Skip to contentSkip to conversation

uio_hv_generic: Enable interrupt for low speed VMBus devices

PATCH In review Saurabh Sengar 3 participants 8 messages Oct 13, 2022 – Oct 18, 2022

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 uio_hv_generic: Enable interrupt for low speed VMBus devices

Hyper-V is adding some "specialty" synthetic devices. Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can handle the device. Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

---
 drivers/uio/uio_hv_generic.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
drivers/uio/uio_hv_generic.c +3 -6
index c08a6cfd119f..8e5aa4a1247f 100644
@@ -84,6 +84,9 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state)
 	dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state;
 	virt_mb();
 
+	if (!dev->channel->offermsg.monitor_allocated && irq_state)
+		vmbus_setevent(dev->channel);
+
 	return 0;
 }
 
@@ -239,12 +242,6 @@ hv_uio_probe(struct hv_device *dev,
 	void *ring_buffer;
 	int ret;
 
-	/* Communicating with host has to be via shared memory not hypercall */
-	if (!channel->offermsg.monitor_allocated) {
-		dev_err(&dev->device, "vmbus channel requires hypercall\n");
-		return -ENOTSUPP;
-	}
-
 	pdata = devm_kzalloc(&dev->device, sizeof(*pdata), GFP_KERNEL);
 	if (!pdata)
 		return -ENOMEM;
Signature · 1 line
2.34.1
maintainer
#1 lore raw
Quoted 1 line · Saurabh Sengar

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 6 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can handle the device. Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device types.

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not a good idea.

Also, if you do do this, you need to list where the source for that userspace code is so that users can get it and have their distros package it up for them. I do not see that here at all.

Quoted 21 lines
---
 drivers/uio/uio_hv_generic.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
drivers/uio/uio_hv_generic.c +3 -0
index c08a6cfd119f..8e5aa4a1247f 100644
@@ -84,6 +84,9 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state)
 	dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state;
 	virt_mb();
 
+	if (!dev->channel->offermsg.monitor_allocated && irq_state)
+		vmbus_setevent(dev->channel);
+
 	return 0;
 }
 

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 244

Quoted 6 lines
drivers/uio/uio_hv_generic.c +0 -3
@@ -239,12 +242,6 @@ hv_uio_probe(struct hv_device *dev,
 	void *ring_buffer;
 	int ret;
 
-	/* Communicating with host has to be via shared memory not hypercall */
-	if (!channel->offermsg.monitor_allocated) {
-		dev_err(&dev->device, "vmbus channel requires hypercall\n");

I do not understand, why is this check not made anymore here? Why constantly make the call above in the irq handler instead? Isn't that going to be massively slow?

thanks,

greg k-h

#2 lore raw
Quoted 15 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 1 line · Saurabh Sengar

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 6 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can handle the device. Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

Quoted 5 lines

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device types.

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not a good idea.

There are some VMBUS drivers currently not implemented in Linux. Out of all VMBUS drivers, two use "monitored bits": they are network and storage drivers. All the rest VMBUS drivers use hypercall for host notification and signal for next interrupt. One example of such driver is to collect process level crash information for diagnostic purposes.

Also, we want to move some existing kernel mode VMBUS drivers to user-space, such as hv_kvp and hv_filecopy. They don't really fit into an existing kernel API, and they create their own devices under /dev and communicates with a user-mode daemon to do most of the work. It's a better model that we can move those drivers entirely into user-mode.

Quoted 32 lines

Also, if you do do this, you need to list where the source for that userspace code is so that users can get it and have their distros package it up for them. I do not see that here at all.

Quoted 21 lines
---
 drivers/uio/uio_hv_generic.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
(quoted diff) +0 -0

b/drivers/uio/uio_hv_generic.c index c08a6cfd119f..8e5aa4a1247f 100644

drivers/uio/uio_hv_generic.c +3 -0
@@ -84,6 +84,9 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state)
 	dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state;
 	virt_mb();
 
+	if (!dev->channel->offermsg.monitor_allocated && irq_state)
+		vmbus_setevent(dev->channel);
+
 	return 0;
 }
 

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 242

Quoted 4 lines
drivers/uio/uio_hv_generic.c +0 -1
@@ -239,12 +242,6 @@ hv_uio_probe(struct hv_device *dev,
 	void *ring_buffer;
 	int ret;
 
-	/* Communicating with host has to be via shared memory not

hypercall */

Quoted 7 lines

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 244

Quoted 2 lines
drivers/uio/uio_hv_generic.c +0 -2
-	if (!channel->offermsg.monitor_allocated) {
-		dev_err(&dev->device, "vmbus channel requires

hypercall\n");

I do not understand, why is this check not made anymore here? Why constantly make the call above in the irq handler instead? Isn't that going to be massively slow?

Some VMBUS devices exposed by the Hyper-V are not modeled as high speed, they use hypercall, not monitored bits. Because they don't fit into other kernel API (as explained above), can we use UIO for those devices?

Thanks,
Long

maintainer
#3 lore raw
Quoted 27 lines · Long Li
Quoted 15 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 1 line · Saurabh Sengar

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 6 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can handle the device. Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

Quoted 5 lines

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device types.

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not a good idea.

There are some VMBUS drivers currently not implemented in Linux. Out of all VMBUS drivers, two use "monitored bits": they are network and storage drivers. All the rest VMBUS drivers use hypercall for host notification and signal for next interrupt. One example of such driver is to collect process level crash information for diagnostic purposes.

Quoted 5 lines

Also, we want to move some existing kernel mode VMBUS drivers to user-space, such as hv_kvp and hv_filecopy. They don't really fit into an existing kernel API, and they create their own devices under /dev and communicates with a user-mode daemon to do most of the work. It's a better model that we can move those drivers entirely into user-mode.

How are you going to be able to remove drivers that export an existing user/kernel api and not break current systems?

Quoted 35 lines
Quoted 35 lines

Also, if you do do this, you need to list where the source for that userspace code is so that users can get it and have their distros package it up for them. I do not see that here at all.

Quoted 21 lines
---
 drivers/uio/uio_hv_generic.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
(quoted diff) +0 -0

b/drivers/uio/uio_hv_generic.c index c08a6cfd119f..8e5aa4a1247f 100644

drivers/uio/uio_hv_generic.c +3 -0
@@ -84,6 +84,9 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state)
 	dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state;
 	virt_mb();
 
+	if (!dev->channel->offermsg.monitor_allocated && irq_state)
+		vmbus_setevent(dev->channel);
+
 	return 0;
 }
 

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 242

Quoted 4 lines
drivers/uio/uio_hv_generic.c +0 -1
@@ -239,12 +242,6 @@ hv_uio_probe(struct hv_device *dev,
 	void *ring_buffer;
 	int ret;
 
-	/* Communicating with host has to be via shared memory not

hypercall */

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 244

Quoted 2 lines
drivers/uio/uio_hv_generic.c +0 -2
-	if (!channel->offermsg.monitor_allocated) {
-		dev_err(&dev->device, "vmbus channel requires

hypercall\n");

Quoted 7 lines
Quoted 3 lines

I do not understand, why is this check not made anymore here? Why constantly make the call above in the irq handler instead? Isn't that going to be massively slow?

Some VMBUS devices exposed by the Hyper-V are not modeled as high speed, they use hypercall, not monitored bits. Because they don't fit into other kernel API (as explained above), can we use UIO for those devices?

UIO is for mmaped memory regions, like PCI devices, how is this a valid Hyper-V api at all?

confused,

greg k-h

#4 lore raw
Quoted 35 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 11 lines · Long Li
Quoted 11 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 1 line · Saurabh Sengar

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 3 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can

handle the device.

Quoted 6 lines
Quoted 6 lines
Quoted 3 lines

Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device

types.

Quoted 3 lines
Quoted 3 lines

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not

a good idea.

Quoted 3 lines

There are some VMBUS drivers currently not implemented in Linux. Out of all VMBUS drivers, two use "monitored bits": they are network and

storage drivers.

Quoted 4 lines

All the rest VMBUS drivers use hypercall for host notification and signal for next interrupt. One example of such driver is to collect process level crash information for diagnostic purposes.

Quoted 8 lines
Quoted 5 lines

Also, we want to move some existing kernel mode VMBUS drivers to user-space, such as hv_kvp and hv_filecopy. They don't really fit into an existing kernel API, and they create their own devices under /dev and communicates with a user-mode daemon to do most of the work. It's a better model that we can move those drivers entirely into user-mode.

How are you going to be able to remove drivers that export an existing user/kernel api and not break current systems?

It will be some configuration challenges, but it's doable. Newer Linux VMs can be configured in a way to use the UIO interfaces. This doesn't break any existing VMs because the old model will continue to work when UIO is not used. Also, this doesn't require any Hyper-V changes.

Quoted 42 lines
Quoted 12 lines
Quoted 12 lines

Also, if you do do this, you need to list where the source for that userspace code is so that users can get it and have their distros package it up for them. I do not see that here at all.

Quoted 6 lines
---
 drivers/uio/uio_hv_generic.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
Quoted 6 lines
Quoted 6 lines

drivers/uio/uio_hv_generic.c hv_uio_irqcontrol(struct uio_info *info, s32

Quoted 6 lines
(quoted diff) +0 -0

b/drivers/uio/uio_hv_generic.c index c08a6cfd119f..8e5aa4a1247f 100644

drivers/uio/uio_hv_generic.c +0 -0
@@ -84,6 +84,9 @@ hv_uio_irqcontrol(struct uio_info *info, s32

irq_state)

Quoted 23 lines
Quoted 22 lines
Quoted 10 lines
drivers/uio/uio_hv_generic.c +3 -0
 	dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state;
 	virt_mb();
 
+	if (!dev->channel->offermsg.monitor_allocated && irq_state)
+		vmbus_setevent(dev->channel);
+
 	return 0;
 }
 

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 242

Quoted 4 lines
drivers/uio/uio_hv_generic.c +0 -1
@@ -239,12 +242,6 @@ hv_uio_probe(struct hv_device *dev,
 	void *ring_buffer;
 	int ret;
 
-	/* Communicating with host has to be via shared memory not

hypercall */

drivers/uio/uio_hv_generic.c hv_uio_probe(struct hv_device *dev, · line 244

Quoted 2 lines
drivers/uio/uio_hv_generic.c +0 -2
-	if (!channel->offermsg.monitor_allocated) {
-		dev_err(&dev->device, "vmbus channel requires

hypercall\n");

I do not understand, why is this check not made anymore here? Why constantly make the call above in the irq handler instead? Isn't that going to be massively slow?

Quoted 7 lines
Quoted 3 lines

Some VMBUS devices exposed by the Hyper-V are not modeled as high speed, they use hypercall, not monitored bits. Because they don't fit into other kernel API (as explained above), can we use UIO for those

devices?

UIO is for mmaped memory regions, like PCI devices, how is this a valid Hyper-V api at all?

Currently uio_hv_generic is used to mmap VMBUS ring buffers to user-mode. The primary use case is for DPDK. However, the same mmap model can be used for all other VMBUS devices, as they all use the same ring buffers model.

Long

maintainer
#5 lore raw
Quoted 41 lines · Long Li
Quoted 41 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 11 lines · Long Li
Quoted 11 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 1 line · Saurabh Sengar

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 3 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can

handle the device.

Quoted 6 lines
Quoted 6 lines
Quoted 3 lines

Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device

types.

Quoted 3 lines
Quoted 3 lines

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not

a good idea.

Quoted 3 lines

There are some VMBUS drivers currently not implemented in Linux. Out of all VMBUS drivers, two use "monitored bits": they are network and

storage drivers.

Quoted 9 lines

All the rest VMBUS drivers use hypercall for host notification and signal for next interrupt. One example of such driver is to collect process level crash information for diagnostic purposes.

Also, we want to move some existing kernel mode VMBUS drivers to user-space, such as hv_kvp and hv_filecopy. They don't really fit into an existing kernel API, and they create their own devices under /dev and communicates with a user-mode daemon to do most of the work. It's a better model that we can move those drivers entirely into user-mode.

Quoted 4 lines
Quoted 2 lines

How are you going to be able to remove drivers that export an existing user/kernel api and not break current systems?

It will be some configuration challenges, but it's doable.

How exactly?

We can not break userspace when you upgrade a kernel version.

Quoted 4 lines

Newer Linux
VMs can be configured in a way to use the UIO interfaces. This doesn't break any existing VMs because the old model will continue to work when UIO is not used. Also, this doesn't require any Hyper-V changes.

Hyper-v changes are not the issue here :)

Again, you have to support the situation where a system upgrades to a new kernel and all the same functionality must be there. How will you do that if you remove drivers from a newer kernel?

Quoted 6 lines
Quoted 2 lines

UIO is for mmaped memory regions, like PCI devices, how is this a valid Hyper-V api at all?

Currently uio_hv_generic is used to mmap VMBUS ring buffers to user-mode. The primary use case is for DPDK. However, the same mmap model can be used for all other VMBUS devices, as they all use the same ring buffers model.

Ok, that feels like overkill...

You need to also post the source for these new userspace drivers somewhere for us to review. I don't see a link to them in the changelog text :(

thanks,

greg k-h

#6 lore raw
Quoted 64 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 39 lines · Long Li
Quoted 39 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 4 lines · Long Li
Quoted 4 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

On Thu, Oct 13, 2022 at 11:29:14AM -0700, Saurabh Sengar wrote:

Quoted 8 lines
Quoted 8 lines
Quoted 1 line

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 4 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can

handle the device.

Quoted 7 lines
Quoted 7 lines
Quoted 3 lines

Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device

types.

Quoted 4 lines
Quoted 4 lines

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not

a good idea.

Quoted 4 lines

There are some VMBUS drivers currently not implemented in Linux. Out of all VMBUS drivers, two use "monitored bits": they are network and

storage drivers.

Quoted 4 lines

All the rest VMBUS drivers use hypercall for host notification and signal for next interrupt. One example of such driver is to collect process level crash information for diagnostic purposes.

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

Also, we want to move some existing kernel mode VMBUS drivers to user-space, such as hv_kvp and hv_filecopy. They don't really fit into an existing kernel API, and they create their own devices under /dev and communicates with a user-mode daemon to do most of the work. It's a better model that we can move those drivers entirely

into user-mode.

Quoted 5 lines
Quoted 3 lines

How are you going to be able to remove drivers that export an existing user/kernel api and not break current systems?

It will be some configuration challenges, but it's doable.

How exactly?

We can not break userspace when you upgrade a kernel version.

Quoted 4 lines

Newer Linux
VMs can be configured in a way to use the UIO interfaces. This doesn't break any existing VMs because the old model will continue to work when UIO is not used. Also, this doesn't require any Hyper-V changes.

Quoted 5 lines

Hyper-v changes are not the issue here :)

Again, you have to support the situation where a system upgrades to a new kernel and all the same functionality must be there. How will you do that if you remove drivers from a newer kernel?

Currently there is a hv_kvp_daemon that interacts with the /dev/device created by the hv_kvp kernel driver, this is the only program interacts with this device. This program is acting like a user-space driver, in a sense.

With the new kernel, if the KVP kernel mode is not present, this kvp daemon will not start. All the KVP functionality is handled by the new UIO driver.

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

UIO is for mmaped memory regions, like PCI devices, how is this a valid Hyper-V api at all?

Currently uio_hv_generic is used to mmap VMBUS ring buffers to user-

mode.

Quoted 2 lines

The primary use case is for DPDK. However, the same mmap model can be used for all other VMBUS devices, as they all use the same ring buffers

Quoted 7 lines

model.

Ok, that feels like overkill...

You need to also post the source for these new userspace drivers somewhere for us to review. I don't see a link to them in the changelog text :(

Yes, we will share the user space VMBUS drivers. What's the concern of creating VMBus driver in user-mode? There are many examples of kernel drivers moving to user-mode. For example, DPDK wants a network driver in user-mode,
SPDK wants a storage driver in user-mode, although they already have corresponding kernel drivers.

Long

maintainer
#7 lore raw
Quoted 70 lines · Long Li
Quoted 64 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 39 lines · Long Li
Quoted 39 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

Quoted 4 lines · Long Li
Quoted 4 lines

Subject: Re: [PATCH] uio_hv_generic: Enable interrupt for low speed VMBus devices

On Thu, Oct 13, 2022 at 11:29:14AM -0700, Saurabh Sengar wrote:

Quoted 8 lines
Quoted 8 lines
Quoted 1 line

Hyper-V is adding some "specialty" synthetic devices.

What devices are those specifically?

Quoted 4 lines

Instead of writing new kernel-level VMBus drivers for these devices, the devices will be presented to user space via this existing Hyper-V generic UIO driver, so that a user space driver can

handle the device.

Quoted 7 lines
Quoted 7 lines
Quoted 3 lines

Since these new synthetic devices are low speed devices, they don't support monitor bits and we must use vmbus_setevent() to enable interrupts from the host.

That is not what the UIO interface is for. Please write real drivers so that they tie into the specific user/kernel apis for those device

types.

Quoted 4 lines
Quoted 4 lines

Without a specific list of what these devices are, I can not recommend that anyone use the UIO api for them as that's probably not

a good idea.

Quoted 4 lines

There are some VMBUS drivers currently not implemented in Linux. Out of all VMBUS drivers, two use "monitored bits": they are network and

storage drivers.

Quoted 4 lines

All the rest VMBUS drivers use hypercall for host notification and signal for next interrupt. One example of such driver is to collect process level crash information for diagnostic purposes.

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

Also, we want to move some existing kernel mode VMBUS drivers to user-space, such as hv_kvp and hv_filecopy. They don't really fit into an existing kernel API, and they create their own devices under /dev and communicates with a user-mode daemon to do most of the work. It's a better model that we can move those drivers entirely

into user-mode.

Quoted 5 lines
Quoted 3 lines

How are you going to be able to remove drivers that export an existing user/kernel api and not break current systems?

It will be some configuration challenges, but it's doable.

How exactly?

We can not break userspace when you upgrade a kernel version.

Quoted 4 lines

Newer Linux
VMs can be configured in a way to use the UIO interfaces. This doesn't break any existing VMs because the old model will continue to work when UIO is not used. Also, this doesn't require any Hyper-V changes.

Quoted 5 lines

Hyper-v changes are not the issue here :)

Again, you have to support the situation where a system upgrades to a new kernel and all the same functionality must be there. How will you do that if you remove drivers from a newer kernel?

Quoted 6 lines

Currently there is a hv_kvp_daemon that interacts with the /dev/device created by the hv_kvp kernel driver, this is the only program interacts with this device. This program is acting like a user-space driver, in a sense.

With the new kernel, if the KVP kernel mode is not present, this kvp daemon will not start. All the KVP functionality is handled by the new UIO driver.

And those changes have already been implemented and pushed out somewhere?

Quoted 15 lines
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

UIO is for mmaped memory regions, like PCI devices, how is this a valid Hyper-V api at all?

Currently uio_hv_generic is used to mmap VMBUS ring buffers to user-

mode.

Quoted 2 lines

The primary use case is for DPDK. However, the same mmap model can be used for all other VMBUS devices, as they all use the same ring buffers

Quoted 7 lines

model.

Ok, that feels like overkill...

You need to also post the source for these new userspace drivers somewhere for us to review. I don't see a link to them in the changelog text :(

Quoted 5 lines

Yes, we will share the user space VMBUS drivers. What's the concern of creating VMBus driver in user-mode? There are many examples of kernel drivers moving to user-mode. For example, DPDK wants a network driver in user-mode,
SPDK wants a storage driver in user-mode, although they already have corresponding kernel drivers.

When moving out of the kernel to userspace, you loose the common and neutral user/kernel api and are now forced to rely on a userspace program for access to that device. no longer can you do something like normal socket calls, but instead, you have to rely on a library for the same functionality.

Same for block devices. Do you really want to give up that common interface to your block device and rely on a library instead of the kernel?

If so, great, but don't take away the functionality from all users, as you will break their systems.

thanks

greg k-h