE) { + rc = -EINVAL; + goto out_put_pt; + } ioas = container_of(pt_obj, struct iommufd_ioas, obj); mutex_lock(&ioas->mutex); hwpt_paging = iommufd_hwpt_paging_alloc(ucmd->ictx, ioas, idev, @@ -195,6 +283,23 @@ int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd) goto out_unlock; } hwpt = &hwpt_paging->common; + } else if (pt_obj->type == IOMMUFD_OBJ_HWPT_PAGING) { + const struct iommu_user_data user_data = { + .type = cmd->data_type, + .uptr = u64_to_user_ptr(cmd->data_uptr), + .len = cmd->data_len, + }; + struct iommufd_hwpt_nested *hwpt_nested; + struct iommufd_hwpt_paging *parent; + + parent = container_of(pt_obj, typeof(*parent), common.obj); + hwpt_nested = iommufd_hwpt_nested_alloc(ucmd->ictx, parent, idev, + cmd->flags, &user_data); + if (IS_ERR(hwpt_nested)) { + rc = PTR_ERR(hwpt_nested); + goto out_unlock; + } + hwpt = &hwpt_nested->common; } else { rc = -EINVAL; goto out_put_pt; diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h index a7ed384717a5..a829458fe762 100644 --- a/drivers/iommu/iommufd/iommufd_private.h +++ b/drivers/iommu/iommufd/iommufd_private.h @@ -114,6 +114,7 @@ enum iommufd_object_type { IOMMUFD_OBJ_ANY = IOMMUFD_OBJ_NONE, IOMMUFD_OBJ_DEVICE, IOMMUFD_OBJ_HWPT_PAGING, + IOMMUFD_OBJ_HWPT_NESTED, IOMMUFD_OBJ_IOAS, IOMMUFD_OBJ_ACCESS, #ifdef CONFIG_IOMMUFD_TEST @@ -243,10 +244,16 @@ struct iommufd_hwpt_paging { bool auto_domain : 1; bool enforce_cache_coherency : 1; bool msi_cookie : 1; + bool nest_parent : 1; /* Head at iommufd_ioas::hwpt_list */ struct list_head hwpt_item; };
+struct iommufd_hwpt_nested { + struct iommufd_hw_pagetable common; + struct iommufd_hwpt_paging *parent; +}; + static inline bool hwpt_is_paging(struct iommufd_hw_pagetable *hwpt) { return hwpt->obj.type == IOMMUFD_OBJ_HWPT_PAGING; @@ -269,18 +276,28 @@ struct iommufd_hw_pagetable * iommufd_hw_pagetable_detach(struct iommufd_device *idev); void iommufd_hwpt_paging_destroy(struct iommufd_object *obj); void iommufd_hwpt_paging_abort(struct iommufd_object *obj); +void iommufd_hwpt_nested_destroy(struct iommufd_object *obj); +void iommufd_hwpt_nested_abort(struct iommufd_object *obj); int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd);
static inline void iommufd_hw_pagetable_put(struct iommufd_ctx *ictx, struct iommufd_hw_pagetable *hwpt) { - struct iommufd_hwpt_paging *hwpt_paging = to_hwpt_paging(hwpt); + if (WARN_ON(hwpt->obj.type != IOMMUFD_OBJ_HWPT_PAGING && + hwpt->obj.type != IOMMUFD_OBJ_HWPT_NESTED)) + return;
- lockdep_assert_not_held(&hwpt_paging->ioas->mutex); - if (hwpt_paging->auto_domain) - iommufd_object_deref_user(ictx, &hwpt->obj); - else - refcount_dec(&hwpt->obj.users); + if (hwpt->obj.type == IOMMUFD_OBJ_HWPT_PAGING) { + struct iommufd_hwpt_paging *hwpt_paging = to_hwpt_paging(hwpt); + + lockdep_assert_not_held(&hwpt_paging->ioas->mutex); + + if (hwpt_paging->auto_domain) { + iommufd_object_deref_user(ictx, &hwpt->obj); + return; + } + } + refcount_dec(&hwpt->obj.users); }
struct iommufd_group { diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c index b4a24ef041e6..f94c14f20918 100644 --- a/drivers/iommu/iommufd/main.c +++ b/drivers/iommu/iommufd/main.c @@ -486,6 +486,10 @@ static const struct iommufd_object_ops iommufd_object_ops[] = { .destroy = iommufd_hwpt_paging_destroy, .abort = iommufd_hwpt_paging_abort, }, + [IOMMUFD_OBJ_HWPT_NESTED] = { + .destroy = iommufd_hwpt_nested_destroy, + .abort = iommufd_hwpt_nested_abort, + }, #ifdef CONFIG_IOMMUFD_TEST [IOMMUFD_OBJ_SELFTEST] = { .destroy = iommufd_selftest_destroy, diff --git a/include/uapi/linux/iommufd.h b/include/uapi/linux/iommufd.h index b459bb7ac579..d82bc663c5cd 100644 --- a/include/uapi/linux/iommufd.h +++ b/include/uapi/linux/iommufd.h @@ -369,15 +369,31 @@ enum iommu_hwpt_data_type { * @size: sizeof(struct iommu_hwpt_alloc) * @flags: Combination of enum iommufd_hwpt_alloc_flags * @dev_id: The device to allocate this HWPT for - * @pt_id: The IOAS to connect this HWPT to + * @pt_id: The IOAS or HWPT to connect this HWPT to * @out_hwpt_id: The ID of the new HWPT * @__reserved: Must be 0 + * @data_type: One of enum iommu_hwpt_data_type + * @data_len: Length of the type specific data + * @data_uptr: User pointer to the type specific data * * Explicitly allocate a hardware page table object. This is the same object * type that is returned by iommufd_device_attach() and represents the * underlying iommu driver's iommu_domain kernel object. * - * A HWPT will be created with the IOVA mappings from the given IOAS. + * A kernel-managed HWPT will be created with the mappings from the given + * IOAS via the @pt_id. The @data_type for this allocation must be set to + * IOMMU_HWPT_DATA_NONE. The HWPT can be allocated as a parent HWPT for a + * nesting configuration by passing IOMMU_HWPT_ALLOC_NEST_PARENT via @flags. + * + * A user-managed nested HWPT will be created from a given parent HWPT via + * @pt_id, in which the parent HWPT must be allocated previously via the + * same ioctl from a given IOAS (@pt_id). In this case, the @data_type + * must be set to a pre-defined type corresponding to an I/O page table + * type supported by the underlying IOMMU hardware. + * + * If the @data_type is set to IOMMU_HWPT_DATA_NONE, @data_len and + * @data_uptr should be zero. Otherwise, both @data_len and @data_uptr + * must be given. */ struct iommu_hwpt_alloc { __u32 size; @@ -386,6 +402,9 @@ struct iommu_hwpt_alloc { __u32 pt_id; __u32 out_hwpt_id; __u32 __reserved; + __u32 data_type; + __u32 data_len; + __aligned_u64 data_uptr; }; #define IOMMU_HWPT_ALLOC _IO(IOMMUFD_TYPE, IOMMUFD_CMD_HWPT_ALLOC)
From: Nicolin Chen nicolinc@nvidia.com
Wrap up the data type/pointer/len sanity and a copy_struct_from_user call for iommu drivers to copy driver specific data via struct iommu_user_data. And expect it to be used in the domain_alloc_user op for example.
Signed-off-by: Nicolin Chen nicolinc@nvidia.com Co-developed-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- include/linux/iommu.h | 41 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 41 insertions(+)
diff --git a/include/linux/iommu.h b/include/linux/iommu.h index cdb989b853a5..48b8a9a03ae7 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -246,6 +246,47 @@ struct iommu_user_data { size_t len; };
+/** + * __iommu_copy_struct_from_user - Copy iommu driver specific user space data + * @dst_data: Pointer to an iommu driver specific user data that is defined in + * include/uapi/linux/iommufd.h + * @src_data: Pointer to a struct iommu_user_data for user space data info + * @data_type: The data type of the @dst_data. Must match with @src_data.type + * @data_len: Length of current user data structure, i.e. sizeof(struct _dst) + * @min_len: Initial length of user data structure for backward compatibility. + * This should be offsetofend using the last member in the user data + * struct that was initially added to include/uapi/linux/iommufd.h + */ +static inline int +__iommu_copy_struct_from_user(void *dst_data, + const struct iommu_user_data *src_data, + unsigned int data_type, size_t data_len, + size_t min_len) +{ + if (src_data->type != data_type) + return -EINVAL; + if (WARN_ON(!dst_data || !src_data)) + return -EINVAL; + if (src_data->len < min_len || data_len < src_data->len) + return -EINVAL; + return copy_struct_from_user(dst_data, data_len, + src_data->uptr, src_data->len); +} + +/** + * iommu_copy_struct_from_user - Copy iommu driver specific user space data + * @kdst: Pointer to an iommu driver specific user data that is defined in + * include/uapi/linux/iommufd.h + * @user_data: Pointer to a struct iommu_user_data for user space data info + * @data_type: The data type of the @kdst. Must match with @user_data->type + * @min_last: The last memember of the data structure @kdst points in the + * initial version. + * Return 0 for success, otherwise -error. + */ +#define iommu_copy_struct_from_user(kdst, user_data, data_type, min_last) \ + __iommu_copy_struct_from_user(kdst, user_data, data_type, sizeof(*kdst), \ + offsetofend(typeof(*kdst), min_last)) + /** * struct iommu_ops - iommu ops and capabilities * @capable: check capability
From: Nicolin Chen nicolinc@nvidia.com
Add nested domain support in the ->domain_alloc_user op with some proper sanity checks. Then, add a domain_nested_ops for all nested domains and split the get_md_pagetable helper into paging and nested helpers.
Also, add an iotlb as a testing property of a nested domain.
Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- drivers/iommu/iommufd/iommufd_test.h | 18 ++++ drivers/iommu/iommufd/selftest.c | 135 +++++++++++++++++++++++---- 2 files changed, 133 insertions(+), 20 deletions(-)
diff --git a/drivers/iommu/iommufd/iommufd_test.h b/drivers/iommu/iommufd/iommufd_test.h index 3f3644375bf1..440ec3d64099 100644 --- a/drivers/iommu/iommufd/iommufd_test.h +++ b/drivers/iommu/iommufd/iommufd_test.h @@ -40,6 +40,11 @@ enum { MOCK_FLAGS_ACCESS_CREATE_NEEDS_PIN_PAGES = 1 << 0, };
+enum { + MOCK_NESTED_DOMAIN_IOTLB_ID_MAX = 3, + MOCK_NESTED_DOMAIN_IOTLB_NUM = 4, +}; + struct iommu_test_cmd { __u32 size; __u32 op; @@ -109,4 +114,17 @@ struct iommu_test_hw_info { __u32 test_reg; };
+/* Should not be equal to any defined value in enum iommu_hwpt_data_type */ +#define IOMMU_HWPT_DATA_SELFTEST 0xdead + +/** + * struct iommu_hwpt_selftest + * + * @iotlb: default mock iotlb value, IOMMU_TEST_IOTLB_DEFAULT + */ +struct iommu_hwpt_selftest { +#define IOMMU_TEST_IOTLB_DEFAULT 0xbadbeef + __u32 iotlb; +}; + #endif diff --git a/drivers/iommu/iommufd/selftest.c b/drivers/iommu/iommufd/selftest.c index eca5b1d0919d..5f513c5d6876 100644 --- a/drivers/iommu/iommufd/selftest.c +++ b/drivers/iommu/iommufd/selftest.c @@ -90,6 +90,12 @@ struct mock_iommu_domain { struct xarray pfns; };
+struct mock_iommu_domain_nested { + struct iommu_domain domain; + struct mock_iommu_domain *parent; + u32 iotlb[MOCK_NESTED_DOMAIN_IOTLB_NUM]; +}; + enum selftest_obj_type { TYPE_IDEV, }; @@ -147,20 +153,16 @@ static void *mock_domain_hw_info(struct device *dev, u32 *length, u32 *type) }
static const struct iommu_ops mock_ops; +static struct iommu_domain_ops domain_nested_ops;
-static struct iommu_domain *mock_domain_alloc(unsigned int iommu_domain_type) +static struct iommu_domain * +__mock_domain_alloc_paging(unsigned int iommu_domain_type) { struct mock_iommu_domain *mock;
- if (iommu_domain_type == IOMMU_DOMAIN_BLOCKED) - return &mock_blocking_domain; - - if (iommu_domain_type != IOMMU_DOMAIN_UNMANAGED) - return NULL; - mock = kzalloc(sizeof(*mock), GFP_KERNEL); if (!mock) - return NULL; + return ERR_PTR(-ENOMEM); mock->domain.geometry.aperture_start = MOCK_APERTURE_START; mock->domain.geometry.aperture_end = MOCK_APERTURE_LAST; mock->domain.pgsize_bitmap = MOCK_IO_PAGE_SIZE; @@ -170,20 +172,72 @@ static struct iommu_domain *mock_domain_alloc(unsigned int iommu_domain_type) return &mock->domain; }
+static struct iommu_domain * +__mock_domain_alloc_nested(struct mock_iommu_domain *mock_parent, + const struct iommu_hwpt_selftest *user_cfg) +{ + struct mock_iommu_domain_nested *mock_nested; + int i; + + mock_nested = kzalloc(sizeof(*mock_nested), GFP_KERNEL); + if (!mock_nested) + return ERR_PTR(-ENOMEM); + mock_nested->parent = mock_parent; + mock_nested->domain.ops = &domain_nested_ops; + mock_nested->domain.type = IOMMU_DOMAIN_NESTED; + for (i = 0; i < MOCK_NESTED_DOMAIN_IOTLB_NUM; i++) + mock_nested->iotlb[i] = user_cfg->iotlb; + return &mock_nested->domain; +} + +static struct iommu_domain *mock_domain_alloc(unsigned int iommu_domain_type) +{ + struct iommu_domain *domain; + + if (iommu_domain_type == IOMMU_DOMAIN_BLOCKED) + return &mock_blocking_domain; + if (iommu_domain_type != IOMMU_DOMAIN_UNMANAGED) + return NULL; + domain = __mock_domain_alloc_paging(iommu_domain_type); + if (IS_ERR(domain)) + domain = NULL; + return domain; +} + static struct iommu_domain * mock_domain_alloc_user(struct device *dev, u32 flags, struct iommu_domain *parent, const struct iommu_user_data *user_data) { - struct iommu_domain *domain; + struct mock_iommu_domain *mock_parent; + struct iommu_hwpt_selftest user_cfg; + int rc;
if (flags & (~IOMMU_HWPT_ALLOC_NEST_PARENT)) return ERR_PTR(-EOPNOTSUPP);
- domain = mock_domain_alloc(IOMMU_DOMAIN_UNMANAGED); - if (!domain) - domain = ERR_PTR(-ENOMEM); - return domain; + if (!user_data) { /* must be mock_domain */ + if (parent) + return ERR_PTR(-EINVAL); + return __mock_domain_alloc_paging(IOMMU_DOMAIN_UNMANAGED); + } + + /* must be mock_domain_nested */ + if (user_data->type != IOMMU_HWPT_DATA_SELFTEST) + return ERR_PTR(-EOPNOTSUPP); + if (!parent || parent->ops != mock_ops.default_domain_ops) + return ERR_PTR(-EINVAL); + + mock_parent = container_of(parent, struct mock_iommu_domain, domain); + if (!mock_parent) + return ERR_PTR(-EINVAL); + + rc = iommu_copy_struct_from_user(&user_cfg, user_data, + IOMMU_HWPT_DATA_SELFTEST, iotlb); + if (rc) + return ERR_PTR(rc); + + return __mock_domain_alloc_nested(mock_parent, &user_cfg); }
static void mock_domain_free(struct iommu_domain *domain) @@ -342,19 +396,41 @@ static const struct iommu_ops mock_ops = { }, };
+static void mock_domain_free_nested(struct iommu_domain *domain) +{ + struct mock_iommu_domain_nested *mock_nested = + container_of(domain, struct mock_iommu_domain_nested, domain); + + kfree(mock_nested); +} + +static struct iommu_domain_ops domain_nested_ops = { + .free = mock_domain_free_nested, + .attach_dev = mock_domain_nop_attach, +}; + static inline struct iommufd_hw_pagetable * -get_md_pagetable(struct iommufd_ucmd *ucmd, u32 mockpt_id, - struct mock_iommu_domain **mock) +__get_md_pagetable(struct iommufd_ucmd *ucmd, u32 mockpt_id, u32 hwpt_type) { - struct iommufd_hw_pagetable *hwpt; struct iommufd_object *obj;
- obj = iommufd_get_object(ucmd->ictx, mockpt_id, - IOMMUFD_OBJ_HWPT_PAGING); + obj = iommufd_get_object(ucmd->ictx, mockpt_id, hwpt_type); if (IS_ERR(obj)) return ERR_CAST(obj); - hwpt = container_of(obj, struct iommufd_hw_pagetable, obj); - if (hwpt->domain->ops != mock_ops.default_domain_ops) { + return container_of(obj, struct iommufd_hw_pagetable, obj); +} + +static inline struct iommufd_hw_pagetable * +get_md_pagetable(struct iommufd_ucmd *ucmd, u32 mockpt_id, + struct mock_iommu_domain **mock) +{ + struct iommufd_hw_pagetable *hwpt; + + hwpt = __get_md_pagetable(ucmd, mockpt_id, IOMMUFD_OBJ_HWPT_PAGING); + if (IS_ERR(hwpt)) + return hwpt; + if (hwpt->domain->type != IOMMU_DOMAIN_UNMANAGED || + hwpt->domain->ops != mock_ops.default_domain_ops) { iommufd_put_object(&hwpt->obj); return ERR_PTR(-EINVAL); } @@ -362,6 +438,25 @@ get_md_pagetable(struct iommufd_ucmd *ucmd, u32 mockpt_id, return hwpt; }
+static inline struct iommufd_hw_pagetable * +get_md_pagetable_nested(struct iommufd_ucmd *ucmd, u32 mockpt_id, + struct mock_iommu_domain_nested **mock_nested) +{ + struct iommufd_hw_pagetable *hwpt; + + hwpt = __get_md_pagetable(ucmd, mockpt_id, IOMMUFD_OBJ_HWPT_NESTED); + if (IS_ERR(hwpt)) + return hwpt; + if (hwpt->domain->type != IOMMU_DOMAIN_NESTED || + hwpt->domain->ops != &domain_nested_ops) { + iommufd_put_object(&hwpt->obj); + return ERR_PTR(-EINVAL); + } + *mock_nested = container_of(hwpt->domain, + struct mock_iommu_domain_nested, domain); + return hwpt; +} + struct mock_bus_type { struct bus_type bus; struct notifier_block nb;
From: Nicolin Chen nicolinc@nvidia.com
The IOMMU_HWPT_ALLOC ioctl now supports passing user_data to allocate a user-managed domain for nested HWPTs. Add its coverage for that. Also, update _test_cmd_hwpt_alloc() and add test_cmd/err_hwpt_alloc_nested().
Signed-off-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com --- tools/testing/selftests/iommu/iommufd.c | 120 ++++++++++++++++++ .../selftests/iommu/iommufd_fail_nth.c | 3 +- tools/testing/selftests/iommu/iommufd_utils.h | 31 ++++- 3 files changed, 146 insertions(+), 8 deletions(-)
diff --git a/tools/testing/selftests/iommu/iommufd.c b/tools/testing/selftests/iommu/iommufd.c index 6323153d277b..7ab9ca06460d 100644 --- a/tools/testing/selftests/iommu/iommufd.c +++ b/tools/testing/selftests/iommu/iommufd.c @@ -264,6 +264,126 @@ TEST_F(iommufd_ioas, ioas_destroy) } }
+TEST_F(iommufd_ioas, alloc_hwpt_nested) +{ + const uint32_t min_data_len = + offsetofend(struct iommu_hwpt_selftest, iotlb); + struct iommu_hwpt_selftest data = { + .iotlb = IOMMU_TEST_IOTLB_DEFAULT, + }; + uint32_t nested_hwpt_id[2] = {}; + uint32_t parent_hwpt_id = 0; + uint32_t parent_hwpt_id_not_work = 0; + uint32_t test_hwpt_id = 0; + + if (self->device_id) { + /* Negative tests */ + test_err_hwpt_alloc(ENOENT, self->ioas_id, self->device_id, + 0, &test_hwpt_id); + test_err_hwpt_alloc(EINVAL, self->device_id, + self->device_id, 0, &test_hwpt_id); + + test_cmd_hwpt_alloc(self->device_id, self->ioas_id, + IOMMU_HWPT_ALLOC_NEST_PARENT, + &parent_hwpt_id); + + test_cmd_hwpt_alloc(self->device_id, self->ioas_id, + 0, &parent_hwpt_id_not_work); + + /* Negative nested tests */ + test_err_hwpt_alloc_nested(EINVAL, + self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_NONE, + &data, sizeof(data)); + test_err_hwpt_alloc_nested(EOPNOTSUPP, + self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST + 1, + &data, sizeof(data)); + test_err_hwpt_alloc_nested(EINVAL, + self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST, + &data, min_data_len - 1); + test_err_hwpt_alloc_nested(EFAULT, + self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST, + NULL, sizeof(data)); + test_err_hwpt_alloc_nested(EOPNOTSUPP, + self->device_id, parent_hwpt_id, + IOMMU_HWPT_ALLOC_NEST_PARENT, + &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + test_err_hwpt_alloc_nested(EINVAL, self->device_id, + parent_hwpt_id_not_work, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + + /* Allocate two nested hwpts sharing one common parent hwpt */ + test_cmd_hwpt_alloc_nested(self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + test_cmd_hwpt_alloc_nested(self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[1], + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + + /* Negative test: a nested hwpt on top of a nested hwpt */ + test_err_hwpt_alloc_nested(EINVAL, + self->device_id, nested_hwpt_id[0], + 0, &test_hwpt_id, + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + /* Negative test: parent hwpt now cannot be freed */ + EXPECT_ERRNO(EBUSY, + _test_ioctl_destroy(self->fd, parent_hwpt_id)); + + /* Attach device to nested_hwpt_id[0] that then will be busy */ + test_cmd_mock_domain_replace(self->stdev_id, + nested_hwpt_id[0]); + EXPECT_ERRNO(EBUSY, + _test_ioctl_destroy(self->fd, nested_hwpt_id[0])); + + /* Switch from nested_hwpt_id[0] to nested_hwpt_id[1] */ + test_cmd_mock_domain_replace(self->stdev_id, + nested_hwpt_id[1]); + EXPECT_ERRNO(EBUSY, + _test_ioctl_destroy(self->fd, nested_hwpt_id[1])); + test_ioctl_destroy(nested_hwpt_id[0]); + + /* Detach from nested_hwpt_id[1] and destroy it */ + test_cmd_mock_domain_replace(self->stdev_id, parent_hwpt_id); + test_ioctl_destroy(nested_hwpt_id[1]); + + /* Detach from the parent hw_pagetable and destroy it */ + test_cmd_mock_domain_replace(self->stdev_id, self->ioas_id); + test_ioctl_destroy(parent_hwpt_id); + test_ioctl_destroy(parent_hwpt_id_not_work); + } else { + test_err_hwpt_alloc(ENOENT, self->device_id, self->ioas_id, + 0, &parent_hwpt_id); + test_err_hwpt_alloc_nested(ENOENT, + self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[0], + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + test_err_hwpt_alloc_nested(ENOENT, + self->device_id, parent_hwpt_id, + 0, &nested_hwpt_id[1], + IOMMU_HWPT_DATA_SELFTEST, + &data, sizeof(data)); + test_err_mock_domain_replace(ENOENT, + self->stdev_id, nested_hwpt_id[0]); + test_err_mock_domain_replace(ENOENT, + self->stdev_id, nested_hwpt_id[1]); + } +} + TEST_F(iommufd_ioas, hwpt_attach) { /* Create a device attached directly to a hwpt */ diff --git a/tools/testing/selftests/iommu/iommufd_fail_nth.c b/tools/testing/selftests/iommu/iommufd_fail_nth.c index 31386be42439..2a3b45d49700 100644 --- a/tools/testing/selftests/iommu/iommufd_fail_nth.c +++ b/tools/testing/selftests/iommu/iommufd_fail_nth.c @@ -615,7 +615,8 @@ TEST_FAIL_NTH(basic_fail_nth, device) if (_test_cmd_get_hw_info(self->fd, idev_id, &info, sizeof(info))) return -1;
- if (_test_cmd_hwpt_alloc(self->fd, idev_id, ioas_id, 0, &hwpt_id)) + if (_test_cmd_hwpt_alloc(self->fd, idev_id, ioas_id, 0, &hwpt_id, + IOMMU_HWPT_DATA_NONE, 0, 0)) return -1;
if (_test_cmd_mock_domain_replace(self->fd, stdev_id, ioas_id2, NULL)) diff --git a/tools/testing/selftests/iommu/iommufd_utils.h b/tools/testing/selftests/iommu/iommufd_utils.h index be4970a84977..1a5fa44b1627 100644 --- a/tools/testing/selftests/iommu/iommufd_utils.h +++ b/tools/testing/selftests/iommu/iommufd_utils.h @@ -103,13 +103,17 @@ static int _test_cmd_mock_domain_replace(int fd, __u32 stdev_id, __u32 pt_id, pt_id, NULL))
static int _test_cmd_hwpt_alloc(int fd, __u32 device_id, __u32 pt_id, - __u32 flags, __u32 *hwpt_id) + __u32 flags, __u32 *hwpt_id, __u32 data_type, + void *data, size_t data_len) { struct iommu_hwpt_alloc cmd = { .size = sizeof(cmd), .flags = flags, .dev_id = device_id, .pt_id = pt_id, + .data_type = data_type, + .data_len = data_len, + .data_uptr = (uint64_t)data, }; int ret;
@@ -121,12 +125,25 @@ static int _test_cmd_hwpt_alloc(int fd, __u32 device_id, __u32 pt_id, return 0; }
-#define test_cmd_hwpt_alloc(device_id, pt_id, flags, hwpt_id) \ - ASSERT_EQ(0, _test_cmd_hwpt_alloc(self->fd, device_id, \ - pt_id, flags, hwpt_id)) -#define test_err_hwpt_alloc(_errno, device_id, pt_id, flags, hwpt_id) \ - EXPECT_ERRNO(_errno, _test_cmd_hwpt_alloc(self->fd, device_id, \ - pt_id, flags, hwpt_id)) +#define test_cmd_hwpt_alloc(device_id, pt_id, flags, hwpt_id) \ + ASSERT_EQ(0, _test_cmd_hwpt_alloc(self->fd, device_id, pt_id, flags, \ + hwpt_id, IOMMU_HWPT_DATA_NONE, \ + NULL, 0)) +#define test_err_hwpt_alloc(_errno, device_id, pt_id, flags, hwpt_id) \ + EXPECT_ERRNO(_errno, _test_cmd_hwpt_alloc(self->fd, device_id, pt_id, \ + flags, hwpt_id, \ + IOMMU_HWPT_DATA_NONE, \ + NULL, 0)) + +#define test_cmd_hwpt_alloc_nested(device_id, pt_id, flags, hwpt_id, \ + data_type, data, data_len) \ + ASSERT_EQ(0, _test_cmd_hwpt_alloc(self->fd, device_id, pt_id, flags, \ + hwpt_id, data_type, data, data_len)) +#define test_err_hwpt_alloc_nested(_errno, device_id, pt_id, flags, hwpt_id, \ + data_type, data, data_len) \ + EXPECT_ERRNO(_errno, \ + _test_cmd_hwpt_alloc(self->fd, device_id, pt_id, flags, \ + hwpt_id, data_type, data, data_len))
static int _test_cmd_access_replace_ioas(int fd, __u32 access_id, unsigned int ioas_id)
linux-kselftest-mirror@lists.linaro.org
-
Yi Liu