vfio-pci support pasid attach/detach
Subsystems (7)
v4 cover vfio-pci support pasid attach/detach
This adds the pasid attach/detach uAPIs for userspace to attach/detach a PASID of a device to/from a given ioas/hwpt. Only vfio-pci driver is enabled in this series. After this series, PASID-capable devices bound with vfio-pci can report PASID capability to userspace and VM to enable PASID usages like Shared Virtual Addressing (SVA).
Based on the discussion about reporting the vPASID to VM [1], it's agreed that we will let the userspace VMM to synthesize the vPASID capability. The VMM needs to figure out a hole to put the vPASID cap. This includes the hidden bits handling for some devices. While, it's up to the userspace, it's not the focus of this series.
This series first adds the helpers for pasid attach in vfio core and then extends the device cdev attach/detach ioctls for pasid attach/detach. In the end of this series, the IOMMU_GET_HW_INFO ioctl is extended to report the PCI PASID capability to the userspace. Userspace should check this before using any PASID related uAPIs provided by VFIO, which is the agreement in [2]. This series depends on the iommufd pasid attach/detach series [3].
The completed code can be found at [4], tested with a hacky Qemu branch [5].
[1] https://lore.kernel.org/kvm/BN9PR11MB5276318969A212AD0649C7BE8CBE2@BN9PR11MB5276.namprd11.prod.outlook.com/
[2] https://lore.kernel.org/kvm/[email protected]/
[3] https://lore.kernel.org/linux-iommu/[email protected]/
[4] https://github.com/yiliu1765/iommufd/tree/iommufd_pasid
[5] https://github.com/yiliu1765/qemu/tree/wip/zhenzhong/iommufd_nesting_rfcv2-test-pasid
Change log:
v4:
- Add acked-by for the ida patch from Matthew
- Add r-b from Kevin and Jason on patch 01, 02 and 04 of v3
- Add common code to copy user data for the user struct with new fields
- Extend the VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT to support pasid, patch 03 is updated per this change. Hence drop r-b of it. (Kevin, Alex)
- Add t-b from Zhangfei for patch 4 of v3
- Nits from Vasant
v3: https://lore.kernel.org/linux-iommu/[email protected]/
- Misc enhancement on patch 01 of v2 (Alex, Jason)
- Add Jason's r-b to patch 03 of v2
- Drop the logic that report PASID via VFIO_DEVICE_FEATURE ioctl
- Extend IOMMU_GET_HW_INFO to report PASID support (Kevin, Jason, Alex)
v2: https://lore.kernel.org/kvm/[email protected]/
- Use IDA to track if PASID is attached or not in VFIO. (Jason)
- Fix the issue of calling pasid_at[de]tach_ioas callback unconditionally (Alex)
- Fix the wrong data copy in vfio_df_ioctl_pasid_detach_pt() (Zhenzhong)
- Minor tweaks in comments (Kevin)
v1: https://lore.kernel.org/kvm/[email protected]/
- Report PASID capability via VFIO_DEVICE_FEATURE (Alex)
rfc: https://lore.kernel.org/linux-iommu/[email protected]/
Regards,
Yi Liu
Yi Liu (4):
ida: Add ida_find_first_range()
vfio-iommufd: Support pasid [at|de]tach for physical VFIO devices
vfio: VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT support pasid
iommufd: Extend IOMMU_GET_HW_INFO to report PASID capability
drivers/iommu/iommufd/device.c | 24 +++++++++++- drivers/pci/ats.c | 33 ++++++++++++++++ drivers/vfio/device_cdev.c | 62 +++++++++++++++++++++--------- drivers/vfio/iommufd.c | 50 ++++++++++++++++++++++++ drivers/vfio/pci/vfio_pci.c | 2 + drivers/vfio/vfio.h | 18 +++++++++ drivers/vfio/vfio_main.c | 55 ++++++++++++++++++++++++++ include/linux/idr.h | 11 ++++++ include/linux/pci-ats.h | 3 ++ include/linux/vfio.h | 11 ++++++ include/uapi/linux/iommufd.h | 14 ++++++- include/uapi/linux/vfio.h | 29 +++++++++----- lib/idr.c | 67 ++++++++++++++++++++++++++++++++ lib/test_ida.c | 70 ++++++++++++++++++++++++++++++++++ 14 files changed, 419 insertions(+), 30 deletions(-)
Signature · 1 line
2.34.1
PATCH v4 1/4 ida: Add ida_find_first_range()
There is no helpers for user to check if a given ID is allocated or not, neither a helper to loop all the allocated IDs in an IDA and do something for cleanup. With the two needs, a helper to get the lowest allocated ID of a range and two variants based on it.
Caller can check if a given ID is allocated or not by:
bool ida_exists(struct ida *ida, unsigned int id)
Caller can iterate all allocated IDs by:
int id;
while ((id = ida_find_first(&pasid_ida)) >= 0) {
//anything to do with the allocated ID
ida_free(pasid_ida, pasid);
}--- include/linux/idr.h | 11 +++++++ lib/idr.c | 67 +++++++++++++++++++++++++++++++++++++++++++ lib/test_ida.c | 70 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 148 insertions(+)
include/linux/idr.h +11 -0
@@ -257,6 +257,7 @@ struct ida {int ida_alloc_range (struct ida *,unsigned int min,unsigned int max,gfp_t );void ida_free (struct ida *,unsigned int id);void ida_destroy (struct ida *ida); +int ida_find_first_range (struct ida *ida,unsigned int min,unsigned int max);/** * ida_alloc() - Allocate an unused ID.
@@ -328,4 +329,14 @@ static inline bool ida_is_empty(const struct ida *ida) {return xa_empty (&ida->xa); } + +static inline bool ida_exists (struct ida *ida,unsigned int id) +{ +return ida_find_first_range (ida, id, id) == id; +} + +static inline int ida_find_first (struct ida *ida) +{ +return ida_find_first_range (ida,0 , ~0 ); +}#endif /* __IDR_H__ */
lib/idr.c +67 -0
73 lines not on this page · raw message
lib/test_ida.c +70 -0
82 lines not on this page · raw message
Signature · 1 line
2.34.1
PATCH v4 2/4 vfio-iommufd: Support pasid [at|de]tach for physical VFIO devices
In reply to Yi LiuThis adds pasid_at|de]tach_ioas ops for attaching hwpt to pasid of a device and the helpers for it. For now, only vfio-pci supports pasid attach/detach.
--- drivers/vfio/iommufd.c | 50 +++++++++++++++++++++++++++++++++++++ drivers/vfio/pci/vfio_pci.c | 2 ++ include/linux/vfio.h | 11 ++++++++ 3 files changed, 63 insertions(+)
drivers/vfio/iommufd.c +50 -0
@@ -119,14 +119,22 @@ int vfio_iommufd_physical_bind(struct vfio_device *vdev,if (IS_ERR (idev))return PTR_ERR (idev); vdev->iommufd_device = idev; +ida_init (&vdev->pasids);return 0 ; }EXPORT_SYMBOL_GPL (vfio_iommufd_physical_bind);void vfio_iommufd_physical_unbind (struct vfio_device *vdev) { +int pasid; +lockdep_assert_held (&vdev->dev_set->lock); +while ((pasid =ida_find_first (&vdev->pasids)) >=0 ) { +iommufd_device_pasid_detach (vdev->iommufd_device, pasid); +ida_free (&vdev->pasids, pasid); + } +if (vdev->iommufd_attached) {iommufd_device_detach (vdev->iommufd_device); vdev->iommufd_attached = false;
@@ -168,6 +176,48 @@ void vfio_iommufd_physical_detach_ioas(struct vfio_device *vdev) }EXPORT_SYMBOL_GPL (vfio_iommufd_physical_detach_ioas); +int vfio_iommufd_physical_pasid_attach_ioas (struct vfio_device *vdev, +u32 pasid,u32 *pt_id) +{ +int rc; + +lockdep_assert_held (&vdev->dev_set->lock); + +if (WARN_ON (!vdev->iommufd_device)) +return -EINVAL ; + +if (ida_exists (&vdev->pasids, pasid)) +return iommufd_device_pasid_replace (vdev->iommufd_device, + pasid, pt_id); + + rc =ida_alloc_range (&vdev->pasids, pasid, pasid,GFP_KERNEL ); +if (rc <0 ) +return rc; + + rc =iommufd_device_pasid_attach (vdev->iommufd_device, pasid, pt_id); +if (rc) +ida_free (&vdev->pasids, pasid); + +return 0 ; +} +EXPORT_SYMBOL_GPL (vfio_iommufd_physical_pasid_attach_ioas); + +void vfio_iommufd_physical_pasid_detach_ioas (struct vfio_device *vdev, +u32 pasid) +{ +lockdep_assert_held (&vdev->dev_set->lock); + +if (WARN_ON (!vdev->iommufd_device)) +return ; + +if (!ida_exists (&vdev->pasids, pasid)) +return ; + +iommufd_device_pasid_detach (vdev->iommufd_device, pasid); +ida_free (&vdev->pasids, pasid); +} +EXPORT_SYMBOL_GPL (vfio_iommufd_physical_pasid_detach_ioas); +/* * The emulated standard ops mean that vfio_device is going to use the * "mdev path" and will call vfio_pin_pages()/vfio_dma_rw(). Drivers using this
drivers/vfio/pci/vfio_pci.c +2 -0
@@ -144,6 +144,8 @@ static const struct vfio_device_ops vfio_pci_ops = { .unbind_iommufd = vfio_iommufd_physical_unbind, .attach_ioas = vfio_iommufd_physical_attach_ioas, .detach_ioas = vfio_iommufd_physical_detach_ioas, + .pasid_attach_ioas = vfio_iommufd_physical_pasid_attach_ioas, + .pasid_detach_ioas = vfio_iommufd_physical_pasid_detach_ioas, };static int vfio_pci_probe (struct pci_dev *pdev,const struct pci_device_id *id)
include/linux/vfio.h +11 -0
@@ -67,6 +67,7 @@ struct vfio_device {struct inode *inode;#if IS_ENABLED (CONFIG_IOMMUFD )struct iommufd_device *iommufd_device; +struct ida pasids; u8 iommufd_attached:1;#endif u8 cdev_opened:1;
@@ -91,6 +92,8 @@ struct vfio_device {* bound iommufd. Undo in unbind_iommufd if @detach_ioas is not * called. * @detach_ioas: Opposite of attach_ioas +* @pasid_attach_ioas: The pasid variation of attach_ioas +* @pasid_detach_ioas: Opposite of pasid_attach_ioas * @open_device: Called when the first file descriptor is opened for this device * @close_device: Opposite of open_device * @read: Perform read(2) on device file descriptor
@@ -115,6 +118,8 @@ struct vfio_device_ops {void (*unbind_iommufd)(struct vfio_device *vdev);int (*attach_ioas)(struct vfio_device *vdev,u32 *pt_id);void (*detach_ioas)(struct vfio_device *vdev); +int (*pasid_attach_ioas)(struct vfio_device *vdev,u32 pasid,u32 *pt_id); +void (*pasid_detach_ioas)(struct vfio_device *vdev,u32 pasid);int (*open_device)(struct vfio_device *vdev);void (*close_device)(struct vfio_device *vdev); ssize_t (*read)(struct vfio_device *vdev, char __user *buf,
@@ -139,6 +144,8 @@ int vfio_iommufd_physical_bind(struct vfio_device *vdev,void vfio_iommufd_physical_unbind (struct vfio_device *vdev);int vfio_iommufd_physical_attach_ioas (struct vfio_device *vdev,u32 *pt_id);void vfio_iommufd_physical_detach_ioas (struct vfio_device *vdev); +int vfio_iommufd_physical_pasid_attach_ioas (struct vfio_device *vdev,u32 pasid,u32 *pt_id); +void vfio_iommufd_physical_pasid_detach_ioas (struct vfio_device *vdev,u32 pasid);int vfio_iommufd_emulated_bind (struct vfio_device *vdev,struct iommufd_ctx *ictx,u32 *out_device_id);void vfio_iommufd_emulated_unbind (struct vfio_device *vdev);
@@ -166,6 +173,10 @@ vfio_iommufd_get_dev_id(struct vfio_device *vdev, struct iommufd_ctx *ictx) ((int (*)(struct vfio_device *vdev, u32 *pt_id)) NULL)#define vfio_iommufd_physical_detach_ioas \ ((void (*)(struct vfio_device *vdev)) NULL) +#define vfio_iommufd_physical_pasid_attach_ioas \ + ((int (*)(struct vfio_device *vdev, u32 pasid, u32 *pt_id)) NULL) +#define vfio_iommufd_physical_pasid_detach_ioas \ + ((void (*)(struct vfio_device *vdev, u32 pasid)) NULL)#define vfio_iommufd_emulated_bind \ ((int (*)(struct vfio_device *vdev, struct iommufd_ctx *ictx, \ u32 *out_device_id)) NULL)
Signature · 1 line
2.34.1
Re: [PATCH v4 2/4] vfio-iommufd: Support pasid [at|de]tach for physical VFIO devices
Quoted 57 lines · On Mon, 4 Nov 2024 05:27:30 -0800 Yi Liu <[email protected]> wrote:
This adds pasid_at|de]tach_ioas ops for attaching hwpt to pasid of a device and the helpers for it. For now, only vfio-pci supports pasid attach/detach.
Signed-off-by Kevin Tian <[email protected]>Reviewed-by Jason Gunthorpe <[email protected]>Signed-off-by Yi Liu <[email protected]>--- drivers/vfio/iommufd.c | 50 +++++++++++++++++++++++++++++++++++++ drivers/vfio/pci/vfio_pci.c | 2 ++ include/linux/vfio.h | 11 ++++++++ 3 files changed, 63 insertions(+)drivers/vfio/iommufd.c +21
-0@@ -119,14 +119,22 @@ int vfio_iommufd_physical_bind(struct vfio_device *vdev,if (IS_ERR (idev))return PTR_ERR (idev); vdev->iommufd_device = idev; +ida_init (&vdev->pasids);return 0 ; }EXPORT_SYMBOL_GPL (vfio_iommufd_physical_bind);void vfio_iommufd_physical_unbind (struct vfio_device *vdev) { +int pasid; +lockdep_assert_held (&vdev->dev_set->lock); +while ((pasid =ida_find_first (&vdev->pasids)) >=0 ) { +iommufd_device_pasid_detach (vdev->iommufd_device, pasid); +ida_free (&vdev->pasids, pasid); + } +if (vdev->iommufd_attached) {iommufd_device_detach (vdev->iommufd_device); vdev->iommufd_attached = false;@@ -168,6 +176,48 @@ void vfio_iommufd_physical_detach_ioas(struct vfio_device *vdev) }EXPORT_SYMBOL_GPL (vfio_iommufd_physical_detach_ioas); +int vfio_iommufd_physical_pasid_attach_ioas (struct vfio_device *vdev, +u32 pasid,u32 *pt_id) +{ +int rc; + +lockdep_assert_held (&vdev->dev_set->lock); + +if (WARN_ON (!vdev->iommufd_device)) +return -EINVAL ; + +if (ida_exists (&vdev->pasids, pasid)) +return iommufd_device_pasid_replace (vdev->iommufd_device, + pasid, pt_id);
drivers/vfio/iommufd.c void vfio_iommufd_physical_detach_ioas(struct vfio_device *vdev) · line 201
Quoted 10 lines
drivers/vfio/iommufd.c +10
-0+ + rc =ida_alloc_range (&vdev->pasids, pasid, pasid,GFP_KERNEL ); +if (rc <0 ) +return rc; + + rc =iommufd_device_pasid_attach (vdev->iommufd_device, pasid, pt_id); +if (rc) +ida_free (&vdev->pasids, pasid); + +return 0 ;
I think you meant to return rc here. Thanks,
Alex
Quoted 85 lines
drivers/vfio/iommufd.c +19
-0+} +EXPORT_SYMBOL_GPL (vfio_iommufd_physical_pasid_attach_ioas); + +void vfio_iommufd_physical_pasid_detach_ioas (struct vfio_device *vdev, +u32 pasid) +{ +lockdep_assert_held (&vdev->dev_set->lock); + +if (WARN_ON (!vdev->iommufd_device)) +return ; + +if (!ida_exists (&vdev->pasids, pasid)) +return ; + +iommufd_device_pasid_detach (vdev->iommufd_device, pasid); +ida_free (&vdev->pasids, pasid); +} +EXPORT_SYMBOL_GPL (vfio_iommufd_physical_pasid_detach_ioas); +/* * The emulated standard ops mean that vfio_device is going to use the * "mdev path" and will call vfio_pin_pages()/vfio_dma_rw(). Drivers using this drivers/vfio/pci/vfio_pci.c +2
-0@@ -144,6 +144,8 @@ static const struct vfio_device_ops vfio_pci_ops = { .unbind_iommufd = vfio_iommufd_physical_unbind, .attach_ioas = vfio_iommufd_physical_attach_ioas, .detach_ioas = vfio_iommufd_physical_detach_ioas, + .pasid_attach_ioas = vfio_iommufd_physical_pasid_attach_ioas, + .pasid_detach_ioas = vfio_iommufd_physical_pasid_detach_ioas, };static int vfio_pci_probe (struct pci_dev *pdev,const struct pci_device_id *id)include/linux/vfio.h +11
-0@@ -67,6 +67,7 @@ struct vfio_device {struct inode *inode;#if IS_ENABLED (CONFIG_IOMMUFD )struct iommufd_device *iommufd_device; +struct ida pasids; u8 iommufd_attached:1;#endif u8 cdev_opened:1;@@ -91,6 +92,8 @@ struct vfio_device {* bound iommufd. Undo in unbind_iommufd if @detach_ioas is not * called. * @detach_ioas: Opposite of attach_ioas +* @pasid_attach_ioas: The pasid variation of attach_ioas +* @pasid_detach_ioas: Opposite of pasid_attach_ioas * @open_device: Called when the first file descriptor is opened for this device * @close_device: Opposite of open_device * @read: Perform read(2) on device file descriptor @@ -115,6 +118,8 @@ struct vfio_device_ops {void (*unbind_iommufd)(struct vfio_device *vdev);int (*attach_ioas)(struct vfio_device *vdev,u32 *pt_id);void (*detach_ioas)(struct vfio_device *vdev); +int (*pasid_attach_ioas)(struct vfio_device *vdev,u32 pasid,u32 *pt_id); +void (*pasid_detach_ioas)(struct vfio_device *vdev,u32 pasid);int (*open_device)(struct vfio_device *vdev);void (*close_device)(struct vfio_device *vdev); ssize_t (*read)(struct vfio_device *vdev, char __user *buf,@@ -139,6 +144,8 @@ int vfio_iommufd_physical_bind(struct vfio_device *vdev,void vfio_iommufd_physical_unbind (struct vfio_device *vdev);int vfio_iommufd_physical_attach_ioas (struct vfio_device *vdev,u32 *pt_id);void vfio_iommufd_physical_detach_ioas (struct vfio_device *vdev); +int vfio_iommufd_physical_pasid_attach_ioas (struct vfio_device *vdev,u32 pasid,u32 *pt_id); +void vfio_iommufd_physical_pasid_detach_ioas (struct vfio_device *vdev,u32 pasid);int vfio_iommufd_emulated_bind (struct vfio_device *vdev,struct iommufd_ctx *ictx,u32 *out_device_id);void vfio_iommufd_emulated_unbind (struct vfio_device *vdev);@@ -166,6 +173,10 @@ vfio_iommufd_get_dev_id(struct vfio_device *vdev, struct iommufd_ctx *ictx) ((int (*)(struct vfio_device *vdev, u32 *pt_id)) NULL)#define vfio_iommufd_physical_detach_ioas \ ((void (*)(struct vfio_device *vdev)) NULL) +#define vfio_iommufd_physical_pasid_attach_ioas \ + ((int (*)(struct vfio_device *vdev, u32 pasid, u32 *pt_id)) NULL) +#define vfio_iommufd_physical_pasid_detach_ioas \ + ((void (*)(struct vfio_device *vdev, u32 pasid)) NULL)#define vfio_iommufd_emulated_bind \ ((int (*)(struct vfio_device *vdev, struct iommufd_ctx *ictx, \ u32 *out_device_id)) NULL)
Re: [PATCH v4 2/4] vfio-iommufd: Support pasid [at|de]tach for physical VFIO devices
Quoted 64 lines · Alex Williamson
Quoted 61 lines · On Mon, 4 Nov 2024 05:27:30 -0800 Yi Liu <[email protected]> wrote:
This adds pasid_at|de]tach_ioas ops for attaching hwpt to pasid of a device and the helpers for it. For now, only vfio-pci supports pasid attach/detach.
Signed-off-by Kevin Tian <[email protected]>Reviewed-by Jason Gunthorpe <[email protected]>Signed-off-by Yi Liu <[email protected]>--- drivers/vfio/iommufd.c | 50 +++++++++++++++++++++++++++++++++++++ drivers/vfio/pci/vfio_pci.c | 2 ++ include/linux/vfio.h | 11 ++++++++ 3 files changed, 63 insertions(+)drivers/vfio/iommufd.c +25
-0@@ -119,14 +119,22 @@ int vfio_iommufd_physical_bind(struct vfio_device *vdev,if (IS_ERR (idev))return PTR_ERR (idev); vdev->iommufd_device = idev; +ida_init (&vdev->pasids);return 0 ; }EXPORT_SYMBOL_GPL (vfio_iommufd_physical_bind);void vfio_iommufd_physical_unbind (struct vfio_device *vdev) { +int pasid; +lockdep_assert_held (&vdev->dev_set->lock); +while ((pasid =ida_find_first (&vdev->pasids)) >=0 ) { +iommufd_device_pasid_detach (vdev->iommufd_device, pasid); +ida_free (&vdev->pasids, pasid); + } +if (vdev->iommufd_attached) {iommufd_device_detach (vdev->iommufd_device); vdev->iommufd_attached = false;@@ -168,6 +176,48 @@ void vfio_iommufd_physical_detach_ioas(struct vfio_device *vdev) }EXPORT_SYMBOL_GPL (vfio_iommufd_physical_detach_ioas); +int vfio_iommufd_physical_pasid_attach_ioas (struct vfio_device *vdev, +u32 pasid,u32 *pt_id) +{ +int rc; + +lockdep_assert_held (&vdev->dev_set->lock); + +if (WARN_ON (!vdev->iommufd_device)) +return -EINVAL ; + +if (ida_exists (&vdev->pasids, pasid)) +return iommufd_device_pasid_replace (vdev->iommufd_device, + pasid, pt_id); + + rc =ida_alloc_range (&vdev->pasids, pasid, pasid,GFP_KERNEL ); +if (rc <0 ) +return rc;
Quoted 8 lines
drivers/vfio/iommufd.c void vfio_iommufd_physical_detach_ioas(struct vfio_device *vdev) · line 201
Quoted 6 lines
drivers/vfio/iommufd.c +6
-0+ + rc =iommufd_device_pasid_attach (vdev->iommufd_device, pasid, pt_id); +if (rc) +ida_free (&vdev->pasids, pasid); + +return 0 ;I think you meant to return rc here. Thanks,
you are absolutely right. :)
Signature · 2 lines
Regards, Yi Liu
PATCH v4 3/4 vfio: VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT support pasid
In reply to Yi LiuThis extends the VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT ioctls to attach/detach a given pasid of a vfio device to/from an IOAS/HWPT.
vfio_copy_from_user() is added to copy the user data for the case in which the existing user struct has introduced new fields. The rule is not breaking the existing usersapce. The kernel only copies the new fields when the corresponding flag is set by the userspace. For the case that has multiple new fields marked by different flags, kernel checks the flags one by one to get the correct size to copy besides the minsz. Such logics can be shared by the other uapi extensions, hence add a helper for it.
--- drivers/vfio/device_cdev.c | 62 +++++++++++++++++++++++++++----------- drivers/vfio/vfio.h | 18 +++++++++++ drivers/vfio/vfio_main.c | 55 +++++++++++++++++++++++++++++++++ include/uapi/linux/vfio.h | 29 ++++++++++++------ 4 files changed, 136 insertions(+), 28 deletions(-)
drivers/vfio/device_cdev.c +44 -18
@@ -159,24 +159,44 @@ void vfio_df_unbind_iommufd(struct vfio_device_file *df)vfio_device_unblock_group (device); } +#define VFIO_ATTACH_FLAGS_MASK VFIO_DEVICE_ATTACH_PASID +static unsigned long +vfio_attach_xends[ilog2 (VFIO_ATTACH_FLAGS_MASK ) +1 ] = { +XEND_SIZE (VFIO_DEVICE_ATTACH_PASID , + struct vfio_device_attach_iommufd_pt, pasid), +}; + +#define VFIO_DETACH_FLAGS_MASK VFIO_DEVICE_DETACH_PASID +static unsigned long +vfio_detach_xends[ilog2 (VFIO_DETACH_FLAGS_MASK ) +1 ] = { +XEND_SIZE (VFIO_DEVICE_DETACH_PASID , + struct vfio_device_detach_iommufd_pt, pasid), +}; +int vfio_df_ioctl_attach_pt (struct vfio_device_file *df,struct vfio_device_attach_iommufd_pt __user *arg) { -struct vfio_device *device = df->device;struct vfio_device_attach_iommufd_pt attach; -unsigned long minsz; +struct vfio_device *device = df->device;int ret; - minsz = offsetofend(struct vfio_device_attach_iommufd_pt, pt_id); - -if (copy_from_user (&attach, arg, minsz)) -return -EFAULT ; + ret =VFIO_COPY_USER_DATA ((void __user *)arg, &attach, + struct vfio_device_attach_iommufd_pt, + pt_id,VFIO_ATTACH_FLAGS_MASK , + vfio_attach_xends); +if (ret) +return ret; -if (attach.argsz < minsz || attach.flags) -return -EINVAL ; +if ((attach.flags &VFIO_DEVICE_ATTACH_PASID ) && + !device->ops->pasid_attach_ioas) +return -EOPNOTSUPP ;mutex_lock (&device->dev_set->lock); - ret = device->ops->attach_ioas (device, &attach.pt_id); +if (attach.flags &VFIO_DEVICE_ATTACH_PASID ) + ret = device->ops->pasid_attach_ioas (device, attach.pasid, + &attach.pt_id); +else + ret = device->ops->attach_ioas (device, &attach.pt_id);if (ret) goto out_unlock;
@@ -198,20 +218,26 @@ int vfio_df_ioctl_attach_pt(struct vfio_device_file *df,int vfio_df_ioctl_detach_pt (struct vfio_device_file *df,struct vfio_device_detach_iommufd_pt __user *arg) { -struct vfio_device *device = df->device;struct vfio_device_detach_iommufd_pt detach; -unsigned long minsz; - - minsz = offsetofend(struct vfio_device_detach_iommufd_pt, flags); +struct vfio_device *device = df->device; +int ret; -if (copy_from_user (&detach, arg, minsz)) -return -EFAULT ; + ret =VFIO_COPY_USER_DATA ((void __user *)arg, &detach, + struct vfio_device_detach_iommufd_pt, + flags,VFIO_DETACH_FLAGS_MASK , + vfio_detach_xends); +if (ret) +return ret; -if (detach.argsz < minsz || detach.flags) -return -EINVAL ; +if ((detach.flags &VFIO_DEVICE_DETACH_PASID ) && + !device->ops->pasid_detach_ioas) +return -EOPNOTSUPP ;mutex_lock (&device->dev_set->lock); - device->ops->detach_ioas (device); +if (detach.flags &VFIO_DEVICE_DETACH_PASID ) + device->ops->pasid_detach_ioas (device, detach.pasid); +else + device->ops->detach_ioas (device);mutex_unlock (&device->dev_set->lock);return 0 ;
drivers/vfio/vfio.h +18 -0
@@ -34,6 +34,24 @@ void vfio_df_close(struct vfio_device_file *df);struct vfio_device_file *vfio_allocate_device_file (struct vfio_device *device); +int vfio_copy_from_user (void *buffer,void __user *arg, +unsigned long minsz,u32 flags_mask, +unsigned long *xend_array); + +#define VFIO_COPY_USER_DATA (_arg, _local_buffer, _struct, _min_last, \ + _flags_mask, _xend_array) \ + vfio_copy_from_user(_local_buffer, _arg, \ + offsetofend(_struct, _min_last) + \ + BUILD_BUG_ON_ZERO(offsetof(_struct, argsz) != \ + 0) + \ + BUILD_BUG_ON_ZERO(offsetof(_struct, flags) != \ + sizeof(u32)), \ + _flags_mask, _xend_array) + +#define XEND_SIZE (_flag, _struct, _xlast) \ + [ilog2(_flag)] = offsetofend(_struct, _xlast) + \ + BUILD_BUG_ON_ZERO(_flag == 0) \ +extern const struct file_operations vfio_device_fops;#ifdef CONFIG_VFIO_NOIOMMU
drivers/vfio/vfio_main.c +55 -0
61 lines not on this page · raw message
include/uapi/linux/vfio.h +19 -10
@@ -931,29 +931,34 @@ struct vfio_device_bind_iommufd {* VFIO_DEVICE_ATTACH_IOMMUFD_PT - _IOW(VFIO_TYPE, VFIO_BASE + 19, * struct vfio_device_attach_iommufd_pt) * @argsz: User filled size of this data. -* @flags: Must be 0. +* @flags: Flags for attach. * @pt_id: Input the target id which can represent an ioas or a hwpt * allocated via iommufd subsystem. * Output the input ioas id or the attached hwpt id which could * be the specified hwpt itself or a hwpt automatically created * for the specified ioas by kernel during the attachment. +* @pasid: The pasid to be attached, only meaningful when +* VFIO_DEVICE_ATTACH_PASID is set in @flags * * Associate the device with an address space within the bound iommufd. * Undo by VFIO_DEVICE_DETACH_IOMMUFD_PT or device fd close. This is only * allowed on cdev fds. * -* If a vfio device is currently attached to a valid hw_pagetable, without doing -* a VFIO_DEVICE_DETACH_IOMMUFD_PT, a second VFIO_DEVICE_ATTACH_IOMMUFD_PT ioctl -* passing in another hw_pagetable (hwpt) id is allowed. This action, also known -* as a hw_pagetable replacement, will replace the device's currently attached -* hw_pagetable with a new hw_pagetable corresponding to the given pt_id. +* If a vfio device or a pasid of this device is currently attached to a valid +* hw_pagetable (hwpt), without doing a VFIO_DEVICE_DETACH_IOMMUFD_PT, a second +* VFIO_DEVICE_ATTACH_IOMMUFD_PT ioctl passing in another hwpt id is allowed. +* This action, also known as a hw_pagetable replacement, will replace the +* currently attached hwpt of the device or the pasid of this device with a new +* hwpt corresponding to the given pt_id. * * Return: 0 on success, -errno on failure. */ struct vfio_device_attach_iommufd_pt {__u32 argsz;__u32 flags; +#define VFIO_DEVICE_ATTACH_PASID (1 << 0)__u32 pt_id; +__u32 pasid; };#define VFIO_DEVICE_ATTACH_IOMMUFD_PT _IO(VFIO_TYPE, VFIO_BASE + 19)
@@ -962,17 +967,21 @@ struct vfio_device_attach_iommufd_pt {* VFIO_DEVICE_DETACH_IOMMUFD_PT - _IOW(VFIO_TYPE, VFIO_BASE + 20, * struct vfio_device_detach_iommufd_pt) * @argsz: User filled size of this data. -* @flags: Must be 0. +* @flags: Flags for detach. +* @pasid: The pasid to be detached, only meaningful when +* VFIO_DEVICE_DETACH_PASID is set in @flags * -* Remove the association of the device and its current associated address -* space. After it, the device should be in a blocking DMA state. This is only -* allowed on cdev fds. +* Remove the association of the device or a pasid of the device and its current +* associated address space. After it, the device or the pasid should be in a +* blocking DMA state. This is only allowed on cdev fds. * * Return: 0 on success, -errno on failure. */ struct vfio_device_detach_iommufd_pt {__u32 argsz;__u32 flags; +#define VFIO_DEVICE_DETACH_PASID (1 << 0) +__u32 pasid; };#define VFIO_DEVICE_DETACH_IOMMUFD_PT _IO(VFIO_TYPE, VFIO_BASE + 20)
Signature · 1 line
2.34.1
Re: [PATCH v4 3/4] vfio: VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT support pasid
Quoted 30 lines · On Mon, 4 Nov 2024 05:27:31 -0800 Yi Liu <[email protected]> wrote:
This extends the VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT ioctls to attach/detach a given pasid of a vfio device to/from an IOAS/HWPT.
vfio_copy_from_user()is added to copy the user data for the case in which the existing user struct has introduced new fields. The rule is not breaking the existing usersapce. The kernel only copies the new fields when the corresponding flag is set by the userspace. For the case that has multiple new fields marked by different flags, kernel checks the flags one by one to get the correct size to copy besides the minsz. Such logics can be shared by the other uapi extensions, hence add a helper for it.Signed-off-by Yi Liu <[email protected]>--- drivers/vfio/device_cdev.c | 62 +++++++++++++++++++++++++++----------- drivers/vfio/vfio.h | 18 +++++++++++ drivers/vfio/vfio_main.c | 55 +++++++++++++++++++++++++++++++++ include/uapi/linux/vfio.h | 29 ++++++++++++------ 4 files changed, 136 insertions(+), 28 deletions(-)drivers/vfio/device_cdev.c +3
-0@@ -159,24 +159,44 @@ void vfio_df_unbind_iommufd(struct vfio_device_file *df)vfio_device_unblock_group (device); } +#define VFIO_ATTACH_FLAGS_MASK VFIO_DEVICE_ATTACH_PASID +static unsigned long +vfio_attach_xends[ilog2 (VFIO_ATTACH_FLAGS_MASK ) +1 ] = {
drivers/vfio/device_cdev.c void vfio_df_unbind_iommufd(struct vfio_device_file *df) · line 174
Quoted 10 lines
drivers/vfio/device_cdev.c +10
-0+XEND_SIZE (VFIO_DEVICE_ATTACH_PASID , + struct vfio_device_attach_iommufd_pt, pasid), +}; + +#define VFIO_DETACH_FLAGS_MASK VFIO_DEVICE_DETACH_PASID +static unsigned long +vfio_detach_xends[ilog2 (VFIO_DETACH_FLAGS_MASK ) +1 ] = { +XEND_SIZE (VFIO_DEVICE_DETACH_PASID , + struct vfio_device_detach_iommufd_pt, pasid), +};
Doesn't this rather imply that every valid flag bit indicates some new structure field?
For example, we start out with:
struct vfio_device_attach_iommufd_pt {
__u32 argsz;
__u32 flags;
__u32 pt_id;
};And then here it becomes:
struct vfio_device_attach_iommufd_pt {
__u32 argsz;
__u32 flags;
#define VFIO_DEVICE_ATTACH_PASID (1 << 0)
__u32 pt_id;
__u32 pasid;
};What if the next flag is simply related to the processing of @pt_id and doesn't require @pasid?
The xend array necessarily expands, but what's the value? Logically it would be offsetofend(, pt_id), so the array becomes { 16, 12 }.
Similarly, rather than pasid we might have reused a previously reserved field, for instance what if we already expanded the structure as:
struct vfio_device_attach_iommufd_pt {
__u32 argsz;
__u32 flags;
#define VFIO_DEVICE_ATTACH_FOO (1 << 0)
__u32 pt_id;
__u32 reserved;
__u64 foo;
};If we then want to add @pasid, we might really prefer to take advantage of that reserved field and the array becomes { 24, 16 }.
I think these can work (see below), but this seems like a pretty complicated generalization. It might make sense to initially open code the handling for @pasid with a follow-on patch with this sort of generalization so we can evaluate them separately.
BTW, don't feel obligated to use "xend" based on my email sample code.
Quoted 84 lines
drivers/vfio/device_cdev.c +31
-18+int vfio_df_ioctl_attach_pt (struct vfio_device_file *df,struct vfio_device_attach_iommufd_pt __user *arg) { -struct vfio_device *device = df->device;struct vfio_device_attach_iommufd_pt attach; -unsigned long minsz; +struct vfio_device *device = df->device;int ret; - minsz = offsetofend(struct vfio_device_attach_iommufd_pt, pt_id); - -if (copy_from_user (&attach, arg, minsz)) -return -EFAULT ; + ret =VFIO_COPY_USER_DATA ((void __user *)arg, &attach, + struct vfio_device_attach_iommufd_pt, + pt_id,VFIO_ATTACH_FLAGS_MASK , + vfio_attach_xends); +if (ret) +return ret; -if (attach.argsz < minsz || attach.flags) -return -EINVAL ; +if ((attach.flags &VFIO_DEVICE_ATTACH_PASID ) && + !device->ops->pasid_attach_ioas) +return -EOPNOTSUPP ;mutex_lock (&device->dev_set->lock); - ret = device->ops->attach_ioas (device, &attach.pt_id); +if (attach.flags &VFIO_DEVICE_ATTACH_PASID ) + ret = device->ops->pasid_attach_ioas (device, attach.pasid, + &attach.pt_id); +else + ret = device->ops->attach_ioas (device, &attach.pt_id);if (ret) goto out_unlock;@@ -198,20 +218,26 @@ int vfio_df_ioctl_attach_pt(struct vfio_device_file *df,int vfio_df_ioctl_detach_pt (struct vfio_device_file *df,struct vfio_device_detach_iommufd_pt __user *arg) { -struct vfio_device *device = df->device;struct vfio_device_detach_iommufd_pt detach; -unsigned long minsz; - - minsz = offsetofend(struct vfio_device_detach_iommufd_pt, flags); +struct vfio_device *device = df->device; +int ret; -if (copy_from_user (&detach, arg, minsz)) -return -EFAULT ; + ret =VFIO_COPY_USER_DATA ((void __user *)arg, &detach, + struct vfio_device_detach_iommufd_pt, + flags,VFIO_DETACH_FLAGS_MASK , + vfio_detach_xends); +if (ret) +return ret; -if (detach.argsz < minsz || detach.flags) -return -EINVAL ; +if ((detach.flags &VFIO_DEVICE_DETACH_PASID ) && + !device->ops->pasid_detach_ioas) +return -EOPNOTSUPP ;mutex_lock (&device->dev_set->lock); - device->ops->detach_ioas (device); +if (detach.flags &VFIO_DEVICE_DETACH_PASID ) + device->ops->pasid_detach_ioas (device, detach.pasid); +else + device->ops->detach_ioas (device);mutex_unlock (&device->dev_set->lock);return 0 ;drivers/vfio/vfio.h +3
-0@@ -34,6 +34,24 @@ void vfio_df_close(struct vfio_device_file *df);struct vfio_device_file *vfio_allocate_device_file (struct vfio_device *device); +int vfio_copy_from_user (void *buffer,void __user *arg, +unsigned long minsz,u32 flags_mask, +unsigned long *xend_array);
drivers/vfio/vfio.h void vfio_df_close(struct vfio_device_file *df); · line 49
Quoted 10 lines
drivers/vfio/vfio.h +10
-0+ +#define VFIO_COPY_USER_DATA (_arg, _local_buffer, _struct, _min_last, \ + _flags_mask, _xend_array) \ + vfio_copy_from_user(_local_buffer, _arg, \ + offsetofend(_struct, _min_last) + \ + BUILD_BUG_ON_ZERO(offsetof(_struct, argsz) != \ + 0) + \ + BUILD_BUG_ON_ZERO(offsetof(_struct, flags) != \ + sizeof(u32)), \ + _flags_mask, _xend_array)
We have a precedence in vfio_alloc_device() that macros wrapping functions don't need to be all caps.
Quoted 45 lines
drivers/vfio/vfio.h +5
-0+ +#define XEND_SIZE (_flag, _struct, _xlast) \ + [ilog2(_flag)] = offsetofend(_struct, _xlast) + \ + BUILD_BUG_ON_ZERO(_flag == 0) \ +extern const struct file_operations vfio_device_fops;#ifdef CONFIG_VFIO_NOIOMMU drivers/vfio/vfio_main.c +29
-0@@ -1694,6 +1694,61 @@ int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, }EXPORT_SYMBOL (vfio_dma_rw); +/** +* vfio_copy_from_user - Copy the user struct that may have extended fields +* +* @buffer: The local buffer to store the data copied from user +* @arg: The user buffer pointer +* @minsz: The minimum size of the user struct, it should never bump up. +* @flags_mask: The combination of all the falgs defined +* @xend_array: The array that stores the xend size for set flags. +* +* This helper requires the user struct put the argsz and flags fields in +* the first 8 bytes. +* +* Return 0 for success, otherwise -errno +*/ +int vfio_copy_from_user (void *buffer,void __user *arg, +unsigned long minsz,u32 flags_mask, +unsigned long *xend_array) +{ +unsigned long xend =0 ; +struct user_header { +u32 argsz; +u32 flags; + } *header; +unsigned long flags; +u32 flag; + +if (copy_from_user (buffer, arg, minsz)) +return -EFAULT ; +
drivers/vfio/vfio_main.c int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, · line 1735
Quoted 10 lines
drivers/vfio/vfio_main.c +10
-0+ header = (struct user_header *)buffer; +if (header->argsz < minsz) +return -EINVAL ; + +if (header->flags & ~flags_mask) +return -EINVAL ; + +/* Loop each set flag to decide the xend */ + flags = header->flags; +for_each_set_bit (flag, &flags,BITS_PER_LONG ) {
I suppose it doesn't matter, but there's a logical inconsistency searching BITS_PER_LONG on a buffer initialized by a u32.
drivers/vfio/vfio_main.c int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, · line 1736
Quoted 1 line
drivers/vfio/vfio_main.c +1
-0+if (xend_array[flag])
Given the earlier concern, this should be:
if (xend_array[flags] > xend)
Thanks,
Alex
Quoted 89 lines
drivers/vfio/vfio_main.c +15
-0+ xend = xend_array[flag]; + } + +if (xend) { +if (header->argsz < xend) +return -EINVAL ; + +if (copy_from_user (buffer + minsz, + arg + minsz, xend - minsz)) +return -EFAULT ; + } + +return 0 ; +} +/* * Module/class support */ include/uapi/linux/vfio.h +19
-10@@ -931,29 +931,34 @@ struct vfio_device_bind_iommufd {* VFIO_DEVICE_ATTACH_IOMMUFD_PT - _IOW(VFIO_TYPE, VFIO_BASE + 19, * struct vfio_device_attach_iommufd_pt) * @argsz: User filled size of this data. -* @flags: Must be 0. +* @flags: Flags for attach. * @pt_id: Input the target id which can represent an ioas or a hwpt * allocated via iommufd subsystem. * Output the input ioas id or the attached hwpt id which could * be the specified hwpt itself or a hwpt automatically created * for the specified ioas by kernel during the attachment. +* @pasid: The pasid to be attached, only meaningful when +* VFIO_DEVICE_ATTACH_PASID is set in @flags * * Associate the device with an address space within the bound iommufd. * Undo by VFIO_DEVICE_DETACH_IOMMUFD_PT or device fd close. This is only * allowed on cdev fds. * -* If a vfio device is currently attached to a valid hw_pagetable, without doing -* a VFIO_DEVICE_DETACH_IOMMUFD_PT, a second VFIO_DEVICE_ATTACH_IOMMUFD_PT ioctl -* passing in another hw_pagetable (hwpt) id is allowed. This action, also known -* as a hw_pagetable replacement, will replace the device's currently attached -* hw_pagetable with a new hw_pagetable corresponding to the given pt_id. +* If a vfio device or a pasid of this device is currently attached to a valid +* hw_pagetable (hwpt), without doing a VFIO_DEVICE_DETACH_IOMMUFD_PT, a second +* VFIO_DEVICE_ATTACH_IOMMUFD_PT ioctl passing in another hwpt id is allowed. +* This action, also known as a hw_pagetable replacement, will replace the +* currently attached hwpt of the device or the pasid of this device with a new +* hwpt corresponding to the given pt_id. * * Return: 0 on success, -errno on failure. */ struct vfio_device_attach_iommufd_pt {__u32 argsz;__u32 flags; +#define VFIO_DEVICE_ATTACH_PASID (1 << 0)__u32 pt_id; +__u32 pasid; };#define VFIO_DEVICE_ATTACH_IOMMUFD_PT _IO(VFIO_TYPE, VFIO_BASE + 19)@@ -962,17 +967,21 @@ struct vfio_device_attach_iommufd_pt {* VFIO_DEVICE_DETACH_IOMMUFD_PT - _IOW(VFIO_TYPE, VFIO_BASE + 20, * struct vfio_device_detach_iommufd_pt) * @argsz: User filled size of this data. -* @flags: Must be 0. +* @flags: Flags for detach. +* @pasid: The pasid to be detached, only meaningful when +* VFIO_DEVICE_DETACH_PASID is set in @flags * -* Remove the association of the device and its current associated address -* space. After it, the device should be in a blocking DMA state. This is only -* allowed on cdev fds. +* Remove the association of the device or a pasid of the device and its current +* associated address space. After it, the device or the pasid should be in a +* blocking DMA state. This is only allowed on cdev fds. * * Return: 0 on success, -errno on failure. */ struct vfio_device_detach_iommufd_pt {__u32 argsz;__u32 flags; +#define VFIO_DEVICE_DETACH_PASID (1 << 0) +__u32 pasid; };#define VFIO_DEVICE_DETACH_IOMMUFD_PT _IO(VFIO_TYPE, VFIO_BASE + 20)
Re: [PATCH v4 3/4] vfio: VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT support pasid
Quoted 65 lines · Alex Williamson
Quoted 30 lines · On Mon, 4 Nov 2024 05:27:31 -0800 Yi Liu <[email protected]> wrote:
This extends the VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT ioctls to attach/detach a given pasid of a vfio device to/from an IOAS/HWPT.
vfio_copy_from_user()is added to copy the user data for the case in which the existing user struct has introduced new fields. The rule is not breaking the existing usersapce. The kernel only copies the new fields when the corresponding flag is set by the userspace. For the case that has multiple new fields marked by different flags, kernel checks the flags one by one to get the correct size to copy besides the minsz. Such logics can be shared by the other uapi extensions, hence add a helper for it.Signed-off-by Yi Liu <[email protected]>--- drivers/vfio/device_cdev.c | 62 +++++++++++++++++++++++++++----------- drivers/vfio/vfio.h | 18 +++++++++++ drivers/vfio/vfio_main.c | 55 +++++++++++++++++++++++++++++++++ include/uapi/linux/vfio.h | 29 ++++++++++++------ 4 files changed, 136 insertions(+), 28 deletions(-)drivers/vfio/device_cdev.c +3
-0@@ -159,24 +159,44 @@ void vfio_df_unbind_iommufd(struct vfio_device_file *df)vfio_device_unblock_group (device); } +#define VFIO_ATTACH_FLAGS_MASK VFIO_DEVICE_ATTACH_PASID +static unsigned long +vfio_attach_xends[ilog2 (VFIO_ATTACH_FLAGS_MASK ) +1 ] = {drivers/vfio/device_cdev.c void vfio_df_unbind_iommufd(struct vfio_device_file *df) · line 174
Quoted 10 lines
drivers/vfio/device_cdev.c +10
-0+XEND_SIZE (VFIO_DEVICE_ATTACH_PASID , + struct vfio_device_attach_iommufd_pt, pasid), +}; + +#define VFIO_DETACH_FLAGS_MASK VFIO_DEVICE_DETACH_PASID +static unsigned long +vfio_detach_xends[ilog2 (VFIO_DETACH_FLAGS_MASK ) +1 ] = { +XEND_SIZE (VFIO_DEVICE_DETACH_PASID , + struct vfio_device_detach_iommufd_pt, pasid), +};Doesn't this rather imply that every valid flag bit indicates some new structure field?
For example, we start out with:
struct vfio_device_attach_iommufd_pt { __u32 argsz; __u32 flags; __u32 pt_id; };And then here it becomes:
struct vfio_device_attach_iommufd_pt { __u32 argsz; __u32 flags; #define VFIO_DEVICE_ATTACH_PASID (1 << 0) __u32 pt_id; __u32 pasid; };
Quoted 5 lines
What if the next flag is simply related to the processing of @pt_id and doesn't require @pasid?
The xend array necessarily expands, but what's the value? Logically it would be
offsetofend(, pt_id), so the array becomes { 16, 12 }.
You are right.
Quoted 14 lines
Similarly, rather than pasid we might have reused a previously reserved field, for instance what if we already expanded the structure as:
struct vfio_device_attach_iommufd_pt { __u32 argsz; __u32 flags; #define VFIO_DEVICE_ATTACH_FOO (1 << 0) __u32 pt_id; __u32 reserved; __u64 foo; };
Quoted 2 lines
If we then want to add @pasid, we might really prefer to take advantage of that reserved field and the array becomes { 24, 16 }.
yes.
Quoted 4 lines
I think these can work (see below), but this seems like a pretty complicated generalization. It might make sense to initially open code the handling for @pasid with a follow-on patch with this sort of generalization so we can evaluate them separately.
sure. If don't mind, I'd like to mention the two examples in the commit message when adding the generalization. To let future people understand how should the array be programmed.
Quoted 2 lines
BTW, don't feel obligated to use "xend" based on my email sample code.
TBH. I don't see a better name besides it yet. :) If you don't mind, I may keep using it in later versions.
Quoted 82 lines
Quoted 82 lines
drivers/vfio/device_cdev.c +31
-18+int vfio_df_ioctl_attach_pt (struct vfio_device_file *df,struct vfio_device_attach_iommufd_pt __user *arg) { -struct vfio_device *device = df->device;struct vfio_device_attach_iommufd_pt attach; -unsigned long minsz; +struct vfio_device *device = df->device;int ret; - minsz = offsetofend(struct vfio_device_attach_iommufd_pt, pt_id); - -if (copy_from_user (&attach, arg, minsz)) -return -EFAULT ; + ret =VFIO_COPY_USER_DATA ((void __user *)arg, &attach, + struct vfio_device_attach_iommufd_pt, + pt_id,VFIO_ATTACH_FLAGS_MASK , + vfio_attach_xends); +if (ret) +return ret; -if (attach.argsz < minsz || attach.flags) -return -EINVAL ; +if ((attach.flags &VFIO_DEVICE_ATTACH_PASID ) && + !device->ops->pasid_attach_ioas) +return -EOPNOTSUPP ;mutex_lock (&device->dev_set->lock); - ret = device->ops->attach_ioas (device, &attach.pt_id); +if (attach.flags &VFIO_DEVICE_ATTACH_PASID ) + ret = device->ops->pasid_attach_ioas (device, attach.pasid, + &attach.pt_id); +else + ret = device->ops->attach_ioas (device, &attach.pt_id);if (ret) goto out_unlock;@@ -198,20 +218,26 @@ int vfio_df_ioctl_attach_pt(struct vfio_device_file *df,int vfio_df_ioctl_detach_pt (struct vfio_device_file *df,struct vfio_device_detach_iommufd_pt __user *arg) { -struct vfio_device *device = df->device;struct vfio_device_detach_iommufd_pt detach; -unsigned long minsz; - - minsz = offsetofend(struct vfio_device_detach_iommufd_pt, flags); +struct vfio_device *device = df->device; +int ret; -if (copy_from_user (&detach, arg, minsz)) -return -EFAULT ; + ret =VFIO_COPY_USER_DATA ((void __user *)arg, &detach, + struct vfio_device_detach_iommufd_pt, + flags,VFIO_DETACH_FLAGS_MASK , + vfio_detach_xends); +if (ret) +return ret; -if (detach.argsz < minsz || detach.flags) -return -EINVAL ; +if ((detach.flags &VFIO_DEVICE_DETACH_PASID ) && + !device->ops->pasid_detach_ioas) +return -EOPNOTSUPP ;mutex_lock (&device->dev_set->lock); - device->ops->detach_ioas (device); +if (detach.flags &VFIO_DEVICE_DETACH_PASID ) + device->ops->pasid_detach_ioas (device, detach.pasid); +else + device->ops->detach_ioas (device);mutex_unlock (&device->dev_set->lock);return 0 ;drivers/vfio/vfio.h +0
-0@@ -34,6 +34,24 @@ void vfio_df_close(struct vfio_device_file *df);struct vfio_device_file *vfio_allocate_device_file (struct vfio_device *device);
Quoted 16 lines
drivers/vfio/vfio.h void vfio_df_close(struct vfio_device_file *df); · line 49
Quoted 13 lines
drivers/vfio/vfio.h +13
-0+int vfio_copy_from_user (void *buffer,void __user *arg, +unsigned long minsz,u32 flags_mask, +unsigned long *xend_array); + +#define VFIO_COPY_USER_DATA (_arg, _local_buffer, _struct, _min_last, \ + _flags_mask, _xend_array) \ + vfio_copy_from_user(_local_buffer, _arg, \ + offsetofend(_struct, _min_last) + \ + BUILD_BUG_ON_ZERO(offsetof(_struct, argsz) != \ + 0) + \ + BUILD_BUG_ON_ZERO(offsetof(_struct, flags) != \ + sizeof(u32)), \ + _flags_mask, _xend_array)We have a precedence in
vfio_alloc_device()that macros wrapping functions don't need to be all caps.
got it. :)
Quoted 51 lines
Quoted 51 lines
drivers/vfio/vfio.h +5
-0+ +#define XEND_SIZE (_flag, _struct, _xlast) \ + [ilog2(_flag)] = offsetofend(_struct, _xlast) + \ + BUILD_BUG_ON_ZERO(_flag == 0) \ +extern const struct file_operations vfio_device_fops;#ifdef CONFIG_VFIO_NOIOMMU drivers/vfio/vfio_main.c +34
-0@@ -1694,6 +1694,61 @@ int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, }EXPORT_SYMBOL (vfio_dma_rw); +/** +* vfio_copy_from_user - Copy the user struct that may have extended fields +* +* @buffer: The local buffer to store the data copied from user +* @arg: The user buffer pointer +* @minsz: The minimum size of the user struct, it should never bump up. +* @flags_mask: The combination of all the falgs defined +* @xend_array: The array that stores the xend size for set flags. +* +* This helper requires the user struct put the argsz and flags fields in +* the first 8 bytes. +* +* Return 0 for success, otherwise -errno +*/ +int vfio_copy_from_user (void *buffer,void __user *arg, +unsigned long minsz,u32 flags_mask, +unsigned long *xend_array) +{ +unsigned long xend =0 ; +struct user_header { +u32 argsz; +u32 flags; + } *header; +unsigned long flags; +u32 flag; + +if (copy_from_user (buffer, arg, minsz)) +return -EFAULT ; + + header = (struct user_header *)buffer; +if (header->argsz < minsz) +return -EINVAL ; + +if (header->flags & ~flags_mask)
Quoted 8 lines
drivers/vfio/vfio_main.c int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, · line 1735
Quoted 5 lines
drivers/vfio/vfio_main.c +5
-0+return -EINVAL ; + +/* Loop each set flag to decide the xend */ + flags = header->flags; +for_each_set_bit (flag, &flags,BITS_PER_LONG ) {I suppose it doesn't matter, but there's a logical inconsistency searching
BITS_PER_LONGon a buffer initialized by a u32.
how about using BITS_PER_TYPE(u32) or hard-code 32 here?
Quoted 6 lines
drivers/vfio/vfio_main.c int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, · line 1736
Quoted 2 lines
drivers/vfio/vfio_main.c +1
-0+if (xend_array[flag])Given the earlier concern, this should be:
if (xend_array[flags] > xend)
yes.
Quoted 8 lines
Thanks,
Alexdrivers/vfio/vfio_main.c int vfio_dma_rw(struct vfio_device *device, dma_addr_t iova, void *data, · line 1740
Quoted 4 lines
drivers/vfio/vfio_main.c +4
-0+ xend = xend_array[flag]; + } + +if (xend) {
I may change this check as the below. Hence we don't bother to do the check and funtion call at all.
if (xend != minsz) {Quoted 14 lines
Quoted 14 lines
drivers/vfio/vfio_main.c +11
-0+if (header->argsz < xend) +return -EINVAL ; + +if (copy_from_user (buffer + minsz, + arg + minsz, xend - minsz)) +return -EFAULT ; + } + +return 0 ; +} +/* * Module/class support */
Signature · 2 lines
Regards, Yi Liu
PATCH v4 4/4 iommufd: Extend IOMMU_GET_HW_INFO to report PASID capability
In reply to Yi LiuPASID usage requires PASID support in both device and IOMMU. Since the iommu drivers always enable the PASID capability for the device if it is supported, so it is reasonable to extend the IOMMU_GET_HW_INFO to report the PASID capability to userspace.
--- drivers/iommu/iommufd/device.c | 24 +++++++++++++++++++++++- drivers/pci/ats.c | 33 +++++++++++++++++++++++++++++++++ include/linux/pci-ats.h | 3 +++ include/uapi/linux/iommufd.h | 14 +++++++++++++- 4 files changed, 72 insertions(+), 2 deletions(-)
drivers/iommu/iommufd/device.c +23 -1
@@ -3,6 +3,8 @@*/ #include <linux/iommu.h> #include <linux/iommufd.h> +#include <linux/pci.h> +#include <linux/pci-ats.h> #include <linux/slab.h> #include <uapi/linux/iommufd.h>
@@ -1248,7 +1250,8 @@ int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)void *data;int rc; -if (cmd->flags || cmd->__reserved) +if (cmd->flags || cmd->__reserved[0 ] || cmd->__reserved[1 ] || + cmd->__reserved[2 ])return -EOPNOTSUPP ; idev =iommufd_get_device (ucmd, cmd->dev_id);
@@ -1305,6 +1308,25 @@ int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)if (device_iommu_capable (idev->dev,IOMMU_CAP_DIRTY_TRACKING )) cmd->out_capabilities |=IOMMU_HW_CAP_DIRTY_TRACKING ; + cmd->out_max_pasid_log2 =0 ; + +if (dev_is_pci (idev->dev)) { +struct pci_dev *pdev =to_pci_dev (idev->dev); +int ctrl; + + ctrl =pci_pasid_ctrl_status (pdev); +if (ctrl >=0 && (ctrl &PCI_PASID_CTRL_ENABLE )) { + cmd->out_max_pasid_log2 = +ilog2 (idev->dev->iommu->max_pasids); +if (ctrl &PCI_PASID_CTRL_EXEC ) + cmd->out_capabilities |= +IOMMU_HW_CAP_PCI_PASID_EXEC ; +if (ctrl &PCI_PASID_CTRL_PRIV ) + cmd->out_capabilities |= +IOMMU_HW_CAP_PCI_PASID_PRIV ; + } + } + rc =iommufd_ucmd_respond (ucmd,sizeof (*cmd)); out_free:kfree (data);
drivers/pci/ats.c +33 -0
@@ -538,4 +538,37 @@ int pci_max_pasids(struct pci_dev *pdev) return (1 << FIELD_GET(PCI_PASID_CAP_WIDTH, supported)); }EXPORT_SYMBOL_GPL (pci_max_pasids); + +/** +* pci_pasid_ctrl_status - Check the PASID status +* @pdev: PCI device structure +* +* Returns a negative value when no PASID capability is present. +* Otherwise the value of the control register is returned. +* Status reported are: +* +* PCI_PASID_CTRL_ENABLE - PASID enabled +* PCI_PASID_CTRL_EXEC - Execute permission enabled +* PCI_PASID_CTRL_PRIV - Privileged mode enabled +*/ +int pci_pasid_ctrl_status (struct pci_dev *pdev) +{ +int pasid; +u16 ctrl; + +if (pdev->is_virtfn) + pdev =pci_physfn (pdev); + + pasid = pdev->pasid_cap; +if (!pasid) +return -EINVAL ; + +pci_read_config_word (pdev, pasid +PCI_PASID_CTRL , &ctrl); + + ctrl &=PCI_PASID_CTRL_ENABLE |PCI_PASID_CTRL_EXEC | +PCI_PASID_CTRL_PRIV ; + +return ctrl; +} +EXPORT_SYMBOL_GPL (pci_pasid_ctrl_status);#endif /* CONFIG_PCI_PASID */
include/linux/pci-ats.h +3 -0
@@ -42,6 +42,7 @@ int pci_enable_pasid(struct pci_dev *pdev, int features);void pci_disable_pasid (struct pci_dev *pdev);int pci_pasid_features (struct pci_dev *pdev);int pci_max_pasids (struct pci_dev *pdev); +int pci_pasid_ctrl_status (struct pci_dev *pdev);#else /* CONFIG_PCI_PASID */ static inline int pci_enable_pasid (struct pci_dev *pdev,int features) {return -EINVAL ; }
@@ -50,6 +51,8 @@ static inline int pci_pasid_features(struct pci_dev *pdev) {return -EINVAL ; }static inline int pci_max_pasids (struct pci_dev *pdev) {return -EINVAL ; } +static inline int pci_pasid_ctrl_status (struct pci_dev *pdev) +{return -EINVAL ; }#endif /* CONFIG_PCI_PASID */ #endif /* LINUX_PCI_ATS_H */
include/uapi/linux/iommufd.h +13 -1
@@ -512,9 +512,17 @@ enum iommu_hw_info_type {* IOMMU_HWPT_GET_DIRTY_BITMAP * IOMMU_HWPT_SET_DIRTY_TRACKING * +* @IOMMU_HW_CAP_PASID_EXEC: Execute Permission Supported, user ignores it +* when the struct iommu_hw_info::out_max_pasid_log2 +* is zero. +* @IOMMU_HW_CAP_PASID_PRIV: Privileged Mode Supported, user ignores it +* when the struct iommu_hw_info::out_max_pasid_log2 +* is zero. */ enum iommufd_hw_capabilities {IOMMU_HW_CAP_DIRTY_TRACKING =1 <<0 , +IOMMU_HW_CAP_PCI_PASID_EXEC =1 <<1 , +IOMMU_HW_CAP_PCI_PASID_PRIV =1 <<2 , };/**
@@ -530,6 +538,9 @@ enum iommufd_hw_capabilities {* iommu_hw_info_type. * @out_capabilities: Output the generic iommu capability info type as defined * in the enum iommu_hw_capabilities. +* @out_max_pasid_log2: Output the width of PASIDs. 0 means no PASID support. +* PCI devices turn to out_capabilities to check if the +* specific capabilities is supported or not. * @__reserved: Must be 0 * * Query an iommu type specific hardware information data from an iommu behind
@@ -553,7 +564,8 @@ struct iommu_hw_info {__u32 data_len;__aligned_u64 data_uptr;__u32 out_data_type; -__u32 __reserved; +__u8 out_max_pasid_log2; +__u8 __reserved[3 ];__aligned_u64 out_capabilities; };#define IOMMU_GET_HW_INFO _IO(IOMMUFD_TYPE, IOMMUFD_CMD_GET_HW_INFO)
Signature · 1 line
2.34.1
Quoted 19 lines · Yi Liu
This adds the pasid attach/detach uAPIs for userspace to attach/detach a PASID of a device to/from a given ioas/hwpt. Only vfio-pci driver is enabled in this series. After this series, PASID-capable devices bound with vfio-pci can report PASID capability to userspace and VM to enable PASID usages like Shared Virtual Addressing (SVA).
Based on the discussion about reporting the vPASID to VM [1], it's agreed that we will let the userspace VMM to synthesize the vPASID capability. The VMM needs to figure out a hole to put the vPASID cap. This includes the hidden bits handling for some devices. While, it's up to the userspace, it's not the focus of this series.
This series first adds the helpers for pasid attach in vfio core and then extends the device cdev attach/detach ioctls for pasid attach/detach. In the end of this series, the
IOMMU_GET_HW_INFOioctl is extended to report the PCI PASID capability to the userspace. Userspace should check this before using any PASID related uAPIs provided by VFIO, which is the agreement in [2]. This series depends on the iommufd pasid attach/detach series [3].
Quoted 5 lines
The completed code can be found at [4], tested with a hacky Qemu branch [5].
[1] https://lore.kernel.org/kvm/BN9PR11MB5276318969A212AD0649C7BE8CBE2@BN9PR11MB5276.namprd11.prod.outlook.com/
[2] https://lore.kernel.org/kvm/[email protected]/
[3] https://lore.kernel.org/linux-iommu/[email protected]/
correct the latest link.
https://lore.kernel.org/linux-iommu/[email protected]/
Quoted 56 lines
[4] https://github.com/yiliu1765/iommufd/tree/iommufd_pasid
[5] https://github.com/yiliu1765/qemu/tree/wip/zhenzhong/iommufd_nesting_rfcv2-test-pasidChange log:
v4:
- Add acked-by for the ida patch from Matthew
- Add r-b from Kevin and Jason on patch 01, 02 and 04 of v3
- Add common code to copy user data for the user struct with new fields
- Extend the VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT to support pasid, patch 03 is updated per this change. Hence drop r-b of it. (Kevin, Alex)
- Add t-b from Zhangfei for patch 4 of v3
- Nits from Vasantv3: https://lore.kernel.org/linux-iommu/[email protected]/
- Misc enhancement on patch 01 of v2 (Alex, Jason)
- Add Jason's r-b to patch 03 of v2
- Drop the logic that report PASID viaVFIO_DEVICE_FEATUREioctl
- ExtendIOMMU_GET_HW_INFOto report PASID support (Kevin, Jason, Alex)v2: https://lore.kernel.org/kvm/[email protected]/
- Use IDA to track if PASID is attached or not in VFIO. (Jason)
- Fix the issue of calling pasid_at[de]tach_ioas callback unconditionally (Alex)
- Fix the wrong data copy invfio_df_ioctl_pasid_detach_pt()(Zhenzhong)
- Minor tweaks in comments (Kevin)v1: https://lore.kernel.org/kvm/[email protected]/
- Report PASID capability viaVFIO_DEVICE_FEATURE(Alex)rfc: https://lore.kernel.org/linux-iommu/[email protected]/
Regards,
Yi LiuYi Liu (4):
ida: Addida_find_first_range()
vfio-iommufd: Support pasid [at|de]tach for physical VFIO devices
vfio: VFIO_DEVICE_[AT|DE]TACH_IOMMUFD_PT support pasid
iommufd: ExtendIOMMU_GET_HW_INFOto report PASID capabilitydrivers/iommu/iommufd/device.c | 24 +++++++++++- drivers/pci/ats.c | 33 ++++++++++++++++ drivers/vfio/device_cdev.c | 62 +++++++++++++++++++++--------- drivers/vfio/iommufd.c | 50 ++++++++++++++++++++++++ drivers/vfio/pci/vfio_pci.c | 2 + drivers/vfio/vfio.h | 18 +++++++++ drivers/vfio/vfio_main.c | 55 ++++++++++++++++++++++++++ include/linux/idr.h | 11 ++++++ include/linux/pci-ats.h | 3 ++ include/linux/vfio.h | 11 ++++++ include/uapi/linux/iommufd.h | 14 ++++++- include/uapi/linux/vfio.h | 29 +++++++++----- lib/idr.c | 67 ++++++++++++++++++++++++++++++++ lib/test_ida.c | 70 ++++++++++++++++++++++++++++++++++ 14 files changed, 419 insertions(+), 30 deletions(-)
Signature · 2 lines
Regards, Yi Liu