Skip to contentSkip to conversation

Deliver vGPU display refresh event to userspace

RFC PATCH In review Kechen Lu 5 participants 22 messages Jul 18, 2019 – Aug 2, 2019

Subsystems (5)
lore
Share & tools
lore

Kechen Lu

author
#0 lore raw

v4 cover Deliver vGPU display refresh event to userspace

This series tries to send the vGPU display refresh event to user land.

Instead of delivering page flip events only or vblank events only, we choose to combine two of them, i.e. post display refresh event at vblanks and skip some of them when no page flip happens. Vblanks as upper bound are safe and skipping no-page-flip vblanks guarantees both trivial performance impacts and good user experience without screen tearing. Plus, we have the mask/unmask mechansim providing user space flexibility to switch between event-notified refresh and classic timer-based refresh.

In addition, there are some cases that guest app only uses one framebuffer for both drawing and display. In such case, guest OS won't do the plane page flip when the framebuffer is updated, thus the user land won't be notified about the updated framebuffer. Hence, in single framebuffer case, we apply a heuristic to determine whether it is the case or not. If it is, notify user land when each vblank event triggers.

v4:
- Deliver page flip event and single framebuffer refresh event bounded by display vblanks. (Kechen)
v3:
- Deliver display vblank event instead of page flip event. (Zhenyu) v2:
- Use VFIO irq chain to get eventfds from userspace instead of adding a new ABI. (Alex)
v1:
- https://patchwork.kernel.org/cover/10962341/

Kechen Lu (2):
drm/i915/gvt: Deliver async primary plane page flip events at vblank drm/i915/gvt: Add cursor plane reg update trap emulation handler

Tina Zhang (4):
vfio: Define device specific irq type capability
vfio: Introduce vGPU display irq type drm/i915/gvt: Register vGPU display event irq drm/i915/gvt: Deliver vGPU refresh event to userspace

 drivers/gpu/drm/i915/gvt/cmd_parser.c |   5 +-
 drivers/gpu/drm/i915/gvt/display.c    |  54 ++++++-
 drivers/gpu/drm/i915/gvt/gvt.h        |  11 ++
 drivers/gpu/drm/i915/gvt/handlers.c   |  32 ++++-
 drivers/gpu/drm/i915/gvt/hypercall.h  |   1 +
 drivers/gpu/drm/i915/gvt/interrupt.c  |   7 +
 drivers/gpu/drm/i915/gvt/interrupt.h  |   3 +
 drivers/gpu/drm/i915/gvt/kvmgt.c      | 197 ++++++++++++++++++++++++--
 drivers/gpu/drm/i915/gvt/mpt.h        |  17 +++
 include/uapi/linux/vfio.h             |  22 ++-
 10 files changed, 330 insertions(+), 19 deletions(-)
Signature · 1 line
2.17.1
author
#1 lore raw

RFC PATCH v4 1/6 vfio: Define device specific irq type capability

From: Tina Zhang <[email protected]>

Cap the number of irqs with fixed indexes and use capability chains to chain device specific irqs.

---
 include/uapi/linux/vfio.h | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)
include/uapi/linux/vfio.h +18 -1
index 8f10748dac79..be6adab4f759 100644
@@ -448,11 +448,27 @@ struct vfio_irq_info {
 #define VFIO_IRQ_INFO_MASKABLE		(1 << 1)
 #define VFIO_IRQ_INFO_AUTOMASKED	(1 << 2)
 #define VFIO_IRQ_INFO_NORESIZE		(1 << 3)
+#define VFIO_IRQ_INFO_FLAG_CAPS		(1 << 4) /* Info supports caps */
 	__u32	index;		/* IRQ index */
 	__u32	count;		/* Number of IRQs within this index */
+	__u32	cap_offset;	/* Offset within info struct of first cap */
 };
 #define VFIO_DEVICE_GET_IRQ_INFO	_IO(VFIO_TYPE, VFIO_BASE + 9)
 
+/*
+ * The irq type capability allows irqs unique to a specific device or
+ * class of devices to be exposed.
+ *
+ * The structures below define version 1 of this capability.
+ */
+#define VFIO_IRQ_INFO_CAP_TYPE      3
+
+struct vfio_irq_info_cap_type {
+	struct vfio_info_cap_header header;
+	__u32 type;     /* global per bus driver */
+	__u32 subtype;  /* type specific */
+};
+
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct vfio_irq_set)
  *
@@ -554,7 +570,8 @@ enum {
 	VFIO_PCI_MSIX_IRQ_INDEX,
 	VFIO_PCI_ERR_IRQ_INDEX,
 	VFIO_PCI_REQ_IRQ_INDEX,
-	VFIO_PCI_NUM_IRQS
+	VFIO_PCI_NUM_IRQS = 5	/* Fixed user ABI, IRQ indexes >=5 use   */
+				/* device specific cap to define content */
 };
 
 /*
Signature · 1 line
2.17.1
#2 lore raw

Re: [RFC PATCH v4 1/6] vfio: Define device specific irq type capability

Quoted 16 lines · Kechen Lu

From: Tina Zhang <[email protected]>

Cap the number of irqs with fixed indexes and use capability chains to chain device specific irqs.

---
 include/uapi/linux/vfio.h | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)
include/uapi/linux/vfio.h +0 -0
index 8f10748dac79..be6adab4f759 100644

include/uapi/linux/vfio.h struct vfio_irq_info { · line 454

Quoted 7 lines
include/uapi/linux/vfio.h +2 -0
@@ -448,11 +448,27 @@ struct vfio_irq_info {
 #define VFIO_IRQ_INFO_MASKABLE		(1 << 1)
 #define VFIO_IRQ_INFO_AUTOMASKED	(1 << 2)
 #define VFIO_IRQ_INFO_NORESIZE		(1 << 3)
+#define VFIO_IRQ_INFO_FLAG_CAPS		(1 << 4) /* Info supports caps */
 	__u32	index;		/* IRQ index */
 	__u32	count;		/* Number of IRQs within this index */
+	__u32	cap_offset;	/* Offset within info struct of first cap */

This still breaks ABI as argsz would be updated with this new field, so it would cause compat issue. I think my last suggestion was to assume cap list starts after vfio_irq_info.

Quoted 33 lines
include/uapi/linux/vfio.h +16 -1
 };
 #define VFIO_DEVICE_GET_IRQ_INFO	_IO(VFIO_TYPE, VFIO_BASE + 9)
 
+/*
+ * The irq type capability allows irqs unique to a specific device or
+ * class of devices to be exposed.
+ *
+ * The structures below define version 1 of this capability.
+ */
+#define VFIO_IRQ_INFO_CAP_TYPE      3
+
+struct vfio_irq_info_cap_type {
+	struct vfio_info_cap_header header;
+	__u32 type;     /* global per bus driver */
+	__u32 subtype;  /* type specific */
+};
+
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct vfio_irq_set)
  *
@@ -554,7 +570,8 @@ enum {
 	VFIO_PCI_MSIX_IRQ_INDEX,
 	VFIO_PCI_ERR_IRQ_INDEX,
 	VFIO_PCI_REQ_IRQ_INDEX,
-	VFIO_PCI_NUM_IRQS
+	VFIO_PCI_NUM_IRQS = 5	/* Fixed user ABI, IRQ indexes >=5 use   */
+				/* device specific cap to define content */
 };
 
 /*
Signature · 1 line
2.17.1
Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
author
#3 lore raw

RE: [RFC PATCH v4 1/6] vfio: Define device specific irq type capability

Hi,

Quoted 35 lines

-----Original Message-----
From: Zhenyu Wang [mailto:[email protected]]
Sent: Friday, July 19, 2019 2:06 PM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>; [email protected]; Eric Auger <[email protected]>
Subject: Re: [RFC PATCH v4 1/6] vfio: Define device specific irq type capability

Quoted 16 lines · Kechen Lu

From: Tina Zhang <[email protected]>

Cap the number of irqs with fixed indexes and use capability chains to chain device specific irqs.

---
 include/uapi/linux/vfio.h | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)
include/uapi/linux/vfio.h +0 -0
index 8f10748dac79..be6adab4f759 100644

include/uapi/linux/vfio.h struct vfio_irq_info { · line 451

Quoted 4 lines
include/uapi/linux/vfio.h +1 -0
@@ -448,11 +448,27 @@ struct vfio_irq_info {
 #define VFIO_IRQ_INFO_MASKABLE		(1 << 1)
 #define VFIO_IRQ_INFO_AUTOMASKED	(1 << 2)
 #define VFIO_IRQ_INFO_NORESIZE		(1 << 3)
+#define VFIO_IRQ_INFO_FLAG_CAPS		(1 << 4) /* Info supports caps

*/

Quoted 8 lines

include/uapi/linux/vfio.h struct vfio_irq_info { · line 454

Quoted 3 lines
include/uapi/linux/vfio.h +1 -0
 	__u32	index;		/* IRQ index */
 	__u32	count;		/* Number of IRQs within this index */
+	__u32	cap_offset;	/* Offset within info struct of first cap */

This still breaks ABI as argsz would be updated with this new field, so it would cause compat issue. I think my last suggestion was to assume cap list starts after vfio_irq_info.

In the common practice, the general logic is first use the "count" as the "minsz" boundary to perform copy from user, and then perform following logic, so that the incompatibility issue would not happen. BTW, this patch has been double checked by Eric Auger before included in his patch-set.

Best Regards,
Kechen

Quoted 40 lines
Quoted 9 lines
include/uapi/linux/vfio.h +6 -0
 };
 #define VFIO_DEVICE_GET_IRQ_INFO	_IO(VFIO_TYPE, VFIO_BASE + 9)
 
+/*
+ * The irq type capability allows irqs unique to a specific device 
+or
+ * class of devices to be exposed.
+ *
+ * The structures below define version 1 of this capability.

include/uapi/linux/vfio.h struct vfio_irq_info {

Quoted 10 lines
include/uapi/linux/vfio.h +8 -0
+ */
+#define VFIO_IRQ_INFO_CAP_TYPE      3
+
+struct vfio_irq_info_cap_type {
+	struct vfio_info_cap_header header;
+	__u32 type;     /* global per bus driver */
+	__u32 subtype;  /* type specific */ };
+
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct

vfio_irq_set)

include/uapi/linux/vfio.h enum { · line 573

Quoted 7 lines
include/uapi/linux/vfio.h +1 -1
  *
@@ -554,7 +570,8 @@ enum {
 	VFIO_PCI_MSIX_IRQ_INDEX,
 	VFIO_PCI_ERR_IRQ_INDEX,
 	VFIO_PCI_REQ_IRQ_INDEX,
-	VFIO_PCI_NUM_IRQS
+	VFIO_PCI_NUM_IRQS = 5	/* Fixed user ABI, IRQ indexes >=5 use

*/

Quoted 7 lines
include/uapi/linux/vfio.h +1 -1
+				/* device specific cap to define content */
 };
 
 /*
--

2.17.1

--
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

#4 lore raw

Re: [RFC PATCH v4 1/6] vfio: Define device specific irq type capability

Quoted 46 lines · Kechen

Hi,

Quoted 35 lines

-----Original Message-----
From: Zhenyu Wang [mailto:[email protected]]
Sent: Friday, July 19, 2019 2:06 PM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>; [email protected]; Eric Auger <[email protected]>
Subject: Re: [RFC PATCH v4 1/6] vfio: Define device specific irq type capability

Quoted 16 lines · Kechen Lu

From: Tina Zhang <[email protected]>

Cap the number of irqs with fixed indexes and use capability chains to chain device specific irqs.

---
 include/uapi/linux/vfio.h | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)
include/uapi/linux/vfio.h +0 -0
index 8f10748dac79..be6adab4f759 100644

include/uapi/linux/vfio.h struct vfio_irq_info { · line 451

Quoted 4 lines
include/uapi/linux/vfio.h +1 -0
@@ -448,11 +448,27 @@ struct vfio_irq_info {
 #define VFIO_IRQ_INFO_MASKABLE		(1 << 1)
 #define VFIO_IRQ_INFO_AUTOMASKED	(1 << 2)
 #define VFIO_IRQ_INFO_NORESIZE		(1 << 3)
+#define VFIO_IRQ_INFO_FLAG_CAPS		(1 << 4) /* Info supports caps

*/

Quoted 8 lines

include/uapi/linux/vfio.h struct vfio_irq_info { · line 454

Quoted 3 lines
include/uapi/linux/vfio.h +1 -0
 	__u32	index;		/* IRQ index */
 	__u32	count;		/* Number of IRQs within this index */
+	__u32	cap_offset;	/* Offset within info struct of first cap */

This still breaks ABI as argsz would be updated with this new field, so it would cause compat issue. I think my last suggestion was to assume cap list starts after vfio_irq_info.

Quoted 2 lines

In the common practice, the general logic is first use the "count" as the "minsz" boundary to perform copy from user, and then perform following logic, so that the incompatibility issue would not happen. BTW, this patch has been double checked by Eric Auger before included in his patch-set.

yeah, sorry I was thinking vfio might fail in that case but it seems current code assume argsz should be larger than minsz for count here, so that's fine.

Quoted 41 lines
Quoted 41 lines
Quoted 10 lines
include/uapi/linux/vfio.h +6 -0
 };
 #define VFIO_DEVICE_GET_IRQ_INFO	_IO(VFIO_TYPE, VFIO_BASE + 9)
 
+/*
+ * The irq type capability allows irqs unique to a specific device 
+or
+ * class of devices to be exposed.
+ *
+ * The structures below define version 1 of this capability.

include/uapi/linux/vfio.h struct vfio_irq_info {

Quoted 10 lines
include/uapi/linux/vfio.h +8 -0
+ */
+#define VFIO_IRQ_INFO_CAP_TYPE      3
+
+struct vfio_irq_info_cap_type {
+	struct vfio_info_cap_header header;
+	__u32 type;     /* global per bus driver */
+	__u32 subtype;  /* type specific */ };
+
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct

vfio_irq_set)

include/uapi/linux/vfio.h enum { · line 573

Quoted 7 lines
include/uapi/linux/vfio.h +1 -1
  *
@@ -554,7 +570,8 @@ enum {
 	VFIO_PCI_MSIX_IRQ_INDEX,
 	VFIO_PCI_ERR_IRQ_INDEX,
 	VFIO_PCI_REQ_IRQ_INDEX,
-	VFIO_PCI_NUM_IRQS
+	VFIO_PCI_NUM_IRQS = 5	/* Fixed user ABI, IRQ indexes >=5 use

*/

Quoted 7 lines
include/uapi/linux/vfio.h +1 -1
+				/* device specific cap to define content */
 };
 
 /*
--

2.17.1

--
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
author
#5 lore raw

RFC PATCH v4 2/6 vfio: Introduce vGPU display irq type

In reply to Kechen Lu

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/vfio.h +3 -0
index be6adab4f759..df28b17a6e2e 100644
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */
 };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct vfio_irq_set)
  *
Signature · 1 line
2.17.1
#6 lore raw

Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 15 lines · On Thu, 18 Jul 2019 23:56:36 +0800 Kechen Lu <[email protected]> wrote:

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/vfio.h +0 -0
index be6adab4f759..df28b17a6e2e 100644

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 474

Quoted 6 lines
include/uapi/linux/vfio.h +3 -0
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */
 };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+

Please include a description defining exactly what this IRQ is intended to signal. For instance, if another vGPU vendor wanted to implement this in their driver and didn't have the QEMU code for reference to what it does with the IRQ, what would they need to know? Thanks,

Alex

Quoted 3 lines
include/uapi/linux/vfio.h +0 -0
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct vfio_irq_set)
  *
author
#7 lore raw

RE: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Hi,

Quoted 41 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Saturday, July 20, 2019 12:25 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 15 lines · On Thu, 18 Jul 2019 23:56:36 +0800 Kechen Lu <[email protected]> wrote:

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/vfio.h +0 -0
index be6adab4f759..df28b17a6e2e 100644

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 474

Quoted 6 lines
include/uapi/linux/vfio.h +3 -0
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */
 };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+

Please include a description defining exactly what this IRQ is intended to signal. For instance, if another vGPU vendor wanted to implement this in their driver and didn't have the QEMU code for reference to what it does with the IRQ, what would they need to know? Thanks,

Quoted 2 lines

Alex

Yes, that makes more sense. I'll add the description for it at next version patch.

BTW, may I have one more question? In the current design ideas, we partitioned the vGPU display eventfd counted 8-byte value into at most 8 events to deliver multiple display events, so we need different increasement counter value to differentiate the events. As this is the exposed thing the QEMU has to know, we plan adds a macro here VFIO_IRQ_SUBTYPE_GFX_DISPLAY_EVENTFD_BASE_SHIFT to make sure the partitions shift in 1 byte, does it make sense putting here? Looking forward to your and Gerd's comments. Thanks!

Best Regards,
Kechen

Quoted 4 lines

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 476

Quoted 2 lines
include/uapi/linux/vfio.h +0 -0
 /**
  * VFIO_DEVICE_SET_IRQS - _IOW(VFIO_TYPE, VFIO_BASE + 10, struct

vfio_irq_set)

Quoted 1 line
include/uapi/linux/vfio.h +0 -0
  *
#8 lore raw

Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 48 lines · Kechen

Hi,

Quoted 41 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Saturday, July 20, 2019 12:25 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 15 lines · On Thu, 18 Jul 2019 23:56:36 +0800 Kechen Lu <[email protected]> wrote:

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/vfio.h +0 -0
index be6adab4f759..df28b17a6e2e 100644

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 474

Quoted 6 lines
include/uapi/linux/vfio.h +3 -0
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */
 };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+  

Please include a description defining exactly what this IRQ is intended to signal. For instance, if another vGPU vendor wanted to implement this in their driver and didn't have the QEMU code for reference to what it does with the IRQ, what would they need to know? Thanks,

Quoted 2 lines

Alex

Yes, that makes more sense. I'll add the description for it at next version patch.

Quoted 7 lines

BTW, may I have one more question? In the current design ideas, we partitioned the vGPU display eventfd counted 8-byte value into at most 8 events to deliver multiple display events, so we need different increasement counter value to differentiate the events. As this is the exposed thing the QEMU has to know, we plan adds a macro here VFIO_IRQ_SUBTYPE_GFX_DISPLAY_EVENTFD_BASE_SHIFT to make sure the partitions shift in 1 byte, does it make sense putting here? Looking forward to your and Gerd's comments. Thanks!

Couldn't you expose this as another capability within the IRQ_INFO return data? If you were to define it as a macro, I assume that means it would be hard coded, in which case this probably becomes an Intel specific IRQ, rather than what appears to be framed as a generic graphics IRQ extension. A new capability could instead allow the vendor to specify their own value, where we could define how userspace should interpret and make use of this value. Thanks,

Alex

#9 lore raw

RE: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 76 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Tuesday, July 23, 2019 3:41 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 39 lines · Kechen

Hi,

Quoted 37 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Saturday, July 20, 2019 12:25 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 15 lines · On Thu, 18 Jul 2019 23:56:36 +0800 Kechen Lu <[email protected]> wrote:

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/vfio.h +0 -0
index be6adab4f759..df28b17a6e2e 100644

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 473

Quoted 5 lines
include/uapi/linux/vfio.h +3 -0
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */  };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+

Please include a description defining exactly what this IRQ is intended to

signal.

Quoted 8 lines
Quoted 6 lines

For instance, if another vGPU vendor wanted to implement this in their driver and didn't have the QEMU code for reference to what it does with the IRQ, what would they need to know? Thanks,

Alex

Yes, that makes more sense. I'll add the description for it at next version

patch.

Quoted 7 lines

BTW, may I have one more question? In the current design ideas, we partitioned the vGPU display eventfd counted 8-byte value into at most 8 events to deliver multiple display events, so we need different increasement counter value to differentiate the events. As this is the exposed thing the QEMU has to know, we plan adds a macro here VFIO_IRQ_SUBTYPE_GFX_DISPLAY_EVENTFD_BASE_SHIFT to make sure

the

Quoted 1 line

partitions shift in 1 byte, does it make sense putting here? Looking forward

to your and Gerd's comments. Thanks!

Quoted 7 lines

Couldn't you expose this as another capability within the IRQ_INFO return data? If you were to define it as a macro, I assume that means it would be hard coded, in which case this probably becomes an Intel specific IRQ, rather than what appears to be framed as a generic graphics IRQ extension. A new capability could instead allow the vendor to specify their own value, where we could define how userspace should interpret and make use of this value. Thanks,

Good suggestion. Currently, vfio_irq_info is used to save one irq info. What we need here is to use it to save several events info. Maybe we could figure out a general layout of this capability so that it can be leveraged by others, not only for display irq/events. Thanks.

BR,
Tina

Quoted 2 lines

Alex

#10 lore raw

Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 81 lines · Tina
Quoted 74 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Tuesday, July 23, 2019 3:41 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 39 lines · Kechen

Hi,

Quoted 37 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Saturday, July 20, 2019 12:25 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 15 lines · On Thu, 18 Jul 2019 23:56:36 +0800 Kechen Lu <[email protected]> wrote:

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/vfio.h +0 -0
index be6adab4f759..df28b17a6e2e 100644

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 473

Quoted 5 lines
include/uapi/linux/vfio.h +3 -0
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */  };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+  

Please include a description defining exactly what this IRQ is intended to

signal.

Quoted 8 lines
Quoted 6 lines

For instance, if another vGPU vendor wanted to implement this in their driver and didn't have the QEMU code for reference to what it does with the IRQ, what would they need to know? Thanks,

Alex

Yes, that makes more sense. I'll add the description for it at next version

patch.

Quoted 7 lines

BTW, may I have one more question? In the current design ideas, we partitioned the vGPU display eventfd counted 8-byte value into at most 8 events to deliver multiple display events, so we need different increasement counter value to differentiate the events. As this is the exposed thing the QEMU has to know, we plan adds a macro here VFIO_IRQ_SUBTYPE_GFX_DISPLAY_EVENTFD_BASE_SHIFT to make sure

the

Quoted 1 line

partitions shift in 1 byte, does it make sense putting here? Looking forward

to your and Gerd's comments. Thanks!

Quoted 7 lines

Couldn't you expose this as another capability within the IRQ_INFO return data? If you were to define it as a macro, I assume that means it would be hard coded, in which case this probably becomes an Intel specific IRQ, rather than what appears to be framed as a generic graphics IRQ extension. A new capability could instead allow the vendor to specify their own value, where we could define how userspace should interpret and make use of this value. Thanks,

Quoted 4 lines

Good suggestion. Currently, vfio_irq_info is used to save one irq info. What we need here is to use it to save several events info. Maybe we could figure out a general layout of this capability so that it can be leveraged by others, not only for display irq/events.

You could also expose a device specific IRQ with count > 1 (ie. similar to MSI/X) and avoid munging the eventfd value, which is not something we do elsewhere, at least in vfio. Thanks,

Alex

#11 lore raw

RE: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 106 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Tuesday, July 23, 2019 9:19 AM
To: Zhang, Tina <[email protected]>
Cc: Lu, Kechen <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 80 lines · Tina
Quoted 80 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Tuesday, July 23, 2019 3:41 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 43 lines · Kechen

Hi,

Quoted 41 lines

-----Original Message-----
From: Alex Williamson [mailto:[email protected]]
Sent: Saturday, July 20, 2019 12:25 AM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>
Subject: Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 17 lines · On Thu, 18 Jul 2019 23:56:36 +0800 Kechen Lu <[email protected]> wrote:

From: Tina Zhang <[email protected]>

Introduce vGPU specific irq type VFIO_IRQ_TYPE_GFX, and VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ as the subtype for vGPU display

---
 include/uapi/linux/vfio.h | 3 +++
 1 file changed, 3 insertions(+)
(quoted diff) +0 -0

b/include/uapi/linux/vfio.h index be6adab4f759..df28b17a6e2e 100644

include/uapi/linux/vfio.h struct vfio_irq_info_cap_type { · line 473

Quoted 5 lines
include/uapi/linux/vfio.h +3 -0
@@ -469,6 +469,9 @@ struct vfio_irq_info_cap_type {
 	__u32 subtype;  /* type specific */  };
 
+#define VFIO_IRQ_TYPE_GFX				(1)
+#define VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ		(1)
+

Please include a description defining exactly what this IRQ is intended to

signal.

Quoted 9 lines
Quoted 6 lines

For instance, if another vGPU vendor wanted to implement this in their driver and didn't have the QEMU code for reference to what it does with the IRQ, what would they need to know? Thanks,

Alex

Yes, that makes more sense. I'll add the description for it at next version

patch.

Quoted 8 lines

BTW, may I have one more question? In the current design ideas, we partitioned the vGPU display eventfd counted 8-byte value into at most
8 events to deliver multiple display events, so we need different increasement counter value to differentiate the events. As this is the exposed thing the QEMU has to know, we plan adds a macro here VFIO_IRQ_SUBTYPE_GFX_DISPLAY_EVENTFD_BASE_SHIFT to make sure

the

Quoted 2 lines

partitions shift in 1 byte, does it make sense putting here? Looking forward

Quoted 8 lines
Quoted 8 lines

to your and Gerd's comments. Thanks!

Couldn't you expose this as another capability within the IRQ_INFO return data? If you were to define it as a macro, I assume that means it would be hard coded, in which case this probably becomes an Intel specific IRQ, rather than what appears to be framed as a generic graphics IRQ extension. A new capability could instead allow the vendor to specify their own value, where we could define how

userspace should interpret and make use of this value.

Quoted 1 line
Quoted 1 line

Thanks,

Quoted 8 lines
Quoted 4 lines

Good suggestion. Currently, vfio_irq_info is used to save one irq info. What we need here is to use it to save several events info. Maybe we could figure out a general layout of this capability so that it can be leveraged by others, not only for display irq/events.

You could also expose a device specific IRQ with count > 1 (ie. similar to MSI/X) and avoid munging the eventfd value, which is not something we do elsewhere, at least in vfio. Thanks,

Actually, we had this implementation before. At that time, we got the suggestion that count > 1 means more than one eventfd which might be not necessary. Anyway, we can consider the "count > 1" again if anyone is agree on this. Thanks

BR,
Tina

Quoted 2 lines

Alex

#12 lore raw

Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

In reply to Alex Williamson

Hi,

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

Couldn't you expose this as another capability within the IRQ_INFO return data? If you were to define it as a macro, I assume that means it would be hard coded, in which case this probably becomes an Intel specific IRQ, rather than what appears to be framed as a generic graphics IRQ extension. A new capability could instead allow the vendor to specify their own value, where we could define how userspace should interpret and make use of this value. Thanks,

Quoted 8 lines
Quoted 4 lines

Good suggestion. Currently, vfio_irq_info is used to save one irq info. What we need here is to use it to save several events info. Maybe we could figure out a general layout of this capability so that it can be leveraged by others, not only for display irq/events.

You could also expose a device specific IRQ with count > 1 (ie. similar to MSI/X) and avoid munging the eventfd value, which is not something we do elsewhere, at least in vfio. Thanks,

Well, the basic idea is to use the eventfd value to signal the kind of changes which did happen, simliar to IRQ status register bits.

So, when the guest changes the primary plane, the mdev driver notes this. Same with the cursor plane. On vblank (when the guests update is actually applied) the mdev driver wakes the eventfd and uses eventfd value to signal whenever primary plane or cursor plane or both did change.

Then userspace knows which planes need an update without an extra VFIO_DEVICE_QUERY_GFX_PLANE roundtrip to the kernel.

Alternatively we could have one eventfd for each change type. But given that these changes are typically applied at the same time (vblank) we would have multiple eventfds being signaled at the same time. Which doesn't look ideal to me ...

cheers,
Gerd

#13 lore raw

Re: [RFC PATCH v4 2/6] vfio: Introduce vGPU display irq type

Quoted 27 lines · On Fri, 2 Aug 2019 15:35:31 +0200 "[email protected]" <[email protected]> wrote:

Hi,

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

Couldn't you expose this as another capability within the IRQ_INFO return data? If you were to define it as a macro, I assume that means it would be hard coded, in which case this probably becomes an Intel specific IRQ, rather than what appears to be framed as a generic graphics IRQ extension. A new capability could instead allow the vendor to specify their own value, where we could define how userspace should interpret and make use of this value. Thanks,

Quoted 8 lines
Quoted 4 lines

Good suggestion. Currently, vfio_irq_info is used to save one irq info. What we need here is to use it to save several events info. Maybe we could figure out a general layout of this capability so that it can be leveraged by others, not only for display irq/events.

You could also expose a device specific IRQ with count > 1 (ie. similar to MSI/X) and avoid munging the eventfd value, which is not something we do elsewhere, at least in vfio. Thanks,

Well, the basic idea is to use the eventfd value to signal the kind of changes which did happen, simliar to IRQ status register bits.

So, when the guest changes the primary plane, the mdev driver notes this. Same with the cursor plane. On vblank (when the guests update is actually applied) the mdev driver wakes the eventfd and uses eventfd value to signal whenever primary plane or cursor plane or both did change.

Quoted 7 lines

Then userspace knows which planes need an update without an extra VFIO_DEVICE_QUERY_GFX_PLANE roundtrip to the kernel.

Alternatively we could have one eventfd for each change type. But given that these changes are typically applied at the same time (vblank) we would have multiple eventfds being signaled at the same time. Which doesn't look ideal to me ...

Good point, looking at the bits in the eventfd value seems better than a flood of concurrent interrupts. Thanks,

Alex

author
#14 lore raw

RFC PATCH v4 3/6 drm/i915/gvt: Register vGPU display event irq

In reply to Kechen Lu

From: Tina Zhang <[email protected]>

Gvt-g emulates and injects the vGPU's display interrupts in kernel space. However the dma-buf based framebuffer consumer in the user land (e.g. Qemu vfio/display) may also need to be notified by this event.

Register the display irq as VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ to each vGPU, so that the display interrupt event can be delivered to userspace through eventfd.

---
 drivers/gpu/drm/i915/gvt/display.c   | 10 +++++--
 drivers/gpu/drm/i915/gvt/gvt.h       |  2 ++
 drivers/gpu/drm/i915/gvt/hypercall.h |  1 +
 drivers/gpu/drm/i915/gvt/kvmgt.c     | 43 ++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/gvt/mpt.h       | 17 +++++++++++
 5 files changed, 71 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +8 -2
index e1c313da6c00..1a0a4ae4826e 100644
@@ -506,16 +506,22 @@ void intel_vgpu_clean_display(struct intel_vgpu *vgpu)
 int intel_vgpu_init_display(struct intel_vgpu *vgpu, u64 resolution)
 {
 	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
+	int ret;
 
 	intel_vgpu_init_i2c_edid(vgpu);
 
 	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) ||
 	    IS_COFFEELAKE(dev_priv))
-		return setup_virtual_dp_monitor(vgpu, PORT_D, GVT_DP_D,
+		ret = setup_virtual_dp_monitor(vgpu, PORT_D, GVT_DP_D,
 						resolution);
 	else
-		return setup_virtual_dp_monitor(vgpu, PORT_B, GVT_DP_B,
+		ret = setup_virtual_dp_monitor(vgpu, PORT_B, GVT_DP_B,
 						resolution);
+
+	if (ret == 0)
+		intel_gvt_hypervisor_register_display_irq(vgpu);
+
+	return ret;
 }
 
 /**
drivers/gpu/drm/i915/gvt/gvt.h +2 -0
index b54f2bdc13a4..64d1c1aaa42a 100644
@@ -201,6 +201,8 @@ struct intel_vgpu {
 		struct mdev_device *mdev;
 		struct vfio_region *region;
 		int num_regions;
+		struct vfio_irq *irq;
+		int num_irqs;
 		struct eventfd_ctx *intx_trigger;
 		struct eventfd_ctx *msi_trigger;
 
drivers/gpu/drm/i915/gvt/hypercall.h +1 -0
index 4862fb12778e..be33f20f3bc1 100644
@@ -68,6 +68,7 @@ struct intel_gvt_mpt {
 			     bool map);
 	int (*set_opregion)(void *vgpu);
 	int (*set_edid)(void *vgpu, int port_num);
+	int (*register_display_irq)(void *vgpu);
 	int (*get_vfio_device)(void *vgpu);
 	void (*put_vfio_device)(void *vgpu);
 	bool (*is_valid_gfn)(unsigned long handle, unsigned long gfn);
drivers/gpu/drm/i915/gvt/kvmgt.c +43 -0

67 lines not on this page · raw message

drivers/gpu/drm/i915/gvt/mpt.h +17 -0
index 0f9440128123..abf4a69920d3 100644
@@ -330,6 +330,23 @@ static inline int intel_gvt_hypervisor_set_edid(struct intel_vgpu *vgpu,
 	return intel_gvt_host.mpt->set_edid(vgpu, port_num);
 }
 
+/**
+ * intel_gvt_hypervisor_set_irq - register vgpu specific irq
+ * @vgpu: a vGPU
+ * @port_num: display port number
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+static inline int intel_gvt_hypervisor_register_display_irq(
+						struct intel_vgpu *vgpu)
+{
+	if (!intel_gvt_host.mpt->register_display_irq)
+		return 0;
+
+	return intel_gvt_host.mpt->register_display_irq(vgpu);
+}
+
 /**
  * intel_gvt_hypervisor_get_vfio_device - increase vfio device ref count
  * @vgpu: a vGPU
Signature · 1 line
2.17.1
author
#15 lore raw

RFC PATCH v4 4/6 drm/i915/gvt: Deliver vGPU refresh event to userspace

In reply to Kechen Lu

From: Tina Zhang <[email protected]>

Deliver the display refresh events to the user land. Userspace can use the irq mask/unmask mechanism to disable or enable the event delivery.

As we know, delivering refresh event at each vblank safely avoids tearing and unexpected event overwhelming, but there are still spaces to optimize.

For handling the normal case, deliver the page flip refresh event at each vblank, in other words, bounded by vblanks. Skipping some events bring performance enhancement while not hurting user experience.

For single framebuffer case, deliver the refresh events to userspace at all vblanks. This heuristic at each vblank leverages pageflip_count incresements to determine if there is no page flip happens after a certain period and so that the case is regarded as single framebuffer one. Although this heuristic makes incorrect decision sometimes and it depends on guest behavior, for example, when no cursor movements happen, the user experience does not harm and front buffer is still correctly acquired. Meanwhile, in actual single framebuffer case, the user experience is enhanced compared with page flip events only.

Addtionally, to mitigate the events delivering footprints, one eventfd and 8 byte eventfd counter partition are leveraged.

---
 drivers/gpu/drm/i915/gvt/display.c |  21 ++++
 drivers/gpu/drm/i915/gvt/gvt.h     |   7 ++
 drivers/gpu/drm/i915/gvt/kvmgt.c   | 154 +++++++++++++++++++++++++++--
 3 files changed, 173 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +21 -0
index 1a0a4ae4826e..036db8199983 100644
@@ -387,6 +387,8 @@ void intel_gvt_check_vblank_emulation(struct intel_gvt *gvt)
 	mutex_unlock(&gvt->lock);
 }
 
+#define PAGEFLIP_INC_COUNT 5
+
 static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 {
 	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
@@ -396,7 +398,10 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		[PIPE_B] = PIPE_B_VBLANK,
 		[PIPE_C] = PIPE_C_VBLANK,
 	};
+	int pri_flip_event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
 	int event;
+	u64 eventfd_signal_val = 0;
+	static int pageflip_count;
 
 	if (pipe < PIPE_A || pipe > PIPE_C)
 		return;
@@ -407,11 +412,27 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		if (!pipe_is_enabled(vgpu, pipe))
 			continue;
 
+		if (event == pri_flip_event) {
+			eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
+	if (--pageflip_count < 0) {
+		eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
+		pageflip_count = 0;
+	}
+
+	if (vgpu->vdev.vblank_trigger && !(vgpu->vdev.display_event_mask
+		& (DISPLAY_PRI_REFRESH_EVENT | DISPLAY_CUR_REFRESH_EVENT)) &&
+		eventfd_signal_val)
+		eventfd_signal(vgpu->vdev.vblank_trigger, eventfd_signal_val);
+
 	if (pipe_is_enabled(vgpu, pipe)) {
 		vgpu_vreg_t(vgpu, PIPE_FRMCOUNT_G4X(pipe))++;
+
 		intel_vgpu_trigger_virtual_event(vgpu, vblank_event[pipe]);
 	}
 }
drivers/gpu/drm/i915/gvt/gvt.h +7 -0
index 64d1c1aaa42a..b654b6fa0663 100644
@@ -165,6 +165,11 @@ struct intel_vgpu_submission {
 	bool active;
 };
 
+#define DISPLAY_PRI_REFRESH_EVENT	(1 << 0)
+#define DISPLAY_PRI_REFRESH_EVENT_INC	(1UL << 56)
+#define DISPLAY_CUR_REFRESH_EVENT	(1 << 1)
+#define DISPLAY_CUR_REFRESH_EVENT_INC	(1UL << 48)
+
 struct intel_vgpu {
 	struct intel_gvt *gvt;
 	struct mutex vgpu_lock;
@@ -205,6 +210,8 @@ struct intel_vgpu {
 		int num_irqs;
 		struct eventfd_ctx *intx_trigger;
 		struct eventfd_ctx *msi_trigger;
+		struct eventfd_ctx *vblank_trigger;
+		u32 display_event_mask;
 
 		/*
 		 * Two caches are used to avoid mapping duplicated pages (eg.
drivers/gpu/drm/i915/gvt/kvmgt.c +145 -9

210 lines not on this page · raw message

Signature · 1 line
2.17.1
#16 lore raw

Re: [RFC PATCH v4 4/6] drm/i915/gvt: Deliver vGPU refresh event to userspace

Quoted 64 lines · Kechen Lu

From: Tina Zhang <[email protected]>

Deliver the display refresh events to the user land. Userspace can use the irq mask/unmask mechanism to disable or enable the event delivery.

As we know, delivering refresh event at each vblank safely avoids tearing and unexpected event overwhelming, but there are still spaces to optimize.

For handling the normal case, deliver the page flip refresh event at each vblank, in other words, bounded by vblanks. Skipping some events bring performance enhancement while not hurting user experience.

For single framebuffer case, deliver the refresh events to userspace at all vblanks. This heuristic at each vblank leverages pageflip_count incresements to determine if there is no page flip happens after a certain period and so that the case is regarded as single framebuffer one. Although this heuristic makes incorrect decision sometimes and it depends on guest behavior, for example, when no cursor movements happen, the user experience does not harm and front buffer is still correctly acquired. Meanwhile, in actual single framebuffer case, the user experience is enhanced compared with page flip events only.

Addtionally, to mitigate the events delivering footprints, one eventfd and 8 byte eventfd counter partition are leveraged.

---
 drivers/gpu/drm/i915/gvt/display.c |  21 ++++
 drivers/gpu/drm/i915/gvt/gvt.h     |   7 ++
 drivers/gpu/drm/i915/gvt/kvmgt.c   | 154 +++++++++++++++++++++++++++--
 3 files changed, 173 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +7 -0
index 1a0a4ae4826e..036db8199983 100644
@@ -387,6 +387,8 @@ void intel_gvt_check_vblank_emulation(struct intel_gvt *gvt)
 	mutex_unlock(&gvt->lock);
 }
 
+#define PAGEFLIP_INC_COUNT 5
+
 static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 {
 	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
@@ -396,7 +398,10 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		[PIPE_B] = PIPE_B_VBLANK,
 		[PIPE_C] = PIPE_C_VBLANK,
 	};
+	int pri_flip_event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
 	int event;
+	u64 eventfd_signal_val = 0;
+	static int pageflip_count;
 
 	if (pipe < PIPE_A || pipe > PIPE_C)
 		return;
@@ -407,11 +412,27 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		if (!pipe_is_enabled(vgpu, pipe))
 			continue;
 
+		if (event == pri_flip_event) {
+			eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe) · line 426

Quoted 10 lines
drivers/gpu/drm/i915/gvt/display.c +7 -0
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
+	if (--pageflip_count < 0) {
+		eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
+		pageflip_count = 0;
+	}

If pageflip_count has been increased to a big number from page flip event for some time, then if guest switch for single buffer render, it would take 5x vblank time to send refresh then..

Quoted 17 lines
drivers/gpu/drm/i915/gvt/display.c +7 -0
+
+	if (vgpu->vdev.vblank_trigger && !(vgpu->vdev.display_event_mask
+		& (DISPLAY_PRI_REFRESH_EVENT | DISPLAY_CUR_REFRESH_EVENT)) &&
+		eventfd_signal_val)
+		eventfd_signal(vgpu->vdev.vblank_trigger, eventfd_signal_val);
+
 	if (pipe_is_enabled(vgpu, pipe)) {
 		vgpu_vreg_t(vgpu, PIPE_FRMCOUNT_G4X(pipe))++;
+
 		intel_vgpu_trigger_virtual_event(vgpu, vblank_event[pipe]);
 	}
 }
drivers/gpu/drm/i915/gvt/gvt.h +0 -0
index 64d1c1aaa42a..b654b6fa0663 100644

drivers/gpu/drm/i915/gvt/gvt.h struct intel_vgpu_submission { · line 172

Quoted 8 lines
drivers/gpu/drm/i915/gvt/gvt.h +5 -0
@@ -165,6 +165,11 @@ struct intel_vgpu_submission {
 	bool active;
 };
 
+#define DISPLAY_PRI_REFRESH_EVENT	(1 << 0)
+#define DISPLAY_PRI_REFRESH_EVENT_INC	(1UL << 56)
+#define DISPLAY_CUR_REFRESH_EVENT	(1 << 1)
+#define DISPLAY_CUR_REFRESH_EVENT_INC	(1UL << 48)
+

As this is for eventfd interface definition, need to put in vfio header instead of gvt's, as this is userspace API. And better reorder for different usage on irq masking and eventfd value.

For eventfd value, this looks like counter for each plane? Or do we just need a flag?

Quoted 235 lines
drivers/gpu/drm/i915/gvt/gvt.h +2 -0
 struct intel_vgpu {
 	struct intel_gvt *gvt;
 	struct mutex vgpu_lock;
@@ -205,6 +210,8 @@ struct intel_vgpu {
 		int num_irqs;
 		struct eventfd_ctx *intx_trigger;
 		struct eventfd_ctx *msi_trigger;
+		struct eventfd_ctx *vblank_trigger;
+		u32 display_event_mask;
 
 		/*
 		 * Two caches are used to avoid mapping duplicated pages (eg.
drivers/gpu/drm/i915/gvt/kvmgt.c +145 -9
index 6fe825763d05..61c634618217 100644
@@ -1222,6 +1222,8 @@ static int intel_vgpu_get_irq_count(struct intel_vgpu *vgpu, int type)
 {
 	if (type == VFIO_PCI_INTX_IRQ_INDEX || type == VFIO_PCI_MSI_IRQ_INDEX)
 		return 1;
+	else if (type < VFIO_PCI_NUM_IRQS + vgpu->vdev.num_irqs)
+		return vgpu->vdev.irq[type - VFIO_PCI_NUM_IRQS].count;
 
 	return 0;
 }
@@ -1269,7 +1271,62 @@ static int intel_vgpu_set_msi_trigger(struct intel_vgpu *vgpu,
 	return 0;
 }
 
-static int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags,
+static int intel_vgu_set_display_irq_mask(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	if (start != 0 || count > 2)
+		return -EINVAL;
+
+	if (flags & VFIO_IRQ_SET_DATA_NONE)
+		vgpu->vdev.display_event_mask |= DISPLAY_PRI_REFRESH_EVENT |
+			DISPLAY_CUR_REFRESH_EVENT;
+
+	return 0;
+}
+
+static int intel_vgu_set_display_irq_unmask(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	if (start != 0 || count > 2)
+		return -EINVAL;
+
+	if (flags & VFIO_IRQ_SET_DATA_NONE)
+		vgpu->vdev.display_event_mask &= ~(DISPLAY_PRI_REFRESH_EVENT |
+			   DISPLAY_CUR_REFRESH_EVENT);
+
+	return 0;
+}
+
+static int intel_vgpu_set_display_event_trigger(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	struct eventfd_ctx *trigger;
+
+	if (flags & VFIO_IRQ_SET_DATA_EVENTFD) {
+		int fd = *(int *)data;
+
+		trigger = eventfd_ctx_fdget(fd);
+		if (IS_ERR(trigger)) {
+			gvt_vgpu_err("eventfd_ctx_fdget failed\n");
+			return PTR_ERR(trigger);
+		}
+		vgpu->vdev.vblank_trigger = trigger;
+		vgpu->vdev.display_event_mask = 0;
+	} else if ((flags & VFIO_IRQ_SET_DATA_NONE) && !count) {
+		trigger = vgpu->vdev.vblank_trigger;
+		if (trigger) {
+			eventfd_ctx_put(trigger);
+			vgpu->vdev.vblank_trigger = NULL;
+		}
+	}
+
+	return 0;
+}
+
+int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags,
 		unsigned int index, unsigned int start, unsigned int count,
 		void *data)
 {
@@ -1302,6 +1359,35 @@ static int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags,
 			break;
 		}
 		break;
+	default:
+	{
+		int i;
+
+		if (index >= VFIO_PCI_NUM_IRQS +
+					vgpu->vdev.num_irqs)
+			return -EINVAL;
+		index =
+			array_index_nospec(index,
+						VFIO_PCI_NUM_IRQS +
+						vgpu->vdev.num_irqs);
+
+		i = index - VFIO_PCI_NUM_IRQS;
+		if (vgpu->vdev.irq[i].type == VFIO_IRQ_TYPE_GFX &&
+		    vgpu->vdev.irq[i].subtype ==
+		    VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ) {
+			switch (flags & VFIO_IRQ_SET_ACTION_TYPE_MASK) {
+			case VFIO_IRQ_SET_ACTION_MASK:
+				func = intel_vgu_set_display_irq_mask;
+				break;
+			case VFIO_IRQ_SET_ACTION_UNMASK:
+				func = intel_vgu_set_display_irq_unmask;
+				break;
+			case VFIO_IRQ_SET_ACTION_TRIGGER:
+				func = intel_vgpu_set_display_event_trigger;
+				break;
+			}
+		}
+	}
 	}
 
 	if (!func)
@@ -1333,7 +1419,7 @@ static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd,
 		info.flags |= VFIO_DEVICE_FLAGS_RESET;
 		info.num_regions = VFIO_PCI_NUM_REGIONS +
 				vgpu->vdev.num_regions;
-		info.num_irqs = VFIO_PCI_NUM_IRQS;
+		info.num_irqs = VFIO_PCI_NUM_IRQS + vgpu->vdev.num_irqs;
 
 		return copy_to_user((void __user *)arg, &info, minsz) ?
 			-EFAULT : 0;
@@ -1493,32 +1579,81 @@ static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd,
 			-EFAULT : 0;
 	} else if (cmd == VFIO_DEVICE_GET_IRQ_INFO) {
 		struct vfio_irq_info info;
+		struct vfio_info_cap caps = { .buf = NULL, .size = 0 };
+		unsigned int i;
+		int ret;
 
 		minsz = offsetofend(struct vfio_irq_info, count);
 
 		if (copy_from_user(&info, (void __user *)arg, minsz))
 			return -EFAULT;
 
-		if (info.argsz < minsz || info.index >= VFIO_PCI_NUM_IRQS)
+		if (info.argsz < minsz)
 			return -EINVAL;
 
 		switch (info.index) {
 		case VFIO_PCI_INTX_IRQ_INDEX:
 		case VFIO_PCI_MSI_IRQ_INDEX:
+			info.flags = VFIO_IRQ_INFO_EVENTFD;
 			break;
-		default:
+		case VFIO_PCI_MSIX_IRQ_INDEX:
+		case VFIO_PCI_ERR_IRQ_INDEX:
+		case VFIO_PCI_REQ_IRQ_INDEX:
 			return -EINVAL;
-		}
+		default:
+		{
+			struct vfio_irq_info_cap_type cap_type = {
+				.header.id = VFIO_IRQ_INFO_CAP_TYPE,
+				.header.version = 1 };
 
-		info.flags = VFIO_IRQ_INFO_EVENTFD;
+			if (info.index >= VFIO_PCI_NUM_IRQS +
+					vgpu->vdev.num_irqs)
+				return -EINVAL;
+			info.index =
+				array_index_nospec(info.index,
+						VFIO_PCI_NUM_IRQS +
+						vgpu->vdev.num_irqs);
+
+			i = info.index - VFIO_PCI_NUM_IRQS;
+
+			info.flags = vgpu->vdev.irq[i].flags;
+			cap_type.type = vgpu->vdev.irq[i].type;
+			cap_type.subtype = vgpu->vdev.irq[i].subtype;
+
+			ret = vfio_info_add_capability(&caps,
+						&cap_type.header,
+						sizeof(cap_type));
+			if (ret)
+				return ret;
+		}
+		}
 
 		info.count = intel_vgpu_get_irq_count(vgpu, info.index);
 
 		if (info.index == VFIO_PCI_INTX_IRQ_INDEX)
 			info.flags |= (VFIO_IRQ_INFO_MASKABLE |
 				       VFIO_IRQ_INFO_AUTOMASKED);
-		else
-			info.flags |= VFIO_IRQ_INFO_NORESIZE;
+
+		if (caps.size) {
+			info.flags |= VFIO_IRQ_INFO_FLAG_CAPS;
+			if (info.argsz < sizeof(info) + caps.size) {
+				info.argsz = sizeof(info) + caps.size;
+				info.cap_offset = 0;
+			} else {
+				vfio_info_cap_shift(&caps, sizeof(info));
+				if (copy_to_user((void __user *)arg +
+						  sizeof(info), caps.buf,
+						  caps.size)) {
+					kfree(caps.buf);
+					return -EFAULT;
+				}
+				info.cap_offset = sizeof(info);
+				if (offsetofend(struct vfio_irq_info, cap_offset) > minsz)
+					minsz = offsetofend(struct vfio_irq_info, cap_offset);
+			}
+
+			kfree(caps.buf);
+		}
 
 		return copy_to_user((void __user *)arg, &info, minsz) ?
 			-EFAULT : 0;
@@ -1537,7 +1672,8 @@ static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd,
 			int max = intel_vgpu_get_irq_count(vgpu, hdr.index);
 
 			ret = vfio_set_irqs_validate_and_prepare(&hdr, max,
-						VFIO_PCI_NUM_IRQS, &data_size);
+					VFIO_PCI_NUM_IRQS + vgpu->vdev.num_irqs,
+								 &data_size);
 			if (ret) {
 				gvt_vgpu_err("intel:vfio_set_irqs_validate_and_prepare failed\n");
 				return -EINVAL;
Signature · 1 line
2.17.1
Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
author
#17 lore raw

RE: [RFC PATCH v4 4/6] drm/i915/gvt: Deliver vGPU refresh event to userspace

Hi,

Quoted 91 lines

-----Original Message-----
From: Zhenyu Wang [mailto:[email protected]]
Sent: Friday, July 19, 2019 2:25 PM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Tina <[email protected]>; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>; [email protected]
Subject: Re: [RFC PATCH v4 4/6] drm/i915/gvt: Deliver vGPU refresh event to userspace

Quoted 13 lines · Kechen Lu

From: Tina Zhang <[email protected]>

Deliver the display refresh events to the user land. Userspace can use the irq mask/unmask mechanism to disable or enable the event delivery.

As we know, delivering refresh event at each vblank safely avoids tearing and unexpected event overwhelming, but there are still spaces to optimize.

For handling the normal case, deliver the page flip refresh event at each vblank, in other words, bounded by vblanks. Skipping some events bring performance enhancement while not hurting user experience.

Quoted 7 lines

For single framebuffer case, deliver the refresh events to userspace at all vblanks. This heuristic at each vblank leverages pageflip_count incresements to determine if there is no page flip happens after a certain period and so that the case is regarded as single framebuffer one. Although this heuristic makes incorrect decision sometimes and it depends on guest behavior, for example, when no cursor movements happen, the user experience does not harm and front buffer is still correctly

acquired.

drivers/gpu/drm/i915/gvt/display.c void intel_gvt_check_vblank_emulation(struct

Quoted 21 lines

Meanwhile, in actual single framebuffer case, the user experience is enhanced compared with page flip events only.

Addtionally, to mitigate the events delivering footprints, one eventfd and
8 byte eventfd counter partition are leveraged.

---
 drivers/gpu/drm/i915/gvt/display.c |  21 ++++
 drivers/gpu/drm/i915/gvt/gvt.h     |   7 ++
 drivers/gpu/drm/i915/gvt/kvmgt.c   | 154 +++++++++++++++++++++++++++--
 3 files changed, 173 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +0 -0
index 1a0a4ae4826e..036db8199983 100644
@@ -387,6 +387,8 @@ void intel_gvt_check_vblank_emulation(struct

intel_gvt *gvt)

drivers/gpu/drm/i915/gvt/display.c void intel_gvt_check_vblank_emulation(struct intel_gvt *gvt)

Quoted 9 lines
drivers/gpu/drm/i915/gvt/display.c +3 -0
 	mutex_unlock(&gvt->lock);
 }
 
+#define PAGEFLIP_INC_COUNT 5
+
 static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 {
 	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; @@ -396,7
+398,10 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int

pipe)

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct

Quoted 11 lines
drivers/gpu/drm/i915/gvt/display.c +3 -0
 		[PIPE_B] = PIPE_B_VBLANK,
 		[PIPE_C] = PIPE_C_VBLANK,
 	};
+	int pri_flip_event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
 	int event;
+	u64 eventfd_signal_val = 0;
+	static int pageflip_count;
 
 	if (pipe < PIPE_A || pipe > PIPE_C)
 		return;
@@ -407,11 +412,27 @@ static void emulate_vblank_on_pipe(struct

intel_vgpu *vgpu, int pipe)

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)

Quoted 5 lines
drivers/gpu/drm/i915/gvt/display.c +2 -0
 		if (!pipe_is_enabled(vgpu, pipe))
 			continue;
 
+		if (event == pri_flip_event) {
+			eventfd_signal_val +=

DISPLAY_PRI_REFRESH_EVENT_INC;

Quoted 6 lines
drivers/gpu/drm/i915/gvt/display.c +3 -0
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
Quoted 9 lines

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe) · line 426

Quoted 4 lines
drivers/gpu/drm/i915/gvt/display.c +4 -0
+	if (--pageflip_count < 0) {
+		eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
+		pageflip_count = 0;
+	}

If pageflip_count has been increased to a big number from page flip event for some time, then if guest switch for single buffer render, it would take 5x vblank time to send refresh then..

Yep. You are right. That could be the corner case. After discussing with Tina, we can change the pageflip_count to nopageflip_count using a changed heuristic to avoid this issue and bound the delay to begin refresh.

Quoted 22 lines

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)

Quoted 3 lines
drivers/gpu/drm/i915/gvt/display.c +3 -0
+
+	if (vgpu->vdev.vblank_trigger && !(vgpu->vdev.display_event_mask
+		& (DISPLAY_PRI_REFRESH_EVENT |

DISPLAY_CUR_REFRESH_EVENT)) &&

Quoted 18 lines
drivers/gpu/drm/i915/gvt/display.c +4 -0
+		eventfd_signal_val)
+		eventfd_signal(vgpu->vdev.vblank_trigger, eventfd_signal_val);
+
 	if (pipe_is_enabled(vgpu, pipe)) {
 		vgpu_vreg_t(vgpu, PIPE_FRMCOUNT_G4X(pipe))++;
+
 		intel_vgpu_trigger_virtual_event(vgpu, vblank_event[pipe]);
 	}
 }
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/gvt/gvt.h index 64d1c1aaa42a..b654b6fa0663 100644

drivers/gpu/drm/i915/gvt/gvt.h +0 -0
@@ -165,6 +165,11 @@ struct intel_vgpu_submission {
 	bool active;
 };
 
Quoted 10 lines

drivers/gpu/drm/i915/gvt/gvt.h struct intel_vgpu_submission { · line 172

Quoted 5 lines
drivers/gpu/drm/i915/gvt/gvt.h +5 -0
+#define DISPLAY_PRI_REFRESH_EVENT	(1 << 0)
+#define DISPLAY_PRI_REFRESH_EVENT_INC	(1UL << 56)
+#define DISPLAY_CUR_REFRESH_EVENT	(1 << 1)
+#define DISPLAY_CUR_REFRESH_EVENT_INC	(1UL << 48)
+

As this is for eventfd interface definition, need to put in vfio header instead of gvt's, as this is userspace API. And better reorder for different usage on irq masking and eventfd value.

I think I made redundant extra flags here. Currently if we are using single irq masking/unmasking, DISPLAY_XXX_REFRESH_EVENT defined is unnecessary. I'll change it in the next version patch.

Quoted 3 lines

For eventfd value, this looks like counter for each plane? Or do we just need a flag?

We need the count val for each plane here in userspace, and we need this to perform partitioned eventfd value checking logic. But it seems too specific to be defined in vfio header. After discussing with Tina, we think we can define a base mask like VFIO_IRQ_EVENTFD_COUNT_BASE_MASK as 0xff, to read/write 1 byte of 8-byte counter of eventfd, and shifted by the drm plane specific flag, but it still needs the discussion with Gerd and Alex. Thanks.

Best Regards,
Kechen

Quoted 254 lines
Quoted 18 lines
drivers/gpu/drm/i915/gvt/gvt.h +2 -0
 struct intel_vgpu {
 	struct intel_gvt *gvt;
 	struct mutex vgpu_lock;
@@ -205,6 +210,8 @@ struct intel_vgpu {
 		int num_irqs;
 		struct eventfd_ctx *intx_trigger;
 		struct eventfd_ctx *msi_trigger;
+		struct eventfd_ctx *vblank_trigger;
+		u32 display_event_mask;
 
 		/*
 		 * Two caches are used to avoid mapping duplicated pages (eg.
drivers/gpu/drm/i915/gvt/kvmgt.c +0 -0
index 6fe825763d05..61c634618217 100644

drivers/gpu/drm/i915/gvt/kvmgt.c static int intel_vgpu_get_irq_count(struct · line 1223

Quoted 2 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +0 -0
@@ -1222,6 +1222,8 @@ static int intel_vgpu_get_irq_count(struct
 intel_vgpu *vgpu, int type)  {
 	if (type == VFIO_PCI_INTX_IRQ_INDEX || type ==

VFIO_PCI_MSI_IRQ_INDEX)

drivers/gpu/drm/i915/gvt/kvmgt.c static int intel_vgpu_set_msi_trigger(struct

Quoted 7 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +2 -0
 		return 1;
+	else if (type < VFIO_PCI_NUM_IRQS + vgpu->vdev.num_irqs)
+		return vgpu->vdev.irq[type - VFIO_PCI_NUM_IRQS].count;
 
 	return 0;
 }
@@ -1269,7 +1271,62 @@ static int intel_vgpu_set_msi_trigger(struct

intel_vgpu *vgpu,

drivers/gpu/drm/i915/gvt/kvmgt.c static int intel_vgpu_set_msi_trigger(struct intel_vgpu *vgpu,

Quoted 13 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +9 -1
 	return 0;
 }
 
-static int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags,
+static int intel_vgu_set_display_irq_mask(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	if (start != 0 || count > 2)
+		return -EINVAL;
+
+	if (flags & VFIO_IRQ_SET_DATA_NONE)
+		vgpu->vdev.display_event_mask |=

DISPLAY_PRI_REFRESH_EVENT |

Quoted 4 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +4 -0
+			DISPLAY_CUR_REFRESH_EVENT;
+
+	return 0;
+}

drivers/gpu/drm/i915/gvt/kvmgt.c static int intel_vgpu_set_msi_trigger(struct intel_vgpu *vgpu,

Quoted 10 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +10 -0
+
+static int intel_vgu_set_display_irq_unmask(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	if (start != 0 || count > 2)
+		return -EINVAL;
+
+	if (flags & VFIO_IRQ_SET_DATA_NONE)
+		vgpu->vdev.display_event_mask &=

~(DISPLAY_PRI_REFRESH_EVENT |

drivers/gpu/drm/i915/gvt/kvmgt.c static int intel_vgpu_set_irqs(struct intel_vgpu

Quoted 37 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +33 -0
+			   DISPLAY_CUR_REFRESH_EVENT);
+
+	return 0;
+}
+
+static int intel_vgpu_set_display_event_trigger(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	struct eventfd_ctx *trigger;
+
+	if (flags & VFIO_IRQ_SET_DATA_EVENTFD) {
+		int fd = *(int *)data;
+
+		trigger = eventfd_ctx_fdget(fd);
+		if (IS_ERR(trigger)) {
+			gvt_vgpu_err("eventfd_ctx_fdget failed\n");
+			return PTR_ERR(trigger);
+		}
+		vgpu->vdev.vblank_trigger = trigger;
+		vgpu->vdev.display_event_mask = 0;
+	} else if ((flags & VFIO_IRQ_SET_DATA_NONE) && !count) {
+		trigger = vgpu->vdev.vblank_trigger;
+		if (trigger) {
+			eventfd_ctx_put(trigger);
+			vgpu->vdev.vblank_trigger = NULL;
+		}
+	}
+
+	return 0;
+}
+
+int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags,
 		unsigned int index, unsigned int start, unsigned int count,
 		void *data)
 {
@@ -1302,6 +1359,35 @@ static int intel_vgpu_set_irqs(struct intel_vgpu

*vgpu, u32 flags,

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device

Quoted 36 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +29 -0
 			break;
 		}
 		break;
+	default:
+	{
+		int i;
+
+		if (index >= VFIO_PCI_NUM_IRQS +
+					vgpu->vdev.num_irqs)
+			return -EINVAL;
+		index =
+			array_index_nospec(index,
+						VFIO_PCI_NUM_IRQS +
+						vgpu->vdev.num_irqs);
+
+		i = index - VFIO_PCI_NUM_IRQS;
+		if (vgpu->vdev.irq[i].type == VFIO_IRQ_TYPE_GFX &&
+		    vgpu->vdev.irq[i].subtype ==
+		    VFIO_IRQ_SUBTYPE_GFX_DISPLAY_IRQ) {
+			switch (flags & VFIO_IRQ_SET_ACTION_TYPE_MASK) {
+			case VFIO_IRQ_SET_ACTION_MASK:
+				func = intel_vgu_set_display_irq_mask;
+				break;
+			case VFIO_IRQ_SET_ACTION_UNMASK:
+				func = intel_vgu_set_display_irq_unmask;
+				break;
+			case VFIO_IRQ_SET_ACTION_TRIGGER:
+				func = intel_vgpu_set_display_event_trigger;
+				break;
+			}
+		}
+	}
 	}
 
 	if (!func)
@@ -1333,7 +1419,7 @@ static long intel_vgpu_ioctl(struct mdev_device

*mdev, unsigned int cmd,

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device

Quoted 9 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +1 -1
 		info.flags |= VFIO_DEVICE_FLAGS_RESET;
 		info.num_regions = VFIO_PCI_NUM_REGIONS +
 				vgpu->vdev.num_regions;
-		info.num_irqs = VFIO_PCI_NUM_IRQS;
+		info.num_irqs = VFIO_PCI_NUM_IRQS + vgpu->vdev.num_irqs;
 
 		return copy_to_user((void __user *)arg, &info, minsz) ?
 			-EFAULT : 0;
@@ -1493,32 +1579,81 @@ static long intel_vgpu_ioctl(struct mdev_device

*mdev, unsigned int cmd,

Quoted 69 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +40 -6
 			-EFAULT : 0;
 	} else if (cmd == VFIO_DEVICE_GET_IRQ_INFO) {
 		struct vfio_irq_info info;
+		struct vfio_info_cap caps = { .buf = NULL, .size = 0 };
+		unsigned int i;
+		int ret;
 
 		minsz = offsetofend(struct vfio_irq_info, count);
 
 		if (copy_from_user(&info, (void __user *)arg, minsz))
 			return -EFAULT;
 
-		if (info.argsz < minsz || info.index >= VFIO_PCI_NUM_IRQS)
+		if (info.argsz < minsz)
 			return -EINVAL;
 
 		switch (info.index) {
 		case VFIO_PCI_INTX_IRQ_INDEX:
 		case VFIO_PCI_MSI_IRQ_INDEX:
+			info.flags = VFIO_IRQ_INFO_EVENTFD;
 			break;
-		default:
+		case VFIO_PCI_MSIX_IRQ_INDEX:
+		case VFIO_PCI_ERR_IRQ_INDEX:
+		case VFIO_PCI_REQ_IRQ_INDEX:
 			return -EINVAL;
-		}
+		default:
+		{
+			struct vfio_irq_info_cap_type cap_type = {
+				.header.id = VFIO_IRQ_INFO_CAP_TYPE,
+				.header.version = 1 };
 
-		info.flags = VFIO_IRQ_INFO_EVENTFD;
+			if (info.index >= VFIO_PCI_NUM_IRQS +
+					vgpu->vdev.num_irqs)
+				return -EINVAL;
+			info.index =
+				array_index_nospec(info.index,
+						VFIO_PCI_NUM_IRQS +
+						vgpu->vdev.num_irqs);
+
+			i = info.index - VFIO_PCI_NUM_IRQS;
+
+			info.flags = vgpu->vdev.irq[i].flags;
+			cap_type.type = vgpu->vdev.irq[i].type;
+			cap_type.subtype = vgpu->vdev.irq[i].subtype;
+
+			ret = vfio_info_add_capability(&caps,
+						&cap_type.header,
+						sizeof(cap_type));
+			if (ret)
+				return ret;
+		}
+		}
 
 		info.count = intel_vgpu_get_irq_count(vgpu, info.index);
 
 		if (info.index == VFIO_PCI_INTX_IRQ_INDEX)
 			info.flags |= (VFIO_IRQ_INFO_MASKABLE |
 				       VFIO_IRQ_INFO_AUTOMASKED);
-		else
-			info.flags |= VFIO_IRQ_INFO_NORESIZE;
+
+		if (caps.size) {
+			info.flags |= VFIO_IRQ_INFO_FLAG_CAPS;
+			if (info.argsz < sizeof(info) + caps.size) {
+				info.argsz = sizeof(info) + caps.size;
+				info.cap_offset = 0;

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd,

Quoted 10 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +10 -0
+			} else {
+				vfio_info_cap_shift(&caps, sizeof(info));
+				if (copy_to_user((void __user *)arg +
+						  sizeof(info), caps.buf,
+						  caps.size)) {
+					kfree(caps.buf);
+					return -EFAULT;
+				}
+				info.cap_offset = sizeof(info);
+				if (offsetofend(struct vfio_irq_info, cap_offset) >

minsz)

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device · line 1652

Quoted 1 line
drivers/gpu/drm/i915/gvt/kvmgt.c +1 -0
+					minsz = offsetofend(struct

vfio_irq_info, cap_offset);

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device

Quoted 8 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +4 -0
+			}
+
+			kfree(caps.buf);
+		}
 
 		return copy_to_user((void __user *)arg, &info, minsz) ?
 			-EFAULT : 0;
@@ -1537,7 +1672,8 @@ static long intel_vgpu_ioctl(struct mdev_device

*mdev, unsigned int cmd,

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd,

Quoted 4 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +0 -1
 			int max = intel_vgpu_get_irq_count(vgpu, hdr.index);
 
 			ret = vfio_set_irqs_validate_and_prepare(&hdr, max,
-						VFIO_PCI_NUM_IRQS,

&data_size);

drivers/gpu/drm/i915/gvt/kvmgt.c static long intel_vgpu_ioctl(struct mdev_device · line 1679

Quoted 5 lines
drivers/gpu/drm/i915/gvt/kvmgt.c +2 -0
+					VFIO_PCI_NUM_IRQS + vgpu-
 vdev.num_irqs,
+								 &data_size);
 			if (ret) {
 
	gvt_vgpu_err("intel:vfio_set_irqs_validate_and_prepare failed\n");
Quoted 4 lines

return -EINVAL;
--
2.17.1

--
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

author
#18 lore raw

RFC PATCH v4 5/6 drm/i915/gvt: Deliver async primary plane page flip events at vblank

In reply to Kechen Lu

Only sync primary plane page flip events are checked and delivered as the display refresh events before, this patch tries to deliver async primary page flip events bounded by vblanks.

To deliver correct async page flip, the new async flip bitmap is introduced and in vblank emulation handler to check bitset.

---
 drivers/gpu/drm/i915/gvt/cmd_parser.c |  5 +++--
 drivers/gpu/drm/i915/gvt/display.c    | 12 ++++++++++++
 drivers/gpu/drm/i915/gvt/gvt.h        |  2 ++
 drivers/gpu/drm/i915/gvt/handlers.c   |  5 +++--
 4 files changed, 20 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/gvt/cmd_parser.c +3 -2
index 5cb59c0b4bbe..e504cc7be559 100644
@@ -1334,9 +1334,10 @@ static int gen8_update_plane_mmio_from_mi_display_flip(
 	if (info->plane == PLANE_PRIMARY)
 		vgpu_vreg_t(vgpu, PIPE_FLIPCOUNT_G4X(info->pipe))++;
 
-	if (info->async_flip)
+	if (info->async_flip) {
 		intel_vgpu_trigger_virtual_event(vgpu, info->event);
-	else
+		set_bit(info->plane, vgpu->display.async_flip_event[info->pipe]);
+	} else
 		set_bit(info->event, vgpu->irq.flip_done_event[info->pipe]);
 
 	return 0;
drivers/gpu/drm/i915/gvt/display.c +12 -0
index 036db8199983..df52e4b4c1b0 100644
@@ -420,6 +420,18 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
+	for_each_set_bit(event, vgpu->display.async_flip_event[pipe],
+			I915_MAX_PLANES) {
+		clear_bit(event, vgpu->display.async_flip_event[pipe]);
+		if (!pipe_is_enabled(vgpu, pipe))
+			continue;
+
+		if (event == PLANE_PRIMARY) {
+			eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+	}
+
 	if (--pageflip_count < 0) {
 		eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
 		pageflip_count = 0;
drivers/gpu/drm/i915/gvt/gvt.h +2 -0
index b654b6fa0663..98a3dc309acb 100644
@@ -128,6 +128,8 @@ struct intel_vgpu_display {
 	struct intel_vgpu_i2c_edid i2c_edid;
 	struct intel_vgpu_port ports[I915_MAX_PORTS];
 	struct intel_vgpu_sbi sbi;
+	DECLARE_BITMAP(async_flip_event[I915_MAX_PIPES],
+		       I915_MAX_PLANES);
 };
 
 struct vgpu_sched_ctl {
drivers/gpu/drm/i915/gvt/handlers.c +3 -2
index e09bd6e0cc4d..6ad29c4f08e5 100644
@@ -758,9 +758,10 @@ static int pri_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
 
 	vgpu_vreg_t(vgpu, PIPE_FLIPCOUNT_G4X(pipe))++;
 
-	if (vgpu_vreg_t(vgpu, DSPCNTR(pipe)) & PLANE_CTL_ASYNC_FLIP)
+	if (vgpu_vreg_t(vgpu, DSPCNTR(pipe)) & PLANE_CTL_ASYNC_FLIP) {
 		intel_vgpu_trigger_virtual_event(vgpu, event);
-	else
+		set_bit(PLANE_PRIMARY, vgpu->display.async_flip_event[pipe]);
+	} else
 		set_bit(event, vgpu->irq.flip_done_event[pipe]);
 
 	return 0;
Signature · 1 line
2.17.1
author
#19 lore raw

RFC PATCH v4 6/6 drm/i915/gvt: Add cursor plane reg update trap emulation handler

In reply to Kechen Lu

This patch adds the cursor plane CURBASE reg update trap handler in order to :

- Deliver the cursor refresh event at each vblank emulation, the flip_done_event bit check is supposed to do here. If cursor plane updates happen, deliver the cursor refresh events.

- Support the sync and async cursor plane updates and corresponding cursor plane flip interrupts reporting.

---
 drivers/gpu/drm/i915/gvt/display.c   | 11 +++++++++++
 drivers/gpu/drm/i915/gvt/handlers.c  | 27 ++++++++++++++++++++++++---
 drivers/gpu/drm/i915/gvt/interrupt.c |  7 +++++++
 drivers/gpu/drm/i915/gvt/interrupt.h |  3 +++
 4 files changed, 45 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +11 -0
index df52e4b4c1b0..a0accc51d44f 100644
@@ -399,6 +399,7 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		[PIPE_C] = PIPE_C_VBLANK,
 	};
 	int pri_flip_event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
+	int cur_flip_event = CURSOR_A_FLIP_DONE + pipe;
 	int event;
 	u64 eventfd_signal_val = 0;
 	static int pageflip_count;
@@ -417,6 +418,11 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 			pageflip_count += PAGEFLIP_INC_COUNT;
 		}
 
+		if (event == cur_flip_event) {
+			eventfd_signal_val += DISPLAY_CUR_REFRESH_EVENT_INC;
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
@@ -430,6 +436,11 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 			eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
 			pageflip_count += PAGEFLIP_INC_COUNT;
 		}
+
+		if (event == PLANE_CURSOR) {
+			eventfd_signal_val += DISPLAY_CUR_REFRESH_EVENT_INC;
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
 	}
 
 	if (--pageflip_count < 0) {
drivers/gpu/drm/i915/gvt/handlers.c +24 -3
index 6ad29c4f08e5..821ff88977d8 100644
@@ -767,6 +767,27 @@ static int pri_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
 	return 0;
 }
 
+#define CURBASE_TO_PIPE(reg) \
+	calc_index(offset, _CURABASE, _CURBBASE, 0, CURBASE(PIPE_C))
+
+static int cur_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
+	u32 pipe = CURBASE_TO_PIPE(offset);
+	int event = CURSOR_A_FLIP_DONE + pipe;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg_t(vgpu, CURCNTR(pipe)) & PLANE_CTL_ASYNC_FLIP) {
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+		set_bit(PLANE_CURSOR, vgpu->display.async_flip_event[pipe]);
+	} else
+		set_bit(event, vgpu->irq.flip_done_event[pipe]);
+
+	return 0;
+}
+
 #define SPRSURF_TO_PIPE(offset) \
 	calc_index(offset, _SPRA_SURF, _SPRB_SURF, 0, SPRSURF(PIPE_C))
 
@@ -1955,9 +1976,9 @@ static int init_generic_mmio_info(struct intel_gvt *gvt)
 	MMIO_D(CURPOS(PIPE_B), D_ALL);
 	MMIO_D(CURPOS(PIPE_C), D_ALL);
 
-	MMIO_D(CURBASE(PIPE_A), D_ALL);
-	MMIO_D(CURBASE(PIPE_B), D_ALL);
-	MMIO_D(CURBASE(PIPE_C), D_ALL);
+	MMIO_DH(CURBASE(PIPE_A), D_ALL, NULL, cur_surf_mmio_write);
+	MMIO_DH(CURBASE(PIPE_B), D_ALL, NULL, cur_surf_mmio_write);
+	MMIO_DH(CURBASE(PIPE_C), D_ALL, NULL, cur_surf_mmio_write);
 
 	MMIO_D(CUR_FBC_CTL(PIPE_A), D_ALL);
 	MMIO_D(CUR_FBC_CTL(PIPE_B), D_ALL);
drivers/gpu/drm/i915/gvt/interrupt.c +7 -0
index 951681813230..9c2b9d2e1529 100644
@@ -113,6 +113,9 @@ static const char * const irq_name[INTEL_GVT_EVENT_MAX] = {
 	[SPRITE_A_FLIP_DONE] = "Sprite Plane A flip done",
 	[SPRITE_B_FLIP_DONE] = "Sprite Plane B flip done",
 	[SPRITE_C_FLIP_DONE] = "Sprite Plane C flip done",
+	[CURSOR_A_FLIP_DONE] = "Cursor Plane A flip done",
+	[CURSOR_B_FLIP_DONE] = "Cursor Plane B flip done",
+	[CURSOR_C_FLIP_DONE] = "Cursor Plane C flip done",
 
 	[PCU_THERMAL] = "PCU Thermal Event",
 	[PCU_PCODE2DRIVER_MAILBOX] = "PCU pcode2driver mailbox event",
@@ -593,6 +596,10 @@ static void gen8_init_irq(
 		SET_BIT_INFO(irq, 4, SPRITE_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
 		SET_BIT_INFO(irq, 4, SPRITE_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
 		SET_BIT_INFO(irq, 4, SPRITE_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+
+		SET_BIT_INFO(irq, 6, CURSOR_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+		SET_BIT_INFO(irq, 6, CURSOR_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+		SET_BIT_INFO(irq, 6, CURSOR_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
 	}
 
 	/* GEN8 interrupt PCU events */
drivers/gpu/drm/i915/gvt/interrupt.h +3 -0
index 5313fb1b33e1..158f1c7a23f2 100644
@@ -92,6 +92,9 @@ enum intel_gvt_event_type {
 	SPRITE_A_FLIP_DONE,
 	SPRITE_B_FLIP_DONE,
 	SPRITE_C_FLIP_DONE,
+	CURSOR_A_FLIP_DONE,
+	CURSOR_B_FLIP_DONE,
+	CURSOR_C_FLIP_DONE,
 
 	PCU_THERMAL,
 	PCU_PCODE2DRIVER_MAILBOX,
Signature · 1 line
2.17.1
#20 lore raw

Re: [RFC PATCH v4 6/6] drm/i915/gvt: Add cursor plane reg update trap emulation handler

Quoted 87 lines · Kechen Lu

This patch adds the cursor plane CURBASE reg update trap handler in order to :

- Deliver the cursor refresh event at each vblank emulation, the flip_done_event bit check is supposed to do here. If cursor plane updates happen, deliver the cursor refresh events.

- Support the sync and async cursor plane updates and corresponding cursor plane flip interrupts reporting.

---
 drivers/gpu/drm/i915/gvt/display.c   | 11 +++++++++++
 drivers/gpu/drm/i915/gvt/handlers.c  | 27 ++++++++++++++++++++++++---
 drivers/gpu/drm/i915/gvt/interrupt.c |  7 +++++++
 drivers/gpu/drm/i915/gvt/interrupt.h |  3 +++
 4 files changed, 45 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +11 -0
index df52e4b4c1b0..a0accc51d44f 100644
@@ -399,6 +399,7 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 		[PIPE_C] = PIPE_C_VBLANK,
 	};
 	int pri_flip_event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
+	int cur_flip_event = CURSOR_A_FLIP_DONE + pipe;
 	int event;
 	u64 eventfd_signal_val = 0;
 	static int pageflip_count;
@@ -417,6 +418,11 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 			pageflip_count += PAGEFLIP_INC_COUNT;
 		}
 
+		if (event == cur_flip_event) {
+			eventfd_signal_val += DISPLAY_CUR_REFRESH_EVENT_INC;
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
@@ -430,6 +436,11 @@ static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
 			eventfd_signal_val += DISPLAY_PRI_REFRESH_EVENT_INC;
 			pageflip_count += PAGEFLIP_INC_COUNT;
 		}
+
+		if (event == PLANE_CURSOR) {
+			eventfd_signal_val += DISPLAY_CUR_REFRESH_EVENT_INC;
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
 	}
 
 	if (--pageflip_count < 0) {
drivers/gpu/drm/i915/gvt/handlers.c +21 -0
index 6ad29c4f08e5..821ff88977d8 100644
@@ -767,6 +767,27 @@ static int pri_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
 	return 0;
 }
 
+#define CURBASE_TO_PIPE(reg) \
+	calc_index(offset, _CURABASE, _CURBBASE, 0, CURBASE(PIPE_C))
+
+static int cur_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
+	u32 pipe = CURBASE_TO_PIPE(offset);
+	int event = CURSOR_A_FLIP_DONE + pipe;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg_t(vgpu, CURCNTR(pipe)) & PLANE_CTL_ASYNC_FLIP) {
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+		set_bit(PLANE_CURSOR, vgpu->display.async_flip_event[pipe]);
+	} else
+		set_bit(event, vgpu->irq.flip_done_event[pipe]);
+
+	return 0;
+}
+
 #define SPRSURF_TO_PIPE(offset) \
 	calc_index(offset, _SPRA_SURF, _SPRB_SURF, 0, SPRSURF(PIPE_C))
 

drivers/gpu/drm/i915/gvt/handlers.c static int init_generic_mmio_info(struct intel_gvt *gvt) · line 1981

Quoted 9 lines
drivers/gpu/drm/i915/gvt/handlers.c +3 -3
@@ -1955,9 +1976,9 @@ static int init_generic_mmio_info(struct intel_gvt *gvt)
 	MMIO_D(CURPOS(PIPE_B), D_ALL);
 	MMIO_D(CURPOS(PIPE_C), D_ALL);
 
-	MMIO_D(CURBASE(PIPE_A), D_ALL);
-	MMIO_D(CURBASE(PIPE_B), D_ALL);
-	MMIO_D(CURBASE(PIPE_C), D_ALL);
+	MMIO_DH(CURBASE(PIPE_A), D_ALL, NULL, cur_surf_mmio_write);
+	MMIO_DH(CURBASE(PIPE_B), D_ALL, NULL, cur_surf_mmio_write);
+	MMIO_DH(CURBASE(PIPE_C), D_ALL, NULL, cur_surf_mmio_write);

I think we should also track cursor pos change right?

Quoted 45 lines
drivers/gpu/drm/i915/gvt/handlers.c +0 -0
 	MMIO_D(CUR_FBC_CTL(PIPE_A), D_ALL);
 	MMIO_D(CUR_FBC_CTL(PIPE_B), D_ALL);
drivers/gpu/drm/i915/gvt/interrupt.c +7 -0
index 951681813230..9c2b9d2e1529 100644
@@ -113,6 +113,9 @@ static const char * const irq_name[INTEL_GVT_EVENT_MAX] = {
 	[SPRITE_A_FLIP_DONE] = "Sprite Plane A flip done",
 	[SPRITE_B_FLIP_DONE] = "Sprite Plane B flip done",
 	[SPRITE_C_FLIP_DONE] = "Sprite Plane C flip done",
+	[CURSOR_A_FLIP_DONE] = "Cursor Plane A flip done",
+	[CURSOR_B_FLIP_DONE] = "Cursor Plane B flip done",
+	[CURSOR_C_FLIP_DONE] = "Cursor Plane C flip done",
 
 	[PCU_THERMAL] = "PCU Thermal Event",
 	[PCU_PCODE2DRIVER_MAILBOX] = "PCU pcode2driver mailbox event",
@@ -593,6 +596,10 @@ static void gen8_init_irq(
 		SET_BIT_INFO(irq, 4, SPRITE_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
 		SET_BIT_INFO(irq, 4, SPRITE_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
 		SET_BIT_INFO(irq, 4, SPRITE_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+
+		SET_BIT_INFO(irq, 6, CURSOR_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+		SET_BIT_INFO(irq, 6, CURSOR_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+		SET_BIT_INFO(irq, 6, CURSOR_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
 	}
 
 	/* GEN8 interrupt PCU events */
drivers/gpu/drm/i915/gvt/interrupt.h +3 -0
index 5313fb1b33e1..158f1c7a23f2 100644
@@ -92,6 +92,9 @@ enum intel_gvt_event_type {
 	SPRITE_A_FLIP_DONE,
 	SPRITE_B_FLIP_DONE,
 	SPRITE_C_FLIP_DONE,
+	CURSOR_A_FLIP_DONE,
+	CURSOR_B_FLIP_DONE,
+	CURSOR_C_FLIP_DONE,
 
 	PCU_THERMAL,
 	PCU_PCODE2DRIVER_MAILBOX,
Signature · 1 line
2.17.1
Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
author
#21 lore raw

RE: [RFC PATCH v4 6/6] drm/i915/gvt: Add cursor plane reg update trap emulation handler

Hi,

Quoted 115 lines

-----Original Message-----
From: Zhenyu Wang [mailto:[email protected]]
Sent: Friday, July 19, 2019 2:35 PM
To: Lu, Kechen <[email protected]>
Cc: [email protected]; [email protected]; linux- [email protected]; [email protected]; [email protected]; Lv, Zhiyuan <[email protected]>; Wang, Zhi A <[email protected]>; Tian, Kevin <[email protected]>; Yuan, Hang <[email protected]>; [email protected]
Subject: Re: [RFC PATCH v4 6/6] drm/i915/gvt: Add cursor plane reg update trap emulation handler

Quoted 24 lines · Kechen Lu

This patch adds the cursor plane CURBASE reg update trap handler in order to :

- Deliver the cursor refresh event at each vblank emulation, the flip_done_event bit check is supposed to do here. If cursor plane updates happen, deliver the cursor refresh events.

- Support the sync and async cursor plane updates and corresponding cursor plane flip interrupts reporting.

---
 drivers/gpu/drm/i915/gvt/display.c   | 11 +++++++++++
 drivers/gpu/drm/i915/gvt/handlers.c  | 27 
++++++++++++++++++++++++--- drivers/gpu/drm/i915/gvt/interrupt.c |  
7 +++++++ drivers/gpu/drm/i915/gvt/interrupt.h |  3 +++
 4 files changed, 45 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/gvt/display.c +0 -0
index df52e4b4c1b0..a0accc51d44f 100644
@@ -399,6 +399,7 @@ static void emulate_vblank_on_pipe(struct 
Quoted 1 line

intel_vgpu

*vgpu, int pipe)

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct

Quoted 8 lines
drivers/gpu/drm/i915/gvt/display.c +1 -0
 		[PIPE_C] = PIPE_C_VBLANK,
 	};
 	int pri_flip_event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
+	int cur_flip_event = CURSOR_A_FLIP_DONE + pipe;
 	int event;
 	u64 eventfd_signal_val = 0;
 	static int pageflip_count;
@@ -417,6 +418,11 @@ static void emulate_vblank_on_pipe(struct

intel_vgpu *vgpu, int pipe)

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)

Quoted 5 lines
drivers/gpu/drm/i915/gvt/display.c +2 -0
 			pageflip_count += PAGEFLIP_INC_COUNT;
 		}
 
+		if (event == cur_flip_event) {
+			eventfd_signal_val +=

DISPLAY_CUR_REFRESH_EVENT_INC;

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct

Quoted 7 lines
drivers/gpu/drm/i915/gvt/display.c +3 -0
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
+
 		intel_vgpu_trigger_virtual_event(vgpu, event);
 	}
 
@@ -430,6 +436,11 @@ static void emulate_vblank_on_pipe(struct

intel_vgpu *vgpu, int pipe)

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct · line 436

Quoted 1 line
drivers/gpu/drm/i915/gvt/display.c +0 -0
 			eventfd_signal_val +=

DISPLAY_PRI_REFRESH_EVENT_INC;

drivers/gpu/drm/i915/gvt/display.c static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)

Quoted 5 lines
drivers/gpu/drm/i915/gvt/display.c +3 -0
 			pageflip_count += PAGEFLIP_INC_COUNT;
 		}
+
+		if (event == PLANE_CURSOR) {
+			eventfd_signal_val +=

DISPLAY_CUR_REFRESH_EVENT_INC;

Quoted 11 lines
drivers/gpu/drm/i915/gvt/display.c +2 -0
+			pageflip_count += PAGEFLIP_INC_COUNT;
+		}
 	}
 
 	if (--pageflip_count < 0) {
drivers/gpu/drm/i915/gvt/handlers.c +0 -0
index 6ad29c4f08e5..821ff88977d8 100644
@@ -767,6 +767,27 @@ static int pri_surf_mmio_write(struct 
Quoted 1 line

intel_vgpu

*vgpu, unsigned int offset,

Quoted 27 lines
drivers/gpu/drm/i915/gvt/handlers.c +20 -0
 	return 0;
 }
 
+#define CURBASE_TO_PIPE(reg) \
+	calc_index(offset, _CURABASE, _CURBBASE, 0, CURBASE(PIPE_C))
+
+static int cur_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes) {
+	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
+	u32 pipe = CURBASE_TO_PIPE(offset);
+	int event = CURSOR_A_FLIP_DONE + pipe;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg_t(vgpu, CURCNTR(pipe)) & PLANE_CTL_ASYNC_FLIP) {
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+		set_bit(PLANE_CURSOR, vgpu->display.async_flip_event[pipe]);
+	} else
+		set_bit(event, vgpu->irq.flip_done_event[pipe]);
+
+	return 0;
+}
+
 #define SPRSURF_TO_PIPE(offset) \
 	calc_index(offset, _SPRA_SURF, _SPRB_SURF, 0, SPRSURF(PIPE_C))
@@ -1955,9 +1976,9 @@ static int init_generic_mmio_info(struct 
Quoted 1 line

intel_gvt

*gvt)

Quoted 3 lines
drivers/gpu/drm/i915/gvt/handlers.c +0 -0
 	MMIO_D(CURPOS(PIPE_B), D_ALL);
 	MMIO_D(CURPOS(PIPE_C), D_ALL);
Quoted 9 lines

drivers/gpu/drm/i915/gvt/handlers.c static int init_generic_mmio_info(struct intel_gvt *gvt) · line 1981

Quoted 6 lines
drivers/gpu/drm/i915/gvt/handlers.c +3 -3
-	MMIO_D(CURBASE(PIPE_A), D_ALL);
-	MMIO_D(CURBASE(PIPE_B), D_ALL);
-	MMIO_D(CURBASE(PIPE_C), D_ALL);
+	MMIO_DH(CURBASE(PIPE_A), D_ALL, NULL, cur_surf_mmio_write);
+	MMIO_DH(CURBASE(PIPE_B), D_ALL, NULL, cur_surf_mmio_write);
+	MMIO_DH(CURBASE(PIPE_C), D_ALL, NULL, cur_surf_mmio_write);

I think we should also track cursor pos change right?

Yes, this may be the case, according to the bspec, the CURPOS regs take into effects at vblanks. But in Linux i915 driver, the CURPOS regs programming companies with CURBASE regs. I'm not sure the Windows driver logic, but I can do some experiments. If that is the case, yes, we also need to trap the cursor pos change. Thanks.

Best Regards,
Kechen

Quoted 60 lines

drivers/gpu/drm/i915/gvt/interrupt.c static const char * const

Quoted 9 lines
drivers/gpu/drm/i915/gvt/handlers.c +0 -0
 	MMIO_D(CUR_FBC_CTL(PIPE_A), D_ALL);
 	MMIO_D(CUR_FBC_CTL(PIPE_B), D_ALL); diff --git 

a/drivers/gpu/drm/i915/gvt/interrupt.c
b/drivers/gpu/drm/i915/gvt/interrupt.c
index 951681813230..9c2b9d2e1529 100644

drivers/gpu/drm/i915/gvt/interrupt.c +0 -0
@@ -113,6 +113,9 @@ static const char * const

irq_name[INTEL_GVT_EVENT_MAX] = {

drivers/gpu/drm/i915/gvt/interrupt.c static const char * const irq_name[INTEL_GVT_EVENT_MAX] = { · line 121

Quoted 9 lines
drivers/gpu/drm/i915/gvt/interrupt.c +3 -0
 	[SPRITE_A_FLIP_DONE] = "Sprite Plane A flip done",
 	[SPRITE_B_FLIP_DONE] = "Sprite Plane B flip done",
 	[SPRITE_C_FLIP_DONE] = "Sprite Plane C flip done",
+	[CURSOR_A_FLIP_DONE] = "Cursor Plane A flip done",
+	[CURSOR_B_FLIP_DONE] = "Cursor Plane B flip done",
+	[CURSOR_C_FLIP_DONE] = "Cursor Plane C flip done",
 
 	[PCU_THERMAL] = "PCU Thermal Event",
 	[PCU_PCODE2DRIVER_MAILBOX] = "PCU pcode2driver mailbox event",

@@

Quoted 2 lines

-593,6 +596,10 @@ static void gen8_init_irq(
SET_BIT_INFO(irq, 4, SPRITE_A_FLIP_DONE,

INTEL_GVT_IRQ_INFO_DE_PIPE_A);

Quoted 1 line

SET_BIT_INFO(irq, 4, SPRITE_B_FLIP_DONE,

INTEL_GVT_IRQ_INFO_DE_PIPE_B);

drivers/gpu/drm/i915/gvt/interrupt.c

Quoted 4 lines

SET_BIT_INFO(irq, 4, SPRITE_C_FLIP_DONE,
INTEL_GVT_IRQ_INFO_DE_PIPE_C);

drivers/gpu/drm/i915/gvt/interrupt.c +2 -0
+
+		SET_BIT_INFO(irq, 6, CURSOR_A_FLIP_DONE,

INTEL_GVT_IRQ_INFO_DE_PIPE_A);

drivers/gpu/drm/i915/gvt/interrupt.c

Quoted 1 line
drivers/gpu/drm/i915/gvt/interrupt.c +1 -0
+		SET_BIT_INFO(irq, 6, CURSOR_B_FLIP_DONE,

INTEL_GVT_IRQ_INFO_DE_PIPE_B);

Quoted 21 lines
drivers/gpu/drm/i915/gvt/interrupt.c +2 -0
+		SET_BIT_INFO(irq, 6, CURSOR_C_FLIP_DONE, 
+INTEL_GVT_IRQ_INFO_DE_PIPE_C);
 	}
 
 	/* GEN8 interrupt PCU events */
drivers/gpu/drm/i915/gvt/interrupt.h +3 -1
index 5313fb1b33e1..158f1c7a23f2 100644
@@ -92,6 +92,9 @@ enum intel_gvt_event_type {
 	SPRITE_A_FLIP_DONE,
 	SPRITE_B_FLIP_DONE,
 	SPRITE_C_FLIP_DONE,
+	CURSOR_A_FLIP_DONE,
+	CURSOR_B_FLIP_DONE,
+	CURSOR_C_FLIP_DONE,
 
 	PCU_THERMAL,
 	PCU_PCODE2DRIVER_MAILBOX,
--
Quoted 2 lines

2.17.1

--
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827