Under nested IOMMU translation, userspace owns the stage-1 translation
table (e.g. the stage-1 page table of Intel VT-d or the context table
of ARM SMMUv3, and etc.). Stage-1 translation tables are vendor specific,
and need to be compatiable with the underlying IOMMU hardware. Hence,
userspace should know the IOMMU hardware capability before creating and
configuring the stage-1 translation table to kernel.

This adds IOMMU_GET_HW_INFO ioctl to query the IOMMU hardware information
(a.k.a capability) for a given device. The returned data is vendor specific,
userspace needs to decode it with the structure mapped by the @out_data_type
field.

As only physical devices have IOMMU hardware, so this will return error
if the given device is not a physical device.

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/iommufd/main.c | 79 ++++++++++++++++++++++++++++++++++++
 include/uapi/linux/iommufd.h | 36 ++++++++++++++++
 2 files changed, 115 insertions(+)

diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
index 94c498b8fdf6..a0302bcaa97c 100644
--- a/drivers/iommu/iommufd/main.c
+++ b/drivers/iommu/iommufd/main.c
@@ -17,6 +17,7 @@
 #include <linux/bug.h>
 #include <uapi/linux/iommufd.h>
 #include <linux/iommufd.h>
+#include "../iommu-priv.h"
 
 #include "io_pagetable.h"
 #include "iommufd_private.h"
@@ -177,6 +178,81 @@ static int iommufd_destroy(struct iommufd_ucmd *ucmd)
 	return 0;
 }
 
+static int iommufd_zero_fill_user(void __user *ptr, size_t bytes)
+{
+	int index = 0;
+
+	for (; index < bytes; index++) {
+		if (put_user(0, (uint8_t __user *)(ptr + index)))
+			return -EFAULT;
+	}
+	return 0;
+}
+
+static int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)
+{
+	u32 hw_info_type = IOMMU_HW_INFO_TYPE_NONE;
+	struct iommu_hw_info *cmd = ucmd->cmd;
+	unsigned int length = 0, data_len;
+	struct iommufd_device *idev;
+	const struct iommu_ops *ops;
+	void __user *user_ptr;
+	void *data = NULL;
+	int rc = 0;
+
+	if (cmd->flags || cmd->__reserved || !cmd->data_len)
+		return -EOPNOTSUPP;
+
+	idev = iommufd_get_device(ucmd, cmd->dev_id);
+	if (IS_ERR(idev))
+		return PTR_ERR(idev);
+
+	user_ptr = u64_to_user_ptr(cmd->data_ptr);
+
+	ops = dev_iommu_ops(idev->dev);
+	if (!ops->hw_info)
+		goto done;
+
+	data = ops->hw_info(idev->dev, &data_len, &hw_info_type);
+	if (IS_ERR(data)) {
+		rc = PTR_ERR(data);
+		goto out_err;
+	}
+
+	/* driver has hw_info callback should have a unique hw_info_type */
+	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
+		rc = -ENODEV;
+		goto out_err;
+	}
+
+	length = min(cmd->data_len, data_len);
+	if (copy_to_user(user_ptr, data, length)) {
+		rc = -EFAULT;
+		goto out_err;
+	}
+
+done:
+	/*
+	 * Zero the trailing bytes if the user buffer is bigger than the
+	 * data size kernel actually has.
+	 */
+	if (length < cmd->data_len) {
+		rc = iommufd_zero_fill_user(user_ptr + length,
+					    cmd->data_len - length);
+		if (rc)
+			goto out_err;
+	}
+
+	cmd->data_len = length;
+	cmd->out_data_type = hw_info_type;
+	rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));
+
+out_err:
+	kfree(data);
+	iommufd_put_object(&idev->obj);
+	return rc;
+}
+
 static int iommufd_fops_open(struct inode *inode, struct file *filp)
 {
 	struct iommufd_ctx *ictx;
@@ -265,6 +341,7 @@ static int iommufd_option(struct iommufd_ucmd *ucmd)
 
 union ucmd_buffer {
 	struct iommu_destroy destroy;
+	struct iommu_hw_info info;
 	struct iommu_hwpt_alloc hwpt;
 	struct iommu_ioas_alloc alloc;
 	struct iommu_ioas_allow_iovas allow_iovas;
@@ -297,6 +374,8 @@ struct iommufd_ioctl_op {
 	}
 static const struct iommufd_ioctl_op iommufd_ioctl_ops[] = {
 	IOCTL_OP(IOMMU_DESTROY, iommufd_destroy, struct iommu_destroy, id),
+	IOCTL_OP(IOMMU_GET_HW_INFO, iommufd_get_hw_info, struct iommu_hw_info,
+		 __reserved),
 	IOCTL_OP(IOMMU_HWPT_ALLOC, iommufd_hwpt_alloc, struct iommu_hwpt_alloc,
 		 __reserved),
 	IOCTL_OP(IOMMU_IOAS_ALLOC, iommufd_ioas_alloc_ioctl,
diff --git a/include/uapi/linux/iommufd.h b/include/uapi/linux/iommufd.h
index 1f616b0f8ae0..4295362e7b44 100644
--- a/include/uapi/linux/iommufd.h
+++ b/include/uapi/linux/iommufd.h
@@ -46,6 +46,7 @@ enum {
 	IOMMUFD_CMD_OPTION,
 	IOMMUFD_CMD_VFIO_IOAS,
 	IOMMUFD_CMD_HWPT_ALLOC,
+	IOMMUFD_CMD_GET_HW_INFO,
 };
 
 /**
@@ -378,4 +379,39 @@ struct iommu_hwpt_alloc {
 enum iommu_hw_info_type {
 	IOMMU_HW_INFO_TYPE_NONE,
 };
+
+/**
+ * struct iommu_hw_info - ioctl(IOMMU_GET_HW_INFO)
+ * @size: sizeof(struct iommu_hw_info)
+ * @flags: Must be 0
+ * @dev_id: The device bound to the iommufd
+ * @data_len: Input the length of the user buffer in bytes. Output the length
+ *            of data filled in the user buffer.
+ * @data_ptr: Pointer to the user buffer
+ * @out_data_type: Output the iommu hardware info type as defined in the enum
+ *                 iommu_hw_info_type.
+ * @__reserved: Must be 0
+ *
+ * Query the hardware information from an iommu behind a given device that has
+ * been bound to iommufd. @data_len is the size of the buffer, which captures an
+ * iommu type specific input data and a filled output data. Trailing bytes will
+ * be zeroed if the user buffer is larger than the data kernel has.
+ *
+ * The type specific data would be used to sync capabilities between the virtual
+ * IOMMU and the hardware IOMMU, e.g. a nested translation setup needs to check
+ * the hardware information, so the guest stage-1 page table will be compatible.
+ *
+ * The @out_data_type will be filled if the ioctl succeeds. It would be used to
+ * decode the data filled in the buffer pointed by @data_ptr.
+ */
+struct iommu_hw_info {
+	__u32 size;
+	__u32 flags;
+	__u32 dev_id;
+	__u32 data_len;
+	__aligned_u64 data_ptr;
+	__u32 out_data_type;
+	__u32 __reserved;
+};
+#define IOMMU_GET_HW_INFO _IO(IOMMUFD_TYPE, IOMMUFD_CMD_GET_HW_INFO)
 #endif
-- 
2.34.1

On 2023/8/3 22:31, Yi Liu wrote:
> Under nested IOMMU translation, userspace owns the stage-1 translation
> table (e.g. the stage-1 page table of Intel VT-d or the context table
> of ARM SMMUv3, and etc.). Stage-1 translation tables are vendor specific,
> and need to be compatiable with the underlying IOMMU hardware. Hence,

                  compatible

> userspace should know the IOMMU hardware capability before creating and
> configuring the stage-1 translation table to kernel.
> 
> This adds IOMMU_GET_HW_INFO ioctl to query the IOMMU hardware information
> (a.k.a capability) for a given device. The returned data is vendor specific,
> userspace needs to decode it with the structure mapped by the @out_data_type
> field.
> 
> As only physical devices have IOMMU hardware, so this will return error
> if the given device is not a physical device.
> 
> 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/iommufd/main.c | 79 ++++++++++++++++++++++++++++++++++++
>   include/uapi/linux/iommufd.h | 36 ++++++++++++++++
>   2 files changed, 115 insertions(+)
> 
> diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
> index 94c498b8fdf6..a0302bcaa97c 100644
> --- a/drivers/iommu/iommufd/main.c
> +++ b/drivers/iommu/iommufd/main.c
> @@ -17,6 +17,7 @@
>   #include <linux/bug.h>
>   #include <uapi/linux/iommufd.h>
>   #include <linux/iommufd.h>
> +#include "../iommu-priv.h"
>   
>   #include "io_pagetable.h"
>   #include "iommufd_private.h"
> @@ -177,6 +178,81 @@ static int iommufd_destroy(struct iommufd_ucmd *ucmd)
>   	return 0;
>   }
>   
> +static int iommufd_zero_fill_user(void __user *ptr, size_t bytes)
> +{
> +	int index = 0;
> +
> +	for (; index < bytes; index++) {
> +		if (put_user(0, (uint8_t __user *)(ptr + index)))
> +			return -EFAULT;
> +	}
> +	return 0;
> +}
> +
> +static int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)
> +{
> +	u32 hw_info_type = IOMMU_HW_INFO_TYPE_NONE;
> +	struct iommu_hw_info *cmd = ucmd->cmd;
> +	unsigned int length = 0, data_len;
> +	struct iommufd_device *idev;
> +	const struct iommu_ops *ops;
> +	void __user *user_ptr;
> +	void *data = NULL;
> +	int rc = 0;
> +
> +	if (cmd->flags || cmd->__reserved || !cmd->data_len)
> +		return -EOPNOTSUPP;
> +
> +	idev = iommufd_get_device(ucmd, cmd->dev_id);
> +	if (IS_ERR(idev))
> +		return PTR_ERR(idev);
> +
> +	user_ptr = u64_to_user_ptr(cmd->data_ptr);
> +
> +	ops = dev_iommu_ops(idev->dev);
> +	if (!ops->hw_info)
> +		goto done;
> +
> +	data = ops->hw_info(idev->dev, &data_len, &hw_info_type);
> +	if (IS_ERR(data)) {
> +		rc = PTR_ERR(data);
> +		goto out_err;

Can kfree() handle a ERR_PTR input? I am afraid not,

/**
  * kfree - free previously allocated memory
  * @object: pointer returned by kmalloc() or kmem_cache_alloc()
  *
  * If @object is NULL, no operation is performed.
  */
void kfree(const void *object)
{
         struct folio *folio;
         struct slab *slab;
         struct kmem_cache *s;

         trace_kfree(_RET_IP_, object);

         if (unlikely(ZERO_OR_NULL_PTR(object)))
                 return;

So, perhaps we should add

	data = NULL;

before
	goto out_err;

?

> +	}
> +
> +	/* driver has hw_info callback should have a unique hw_info_type */
> +	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
> +		rc = -ENODEV;
> +		goto out_err;
> +	}
> +
> +	length = min(cmd->data_len, data_len);
> +	if (copy_to_user(user_ptr, data, length)) {
> +		rc = -EFAULT;
> +		goto out_err;
> +	}
> +
> +done:
> +	/*
> +	 * Zero the trailing bytes if the user buffer is bigger than the
> +	 * data size kernel actually has.
> +	 */
> +	if (length < cmd->data_len) {
> +		rc = iommufd_zero_fill_user(user_ptr + length,
> +					    cmd->data_len - length);
> +		if (rc)
> +			goto out_err;
> +	}
> +
> +	cmd->data_len = length;
> +	cmd->out_data_type = hw_info_type;
> +	rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));
> +
> +out_err:
> +	kfree(data);
> +	iommufd_put_object(&idev->obj);
> +	return rc;
> +}
> +

Others look good to me, feel free to add:

Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>

after above are addressed.

Best regards,
baolu

On Fri, Aug 04, 2023 at 09:59:11AM +0800, Baolu Lu wrote:
> On 2023/8/3 22:31, Yi Liu wrote:
> > Under nested IOMMU translation, userspace owns the stage-1 translation
> > table (e.g. the stage-1 page table of Intel VT-d or the context table
> > of ARM SMMUv3, and etc.). Stage-1 translation tables are vendor specific,
> > and need to be compatiable with the underlying IOMMU hardware. Hence,
> 
>                  compatible
> 
> > userspace should know the IOMMU hardware capability before creating and
> > configuring the stage-1 translation table to kernel.
> > 
> > This adds IOMMU_GET_HW_INFO ioctl to query the IOMMU hardware information
> > (a.k.a capability) for a given device. The returned data is vendor specific,
> > userspace needs to decode it with the structure mapped by the @out_data_type
> > field.
> > 
> > As only physical devices have IOMMU hardware, so this will return error
> > if the given device is not a physical device.
> > 
> > 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/iommufd/main.c | 79 ++++++++++++++++++++++++++++++++++++
> >   include/uapi/linux/iommufd.h | 36 ++++++++++++++++
> >   2 files changed, 115 insertions(+)
> > 
> > diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
> > index 94c498b8fdf6..a0302bcaa97c 100644
> > --- a/drivers/iommu/iommufd/main.c
> > +++ b/drivers/iommu/iommufd/main.c
> > @@ -17,6 +17,7 @@
> >   #include <linux/bug.h>
> >   #include <uapi/linux/iommufd.h>
> >   #include <linux/iommufd.h>
> > +#include "../iommu-priv.h"
> >   #include "io_pagetable.h"
> >   #include "iommufd_private.h"
> > @@ -177,6 +178,81 @@ static int iommufd_destroy(struct iommufd_ucmd *ucmd)
> >   	return 0;
> >   }
> > +static int iommufd_zero_fill_user(void __user *ptr, size_t bytes)
> > +{
> > +	int index = 0;
> > +
> > +	for (; index < bytes; index++) {
> > +		if (put_user(0, (uint8_t __user *)(ptr + index)))
> > +			return -EFAULT;
> > +	}
> > +	return 0;
> > +}
> > +
> > +static int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)
> > +{
> > +	u32 hw_info_type = IOMMU_HW_INFO_TYPE_NONE;
> > +	struct iommu_hw_info *cmd = ucmd->cmd;
> > +	unsigned int length = 0, data_len;
> > +	struct iommufd_device *idev;
> > +	const struct iommu_ops *ops;
> > +	void __user *user_ptr;
> > +	void *data = NULL;
> > +	int rc = 0;
> > +
> > +	if (cmd->flags || cmd->__reserved || !cmd->data_len)
> > +		return -EOPNOTSUPP;
> > +
> > +	idev = iommufd_get_device(ucmd, cmd->dev_id);
> > +	if (IS_ERR(idev))
> > +		return PTR_ERR(idev);
> > +
> > +	user_ptr = u64_to_user_ptr(cmd->data_ptr);
> > +
> > +	ops = dev_iommu_ops(idev->dev);
> > +	if (!ops->hw_info)
> > +		goto done;
> > +
> > +	data = ops->hw_info(idev->dev, &data_len, &hw_info_type);
> > +	if (IS_ERR(data)) {
> > +		rc = PTR_ERR(data);
> > +		goto out_err;
> 
> Can kfree() handle a ERR_PTR input? I am afraid not,
> 
> /**
>  * kfree - free previously allocated memory
>  * @object: pointer returned by kmalloc() or kmem_cache_alloc()
>  *
>  * If @object is NULL, no operation is performed.
>  */
> void kfree(const void *object)
> {
>         struct folio *folio;
>         struct slab *slab;
>         struct kmem_cache *s;
> 
>         trace_kfree(_RET_IP_, object);
> 
>         if (unlikely(ZERO_OR_NULL_PTR(object)))
>                 return;
> 
> So, perhaps we should add
> 
> 	data = NULL;
> 
> before
> 	goto out_err;

No, use more labels.

> > +	}
> > +
> > +	/* driver has hw_info callback should have a unique hw_info_type */
> > +	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
> > +		rc = -ENODEV;
> > +		goto out_err;
> > +	}
> > +
> > +	length = min(cmd->data_len, data_len);
> > +	if (copy_to_user(user_ptr, data, length)) {
> > +		rc = -EFAULT;
> > +		goto out_err;
> > +	}
> > +
> > +done:

No label here, put it in a if {}

> > +	/*
> > +	 * Zero the trailing bytes if the user buffer is bigger than the
> > +	 * data size kernel actually has.
> > +	 */
> > +	if (length < cmd->data_len) {
> > +		rc = iommufd_zero_fill_user(user_ptr + length,
> > +					    cmd->data_len - length);
> > +		if (rc)
> > +			goto out_err;
> > +	}
> > +
> > +	cmd->data_len = length;
> > +	cmd->out_data_type = hw_info_type;
> > +	rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));
> > +
> > +out_err:

err_free:
> > +	kfree(data);

err_put:

> > +	iommufd_put_object(&idev->obj);
> > +	return rc;
> > +}

Jason

> From: Jason Gunthorpe <jgg@nvidia.com>
> Sent: Friday, August 4, 2023 10:12 PM
> 
> On Fri, Aug 04, 2023 at 09:59:11AM +0800, Baolu Lu wrote:
> > On 2023/8/3 22:31, Yi Liu wrote:
> > > Under nested IOMMU translation, userspace owns the stage-1 translation
> > > table (e.g. the stage-1 page table of Intel VT-d or the context table
> > > of ARM SMMUv3, and etc.). Stage-1 translation tables are vendor specific,
> > > and need to be compatiable with the underlying IOMMU hardware. Hence,
> >
> >                  compatible
> >
> > > userspace should know the IOMMU hardware capability before creating and
> > > configuring the stage-1 translation table to kernel.
> > >
> > > This adds IOMMU_GET_HW_INFO ioctl to query the IOMMU hardware information
> > > (a.k.a capability) for a given device. The returned data is vendor specific,
> > > userspace needs to decode it with the structure mapped by the @out_data_type
> > > field.
> > >
> > > As only physical devices have IOMMU hardware, so this will return error
> > > if the given device is not a physical device.
> > >
> > > 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/iommufd/main.c | 79 ++++++++++++++++++++++++++++++++++++
> > >   include/uapi/linux/iommufd.h | 36 ++++++++++++++++
> > >   2 files changed, 115 insertions(+)
> > >
> > > diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
> > > index 94c498b8fdf6..a0302bcaa97c 100644
> > > --- a/drivers/iommu/iommufd/main.c
> > > +++ b/drivers/iommu/iommufd/main.c
> > > @@ -17,6 +17,7 @@
> > >   #include <linux/bug.h>
> > >   #include <uapi/linux/iommufd.h>
> > >   #include <linux/iommufd.h>
> > > +#include "../iommu-priv.h"
> > >   #include "io_pagetable.h"
> > >   #include "iommufd_private.h"
> > > @@ -177,6 +178,81 @@ static int iommufd_destroy(struct iommufd_ucmd *ucmd)
> > >   	return 0;
> > >   }
> > > +static int iommufd_zero_fill_user(void __user *ptr, size_t bytes)
> > > +{
> > > +	int index = 0;
> > > +
> > > +	for (; index < bytes; index++) {
> > > +		if (put_user(0, (uint8_t __user *)(ptr + index)))
> > > +			return -EFAULT;
> > > +	}
> > > +	return 0;
> > > +}
> > > +
> > > +static int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)
> > > +{
> > > +	u32 hw_info_type = IOMMU_HW_INFO_TYPE_NONE;
> > > +	struct iommu_hw_info *cmd = ucmd->cmd;
> > > +	unsigned int length = 0, data_len;
> > > +	struct iommufd_device *idev;
> > > +	const struct iommu_ops *ops;
> > > +	void __user *user_ptr;
> > > +	void *data = NULL;
> > > +	int rc = 0;
> > > +
> > > +	if (cmd->flags || cmd->__reserved || !cmd->data_len)
> > > +		return -EOPNOTSUPP;
> > > +
> > > +	idev = iommufd_get_device(ucmd, cmd->dev_id);
> > > +	if (IS_ERR(idev))
> > > +		return PTR_ERR(idev);
> > > +
> > > +	user_ptr = u64_to_user_ptr(cmd->data_ptr);
> > > +
> > > +	ops = dev_iommu_ops(idev->dev);
> > > +	if (!ops->hw_info)
> > > +		goto done;
> > > +
> > > +	data = ops->hw_info(idev->dev, &data_len, &hw_info_type);
> > > +	if (IS_ERR(data)) {
> > > +		rc = PTR_ERR(data);
> > > +		goto out_err;
> >
> > Can kfree() handle a ERR_PTR input? I am afraid not,
> >
> > /**
> >  * kfree - free previously allocated memory
> >  * @object: pointer returned by kmalloc() or kmem_cache_alloc()
> >  *
> >  * If @object is NULL, no operation is performed.
> >  */
> > void kfree(const void *object)
> > {
> >         struct folio *folio;
> >         struct slab *slab;
> >         struct kmem_cache *s;
> >
> >         trace_kfree(_RET_IP_, object);
> >
> >         if (unlikely(ZERO_OR_NULL_PTR(object)))
> >                 return;
> >
> > So, perhaps we should add
> >
> > 	data = NULL;
> >
> > before
> > 	goto out_err;
> 
> No, use more labels.
> 
> > > +	}
> > > +
> > > +	/* driver has hw_info callback should have a unique hw_info_type */
> > > +	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
> > > +		rc = -ENODEV;
> > > +		goto out_err;
> > > +	}
> > > +
> > > +	length = min(cmd->data_len, data_len);
> > > +	if (copy_to_user(user_ptr, data, length)) {
> > > +		rc = -EFAULT;
> > > +		goto out_err;
> > > +	}
> > > +
> > > +done:
> 
> No label here, put it in a if {}

How about below?

diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
index 94c498b8fdf6..a8bb7da87036 100644
--- a/drivers/iommu/iommufd/main.c
+++ b/drivers/iommu/iommufd/main.c
@@ -17,6 +17,7 @@
 #include <linux/bug.h>
 #include <uapi/linux/iommufd.h>
 #include <linux/iommufd.h>
+#include "../iommu-priv.h"
 
 #include "io_pagetable.h"
 #include "iommufd_private.h"
@@ -177,6 +178,80 @@ static int iommufd_destroy(struct iommufd_ucmd *ucmd)
 	return 0;
 }
 
+static int iommufd_zero_fill_user(void __user *ptr, size_t bytes)
+{
+	int index = 0;
+
+	for (; index < bytes; index++) {
+		if (put_user(0, (uint8_t __user *)(ptr + index)))
+			return -EFAULT;
+	}
+	return 0;
+}
+
+static int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)
+{
+	u32 hw_info_type = IOMMU_HW_INFO_TYPE_NONE;
+	struct iommu_hw_info *cmd = ucmd->cmd;
+	unsigned int length = 0, data_len;
+	struct iommufd_device *idev;
+	const struct iommu_ops *ops;
+	void __user *user_ptr;
+	void *data = NULL;
+	int rc = 0;
+
+	if (cmd->flags || cmd->__reserved || !cmd->data_len)
+		return -EOPNOTSUPP;
+
+	idev = iommufd_get_device(ucmd, cmd->dev_id);
+	if (IS_ERR(idev))
+		return PTR_ERR(idev);
+
+	user_ptr = u64_to_user_ptr(cmd->data_ptr);
+
+	ops = dev_iommu_ops(idev->dev);
+	if (ops->hw_info) {
+		data = ops->hw_info(idev->dev, &data_len, &hw_info_type);
+		if (IS_ERR(data)) {
+			rc = PTR_ERR(data);
+			goto err_put;
+		}
+
+		/* driver has hw_info callback should have a unique hw_info_type */
+		if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
+			rc = -ENODEV;
+			goto err_free;
+		}
+
+		length = min(cmd->data_len, data_len);
+		if (copy_to_user(user_ptr, data, length)) {
+			rc = -EFAULT;
+			goto err_free;
+		}
+	}
+
+	/*
+	 * Zero the trailing bytes if the user buffer is bigger than the
+	 * data size kernel actually has.
+	 */
+	if (length < cmd->data_len) {
+		rc = iommufd_zero_fill_user(user_ptr + length,
+					    cmd->data_len - length);
+		if (rc)
+			goto err_free;
+	}
+
+	cmd->data_len = length;
+	cmd->out_data_type = hw_info_type;
+	rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));
+
+err_free:
+	kfree(data);
+err_put:
+	iommufd_put_object(&idev->obj);
+	return rc;
+}
+
 static int iommufd_fops_open(struct inode *inode, struct file *filp)
 {
 	struct iommufd_ctx *ictx;

Regards,
Yi Liu

> > > +	/*
> > > +	 * Zero the trailing bytes if the user buffer is bigger than the
> > > +	 * data size kernel actually has.
> > > +	 */
> > > +	if (length < cmd->data_len) {
> > > +		rc = iommufd_zero_fill_user(user_ptr + length,
> > > +					    cmd->data_len - length);
> > > +		if (rc)
> > > +			goto out_err;
> > > +	}
> > > +
> > > +	cmd->data_len = length;
> > > +	cmd->out_data_type = hw_info_type;
> > > +	rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));
> > > +
> > > +out_err:
> 
> err_free:
> > > +	kfree(data);
> 
> err_put:
> 
> > > +	iommufd_put_object(&idev->obj);
> > > +	return rc;
> > > +}
> 
> Jason

On Mon, Aug 07, 2023 at 08:49:27AM +0000, Liu, Yi L wrote:

> > > > +	/* driver has hw_info callback should have a unique hw_info_type */
> > > > +	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
> > > > +		rc = -ENODEV;
> > > > +		goto out_err;
> > > > +	}
> > > > +
> > > > +	length = min(cmd->data_len, data_len);
> > > > +	if (copy_to_user(user_ptr, data, length)) {
> > > > +		rc = -EFAULT;
> > > > +		goto out_err;
> > > > +	}
> > > > +
> > > > +done:
> > 
> > No label here, put it in a if {}
> 
> How about below?

Looks better

You could also consider putting the content of the if {} in a function

Jason

> From: Jason Gunthorpe <jgg@nvidia.com>
> Sent: Tuesday, August 8, 2023 9:03 PM
> 
> On Mon, Aug 07, 2023 at 08:49:27AM +0000, Liu, Yi L wrote:
> 
> > > > > +	/* driver has hw_info callback should have a unique hw_info_type */
> > > > > +	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE))
> {
> > > > > +		rc = -ENODEV;
> > > > > +		goto out_err;
> > > > > +	}
> > > > > +
> > > > > +	length = min(cmd->data_len, data_len);
> > > > > +	if (copy_to_user(user_ptr, data, length)) {
> > > > > +		rc = -EFAULT;
> > > > > +		goto out_err;
> > > > > +	}
> > > > > +
> > > > > +done:
> > >
> > > No label here, put it in a if {}
> >
> > How about below?
> 
> Looks better
> 
> You could also consider putting the content of the if {} in a function

Sure.

Regards,
Yi Liu

> From: Baolu Lu <baolu.lu@linux.intel.com>
> Sent: Friday, August 4, 2023 9:59 AM
> 
> On 2023/8/3 22:31, Yi Liu wrote:
> > Under nested IOMMU translation, userspace owns the stage-1 translation
> > table (e.g. the stage-1 page table of Intel VT-d or the context table
> > of ARM SMMUv3, and etc.). Stage-1 translation tables are vendor specific,
> > and need to be compatiable with the underlying IOMMU hardware. Hence,
> 
>                   compatible

Yes.

> > userspace should know the IOMMU hardware capability before creating and
> > configuring the stage-1 translation table to kernel.
> >
> > This adds IOMMU_GET_HW_INFO ioctl to query the IOMMU hardware information
> > (a.k.a capability) for a given device. The returned data is vendor specific,
> > userspace needs to decode it with the structure mapped by the @out_data_type
> > field.
> >
> > As only physical devices have IOMMU hardware, so this will return error
> > if the given device is not a physical device.
> >
> > 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/iommufd/main.c | 79 ++++++++++++++++++++++++++++++++++++
> >   include/uapi/linux/iommufd.h | 36 ++++++++++++++++
> >   2 files changed, 115 insertions(+)
> >
> > diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
> > index 94c498b8fdf6..a0302bcaa97c 100644
> > --- a/drivers/iommu/iommufd/main.c
> > +++ b/drivers/iommu/iommufd/main.c
> > @@ -17,6 +17,7 @@
> >   #include <linux/bug.h>
> >   #include <uapi/linux/iommufd.h>
> >   #include <linux/iommufd.h>
> > +#include "../iommu-priv.h"
> >
> >   #include "io_pagetable.h"
> >   #include "iommufd_private.h"
> > @@ -177,6 +178,81 @@ static int iommufd_destroy(struct iommufd_ucmd *ucmd)
> >   	return 0;
> >   }
> >
> > +static int iommufd_zero_fill_user(void __user *ptr, size_t bytes)
> > +{
> > +	int index = 0;
> > +
> > +	for (; index < bytes; index++) {
> > +		if (put_user(0, (uint8_t __user *)(ptr + index)))
> > +			return -EFAULT;
> > +	}
> > +	return 0;
> > +}
> > +
> > +static int iommufd_get_hw_info(struct iommufd_ucmd *ucmd)
> > +{
> > +	u32 hw_info_type = IOMMU_HW_INFO_TYPE_NONE;
> > +	struct iommu_hw_info *cmd = ucmd->cmd;
> > +	unsigned int length = 0, data_len;
> > +	struct iommufd_device *idev;
> > +	const struct iommu_ops *ops;
> > +	void __user *user_ptr;
> > +	void *data = NULL;
> > +	int rc = 0;
> > +
> > +	if (cmd->flags || cmd->__reserved || !cmd->data_len)
> > +		return -EOPNOTSUPP;
> > +
> > +	idev = iommufd_get_device(ucmd, cmd->dev_id);
> > +	if (IS_ERR(idev))
> > +		return PTR_ERR(idev);
> > +
> > +	user_ptr = u64_to_user_ptr(cmd->data_ptr);
> > +
> > +	ops = dev_iommu_ops(idev->dev);
> > +	if (!ops->hw_info)
> > +		goto done;
> > +
> > +	data = ops->hw_info(idev->dev, &data_len, &hw_info_type);
> > +	if (IS_ERR(data)) {
> > +		rc = PTR_ERR(data);
> > +		goto out_err;
> 
> Can kfree() handle a ERR_PTR input? I am afraid not,
> 
> /**
>   * kfree - free previously allocated memory
>   * @object: pointer returned by kmalloc() or kmem_cache_alloc()
>   *
>   * If @object is NULL, no operation is performed.
>   */
> void kfree(const void *object)
> {
>          struct folio *folio;
>          struct slab *slab;
>          struct kmem_cache *s;
> 
>          trace_kfree(_RET_IP_, object);
> 
>          if (unlikely(ZERO_OR_NULL_PTR(object)))
>                  return;
> 
> So, perhaps we should add
> 
> 	data = NULL;
> 
> before
> 	goto out_err;
> 
> ?

Yes.

> 
> > +	}
> > +
> > +	/* driver has hw_info callback should have a unique hw_info_type */
> > +	if (WARN_ON_ONCE(hw_info_type == IOMMU_HW_INFO_TYPE_NONE)) {
> > +		rc = -ENODEV;
> > +		goto out_err;
> > +	}
> > +
> > +	length = min(cmd->data_len, data_len);
> > +	if (copy_to_user(user_ptr, data, length)) {
> > +		rc = -EFAULT;
> > +		goto out_err;
> > +	}
> > +
> > +done:
> > +	/*
> > +	 * Zero the trailing bytes if the user buffer is bigger than the
> > +	 * data size kernel actually has.
> > +	 */
> > +	if (length < cmd->data_len) {
> > +		rc = iommufd_zero_fill_user(user_ptr + length,
> > +					    cmd->data_len - length);
> > +		if (rc)
> > +			goto out_err;
> > +	}
> > +
> > +	cmd->data_len = length;
> > +	cmd->out_data_type = hw_info_type;
> > +	rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));
> > +
> > +out_err:
> > +	kfree(data);
> > +	iommufd_put_object(&idev->obj);
> > +	return rc;
> > +}
> > +
> 
> Others look good to me, feel free to add:
> 
> Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
> 
> after above are addressed.

Regards,
Yi Liu