[PATCH v2 1/2] vfio: add edid api for display (vgpu) devices.

[PATCH v2 1/2] vfio: add edid api for display (vgpu) devices.

[Gerd Hoffmann]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 include/uapi/linux/vfio.h | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)

diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
index 1aa7b82e81..78e5a37d83 100644
--- a/include/uapi/linux/vfio.h
+++ b/include/uapi/linux/vfio.h
@@ -301,6 +301,45 @@ struct vfio_region_info_cap_type {
 #define VFIO_REGION_SUBTYPE_INTEL_IGD_HOST_CFG	(2)
 #define VFIO_REGION_SUBTYPE_INTEL_IGD_LPC_CFG	(3)
 
+#define VFIO_REGION_TYPE_PCI_GFX                (1)
+#define VFIO_REGION_SUBTYPE_GFX_EDID            (1)
+
+/**
+ * Set display link state and edid blob.
+ *
+ * For the edid blob spec look here:
+ * https://en.wikipedia.org/wiki/Extended_Display_Identification_Data
+ *
+ * The guest should be notified about edid changes, for example by
+ * setting the link status to down temporarely (emulate monitor
+ * hotplug).
+ *
+ * @link_state:
+ * VFIO_DEVICE_GFX_LINK_STATE_UP: Monitor is turned on.
+ * VFIO_DEVICE_GFX_LINK_STATE_DOWN: Monitor is turned off.
+ *
+ * @edid_size: Size of the edid data blob.
+ * @edid_blob: The actual edid data.
+ *
+ * Returns 0 on success, error code (such as -EINVAL) on failure.
+ */
+struct vfio_region_gfx_edid {
+	/* device capability hints (read only) */
+	__u32 max_xres;
+	__u32 max_yres;
+	__u32 __reserved1[6];
+
+	/* device state (read/write) */
+	__u32 link_state;
+#define VFIO_DEVICE_GFX_LINK_STATE_UP    1
+#define VFIO_DEVICE_GFX_LINK_STATE_DOWN  2
+	__u32 edid_size;
+	__u32 __reserved2[6];
+
+	/* edid blob (read/write) */
+	__u8 edid_blob[512];
+};
+
 /*
  * The MSIX mappable capability informs that MSIX data of a BAR can be mmapped
  * which allows direct access to non-MSIX registers which happened to be within
-- 
2.9.3

[PATCH v2 2/2] vfio: add edid support to mbochs sample driver

[Gerd Hoffmann]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 samples/vfio-mdev/mbochs.c | 115 +++++++++++++++++++++++++++++++++++++--------
 1 file changed, 96 insertions(+), 19 deletions(-)

diff --git a/samples/vfio-mdev/mbochs.c b/samples/vfio-mdev/mbochs.c
index 2535c3677c..4e72d36ff3 100644
--- a/samples/vfio-mdev/mbochs.c
+++ b/samples/vfio-mdev/mbochs.c
@@ -71,11 +71,17 @@
 #define MBOCHS_NAME		  "mbochs"
 #define MBOCHS_CLASS_NAME	  "mbochs"
 
+#define MBOCHS_EDID_REGION_INDEX  VFIO_PCI_NUM_REGIONS
+#define MBOCHS_NUM_REGIONS        (MBOCHS_EDID_REGION_INDEX+1)
+
 #define MBOCHS_CONFIG_SPACE_SIZE  0xff
 #define MBOCHS_MMIO_BAR_OFFSET	  PAGE_SIZE
 #define MBOCHS_MMIO_BAR_SIZE	  PAGE_SIZE
-#define MBOCHS_MEMORY_BAR_OFFSET  (MBOCHS_MMIO_BAR_OFFSET + \
+#define MBOCHS_EDID_OFFSET	  (MBOCHS_MMIO_BAR_OFFSET +	\
 				   MBOCHS_MMIO_BAR_SIZE)
+#define MBOCHS_EDID_SIZE	  PAGE_SIZE
+#define MBOCHS_MEMORY_BAR_OFFSET  (MBOCHS_EDID_OFFSET + \
+				   MBOCHS_EDID_SIZE)
 
 #define STORE_LE16(addr, val)	(*(u16 *)addr = val)
 #define STORE_LE32(addr, val)	(*(u32 *)addr = val)
@@ -95,16 +101,24 @@ MODULE_PARM_DESC(mem, "megabytes available to " MBOCHS_NAME " devices");
 static const struct mbochs_type {
 	const char *name;
 	u32 mbytes;
+	u32 max_x;
+	u32 max_y;
 } mbochs_types[] = {
 	{
 		.name	= MBOCHS_CLASS_NAME "-" MBOCHS_TYPE_1,
 		.mbytes = 4,
+		.max_x  = 800,
+		.max_y  = 600,
 	}, {
 		.name	= MBOCHS_CLASS_NAME "-" MBOCHS_TYPE_2,
 		.mbytes = 16,
+		.max_x  = 1920,
+		.max_y  = 1440,
 	}, {
 		.name	= MBOCHS_CLASS_NAME "-" MBOCHS_TYPE_3,
 		.mbytes = 64,
+		.max_x  = 0,
+		.max_y  = 0,
 	},
 };
 
@@ -115,6 +129,11 @@ static struct cdev	mbochs_cdev;
 static struct device	mbochs_dev;
 static int		mbochs_used_mbytes;
 
+struct vfio_region_info_ext {
+	struct vfio_region_info          base;
+	struct vfio_region_info_cap_type type;
+};
+
 struct mbochs_mode {
 	u32 drm_format;
 	u32 bytepp;
@@ -144,13 +163,13 @@ struct mdev_state {
 	u32 memory_bar_mask;
 	struct mutex ops_lock;
 	struct mdev_device *mdev;
-	struct vfio_device_info dev_info;
 
 	const struct mbochs_type *type;
 	u16 vbe[VBE_DISPI_INDEX_COUNT];
 	u64 memsize;
 	struct page **pages;
 	pgoff_t pagecount;
+	struct vfio_region_gfx_edid edid_region;
 
 	struct list_head dmabufs;
 	u32 active_id;
@@ -342,10 +361,22 @@ static void handle_mmio_read(struct mdev_state *mdev_state, u16 offset,
 			     char *buf, u32 count)
 {
 	struct device *dev = mdev_dev(mdev_state->mdev);
+	struct vfio_region_gfx_edid *edid;
 	u16 reg16 = 0;
 	int index;
 
 	switch (offset) {
+	case 0x000 ... 0x3ff: /* edid block */
+		edid = &mdev_state->edid_region;
+		if (offset + count > sizeof(edid->edid_blob))
+			goto unhandled;
+		if (edid->link_state != VFIO_DEVICE_GFX_LINK_STATE_UP ||
+		    offset >= edid->edid_size) {
+			memset(buf, 0, count);
+			break;
+		}
+		memcpy(buf, edid->edid_blob + offset, count);
+		break;
 	case 0x500 ... 0x515: /* bochs dispi interface */
 		if (count != 2)
 			goto unhandled;
@@ -365,6 +396,28 @@ static void handle_mmio_read(struct mdev_state *mdev_state, u16 offset,
 	}
 }
 
+static void handle_edid_write(struct mdev_state *mdev_state, u16 offset,
+			      char *buf, u32 count)
+{
+	char *region = (void *)&mdev_state->edid_region;
+
+	if (offset < offsetof(typeof(mdev_state->edid_region), link_state))
+		return;
+	if (offset + count > sizeof(mdev_state->edid_region))
+		return;
+	memcpy(region + offset, buf, count);
+}
+
+static void handle_edid_read(struct mdev_state *mdev_state, u16 offset,
+			     char *buf, u32 count)
+{
+	char *region = (void *)&mdev_state->edid_region;
+
+	if (offset + count > sizeof(mdev_state->edid_region))
+		return;
+	memcpy(buf, region + offset, count);
+}
+
 static ssize_t mdev_access(struct mdev_device *mdev, char *buf, size_t count,
 			   loff_t pos, bool is_write)
 {
@@ -384,13 +437,23 @@ static ssize_t mdev_access(struct mdev_device *mdev, char *buf, size_t count,
 			memcpy(buf, (mdev_state->vconfig + pos), count);
 
 	} else if (pos >= MBOCHS_MMIO_BAR_OFFSET &&
-		   pos + count <= MBOCHS_MEMORY_BAR_OFFSET) {
+		   pos + count <= (MBOCHS_MMIO_BAR_OFFSET +
+				   MBOCHS_MMIO_BAR_SIZE)) {
 		pos -= MBOCHS_MMIO_BAR_OFFSET;
 		if (is_write)
 			handle_mmio_write(mdev_state, pos, buf, count);
 		else
 			handle_mmio_read(mdev_state, pos, buf, count);
 
+	} else if (pos >= MBOCHS_EDID_OFFSET &&
+		   pos + count <= (MBOCHS_EDID_OFFSET +
+				   MBOCHS_EDID_SIZE)) {
+		pos -= MBOCHS_EDID_OFFSET;
+		if (is_write)
+			handle_edid_write(mdev_state, pos, buf, count);
+		else
+			handle_edid_read(mdev_state, pos, buf, count);
+
 	} else if (pos >= MBOCHS_MEMORY_BAR_OFFSET &&
 		   pos + count <=
 		   MBOCHS_MEMORY_BAR_OFFSET + mdev_state->memsize) {
@@ -471,6 +534,8 @@ static int mbochs_create(struct kobject *kobj, struct mdev_device *mdev)
 	mdev_state->next_id = 1;
 
 	mdev_state->type = type;
+	mdev_state->edid_region.max_xres = type->max_x;
+	mdev_state->edid_region.max_yres = type->max_y;
 	mbochs_create_config_space(mdev_state);
 	mbochs_reset(mdev);
 
@@ -932,16 +997,16 @@ static int mbochs_dmabuf_export(struct mbochs_dmabuf *dmabuf)
 }
 
 static int mbochs_get_region_info(struct mdev_device *mdev,
-				  struct vfio_region_info *region_info,
-				  u16 *cap_type_id, void **cap_type)
+				  struct vfio_region_info_ext *ext)
 {
+	struct vfio_region_info *region_info = &ext->base;
 	struct mdev_state *mdev_state;
 
 	mdev_state = mdev_get_drvdata(mdev);
 	if (!mdev_state)
 		return -EINVAL;
 
-	if (region_info->index >= VFIO_PCI_NUM_REGIONS)
+	if (region_info->index >= MBOCHS_NUM_REGIONS)
 		return -EINVAL;
 
 	switch (region_info->index) {
@@ -964,6 +1029,20 @@ static int mbochs_get_region_info(struct mdev_device *mdev,
 		region_info->flags  = (VFIO_REGION_INFO_FLAG_READ  |
 				       VFIO_REGION_INFO_FLAG_WRITE);
 		break;
+	case MBOCHS_EDID_REGION_INDEX:
+		ext->base.argsz = sizeof(*ext);
+		ext->base.offset = MBOCHS_EDID_OFFSET;
+		ext->base.size = MBOCHS_EDID_SIZE;
+		ext->base.flags = (VFIO_REGION_INFO_FLAG_READ  |
+				   VFIO_REGION_INFO_FLAG_WRITE |
+				   VFIO_REGION_INFO_FLAG_CAPS);
+		ext->base.cap_offset = offsetof(typeof(*ext), type);
+		ext->type.header.id = VFIO_REGION_INFO_CAP_TYPE;
+		ext->type.header.version = 1;
+		ext->type.header.next = 0;
+		ext->type.type = VFIO_REGION_TYPE_PCI_GFX;
+		ext->type.subtype = VFIO_REGION_SUBTYPE_GFX_EDID;
+		break;
 	default:
 		region_info->size   = 0;
 		region_info->offset = 0;
@@ -984,7 +1063,7 @@ static int mbochs_get_device_info(struct mdev_device *mdev,
 				  struct vfio_device_info *dev_info)
 {
 	dev_info->flags = VFIO_DEVICE_FLAGS_PCI;
-	dev_info->num_regions = VFIO_PCI_NUM_REGIONS;
+	dev_info->num_regions = MBOCHS_NUM_REGIONS;
 	dev_info->num_irqs = VFIO_PCI_NUM_IRQS;
 	return 0;
 }
@@ -1084,7 +1163,7 @@ static long mbochs_ioctl(struct mdev_device *mdev, unsigned int cmd,
 			unsigned long arg)
 {
 	int ret = 0;
-	unsigned long minsz;
+	unsigned long minsz, outsz;
 	struct mdev_state *mdev_state;
 
 	mdev_state = mdev_get_drvdata(mdev);
@@ -1106,8 +1185,6 @@ static long mbochs_ioctl(struct mdev_device *mdev, unsigned int cmd,
 		if (ret)
 			return ret;
 
-		memcpy(&mdev_state->dev_info, &info, sizeof(info));
-
 		if (copy_to_user((void __user *)arg, &info, minsz))
 			return -EFAULT;
 
@@ -1115,24 +1192,24 @@ static long mbochs_ioctl(struct mdev_device *mdev, unsigned int cmd,
 	}
 	case VFIO_DEVICE_GET_REGION_INFO:
 	{
-		struct vfio_region_info info;
-		u16 cap_type_id = 0;
-		void *cap_type = NULL;
+		struct vfio_region_info_ext info;
 
-		minsz = offsetofend(struct vfio_region_info, offset);
+		minsz = offsetofend(typeof(info), base.offset);
 
 		if (copy_from_user(&info, (void __user *)arg, minsz))
 			return -EFAULT;
 
-		if (info.argsz < minsz)
+		outsz = info.base.argsz;
+		if (outsz < minsz)
+			return -EINVAL;
+		if (outsz > sizeof(info))
 			return -EINVAL;
 
-		ret = mbochs_get_region_info(mdev, &info, &cap_type_id,
-					   &cap_type);
+		ret = mbochs_get_region_info(mdev, &info);
 		if (ret)
 			return ret;
 
-		if (copy_to_user((void __user *)arg, &info, minsz))
+		if (copy_to_user((void __user *)arg, &info, outsz))
 			return -EFAULT;
 
 		return 0;
@@ -1148,7 +1225,7 @@ static long mbochs_ioctl(struct mdev_device *mdev, unsigned int cmd,
 			return -EFAULT;
 
 		if ((info.argsz < minsz) ||
-		    (info.index >= mdev_state->dev_info.num_irqs))
+		    (info.index >= VFIO_PCI_NUM_IRQS))
 			return -EINVAL;
 
 		ret = mbochs_get_irq_info(mdev, &info);
-- 
2.9.3

Re: [PATCH v2 1/2] vfio: add edid api for display (vgpu) devices.

[Alex Williamson]

On Tue, 18 Sep 2018 15:38:12 +0200
Gerd Hoffmann <kraxel@redhat.com> wrote:

No empty commit logs please.  There must be something to say about the
goal or motivation beyond the subject.

> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> ---
>  include/uapi/linux/vfio.h | 39 +++++++++++++++++++++++++++++++++++++++
>  1 file changed, 39 insertions(+)
> 
> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
> index 1aa7b82e81..78e5a37d83 100644
> --- a/include/uapi/linux/vfio.h
> +++ b/include/uapi/linux/vfio.h
> @@ -301,6 +301,45 @@ struct vfio_region_info_cap_type {
>  #define VFIO_REGION_SUBTYPE_INTEL_IGD_HOST_CFG	(2)
>  #define VFIO_REGION_SUBTYPE_INTEL_IGD_LPC_CFG	(3)
>  
> +#define VFIO_REGION_TYPE_PCI_GFX                (1)

nit, what's the PCI dependency?

> +#define VFIO_REGION_SUBTYPE_GFX_EDID            (1)
> +
> +/**
> + * Set display link state and edid blob.
> + *
> + * For the edid blob spec look here:
> + * https://en.wikipedia.org/wiki/Extended_Display_Identification_Data
> + *
> + * The guest should be notified about edid changes, for example by
> + * setting the link status to down temporarely (emulate monitor
> + * hotplug).

Who is responsible for this notification, the user interacting with
this region or the driver providing the region when a new edid is
provided?  This comment needs to state the expected API as clearly as
possible.

> + *
> + * @link_state:
> + * VFIO_DEVICE_GFX_LINK_STATE_UP: Monitor is turned on.
> + * VFIO_DEVICE_GFX_LINK_STATE_DOWN: Monitor is turned off.
> + *
> + * @edid_size: Size of the edid data blob.
> + * @edid_blob: The actual edid data.

What signals that the user edid_blob update is complete?  Should the
size be written before or after the blob?  Is the user required to
update the entire blob in a single write or can it be written
incrementally?

It might also be worth defining access widths, I see that you use
memcpy to support any width in mbochs, but we could define only native
field accesses for discrete registers if it makes the implementation
easier.

> + *
> + * Returns 0 on success, error code (such as -EINVAL) on failure.

Left over from ioctl.

> + */
> +struct vfio_region_gfx_edid {
> +	/* device capability hints (read only) */
> +	__u32 max_xres;
> +	__u32 max_yres;
> +	__u32 __reserved1[6];

Is the plan to use the version field within vfio_info_cap_header to
make use of these reserved fields later, ie. version 2 might define a
field from this reserved block?

> +
> +	/* device state (read/write) */
> +	__u32 link_state;
> +#define VFIO_DEVICE_GFX_LINK_STATE_UP    1
> +#define VFIO_DEVICE_GFX_LINK_STATE_DOWN  2
> +	__u32 edid_size;
> +	__u32 __reserved2[6];
> +
> +	/* edid blob (read/write) */
> +	__u8 edid_blob[512];

It seems the placement of this blob is what makes us feel like we need
to introduce reserved fields for later use, but we could instead define
an edid_offset read-only field so that the blob is always at the end of
whatever discrete fields we define.  Perhaps then we wouldn't even need
a read-only vs read-write section, simply define it per virtual
register.

Overall, I prefer this approach rather than adding yet more ioctls for
every feature and extension we add, thanks for implementing it.  What's
your impression vs ioctls?  Thanks,

Alex

Re: [PATCH v2 1/2] vfio: add edid api for display (vgpu) devices.

[Gerd Hoffmann]

On Wed, Sep 19, 2018 at 01:52:19PM -0600, Alex Williamson wrote:
> On Tue, 18 Sep 2018 15:38:12 +0200
> Gerd Hoffmann <kraxel@redhat.com> wrote:
> 
> No empty commit logs please.  There must be something to say about the
> goal or motivation beyond the subject.
> 
> > Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> > ---
> >  include/uapi/linux/vfio.h | 39 +++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 39 insertions(+)
> > 
> > diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
> > index 1aa7b82e81..78e5a37d83 100644
> > --- a/include/uapi/linux/vfio.h
> > +++ b/include/uapi/linux/vfio.h
> > @@ -301,6 +301,45 @@ struct vfio_region_info_cap_type {
> >  #define VFIO_REGION_SUBTYPE_INTEL_IGD_HOST_CFG	(2)
> >  #define VFIO_REGION_SUBTYPE_INTEL_IGD_LPC_CFG	(3)
> >  
> > +#define VFIO_REGION_TYPE_PCI_GFX                (1)
> 
> nit, what's the PCI dependency?

None I guess, just copy&paste from the (pci) vendor cap above ...
Will rename it.

> > + * The guest should be notified about edid changes, for example by
> > + * setting the link status to down temporarely (emulate monitor
> > + * hotplug).
> 
> Who is responsible for this notification, the user interacting with
> this region or the driver providing the region when a new edid is
> provided?  This comment needs to state the expected API as clearly as
> possible.

> > + * @link_state:
> > + * VFIO_DEVICE_GFX_LINK_STATE_UP: Monitor is turned on.
> > + * VFIO_DEVICE_GFX_LINK_STATE_DOWN: Monitor is turned off.
> > + *
> > + * @edid_size: Size of the edid data blob.
> > + * @edid_blob: The actual edid data.
> 
> What signals that the user edid_blob update is complete?  Should the
> size be written before or after the blob?  Is the user required to
> update the entire blob in a single write or can it be written
> incrementally?

My qemu test code first sets link-state = down, then updates edid, then
sets link-state = up.  I can document that as official update protocol.

Drivers should be able to map it do emulation events pretty straight
forward.  Except maybe for the final down-up transition, drivers might
have to delay this to make sure the guest has seen the up-down
transition.

> It might also be worth defining access widths, I see that you use
> memcpy to support any width in mbochs, but we could define only native
> field accesses for discrete registers if it makes the implementation
> easier.

Makes sense for the u32 registers (especially the writable ones).

For mbochs the virtual hardware has no link state notification, so I can
just store whatever I get, then check link-state when the guest attempts
to read the edid.

When actually implementing link-state notification it is probably useful
to not have to deal with a half-written link-state field ...

> > + */
> > +struct vfio_region_gfx_edid {
> > +	/* device capability hints (read only) */
> > +	__u32 max_xres;
> > +	__u32 max_yres;
> > +	__u32 __reserved1[6];
> 
> Is the plan to use the version field within vfio_info_cap_header to
> make use of these reserved fields later, ie. version 2 might define a
> field from this reserved block?

Yes.

> > +	/* device state (read/write) */
> > +	__u32 link_state;
> > +#define VFIO_DEVICE_GFX_LINK_STATE_UP    1
> > +#define VFIO_DEVICE_GFX_LINK_STATE_DOWN  2
> > +	__u32 edid_size;
> > +	__u32 __reserved2[6];
> > +
> > +	/* edid blob (read/write) */
> > +	__u8 edid_blob[512];
> 
> It seems the placement of this blob is what makes us feel like we need
> to introduce reserved fields for later use, but we could instead define
> an edid_offset read-only field so that the blob is always at the end of
> whatever discrete fields we define.

Yes, sounds sensible.

> Perhaps then we wouldn't even need a read-only vs read-write section,
> simply define it per virtual register.

I'll look into this.

> Overall, I prefer this approach rather than adding yet more ioctls for
> every feature and extension we add, thanks for implementing it.  What's
> your impression vs ioctls?

I'd still prefer ioctls, but this is workable too.  So if you like it
better this way lets go for it.

cheers,
  Gerd