[PATCHv7 05/26] vfio: introduce the VFIO_DMA_MAP_FLAG_NOEXEC flag

We introduce the VFIO_DMA_MAP_FLAG_NOEXEC flag to the VFIO dma map call,
and expose its availability via the capability VFIO_IOMMU_PROT_NOEXEC.
This way the user can control whether the XN flag will be set on the
requested mappings. The IOMMU_NOEXEC flag needs to be available for all
the IOMMUs of the container used.

Signed-off-by: Antonios Motakis <a.motakis-lrHrjnjw1UfHK3s98zE1ajGjJy/sRE9J@public.gmane.org>
---
 include/uapi/linux/vfio.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
index 6612974..30f630c 100644
--- a/include/uapi/linux/vfio.h
+++ b/include/uapi/linux/vfio.h

@@ -29,6 +29,7 @@
  * capability is subject to change as groups are added or removed.
  */
 #define VFIO_DMA_CC_IOMMU		4
+#define VFIO_IOMMU_PROT_NOEXEC		5
 
 /* Check if EEH is supported */
 #define VFIO_EEH			5

@@ -401,6 +402,7 @@ struct vfio_iommu_type1_dma_map {
 	__u32	flags;
 #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
 #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
+#define VFIO_DMA_MAP_FLAG_NOEXEC (1 << 2)	/* not executable from device */
 	__u64	vaddr;				/* Process virtual address */
 	__u64	iova;				/* IO virtual address */
 	__u64	size;				/* Size of mapping (bytes) */

-- 
1.8.3.2

On Tue, 2014-09-23 at 16:46 +0200, Antonios Motakis wrote:

Can't we advertise this as a flag bit in vfio_iommu_type1_info instead?
Also, EEH already took 5 as seen immediately below.

On Tue, 2014-09-23 at 16:21 -0600, Alex Williamson wrote:

On Tue, 2014-09-23 at 16:46 +0200, Antonios Motakis wrote:

quoted

We introduce the VFIO_DMA_MAP_FLAG_NOEXEC flag to the VFIO dma map call,
and expose its availability via the capability VFIO_IOMMU_PROT_NOEXEC.
This way the user can control whether the XN flag will be set on the
requested mappings. The IOMMU_NOEXEC flag needs to be available for all
the IOMMUs of the container used.

Signed-off-by: Antonios Motakis <a.motakis-lrHrjnjw1UfHK3s98zE1ajGjJy/sRE9J@public.gmane.org>
---
 include/uapi/linux/vfio.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
index 6612974..30f630c 100644
--- a/include/uapi/linux/vfio.h
+++ b/include/uapi/linux/vfio.h

@@ -29,6 +29,7 @@
  * capability is subject to change as groups are added or removed.
  */
 #define VFIO_DMA_CC_IOMMU		4
+#define VFIO_IOMMU_PROT_NOEXEC		5

Can't we advertise this as a flag bit in vfio_iommu_type1_info instead?

Ok, I see in the next patch that it's pretty similar to
VFIO_DMA_CC_IOMMU, so the check extension is probably correct for
determining the current state.  Maybe we could name it more similarly,
VFIO_DMA_NOEXEC_IOMMU.  I guess the intended usage is that once a user
attaches a group to the container they can query whether the
VFIO_DMA_MAP_FLAG_NOEXEC is valid.  Ok.  Thanks,

Alex

Also, EEH already took 5 as seen immediately below.

quoted

 
 /* Check if EEH is supported */
 #define VFIO_EEH			5

@@ -401,6 +402,7 @@ struct vfio_iommu_type1_dma_map {
 	__u32	flags;
 #define VFIO_DMA_MAP_FLAG_READ (1 << 0)		/* readable from device */
 #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)	/* writable from device */
+#define VFIO_DMA_MAP_FLAG_NOEXEC (1 << 2)	/* not executable from device */
 	__u64	vaddr;				/* Process virtual address */
 	__u64	iova;				/* IO virtual address */
 	__u64	size;				/* Size of mapping (bytes) */

On Wed, Sep 24, 2014 at 12:30 AM, Alex Williamson
[off-list ref] wrote:

On Tue, 2014-09-23 at 16:21 -0600, Alex Williamson wrote:

quoted

On Tue, 2014-09-23 at 16:46 +0200, Antonios Motakis wrote:

quoted

We introduce the VFIO_DMA_MAP_FLAG_NOEXEC flag to the VFIO dma map call,
and expose its availability via the capability VFIO_IOMMU_PROT_NOEXEC.
This way the user can control whether the XN flag will be set on the
requested mappings. The IOMMU_NOEXEC flag needs to be available for all
the IOMMUs of the container used.

Signed-off-by: Antonios Motakis <a.motakis-lrHrjnjw1UfHK3s98zE1ajGjJy/sRE9J@public.gmane.org>
---
 include/uapi/linux/vfio.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
index 6612974..30f630c 100644
--- a/include/uapi/linux/vfio.h
+++ b/include/uapi/linux/vfio.h

@@ -29,6 +29,7 @@
  * capability is subject to change as groups are added or removed.
  */
 #define VFIO_DMA_CC_IOMMU          4
+#define VFIO_IOMMU_PROT_NOEXEC             5

Can't we advertise this as a flag bit in vfio_iommu_type1_info instead?

Ok, I see in the next patch that it's pretty similar to
VFIO_DMA_CC_IOMMU, so the check extension is probably correct for
determining the current state.  Maybe we could name it more similarly,
VFIO_DMA_NOEXEC_IOMMU.  I guess the intended usage is that once a user
attaches a group to the container they can query whether the
VFIO_DMA_MAP_FLAG_NOEXEC is valid.  Ok.  Thanks,

Sorry for the numbering, ack and ack :)

Alex

quoted

Also, EEH already took 5 as seen immediately below.

quoted

 /* Check if EEH is supported */
 #define VFIO_EEH                   5

@@ -401,6 +402,7 @@ struct vfio_iommu_type1_dma_map {
    __u32   flags;
 #define VFIO_DMA_MAP_FLAG_READ (1 << 0)            /* readable from device */
 #define VFIO_DMA_MAP_FLAG_WRITE (1 << 1)   /* writable from device */
+#define VFIO_DMA_MAP_FLAG_NOEXEC (1 << 2)  /* not executable from device */
    __u64   vaddr;                          /* Process virtual address */
    __u64   iova;                           /* IO virtual address */
    __u64   size;                           /* Size of mapping (bytes) */

-- 
Antonios Motakis
Virtual Open Systems