Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1394182 > unrolled thread

[PATCH v9 0/7] KVM PCIe/MSI passthrough on ARM/ARM64: kernel part 3/3: vfio changes

Back to article view | Back to linux.kernel

Contents

  [PATCH v9 0/7] KVM PCIe/MSI passthrough on ARM/ARM64: kernel part 3/3: vfio changes Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
    [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at msi domain Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
      Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain "Chalamarla, Tirumalesh" <Tirumalesh.Chalamarla@caviumnetworks.com> - 2016-05-05 21:30 +0200
        Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Eric Auger <eric.auger@linaro.org> - 2016-05-09 10:10 +0200
      Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability  at msi domain Alex Williamson <alex.williamson@redhat.com> - 2016-05-10 00:50 +0200
        Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Eric Auger <eric.auger@linaro.org> - 2016-05-10 18:20 +0200
          Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Robin Murphy <robin.murphy@arm.com> - 2016-05-10 19:30 +0200
            Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Eric Auger <eric.auger@linaro.org> - 2016-05-11 10:50 +0200
              Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Robin Murphy <robin.murphy@arm.com> - 2016-05-11 11:40 +0200
                Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Eric Auger <eric.auger@linaro.org> - 2016-05-11 11:50 +0200
                  Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Robin Murphy <robin.murphy@arm.com> - 2016-05-11 15:50 +0200
                    Re: [PATCH v9 5/7] vfio/type1: also check IRQ remapping capability at  msi domain Eric Auger <eric.auger@linaro.org> - 2016-05-11 16:40 +0200
    [PATCH v9 6/7] iommu/arm-smmu: do not advertise IOMMU_CAP_INTR_REMAP Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
    [PATCH v9 1/7] vfio: introduce a vfio_dma type field Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
    [PATCH v9 7/7] vfio/type1: return MSI geometry through VFIO_IOMMU_GET_INFO capability chains Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
      Re: [PATCH v9 7/7] vfio/type1: return MSI geometry through  VFIO_IOMMU_GET_INFO capability chains Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:10 +0200
        Re: [PATCH v9 7/7] vfio/type1: return MSI geometry through  VFIO_IOMMU_GET_INFO capability chains Alex Williamson <alex.williamson@redhat.com> - 2016-05-10 01:10 +0200
          Re: [PATCH v9 7/7] vfio/type1: return MSI geometry through  VFIO_IOMMU_GET_INFO capability chains Eric Auger <eric.auger@linaro.org> - 2016-05-10 19:00 +0200
      Re: [PATCH v9 7/7] vfio/type1: return MSI geometry through  VFIO_IOMMU_GET_INFO capability chains Alex Williamson <alex.williamson@redhat.com> - 2016-05-10 01:00 +0200
        Re: [PATCH v9 7/7] vfio/type1: return MSI geometry through  VFIO_IOMMU_GET_INFO capability chains Eric Auger <eric.auger@linaro.org> - 2016-05-10 18:40 +0200
    [PATCH v9 2/7] vfio/type1: vfio_find_dma accepting a type argument Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
      Re: [PATCH v9 2/7] vfio/type1: vfio_find_dma accepting a type  argument Alex Williamson <alex.williamson@redhat.com> - 2016-05-10 00:50 +0200
        Re: [PATCH v9 2/7] vfio/type1: vfio_find_dma accepting a type  argument Eric Auger <eric.auger@linaro.org> - 2016-05-10 17:00 +0200
    [PATCH v9 4/7] vfio: allow reserved msi iova registration Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
      Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration "Chalamarla, Tirumalesh" <Tirumalesh.Chalamarla@caviumnetworks.com> - 2016-05-05 21:40 +0200
        Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration Eric Auger <eric.auger@linaro.org> - 2016-05-09 10:00 +0200
      Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration Alex Williamson <alex.williamson@redhat.com> - 2016-05-10 17:40 +0200
        Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration Eric Auger <eric.auger@linaro.org> - 2016-05-10 17:40 +0200
    [PATCH v9 3/7] vfio/type1: bypass unmap/unpin and replay for VFIO_IOVA_RESERVED slots Eric Auger <eric.auger@linaro.org> - 2016-05-04 14:00 +0200
      Re: [PATCH v9 3/7] vfio/type1: bypass unmap/unpin and replay for  VFIO_IOVA_RESERVED slots Alex Williamson <alex.williamson@redhat.com> - 2016-05-10 00:50 +0200
        Re: [PATCH v9 3/7] vfio/type1: bypass unmap/unpin and replay for  VFIO_IOVA_RESERVED slots Eric Auger <eric.auger@linaro.org> - 2016-05-11 15:00 +0200

Page 2 of 2 — ← Prev page 1 [2]

#1394190 — [PATCH v9 2/7] vfio/type1: vfio_find_dma accepting a type argument

In our RB-tree we now have slots of different types (USER and RESERVED).
It becomes useful to be able to search for dma slots of a specific type or
any type. This patch proposes an implementation for that modality and also
changes the existing callers using the USER type.

Signed-off-by: Eric Auger <eric.auger@linaro.org>
---
 drivers/vfio/vfio_iommu_type1.c | 63 ++++++++++++++++++++++++++++++++++-------
 1 file changed, 53 insertions(+), 10 deletions(-)

diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
index aaf5a6c..2d769d4 100644
--- a/drivers/vfio/vfio_iommu_type1.c
+++ b/drivers/vfio/vfio_iommu_type1.c
@@ -98,23 +98,64 @@ struct vfio_group {
  * into DMA'ble space using the IOMMU
  */
 
-static struct vfio_dma *vfio_find_dma(struct vfio_iommu *iommu,
-				      dma_addr_t start, size_t size)
+/**
+ * vfio_find_dma_from_node: looks for a dma slot intersecting a window
+ * from a given rb tree node
+ * @top: top rb tree node where the search starts (including this node)
+ * @start: window start
+ * @size: window size
+ * @type: window type
+ */
+static struct vfio_dma *vfio_find_dma_from_node(struct rb_node *top,
+						dma_addr_t start, size_t size,
+						enum vfio_iova_type type)
 {
-	struct rb_node *node = iommu->dma_list.rb_node;
+	struct rb_node *node = top;
+	struct vfio_dma *dma;
 
 	while (node) {
-		struct vfio_dma *dma = rb_entry(node, struct vfio_dma, node);
-
+		dma = rb_entry(node, struct vfio_dma, node);
 		if (start + size <= dma->iova)
 			node = node->rb_left;
 		else if (start >= dma->iova + dma->size)
 			node = node->rb_right;
 		else
+			break;
+	}
+	if (!node)
+		return NULL;
+
+	/* a dma slot intersects our window, check the type also matches */
+	if (type == VFIO_IOVA_ANY || dma->type == type)
+		return dma;
+
+	/* restart 2 searches skipping the current node */
+	if (start < dma->iova) {
+		dma = vfio_find_dma_from_node(node->rb_left, start,
+					      size, type);
+		if (dma)
 			return dma;
 	}
+	if (start + size > dma->iova + dma->size)
+		dma = vfio_find_dma_from_node(node->rb_right, start,
+					      size, type);
+	return dma;
+}
+
+/**
+ * vfio_find_dma: find a dma slot intersecting a given window
+ * @iommu: vfio iommu handle
+ * @start: window base iova
+ * @size: window size
+ * @type: window type
+ */
+static struct vfio_dma *vfio_find_dma(struct vfio_iommu *iommu,
+				      dma_addr_t start, size_t size,
+				      enum vfio_iova_type type)
+{
+	struct rb_node *top_node = iommu->dma_list.rb_node;
 
-	return NULL;
+	return vfio_find_dma_from_node(top_node, start, size, type);
 }
 
 static void vfio_link_dma(struct vfio_iommu *iommu, struct vfio_dma *new)
@@ -488,19 +529,21 @@ static int vfio_dma_do_unmap(struct vfio_iommu *iommu,
 	 * mappings within the range.
 	 */
 	if (iommu->v2) {
-		dma = vfio_find_dma(iommu, unmap->iova, 0);
+		dma = vfio_find_dma(iommu, unmap->iova, 0, VFIO_IOVA_USER);
 		if (dma && dma->iova != unmap->iova) {
 			ret = -EINVAL;
 			goto unlock;
 		}
-		dma = vfio_find_dma(iommu, unmap->iova + unmap->size - 1, 0);
+		dma = vfio_find_dma(iommu, unmap->iova + unmap->size - 1, 0,
+				    VFIO_IOVA_USER);
 		if (dma && dma->iova + dma->size != unmap->iova + unmap->size) {
 			ret = -EINVAL;
 			goto unlock;
 		}
 	}
 
-	while ((dma = vfio_find_dma(iommu, unmap->iova, unmap->size))) {
+	while ((dma = vfio_find_dma(iommu, unmap->iova, unmap->size,
+				    VFIO_IOVA_USER))) {
 		if (!iommu->v2 && unmap->iova > dma->iova)
 			break;
 		unmapped += dma->size;
@@ -604,7 +647,7 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
 
 	mutex_lock(&iommu->lock);
 
-	if (vfio_find_dma(iommu, iova, size)) {
+	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
 		mutex_unlock(&iommu->lock);
 		return -EEXIST;
 	}
-- 
1.9.1

[toc] | [prev] | [next] | [standalone]

#1397492 — Re: [PATCH v9 2/7] vfio/type1: vfio_find_dma accepting a type argument

On Wed,  4 May 2016 11:54:13 +0000
Eric Auger <eric.auger@linaro.org> wrote:

> In our RB-tree we now have slots of different types (USER and RESERVED).
> It becomes useful to be able to search for dma slots of a specific type or
> any type. This patch proposes an implementation for that modality and also
> changes the existing callers using the USER type.
> 
> Signed-off-by: Eric Auger <eric.auger@linaro.org>
> ---
>  drivers/vfio/vfio_iommu_type1.c | 63 ++++++++++++++++++++++++++++++++++-------
>  1 file changed, 53 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
> index aaf5a6c..2d769d4 100644
> --- a/drivers/vfio/vfio_iommu_type1.c
> +++ b/drivers/vfio/vfio_iommu_type1.c
> @@ -98,23 +98,64 @@ struct vfio_group {
>   * into DMA'ble space using the IOMMU
>   */
>  
> -static struct vfio_dma *vfio_find_dma(struct vfio_iommu *iommu,
> -				      dma_addr_t start, size_t size)
> +/**
> + * vfio_find_dma_from_node: looks for a dma slot intersecting a window
> + * from a given rb tree node
> + * @top: top rb tree node where the search starts (including this node)
> + * @start: window start
> + * @size: window size
> + * @type: window type
> + */
> +static struct vfio_dma *vfio_find_dma_from_node(struct rb_node *top,
> +						dma_addr_t start, size_t size,
> +						enum vfio_iova_type type)
>  {
> -	struct rb_node *node = iommu->dma_list.rb_node;
> +	struct rb_node *node = top;
> +	struct vfio_dma *dma;
>  
>  	while (node) {
> -		struct vfio_dma *dma = rb_entry(node, struct vfio_dma, node);
> -
> +		dma = rb_entry(node, struct vfio_dma, node);
>  		if (start + size <= dma->iova)
>  			node = node->rb_left;
>  		else if (start >= dma->iova + dma->size)
>  			node = node->rb_right;
>  		else
> +			break;
> +	}
> +	if (!node)
> +		return NULL;
> +
> +	/* a dma slot intersects our window, check the type also matches */
> +	if (type == VFIO_IOVA_ANY || dma->type == type)
> +		return dma;
> +
> +	/* restart 2 searches skipping the current node */
> +	if (start < dma->iova) {
> +		dma = vfio_find_dma_from_node(node->rb_left, start,
> +					      size, type);
> +		if (dma)
>  			return dma;
>  	}
> +	if (start + size > dma->iova + dma->size)
> +		dma = vfio_find_dma_from_node(node->rb_right, start,
> +					      size, type);

It might be nice to split this little bit of trickery out to a separate
patch so that you can better explain what it does.  It's really not
needed until we start creating user-type entries, so it shouldn't hurt
bisection.

> +	return dma;
> +}
> +
> +/**
> + * vfio_find_dma: find a dma slot intersecting a given window
> + * @iommu: vfio iommu handle
> + * @start: window base iova
> + * @size: window size
> + * @type: window type
> + */
> +static struct vfio_dma *vfio_find_dma(struct vfio_iommu *iommu,
> +				      dma_addr_t start, size_t size,
> +				      enum vfio_iova_type type)
> +{
> +	struct rb_node *top_node = iommu->dma_list.rb_node;
>  
> -	return NULL;
> +	return vfio_find_dma_from_node(top_node, start, size, type);
>  }
>  
>  static void vfio_link_dma(struct vfio_iommu *iommu, struct vfio_dma *new)
> @@ -488,19 +529,21 @@ static int vfio_dma_do_unmap(struct vfio_iommu *iommu,
>  	 * mappings within the range.
>  	 */
>  	if (iommu->v2) {
> -		dma = vfio_find_dma(iommu, unmap->iova, 0);
> +		dma = vfio_find_dma(iommu, unmap->iova, 0, VFIO_IOVA_USER);
>  		if (dma && dma->iova != unmap->iova) {
>  			ret = -EINVAL;
>  			goto unlock;
>  		}
> -		dma = vfio_find_dma(iommu, unmap->iova + unmap->size - 1, 0);
> +		dma = vfio_find_dma(iommu, unmap->iova + unmap->size - 1, 0,
> +				    VFIO_IOVA_USER);
>  		if (dma && dma->iova + dma->size != unmap->iova + unmap->size) {
>  			ret = -EINVAL;
>  			goto unlock;
>  		}
>  	}
>  
> -	while ((dma = vfio_find_dma(iommu, unmap->iova, unmap->size))) {
> +	while ((dma = vfio_find_dma(iommu, unmap->iova, unmap->size,
> +				    VFIO_IOVA_USER))) {
>  		if (!iommu->v2 && unmap->iova > dma->iova)
>  			break;
>  		unmapped += dma->size;
> @@ -604,7 +647,7 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
>  
>  	mutex_lock(&iommu->lock);
>  
> -	if (vfio_find_dma(iommu, iova, size)) {
> +	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
>  		mutex_unlock(&iommu->lock);
>  		return -EEXIST;
>  	}

[toc] | [prev] | [next] | [standalone]

#1398172 — Re: [PATCH v9 2/7] vfio/type1: vfio_find_dma accepting a type argument

Hi Alex,
On 05/10/2016 12:49 AM, Alex Williamson wrote:
> On Wed,  4 May 2016 11:54:13 +0000
> Eric Auger <eric.auger@linaro.org> wrote:
> 
>> In our RB-tree we now have slots of different types (USER and RESERVED).
>> It becomes useful to be able to search for dma slots of a specific type or
>> any type. This patch proposes an implementation for that modality and also
>> changes the existing callers using the USER type.
>>
>> Signed-off-by: Eric Auger <eric.auger@linaro.org>
>> ---
>>  drivers/vfio/vfio_iommu_type1.c | 63 ++++++++++++++++++++++++++++++++++-------
>>  1 file changed, 53 insertions(+), 10 deletions(-)
>>
>> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
>> index aaf5a6c..2d769d4 100644
>> --- a/drivers/vfio/vfio_iommu_type1.c
>> +++ b/drivers/vfio/vfio_iommu_type1.c
>> @@ -98,23 +98,64 @@ struct vfio_group {
>>   * into DMA'ble space using the IOMMU
>>   */
>>  
>> -static struct vfio_dma *vfio_find_dma(struct vfio_iommu *iommu,
>> -				      dma_addr_t start, size_t size)
>> +/**
>> + * vfio_find_dma_from_node: looks for a dma slot intersecting a window
>> + * from a given rb tree node
>> + * @top: top rb tree node where the search starts (including this node)
>> + * @start: window start
>> + * @size: window size
>> + * @type: window type
>> + */
>> +static struct vfio_dma *vfio_find_dma_from_node(struct rb_node *top,
>> +						dma_addr_t start, size_t size,
>> +						enum vfio_iova_type type)
>>  {
>> -	struct rb_node *node = iommu->dma_list.rb_node;
>> +	struct rb_node *node = top;
>> +	struct vfio_dma *dma;
>>  
>>  	while (node) {
>> -		struct vfio_dma *dma = rb_entry(node, struct vfio_dma, node);
>> -
>> +		dma = rb_entry(node, struct vfio_dma, node);
>>  		if (start + size <= dma->iova)
>>  			node = node->rb_left;
>>  		else if (start >= dma->iova + dma->size)
>>  			node = node->rb_right;
>>  		else
>> +			break;
>> +	}
>> +	if (!node)
>> +		return NULL;
>> +
>> +	/* a dma slot intersects our window, check the type also matches */
>> +	if (type == VFIO_IOVA_ANY || dma->type == type)
>> +		return dma;
>> +
>> +	/* restart 2 searches skipping the current node */
>> +	if (start < dma->iova) {
>> +		dma = vfio_find_dma_from_node(node->rb_left, start,
>> +					      size, type);
>> +		if (dma)
>>  			return dma;
>>  	}
>> +	if (start + size > dma->iova + dma->size)
>> +		dma = vfio_find_dma_from_node(node->rb_right, start,
>> +					      size, type);
> 
> It might be nice to split this little bit of trickery out to a separate
> patch so that you can better explain what it does.  It's really not
> needed until we start creating user-type entries, so it shouldn't hurt
> bisection.
OK I will

Regards

Eric
> 
>> +	return dma;
>> +}
>> +
>> +/**
>> + * vfio_find_dma: find a dma slot intersecting a given window
>> + * @iommu: vfio iommu handle
>> + * @start: window base iova
>> + * @size: window size
>> + * @type: window type
>> + */
>> +static struct vfio_dma *vfio_find_dma(struct vfio_iommu *iommu,
>> +				      dma_addr_t start, size_t size,
>> +				      enum vfio_iova_type type)
>> +{
>> +	struct rb_node *top_node = iommu->dma_list.rb_node;
>>  
>> -	return NULL;
>> +	return vfio_find_dma_from_node(top_node, start, size, type);
>>  }
>>  
>>  static void vfio_link_dma(struct vfio_iommu *iommu, struct vfio_dma *new)
>> @@ -488,19 +529,21 @@ static int vfio_dma_do_unmap(struct vfio_iommu *iommu,
>>  	 * mappings within the range.
>>  	 */
>>  	if (iommu->v2) {
>> -		dma = vfio_find_dma(iommu, unmap->iova, 0);
>> +		dma = vfio_find_dma(iommu, unmap->iova, 0, VFIO_IOVA_USER);
>>  		if (dma && dma->iova != unmap->iova) {
>>  			ret = -EINVAL;
>>  			goto unlock;
>>  		}
>> -		dma = vfio_find_dma(iommu, unmap->iova + unmap->size - 1, 0);
>> +		dma = vfio_find_dma(iommu, unmap->iova + unmap->size - 1, 0,
>> +				    VFIO_IOVA_USER);
>>  		if (dma && dma->iova + dma->size != unmap->iova + unmap->size) {
>>  			ret = -EINVAL;
>>  			goto unlock;
>>  		}
>>  	}
>>  
>> -	while ((dma = vfio_find_dma(iommu, unmap->iova, unmap->size))) {
>> +	while ((dma = vfio_find_dma(iommu, unmap->iova, unmap->size,
>> +				    VFIO_IOVA_USER))) {
>>  		if (!iommu->v2 && unmap->iova > dma->iova)
>>  			break;
>>  		unmapped += dma->size;
>> @@ -604,7 +647,7 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
>>  
>>  	mutex_lock(&iommu->lock);
>>  
>> -	if (vfio_find_dma(iommu, iova, size)) {
>> +	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
>>  		mutex_unlock(&iommu->lock);
>>  		return -EEXIST;
>>  	}
> 

[toc] | [prev] | [next] | [standalone]

#1394197 — [PATCH v9 4/7] vfio: allow reserved msi iova registration

The user is allowed to register a reserved MSI IOVA range by using the
DMA MAP API and setting the new flag: VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA.
This region is stored in the vfio_dma rb tree. At that point the iova
range is not mapped to any target address yet. The host kernel will use
those iova when needed, typically when MSIs are allocated.

Signed-off-by: Eric Auger <eric.auger@linaro.org>
Signed-off-by: Bharat Bhushan <Bharat.Bhushan@freescale.com>

---
v7 -> v8:
- use iommu_msi_set_aperture function. There is no notion of
  unregistration anymore since the reserved msi slot remains
  until the container gets closed.

v6 -> v7:
- use iommu_free_reserved_iova_domain
- convey prot attributes downto dma-reserved-iommu iova domain creation
- reserved bindings teardown now performed on iommu domain destruction
- rename VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA into
         VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA
- change title
- pass the protection attribute to dma-reserved-iommu API

v3 -> v4:
- use iommu_alloc/free_reserved_iova_domain exported by dma-reserved-iommu
- protect vfio_register_reserved_iova_range implementation with
  CONFIG_IOMMU_DMA_RESERVED
- handle unregistration by user-space and on vfio_iommu_type1 release

v1 -> v2:
- set returned value according to alloc_reserved_iova_domain result
- free the iova domains in case any error occurs

RFC v1 -> v1:
- takes into account Alex comments, based on
  [RFC PATCH 1/6] vfio: Add interface for add/del reserved iova region:
- use the existing dma map/unmap ioctl interface with a flag to register
  a reserved IOVA range. A single reserved iova region is allowed.
---
 drivers/vfio/vfio_iommu_type1.c | 78 ++++++++++++++++++++++++++++++++++++++++-
 include/uapi/linux/vfio.h       | 10 +++++-
 2 files changed, 86 insertions(+), 2 deletions(-)

diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
index 94a9916..4d3a6f1 100644
--- a/drivers/vfio/vfio_iommu_type1.c
+++ b/drivers/vfio/vfio_iommu_type1.c
@@ -36,6 +36,7 @@
 #include <linux/uaccess.h>
 #include <linux/vfio.h>
 #include <linux/workqueue.h>
+#include <linux/msi-iommu.h>
 
 #define DRIVER_VERSION  "0.2"
 #define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
@@ -445,6 +446,20 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
 	vfio_lock_acct(-unlocked);
 }
 
+static int vfio_set_msi_aperture(struct vfio_iommu *iommu,
+				dma_addr_t iova, size_t size)
+{
+	struct vfio_domain *d;
+	int ret = 0;
+
+	list_for_each_entry(d, &iommu->domain_list, next) {
+		ret = iommu_msi_set_aperture(d->domain, iova, iova + size - 1);
+		if (ret)
+			break;
+	}
+	return ret;
+}
+
 static void vfio_remove_dma(struct vfio_iommu *iommu, struct vfio_dma *dma)
 {
 	vfio_unmap_unpin(iommu, dma);
@@ -693,6 +708,63 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
 	return ret;
 }
 
+static int vfio_register_msi_range(struct vfio_iommu *iommu,
+				   struct vfio_iommu_type1_dma_map *map)
+{
+	dma_addr_t iova = map->iova;
+	size_t size = map->size;
+	int ret = 0;
+	struct vfio_dma *dma;
+	unsigned long order;
+	uint64_t mask;
+
+	/* Verify that none of our __u64 fields overflow */
+	if (map->size != size || map->iova != iova)
+		return -EINVAL;
+
+	order =  __ffs(vfio_pgsize_bitmap(iommu));
+	mask = ((uint64_t)1 << order) - 1;
+
+	WARN_ON(mask & PAGE_MASK);
+
+	if (!size || (size | iova) & mask)
+		return -EINVAL;
+
+	/* Don't allow IOVA address wrap */
+	if (iova + size - 1 < iova)
+		return -EINVAL;
+
+	mutex_lock(&iommu->lock);
+
+	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
+		ret =  -EEXIST;
+		goto unlock;
+	}
+
+	dma = kzalloc(sizeof(*dma), GFP_KERNEL);
+	if (!dma) {
+		ret = -ENOMEM;
+		goto unlock;
+	}
+
+	dma->iova = iova;
+	dma->size = size;
+	dma->type = VFIO_IOVA_RESERVED;
+
+	ret = vfio_set_msi_aperture(iommu, iova, size);
+	if (ret)
+		goto free_unlock;
+
+	vfio_link_dma(iommu, dma);
+	goto unlock;
+
+free_unlock:
+	kfree(dma);
+unlock:
+	mutex_unlock(&iommu->lock);
+	return ret;
+}
+
 static int vfio_bus_type(struct device *dev, void *data)
 {
 	struct bus_type **bus = data;
@@ -1062,7 +1134,8 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
 	} else if (cmd == VFIO_IOMMU_MAP_DMA) {
 		struct vfio_iommu_type1_dma_map map;
 		uint32_t mask = VFIO_DMA_MAP_FLAG_READ |
-				VFIO_DMA_MAP_FLAG_WRITE;
+				VFIO_DMA_MAP_FLAG_WRITE |
+				VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA;
 
 		minsz = offsetofend(struct vfio_iommu_type1_dma_map, size);
 
@@ -1072,6 +1145,9 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
 		if (map.argsz < minsz || map.flags & ~mask)
 			return -EINVAL;
 
+		if (map.flags & VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA)
+			return vfio_register_msi_range(iommu, &map);
+
 		return vfio_dma_do_map(iommu, &map);
 
 	} else if (cmd == VFIO_IOMMU_UNMAP_DMA) {
diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
index 255a211..4a9dbc2 100644
--- a/include/uapi/linux/vfio.h
+++ b/include/uapi/linux/vfio.h
@@ -498,12 +498,19 @@ struct vfio_iommu_type1_info {
  *
  * Map process virtual addresses to IO virtual addresses using the
  * provided struct vfio_dma_map. Caller sets argsz. READ &/ WRITE required.
+ *
+ * In case RESERVED_MSI_IOVA flag is set, the API only aims at registering an
+ * IOVA region that will be used on some platforms to map the host MSI frames.
+ * In that specific case, vaddr is ignored. Once registered, an MSI reserved
+ * IOVA region stays until the container is closed.
  */
 struct vfio_iommu_type1_dma_map {
 	__u32	argsz;
 	__u32	flags;
 #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
 #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
+/* reserved iova for MSI vectors*/
+#define VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA (1 << 2)
 	__u64	vaddr;				/* Process virtual address */
 	__u64	iova;				/* IO virtual address */
 	__u64	size;				/* Size of mapping (bytes) */
@@ -519,7 +526,8 @@ struct vfio_iommu_type1_dma_map {
  * Caller sets argsz.  The actual unmapped size is returned in the size
  * field.  No guarantee is made to the user that arbitrary unmaps of iova
  * or size different from those used in the original mapping call will
- * succeed.
+ * succeed. Once registered, an MSI region cannot be unmapped and stays
+ * until the container is closed.
  */
 struct vfio_iommu_type1_dma_unmap {
 	__u32	argsz;
-- 
1.9.1

[toc] | [prev] | [next] | [standalone]

#1395321 — Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration

On 5/4/16, 4:54 AM, "linux-arm-kernel on behalf of Eric Auger" <linux-arm-kernel-bounces@lists.infradead.org on behalf of eric.auger@linaro.org> wrote:

>The user is allowed to register a reserved MSI IOVA range by using the
>DMA MAP API and setting the new flag: VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA.
>This region is stored in the vfio_dma rb tree. At that point the iova
>range is not mapped to any target address yet. The host kernel will use
>those iova when needed, typically when MSIs are allocated.
>
>Signed-off-by: Eric Auger <eric.auger@linaro.org>
>Signed-off-by: Bharat Bhushan <Bharat.Bhushan@freescale.com>
>
>---
>v7 -> v8:
>- use iommu_msi_set_aperture function. There is no notion of
>  unregistration anymore since the reserved msi slot remains
>  until the container gets closed.
>
>v6 -> v7:
>- use iommu_free_reserved_iova_domain
>- convey prot attributes downto dma-reserved-iommu iova domain creation
>- reserved bindings teardown now performed on iommu domain destruction
>- rename VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA into
>         VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA
>- change title
>- pass the protection attribute to dma-reserved-iommu API
>
>v3 -> v4:
>- use iommu_alloc/free_reserved_iova_domain exported by dma-reserved-iommu
>- protect vfio_register_reserved_iova_range implementation with
>  CONFIG_IOMMU_DMA_RESERVED
>- handle unregistration by user-space and on vfio_iommu_type1 release
>
>v1 -> v2:
>- set returned value according to alloc_reserved_iova_domain result
>- free the iova domains in case any error occurs
>
>RFC v1 -> v1:
>- takes into account Alex comments, based on
>  [RFC PATCH 1/6] vfio: Add interface for add/del reserved iova region:
>- use the existing dma map/unmap ioctl interface with a flag to register
>  a reserved IOVA range. A single reserved iova region is allowed.
>---
> drivers/vfio/vfio_iommu_type1.c | 78 ++++++++++++++++++++++++++++++++++++++++-
> include/uapi/linux/vfio.h       | 10 +++++-
> 2 files changed, 86 insertions(+), 2 deletions(-)
>
>diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
>index 94a9916..4d3a6f1 100644
>--- a/drivers/vfio/vfio_iommu_type1.c
>+++ b/drivers/vfio/vfio_iommu_type1.c
>@@ -36,6 +36,7 @@
> #include <linux/uaccess.h>
> #include <linux/vfio.h>
> #include <linux/workqueue.h>
>+#include <linux/msi-iommu.h>
> 
> #define DRIVER_VERSION  "0.2"
> #define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
>@@ -445,6 +446,20 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
> 	vfio_lock_acct(-unlocked);
> }
> 
>+static int vfio_set_msi_aperture(struct vfio_iommu *iommu,
>+				dma_addr_t iova, size_t size)
>+{
>+	struct vfio_domain *d;
>+	int ret = 0;
>+
>+	list_for_each_entry(d, &iommu->domain_list, next) {
>+		ret = iommu_msi_set_aperture(d->domain, iova, iova + size - 1);
>+		if (ret)
>+			break;
>+	}
>+	return ret;
>+}
>+
> static void vfio_remove_dma(struct vfio_iommu *iommu, struct vfio_dma *dma)
> {
> 	vfio_unmap_unpin(iommu, dma);
>@@ -693,6 +708,63 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
> 	return ret;
> }
> 
>+static int vfio_register_msi_range(struct vfio_iommu *iommu,
>+				   struct vfio_iommu_type1_dma_map *map)
>+{
>+	dma_addr_t iova = map->iova;
>+	size_t size = map->size;
>+	int ret = 0;
>+	struct vfio_dma *dma;
>+	unsigned long order;
>+	uint64_t mask;
>+
>+	/* Verify that none of our __u64 fields overflow */
>+	if (map->size != size || map->iova != iova)
>+		return -EINVAL;
>+
>+	order =  __ffs(vfio_pgsize_bitmap(iommu));
>+	mask = ((uint64_t)1 << order) - 1;
>+
>+	WARN_ON(mask & PAGE_MASK);
>+
>+	if (!size || (size | iova) & mask)
>+		return -EINVAL;
>+
>+	/* Don't allow IOVA address wrap */
>+	if (iova + size - 1 < iova)
>+		return -EINVAL;
>+
>+	mutex_lock(&iommu->lock);
>+
>+	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
>+		ret =  -EEXIST;
>+		goto unlock;
>+	}
>+
>+	dma = kzalloc(sizeof(*dma), GFP_KERNEL);
>+	if (!dma) {
>+		ret = -ENOMEM;
>+		goto unlock;
>+	}
>+
>+	dma->iova = iova;
>+	dma->size = size;
>+	dma->type = VFIO_IOVA_RESERVED;
>+
>+	ret = vfio_set_msi_aperture(iommu, iova, size);
>+	if (ret)
>+		goto free_unlock;
>+
>+	vfio_link_dma(iommu, dma);
>+	goto unlock;
>+
>+free_unlock:
>+	kfree(dma);
>+unlock:
>+	mutex_unlock(&iommu->lock);
>+	return ret;
>+}
>+
> static int vfio_bus_type(struct device *dev, void *data)
> {
> 	struct bus_type **bus = data;
>@@ -1062,7 +1134,8 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
> 	} else if (cmd == VFIO_IOMMU_MAP_DMA) {
> 		struct vfio_iommu_type1_dma_map map;
> 		uint32_t mask = VFIO_DMA_MAP_FLAG_READ |
>-				VFIO_DMA_MAP_FLAG_WRITE;
>+				VFIO_DMA_MAP_FLAG_WRITE |
>+				VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA;
> 
> 		minsz = offsetofend(struct vfio_iommu_type1_dma_map, size);
> 
>@@ -1072,6 +1145,9 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
> 		if (map.argsz < minsz || map.flags & ~mask)
> 			return -EINVAL;
> 
>+		if (map.flags & VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA)
>+			return vfio_register_msi_range(iommu, &map);
>+

How are we getting the PA and any validation needed?
> 		return vfio_dma_do_map(iommu, &map);
> 
> 	} else if (cmd == VFIO_IOMMU_UNMAP_DMA) {
>diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
>index 255a211..4a9dbc2 100644
>--- a/include/uapi/linux/vfio.h
>+++ b/include/uapi/linux/vfio.h
>@@ -498,12 +498,19 @@ struct vfio_iommu_type1_info {
>  *
>  * Map process virtual addresses to IO virtual addresses using the
>  * provided struct vfio_dma_map. Caller sets argsz. READ &/ WRITE required.
>+ *
>+ * In case RESERVED_MSI_IOVA flag is set, the API only aims at registering an
>+ * IOVA region that will be used on some platforms to map the host MSI frames.
>+ * In that specific case, vaddr is ignored. Once registered, an MSI reserved
>+ * IOVA region stays until the container is closed.
>  */
> struct vfio_iommu_type1_dma_map {
> 	__u32	argsz;
> 	__u32	flags;
> #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
> #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
>+/* reserved iova for MSI vectors*/
>+#define VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA (1 << 2)
> 	__u64	vaddr;				/* Process virtual address */
> 	__u64	iova;				/* IO virtual address */
> 	__u64	size;				/* Size of mapping (bytes) */
>@@ -519,7 +526,8 @@ struct vfio_iommu_type1_dma_map {
>  * Caller sets argsz.  The actual unmapped size is returned in the size
>  * field.  No guarantee is made to the user that arbitrary unmaps of iova
>  * or size different from those used in the original mapping call will
>- * succeed.
>+ * succeed. Once registered, an MSI region cannot be unmapped and stays
>+ * until the container is closed.
>  */
> struct vfio_iommu_type1_dma_unmap {
> 	__u32	argsz;
>-- 
>1.9.1
>
>
>_______________________________________________
>linux-arm-kernel mailing list
>linux-arm-kernel@lists.infradead.org
>http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[toc] | [prev] | [next] | [standalone]

#1396805 — Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration

Hi Chalamarla,
On 05/05/2016 09:22 PM, Chalamarla, Tirumalesh wrote:
> 
> 
> 
> 
> 
> On 5/4/16, 4:54 AM, "linux-arm-kernel on behalf of Eric Auger" <linux-arm-kernel-bounces@lists.infradead.org on behalf of eric.auger@linaro.org> wrote:
> 
>> The user is allowed to register a reserved MSI IOVA range by using the
>> DMA MAP API and setting the new flag: VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA.
>> This region is stored in the vfio_dma rb tree. At that point the iova
>> range is not mapped to any target address yet. The host kernel will use
>> those iova when needed, typically when MSIs are allocated.
>>
>> Signed-off-by: Eric Auger <eric.auger@linaro.org>
>> Signed-off-by: Bharat Bhushan <Bharat.Bhushan@freescale.com>
>>
>> ---
>> v7 -> v8:
>> - use iommu_msi_set_aperture function. There is no notion of
>>  unregistration anymore since the reserved msi slot remains
>>  until the container gets closed.
>>
>> v6 -> v7:
>> - use iommu_free_reserved_iova_domain
>> - convey prot attributes downto dma-reserved-iommu iova domain creation
>> - reserved bindings teardown now performed on iommu domain destruction
>> - rename VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA into
>>         VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA
>> - change title
>> - pass the protection attribute to dma-reserved-iommu API
>>
>> v3 -> v4:
>> - use iommu_alloc/free_reserved_iova_domain exported by dma-reserved-iommu
>> - protect vfio_register_reserved_iova_range implementation with
>>  CONFIG_IOMMU_DMA_RESERVED
>> - handle unregistration by user-space and on vfio_iommu_type1 release
>>
>> v1 -> v2:
>> - set returned value according to alloc_reserved_iova_domain result
>> - free the iova domains in case any error occurs
>>
>> RFC v1 -> v1:
>> - takes into account Alex comments, based on
>>  [RFC PATCH 1/6] vfio: Add interface for add/del reserved iova region:
>> - use the existing dma map/unmap ioctl interface with a flag to register
>>  a reserved IOVA range. A single reserved iova region is allowed.
>> ---
>> drivers/vfio/vfio_iommu_type1.c | 78 ++++++++++++++++++++++++++++++++++++++++-
>> include/uapi/linux/vfio.h       | 10 +++++-
>> 2 files changed, 86 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
>> index 94a9916..4d3a6f1 100644
>> --- a/drivers/vfio/vfio_iommu_type1.c
>> +++ b/drivers/vfio/vfio_iommu_type1.c
>> @@ -36,6 +36,7 @@
>> #include <linux/uaccess.h>
>> #include <linux/vfio.h>
>> #include <linux/workqueue.h>
>> +#include <linux/msi-iommu.h>
>>
>> #define DRIVER_VERSION  "0.2"
>> #define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
>> @@ -445,6 +446,20 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
>> 	vfio_lock_acct(-unlocked);
>> }
>>
>> +static int vfio_set_msi_aperture(struct vfio_iommu *iommu,
>> +				dma_addr_t iova, size_t size)
>> +{
>> +	struct vfio_domain *d;
>> +	int ret = 0;
>> +
>> +	list_for_each_entry(d, &iommu->domain_list, next) {
>> +		ret = iommu_msi_set_aperture(d->domain, iova, iova + size - 1);
>> +		if (ret)
>> +			break;
>> +	}
>> +	return ret;
>> +}
>> +
>> static void vfio_remove_dma(struct vfio_iommu *iommu, struct vfio_dma *dma)
>> {
>> 	vfio_unmap_unpin(iommu, dma);
>> @@ -693,6 +708,63 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
>> 	return ret;
>> }
>>
>> +static int vfio_register_msi_range(struct vfio_iommu *iommu,
>> +				   struct vfio_iommu_type1_dma_map *map)
>> +{
>> +	dma_addr_t iova = map->iova;
>> +	size_t size = map->size;
>> +	int ret = 0;
>> +	struct vfio_dma *dma;
>> +	unsigned long order;
>> +	uint64_t mask;
>> +
>> +	/* Verify that none of our __u64 fields overflow */
>> +	if (map->size != size || map->iova != iova)
>> +		return -EINVAL;
>> +
>> +	order =  __ffs(vfio_pgsize_bitmap(iommu));
>> +	mask = ((uint64_t)1 << order) - 1;
>> +
>> +	WARN_ON(mask & PAGE_MASK);
>> +
>> +	if (!size || (size | iova) & mask)
>> +		return -EINVAL;
>> +
>> +	/* Don't allow IOVA address wrap */
>> +	if (iova + size - 1 < iova)
>> +		return -EINVAL;
>> +
>> +	mutex_lock(&iommu->lock);
>> +
>> +	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
>> +		ret =  -EEXIST;
>> +		goto unlock;
>> +	}
>> +
>> +	dma = kzalloc(sizeof(*dma), GFP_KERNEL);
>> +	if (!dma) {
>> +		ret = -ENOMEM;
>> +		goto unlock;
>> +	}
>> +
>> +	dma->iova = iova;
>> +	dma->size = size;
>> +	dma->type = VFIO_IOVA_RESERVED;
>> +
>> +	ret = vfio_set_msi_aperture(iommu, iova, size);
>> +	if (ret)
>> +		goto free_unlock;
>> +
>> +	vfio_link_dma(iommu, dma);
>> +	goto unlock;
>> +
>> +free_unlock:
>> +	kfree(dma);
>> +unlock:
>> +	mutex_unlock(&iommu->lock);
>> +	return ret;
>> +}
>> +
>> static int vfio_bus_type(struct device *dev, void *data)
>> {
>> 	struct bus_type **bus = data;
>> @@ -1062,7 +1134,8 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
>> 	} else if (cmd == VFIO_IOMMU_MAP_DMA) {
>> 		struct vfio_iommu_type1_dma_map map;
>> 		uint32_t mask = VFIO_DMA_MAP_FLAG_READ |
>> -				VFIO_DMA_MAP_FLAG_WRITE;
>> +				VFIO_DMA_MAP_FLAG_WRITE |
>> +				VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA;
>>
>> 		minsz = offsetofend(struct vfio_iommu_type1_dma_map, size);
>>
>> @@ -1072,6 +1145,9 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
>> 		if (map.argsz < minsz || map.flags & ~mask)
>> 			return -EINVAL;
>>
>> +		if (map.flags & VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA)
>> +			return vfio_register_msi_range(iommu, &map);
>> +
> 
> How are we getting the PA and any validation needed?
Sorry I am not sure I get your question? The principle is the userspace
registers a spare Guest PA range that can be safely used by the kernel
to map host MSI frames. In QEMU integration we use platform bus unmapped
GPA to make things transparent for the virt machine.

Any validation: we currently make sure the registered GPA was not
already registered as RAM regions. The kernel also checks it is properly
aligned according to the IOMMU page size.

Please let me know if I answered your question.

Best Regards

Eric

>> 		return vfio_dma_do_map(iommu, &map);
>>
>> 	} else if (cmd == VFIO_IOMMU_UNMAP_DMA) {
>> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
>> index 255a211..4a9dbc2 100644
>> --- a/include/uapi/linux/vfio.h
>> +++ b/include/uapi/linux/vfio.h
>> @@ -498,12 +498,19 @@ struct vfio_iommu_type1_info {
>>  *
>>  * Map process virtual addresses to IO virtual addresses using the
>>  * provided struct vfio_dma_map. Caller sets argsz. READ &/ WRITE required.
>> + *
>> + * In case RESERVED_MSI_IOVA flag is set, the API only aims at registering an
>> + * IOVA region that will be used on some platforms to map the host MSI frames.
>> + * In that specific case, vaddr is ignored. Once registered, an MSI reserved
>> + * IOVA region stays until the container is closed.
>>  */
>> struct vfio_iommu_type1_dma_map {
>> 	__u32	argsz;
>> 	__u32	flags;
>> #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
>> #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
>> +/* reserved iova for MSI vectors*/
>> +#define VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA (1 << 2)
>> 	__u64	vaddr;				/* Process virtual address */
>> 	__u64	iova;				/* IO virtual address */
>> 	__u64	size;				/* Size of mapping (bytes) */
>> @@ -519,7 +526,8 @@ struct vfio_iommu_type1_dma_map {
>>  * Caller sets argsz.  The actual unmapped size is returned in the size
>>  * field.  No guarantee is made to the user that arbitrary unmaps of iova
>>  * or size different from those used in the original mapping call will
>> - * succeed.
>> + * succeed. Once registered, an MSI region cannot be unmapped and stays
>> + * until the container is closed.
>>  */
>> struct vfio_iommu_type1_dma_unmap {
>> 	__u32	argsz;
>> -- 
>> 1.9.1
>>
>>
>> _______________________________________________
>> linux-arm-kernel mailing list
>> linux-arm-kernel@lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[toc] | [prev] | [next] | [standalone]

#1398251 — Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration

On Wed,  4 May 2016 11:54:15 +0000
Eric Auger <eric.auger@linaro.org> wrote:

> The user is allowed to register a reserved MSI IOVA range by using the
> DMA MAP API and setting the new flag: VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA.
> This region is stored in the vfio_dma rb tree. At that point the iova
> range is not mapped to any target address yet. The host kernel will use
> those iova when needed, typically when MSIs are allocated.
> 
> Signed-off-by: Eric Auger <eric.auger@linaro.org>
> Signed-off-by: Bharat Bhushan <Bharat.Bhushan@freescale.com>
> 
> ---
> v7 -> v8:
> - use iommu_msi_set_aperture function. There is no notion of
>   unregistration anymore since the reserved msi slot remains
>   until the container gets closed.
> 
> v6 -> v7:
> - use iommu_free_reserved_iova_domain
> - convey prot attributes downto dma-reserved-iommu iova domain creation
> - reserved bindings teardown now performed on iommu domain destruction
> - rename VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA into
>          VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA
> - change title
> - pass the protection attribute to dma-reserved-iommu API
> 
> v3 -> v4:
> - use iommu_alloc/free_reserved_iova_domain exported by dma-reserved-iommu
> - protect vfio_register_reserved_iova_range implementation with
>   CONFIG_IOMMU_DMA_RESERVED
> - handle unregistration by user-space and on vfio_iommu_type1 release
> 
> v1 -> v2:
> - set returned value according to alloc_reserved_iova_domain result
> - free the iova domains in case any error occurs
> 
> RFC v1 -> v1:
> - takes into account Alex comments, based on
>   [RFC PATCH 1/6] vfio: Add interface for add/del reserved iova region:
> - use the existing dma map/unmap ioctl interface with a flag to register
>   a reserved IOVA range. A single reserved iova region is allowed.
> ---
>  drivers/vfio/vfio_iommu_type1.c | 78 ++++++++++++++++++++++++++++++++++++++++-
>  include/uapi/linux/vfio.h       | 10 +++++-
>  2 files changed, 86 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
> index 94a9916..4d3a6f1 100644
> --- a/drivers/vfio/vfio_iommu_type1.c
> +++ b/drivers/vfio/vfio_iommu_type1.c
> @@ -36,6 +36,7 @@
>  #include <linux/uaccess.h>
>  #include <linux/vfio.h>
>  #include <linux/workqueue.h>
> +#include <linux/msi-iommu.h>
>  
>  #define DRIVER_VERSION  "0.2"
>  #define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
> @@ -445,6 +446,20 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
>  	vfio_lock_acct(-unlocked);
>  }
>  
> +static int vfio_set_msi_aperture(struct vfio_iommu *iommu,
> +				dma_addr_t iova, size_t size)
> +{
> +	struct vfio_domain *d;
> +	int ret = 0;
> +
> +	list_for_each_entry(d, &iommu->domain_list, next) {
> +		ret = iommu_msi_set_aperture(d->domain, iova, iova + size - 1);
> +		if (ret)
> +			break;
> +	}
> +	return ret;
> +}
> +
>  static void vfio_remove_dma(struct vfio_iommu *iommu, struct vfio_dma *dma)
>  {
>  	vfio_unmap_unpin(iommu, dma);
> @@ -693,6 +708,63 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
>  	return ret;
>  }
>  
> +static int vfio_register_msi_range(struct vfio_iommu *iommu,
> +				   struct vfio_iommu_type1_dma_map *map)
> +{
> +	dma_addr_t iova = map->iova;
> +	size_t size = map->size;
> +	int ret = 0;
> +	struct vfio_dma *dma;
> +	unsigned long order;
> +	uint64_t mask;
> +
> +	/* Verify that none of our __u64 fields overflow */
> +	if (map->size != size || map->iova != iova)
> +		return -EINVAL;
> +
> +	order =  __ffs(vfio_pgsize_bitmap(iommu));
> +	mask = ((uint64_t)1 << order) - 1;
> +
> +	WARN_ON(mask & PAGE_MASK);
> +
> +	if (!size || (size | iova) & mask)
> +		return -EINVAL;
> +
> +	/* Don't allow IOVA address wrap */
> +	if (iova + size - 1 < iova)
> +		return -EINVAL;
> +
> +	mutex_lock(&iommu->lock);
> +
> +	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
> +		ret =  -EEXIST;
> +		goto unlock;
> +	}
> +
> +	dma = kzalloc(sizeof(*dma), GFP_KERNEL);
> +	if (!dma) {
> +		ret = -ENOMEM;
> +		goto unlock;
> +	}
> +
> +	dma->iova = iova;
> +	dma->size = size;
> +	dma->type = VFIO_IOVA_RESERVED;

[oops, forgot to send this reply with the others]

I'm tempted to suggest we set type explicitly in the USER case too just
to make that abundantly clear rather than taking advantage of the
kzalloc struct.

> +
> +	ret = vfio_set_msi_aperture(iommu, iova, size);
> +	if (ret)
> +		goto free_unlock;
> +
> +	vfio_link_dma(iommu, dma);
> +	goto unlock;
> +
> +free_unlock:
> +	kfree(dma);
> +unlock:
> +	mutex_unlock(&iommu->lock);
> +	return ret;
> +}
> +
>  static int vfio_bus_type(struct device *dev, void *data)
>  {
>  	struct bus_type **bus = data;
> @@ -1062,7 +1134,8 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
>  	} else if (cmd == VFIO_IOMMU_MAP_DMA) {
>  		struct vfio_iommu_type1_dma_map map;
>  		uint32_t mask = VFIO_DMA_MAP_FLAG_READ |
> -				VFIO_DMA_MAP_FLAG_WRITE;
> +				VFIO_DMA_MAP_FLAG_WRITE |
> +				VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA;
>  
>  		minsz = offsetofend(struct vfio_iommu_type1_dma_map, size);
>  
> @@ -1072,6 +1145,9 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
>  		if (map.argsz < minsz || map.flags & ~mask)
>  			return -EINVAL;
>  
> +		if (map.flags & VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA)
> +			return vfio_register_msi_range(iommu, &map);
> +
>  		return vfio_dma_do_map(iommu, &map);
>  
>  	} else if (cmd == VFIO_IOMMU_UNMAP_DMA) {
> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
> index 255a211..4a9dbc2 100644
> --- a/include/uapi/linux/vfio.h
> +++ b/include/uapi/linux/vfio.h
> @@ -498,12 +498,19 @@ struct vfio_iommu_type1_info {
>   *
>   * Map process virtual addresses to IO virtual addresses using the
>   * provided struct vfio_dma_map. Caller sets argsz. READ &/ WRITE required.
> + *
> + * In case RESERVED_MSI_IOVA flag is set, the API only aims at registering an
> + * IOVA region that will be used on some platforms to map the host MSI frames.
> + * In that specific case, vaddr is ignored. Once registered, an MSI reserved
> + * IOVA region stays until the container is closed.
>   */
>  struct vfio_iommu_type1_dma_map {
>  	__u32	argsz;
>  	__u32	flags;
>  #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
>  #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
> +/* reserved iova for MSI vectors*/
> +#define VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA (1 << 2)
>  	__u64	vaddr;				/* Process virtual address */
>  	__u64	iova;				/* IO virtual address */
>  	__u64	size;				/* Size of mapping (bytes) */
> @@ -519,7 +526,8 @@ struct vfio_iommu_type1_dma_map {
>   * Caller sets argsz.  The actual unmapped size is returned in the size
>   * field.  No guarantee is made to the user that arbitrary unmaps of iova
>   * or size different from those used in the original mapping call will
> - * succeed.
> + * succeed. Once registered, an MSI region cannot be unmapped and stays
> + * until the container is closed.
>   */
>  struct vfio_iommu_type1_dma_unmap {
>  	__u32	argsz;

[toc] | [prev] | [next] | [standalone]

#1398252 — Re: [PATCH v9 4/7] vfio: allow reserved msi iova registration

Hi Alex,
On 05/10/2016 05:29 PM, Alex Williamson wrote:
> On Wed,  4 May 2016 11:54:15 +0000
> Eric Auger <eric.auger@linaro.org> wrote:
> 
>> The user is allowed to register a reserved MSI IOVA range by using the
>> DMA MAP API and setting the new flag: VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA.
>> This region is stored in the vfio_dma rb tree. At that point the iova
>> range is not mapped to any target address yet. The host kernel will use
>> those iova when needed, typically when MSIs are allocated.
>>
>> Signed-off-by: Eric Auger <eric.auger@linaro.org>
>> Signed-off-by: Bharat Bhushan <Bharat.Bhushan@freescale.com>
>>
>> ---
>> v7 -> v8:
>> - use iommu_msi_set_aperture function. There is no notion of
>>   unregistration anymore since the reserved msi slot remains
>>   until the container gets closed.
>>
>> v6 -> v7:
>> - use iommu_free_reserved_iova_domain
>> - convey prot attributes downto dma-reserved-iommu iova domain creation
>> - reserved bindings teardown now performed on iommu domain destruction
>> - rename VFIO_DMA_MAP_FLAG_MSI_RESERVED_IOVA into
>>          VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA
>> - change title
>> - pass the protection attribute to dma-reserved-iommu API
>>
>> v3 -> v4:
>> - use iommu_alloc/free_reserved_iova_domain exported by dma-reserved-iommu
>> - protect vfio_register_reserved_iova_range implementation with
>>   CONFIG_IOMMU_DMA_RESERVED
>> - handle unregistration by user-space and on vfio_iommu_type1 release
>>
>> v1 -> v2:
>> - set returned value according to alloc_reserved_iova_domain result
>> - free the iova domains in case any error occurs
>>
>> RFC v1 -> v1:
>> - takes into account Alex comments, based on
>>   [RFC PATCH 1/6] vfio: Add interface for add/del reserved iova region:
>> - use the existing dma map/unmap ioctl interface with a flag to register
>>   a reserved IOVA range. A single reserved iova region is allowed.
>> ---
>>  drivers/vfio/vfio_iommu_type1.c | 78 ++++++++++++++++++++++++++++++++++++++++-
>>  include/uapi/linux/vfio.h       | 10 +++++-
>>  2 files changed, 86 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
>> index 94a9916..4d3a6f1 100644
>> --- a/drivers/vfio/vfio_iommu_type1.c
>> +++ b/drivers/vfio/vfio_iommu_type1.c
>> @@ -36,6 +36,7 @@
>>  #include <linux/uaccess.h>
>>  #include <linux/vfio.h>
>>  #include <linux/workqueue.h>
>> +#include <linux/msi-iommu.h>
>>  
>>  #define DRIVER_VERSION  "0.2"
>>  #define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
>> @@ -445,6 +446,20 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
>>  	vfio_lock_acct(-unlocked);
>>  }
>>  
>> +static int vfio_set_msi_aperture(struct vfio_iommu *iommu,
>> +				dma_addr_t iova, size_t size)
>> +{
>> +	struct vfio_domain *d;
>> +	int ret = 0;
>> +
>> +	list_for_each_entry(d, &iommu->domain_list, next) {
>> +		ret = iommu_msi_set_aperture(d->domain, iova, iova + size - 1);
>> +		if (ret)
>> +			break;
>> +	}
>> +	return ret;
>> +}
>> +
>>  static void vfio_remove_dma(struct vfio_iommu *iommu, struct vfio_dma *dma)
>>  {
>>  	vfio_unmap_unpin(iommu, dma);
>> @@ -693,6 +708,63 @@ static int vfio_dma_do_map(struct vfio_iommu *iommu,
>>  	return ret;
>>  }
>>  
>> +static int vfio_register_msi_range(struct vfio_iommu *iommu,
>> +				   struct vfio_iommu_type1_dma_map *map)
>> +{
>> +	dma_addr_t iova = map->iova;
>> +	size_t size = map->size;
>> +	int ret = 0;
>> +	struct vfio_dma *dma;
>> +	unsigned long order;
>> +	uint64_t mask;
>> +
>> +	/* Verify that none of our __u64 fields overflow */
>> +	if (map->size != size || map->iova != iova)
>> +		return -EINVAL;
>> +
>> +	order =  __ffs(vfio_pgsize_bitmap(iommu));
>> +	mask = ((uint64_t)1 << order) - 1;
>> +
>> +	WARN_ON(mask & PAGE_MASK);
>> +
>> +	if (!size || (size | iova) & mask)
>> +		return -EINVAL;
>> +
>> +	/* Don't allow IOVA address wrap */
>> +	if (iova + size - 1 < iova)
>> +		return -EINVAL;
>> +
>> +	mutex_lock(&iommu->lock);
>> +
>> +	if (vfio_find_dma(iommu, iova, size, VFIO_IOVA_ANY)) {
>> +		ret =  -EEXIST;
>> +		goto unlock;
>> +	}
>> +
>> +	dma = kzalloc(sizeof(*dma), GFP_KERNEL);
>> +	if (!dma) {
>> +		ret = -ENOMEM;
>> +		goto unlock;
>> +	}
>> +
>> +	dma->iova = iova;
>> +	dma->size = size;
>> +	dma->type = VFIO_IOVA_RESERVED;
> 
> [oops, forgot to send this reply with the others]
> 
> I'm tempted to suggest we set type explicitly in the USER case too just
> to make that abundantly clear rather than taking advantage of the
> kzalloc struct.

no worries. OK I will set the dma type explicitly in vfio_dma_do_map too.

Thanks

Eric
> 
>> +
>> +	ret = vfio_set_msi_aperture(iommu, iova, size);
>> +	if (ret)
>> +		goto free_unlock;
>> +
>> +	vfio_link_dma(iommu, dma);
>> +	goto unlock;
>> +
>> +free_unlock:
>> +	kfree(dma);
>> +unlock:
>> +	mutex_unlock(&iommu->lock);
>> +	return ret;
>> +}
>> +
>>  static int vfio_bus_type(struct device *dev, void *data)
>>  {
>>  	struct bus_type **bus = data;
>> @@ -1062,7 +1134,8 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
>>  	} else if (cmd == VFIO_IOMMU_MAP_DMA) {
>>  		struct vfio_iommu_type1_dma_map map;
>>  		uint32_t mask = VFIO_DMA_MAP_FLAG_READ |
>> -				VFIO_DMA_MAP_FLAG_WRITE;
>> +				VFIO_DMA_MAP_FLAG_WRITE |
>> +				VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA;
>>  
>>  		minsz = offsetofend(struct vfio_iommu_type1_dma_map, size);
>>  
>> @@ -1072,6 +1145,9 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
>>  		if (map.argsz < minsz || map.flags & ~mask)
>>  			return -EINVAL;
>>  
>> +		if (map.flags & VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA)
>> +			return vfio_register_msi_range(iommu, &map);
>> +
>>  		return vfio_dma_do_map(iommu, &map);
>>  
>>  	} else if (cmd == VFIO_IOMMU_UNMAP_DMA) {
>> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
>> index 255a211..4a9dbc2 100644
>> --- a/include/uapi/linux/vfio.h
>> +++ b/include/uapi/linux/vfio.h
>> @@ -498,12 +498,19 @@ struct vfio_iommu_type1_info {
>>   *
>>   * Map process virtual addresses to IO virtual addresses using the
>>   * provided struct vfio_dma_map. Caller sets argsz. READ &/ WRITE required.
>> + *
>> + * In case RESERVED_MSI_IOVA flag is set, the API only aims at registering an
>> + * IOVA region that will be used on some platforms to map the host MSI frames.
>> + * In that specific case, vaddr is ignored. Once registered, an MSI reserved
>> + * IOVA region stays until the container is closed.
>>   */
>>  struct vfio_iommu_type1_dma_map {
>>  	__u32	argsz;
>>  	__u32	flags;
>>  #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
>>  #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
>> +/* reserved iova for MSI vectors*/
>> +#define VFIO_DMA_MAP_FLAG_RESERVED_MSI_IOVA (1 << 2)
>>  	__u64	vaddr;				/* Process virtual address */
>>  	__u64	iova;				/* IO virtual address */
>>  	__u64	size;				/* Size of mapping (bytes) */
>> @@ -519,7 +526,8 @@ struct vfio_iommu_type1_dma_map {
>>   * Caller sets argsz.  The actual unmapped size is returned in the size
>>   * field.  No guarantee is made to the user that arbitrary unmaps of iova
>>   * or size different from those used in the original mapping call will
>> - * succeed.
>> + * succeed. Once registered, an MSI region cannot be unmapped and stays
>> + * until the container is closed.
>>   */
>>  struct vfio_iommu_type1_dma_unmap {
>>  	__u32	argsz;
> 

[toc] | [prev] | [next] | [standalone]

#1394200 — [PATCH v9 3/7] vfio/type1: bypass unmap/unpin and replay for VFIO_IOVA_RESERVED slots

Before allowing the end-user to create VFIO_IOVA_RESERVED dma slots,
let's implement the expected behavior for removal and replay. As opposed
to user dma slots, IOVAs are not systematically bound to PAs and PAs are
not pinned. VFIO just initializes the IOVA "aperture". IOVAs are allocated
outside of the VFIO framework, typically the MSI layer which is
responsible to free and unmap them. The MSI mapping resources are freeed
by the IOMMU driver on domain destruction.

Signed-off-by: Eric Auger <eric.auger@linaro.org>

---

v7 -> v8:
- do no destroy anything anymore, just bypass unmap/unpin and iommu_map
  on replay
---
 drivers/vfio/vfio_iommu_type1.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
index 2d769d4..94a9916 100644
--- a/drivers/vfio/vfio_iommu_type1.c
+++ b/drivers/vfio/vfio_iommu_type1.c
@@ -391,7 +391,7 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
 	struct vfio_domain *domain, *d;
 	long unlocked = 0;
 
-	if (!dma->size)
+	if (!dma->size || dma->type != VFIO_IOVA_USER)
 		return;
 	/*
 	 * We use the IOMMU to track the physical addresses, otherwise we'd
@@ -727,6 +727,9 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu,
 		dma = rb_entry(n, struct vfio_dma, node);
 		iova = dma->iova;
 
+		if (dma->type == VFIO_IOVA_RESERVED)
+			continue;
+
 		while (iova < dma->iova + dma->size) {
 			phys_addr_t phys = iommu_iova_to_phys(d->domain, iova);
 			size_t size;
-- 
1.9.1

[toc] | [prev] | [next] | [standalone]

#1397489 — Re: [PATCH v9 3/7] vfio/type1: bypass unmap/unpin and replay for VFIO_IOVA_RESERVED slots

On Wed,  4 May 2016 11:54:14 +0000
Eric Auger <eric.auger@linaro.org> wrote:

> Before allowing the end-user to create VFIO_IOVA_RESERVED dma slots,
> let's implement the expected behavior for removal and replay. As opposed
> to user dma slots, IOVAs are not systematically bound to PAs and PAs are
> not pinned. VFIO just initializes the IOVA "aperture". IOVAs are allocated
> outside of the VFIO framework, typically the MSI layer which is
> responsible to free and unmap them. The MSI mapping resources are freeed
> by the IOMMU driver on domain destruction.
> 
> Signed-off-by: Eric Auger <eric.auger@linaro.org>
> 
> ---
> 
> v7 -> v8:
> - do no destroy anything anymore, just bypass unmap/unpin and iommu_map
>   on replay
> ---
>  drivers/vfio/vfio_iommu_type1.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
> index 2d769d4..94a9916 100644
> --- a/drivers/vfio/vfio_iommu_type1.c
> +++ b/drivers/vfio/vfio_iommu_type1.c
> @@ -391,7 +391,7 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
>  	struct vfio_domain *domain, *d;
>  	long unlocked = 0;
>  
> -	if (!dma->size)
> +	if (!dma->size || dma->type != VFIO_IOVA_USER)
>  		return;
>  	/*
>  	 * We use the IOMMU to track the physical addresses, otherwise we'd
> @@ -727,6 +727,9 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu,
>  		dma = rb_entry(n, struct vfio_dma, node);
>  		iova = dma->iova;
>  
> +		if (dma->type == VFIO_IOVA_RESERVED)
> +			continue;
> +

But you do still need some sort of replay mechanism, right?  Not to
replay the IOVA to PA mapping, but to call iommu_msi_set_aperture() for
the new domain.  How will you know that this entry is an MSI reserved
range or something else?  Perhaps we can't have a generic "reserved"
type here.

>  		while (iova < dma->iova + dma->size) {
>  			phys_addr_t phys = iommu_iova_to_phys(d->domain, iova);
>  			size_t size;

[toc] | [prev] | [next] | [standalone]

#1399053 — Re: [PATCH v9 3/7] vfio/type1: bypass unmap/unpin and replay for VFIO_IOVA_RESERVED slots

Hi Alex,
On 05/10/2016 12:49 AM, Alex Williamson wrote:
> On Wed,  4 May 2016 11:54:14 +0000
> Eric Auger <eric.auger@linaro.org> wrote:
> 
>> Before allowing the end-user to create VFIO_IOVA_RESERVED dma slots,
>> let's implement the expected behavior for removal and replay. As opposed
>> to user dma slots, IOVAs are not systematically bound to PAs and PAs are
>> not pinned. VFIO just initializes the IOVA "aperture". IOVAs are allocated
>> outside of the VFIO framework, typically the MSI layer which is
>> responsible to free and unmap them. The MSI mapping resources are freeed
>> by the IOMMU driver on domain destruction.
>>
>> Signed-off-by: Eric Auger <eric.auger@linaro.org>
>>
>> ---
>>
>> v7 -> v8:
>> - do no destroy anything anymore, just bypass unmap/unpin and iommu_map
>>   on replay
>> ---
>>  drivers/vfio/vfio_iommu_type1.c | 5 ++++-
>>  1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
>> index 2d769d4..94a9916 100644
>> --- a/drivers/vfio/vfio_iommu_type1.c
>> +++ b/drivers/vfio/vfio_iommu_type1.c
>> @@ -391,7 +391,7 @@ static void vfio_unmap_unpin(struct vfio_iommu *iommu, struct vfio_dma *dma)
>>  	struct vfio_domain *domain, *d;
>>  	long unlocked = 0;
>>  
>> -	if (!dma->size)
>> +	if (!dma->size || dma->type != VFIO_IOVA_USER)
>>  		return;
>>  	/*
>>  	 * We use the IOMMU to track the physical addresses, otherwise we'd
>> @@ -727,6 +727,9 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu,
>>  		dma = rb_entry(n, struct vfio_dma, node);
>>  		iova = dma->iova;
>>  
>> +		if (dma->type == VFIO_IOVA_RESERVED)
>> +			continue;
>> +
> 
> But you do still need some sort of replay mechanism, right?  Not to
> replay the IOVA to PA mapping, but to call iommu_msi_set_aperture() for
> the new domain.  How will you know that this entry is an MSI reserved
> range or something else?  Perhaps we can't have a generic "reserved"
> type here.
Thanks for spotting this bug. I was not testing this case yet and
effectively I need to replay the set_aperture.

Thanks for the review

Best Regards

Eric

> 
>>  		while (iova < dma->iova + dma->size) {
>>  			phys_addr_t phys = iommu_iova_to_phys(d->domain, iova);
>>  			size_t size;
> 

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | linux.kernel

csiph-web