Skip to contentSkip to conversation

hv_sock: Hardening changes

RFC PATCH In review Andrea Parri (Microsoft) 2 participants 24 messages Apr 13, 2022 – Apr 15, 2022

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

cover hv_sock: Hardening changes

Miscellaneous changes to "harden" the hv_sock driver and enable it in isolated guests. The diff in ring_buffer.c, hyperv.h is due to a consequent refactoring/code elimination (patch #6).

Applies to v5.18-rc2.

Thanks,
Andrea

Andrea Parri (Microsoft) (6):
hv_sock: Check hv_pkt_iter_first_raw()'s return value
hv_sock: Copy packets sent by Hyper-V out of the ring buffer
hv_sock: Add validation for untrusted Hyper-V values
hv_sock: Initialize send_buf in hvs_stream_enqueue()
Drivers: hv: vmbus: Accept hv_sock offers in isolated guests
Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

 drivers/hv/channel_mgmt.c        |  9 ++++--
 drivers/hv/ring_buffer.c         | 11 ++++----
 include/linux/hyperv.h           | 48 ++++++++++----------------------
 net/vmw_vsock/hyperv_transport.c | 24 ++++++++++++----
 4 files changed, 46 insertions(+), 46 deletions(-)
Signature · 1 line
2.25.1
#1 lore raw

RFC PATCH 1/6 hv_sock: Check hv_pkt_iter_first_raw()'s return value

The function returns NULL if the ring buffer has no enough space available for a packet descriptor. The ring buffer's write_index is in memory which is shared with the Hyper-V host, its value is thus subject to being changed at any time.

---
 net/vmw_vsock/hyperv_transport.c | 2 ++
 1 file changed, 2 insertions(+)
net/vmw_vsock/hyperv_transport.c +2 -0
index e111e13b66604..943352530936e 100644
@@ -603,6 +603,8 @@ static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk, struct msghdr *msg,
 
 	if (need_refill) {
 		hvs->recv_desc = hv_pkt_iter_first_raw(hvs->chan);
+		if (!hvs->recv_desc)
+			return -ENOBUFS;
 		ret = hvs_update_recv_data(hvs);
 		if (ret)
 			return ret;
Signature · 1 line
2.25.1
#2 lore raw

RE: [RFC PATCH 1/6] hv_sock: Check hv_pkt_iter_first_raw()'s return value

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 3 lines

The function returns NULL if the ring buffer has no enough space available for a packet descriptor. The ring buffer's write_index

The first sentence wording is a bit scrambled. I think you mean the ring buffer doesn't contain enough readable bytes to constitute a packet descriptor.

Quoted 2 lines

is in memory which is shared with the Hyper-V host, its value is thus subject to being changed at any time.

This second sentence is true, but I'm not making the connection
with the code change below. Evidently, there is some previous
check made to ensure that enough bytes are available to be received when hvs_stream_dequeue() is called, so we assumed that NULL could never be returned? I looked but didn't find such a check, so maybe I didn't look carefully enough. But now we are assuming that Hyper-V might have invalidated that previous check by subsequently changing the write_index in a bogus way? So now, NULL could be returned when previously we assumed it couldn't.

Michael

Quoted 19 lines
---
 net/vmw_vsock/hyperv_transport.c | 2 ++
 1 file changed, 2 insertions(+)
net/vmw_vsock/hyperv_transport.c +2 -0
index e111e13b66604..943352530936e 100644
@@ -603,6 +603,8 @@ static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk,
 struct msghdr *msg,
 
 	if (need_refill) {
 		hvs->recv_desc = hv_pkt_iter_first_raw(hvs->chan);
+		if (!hvs->recv_desc)
+			return -ENOBUFS;
 		ret = hvs_update_recv_data(hvs);
 		if (ret)
Quoted 3 lines

return ret;
--
2.25.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#3 lore raw

Re: [RFC PATCH 1/6] hv_sock: Check hv_pkt_iter_first_raw()'s return value

Quoted 8 lines · Michael Kelley (LINUX)

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 3 lines

The function returns NULL if the ring buffer has no enough space available for a packet descriptor. The ring buffer's write_index

The first sentence wording is a bit scrambled. I think you mean the ring buffer doesn't contain enough readable bytes to constitute a packet descriptor.

Indeed, replaced with your working.

Quoted 12 lines
Quoted 2 lines

is in memory which is shared with the Hyper-V host, its value is thus subject to being changed at any time.

This second sentence is true, but I'm not making the connection
with the code change below. Evidently, there is some previous
check made to ensure that enough bytes are available to be received when hvs_stream_dequeue() is called, so we assumed that NULL could never be returned? I looked but didn't find such a check, so maybe I didn't look carefully enough. But now we are assuming that Hyper-V might have invalidated that previous check by subsequently changing the write_index in a bogus way? So now, NULL could be returned when previously we assumed it couldn't.

I think you're looking for hvs_stream_has_data(). (Previous checks apart, hvs_stream_dequeue() will "dereference" the pointer so...)

Thanks,
Andrea

#4 lore raw

RE: [RFC PATCH 1/6] hv_sock: Check hv_pkt_iter_first_raw()'s return value

From: Andrea Parri <[email protected]> Sent: Thursday, April 14, 2022 11:42 PM

Quoted 28 lines · Michael Kelley (LINUX)
Quoted 1 line

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13,

2022 1:48 PM

Quoted 7 lines
Quoted 3 lines

The function returns NULL if the ring buffer has no enough space available for a packet descriptor. The ring buffer's write_index

The first sentence wording is a bit scrambled. I think you mean the ring buffer doesn't contain enough readable bytes to constitute a packet descriptor.

Indeed, replaced with your working.

Quoted 12 lines
Quoted 2 lines

is in memory which is shared with the Hyper-V host, its value is thus subject to being changed at any time.

This second sentence is true, but I'm not making the connection
with the code change below. Evidently, there is some previous
check made to ensure that enough bytes are available to be received when hvs_stream_dequeue() is called, so we assumed that NULL could never be returned? I looked but didn't find such a check, so maybe I didn't look carefully enough. But now we are assuming that Hyper-V might have invalidated that previous check by subsequently changing the write_index in a bogus way? So now, NULL could be returned when previously we assumed it couldn't.

Quoted 2 lines

I think you're looking for hvs_stream_has_data(). (Previous checks apart, hvs_stream_dequeue() will "dereference" the pointer so...)

Agreed. I didn't say this explicitly, but I was wondering about the risk in the current code (without these hardening patches) of getting a NULL pointer from hv_pkt_iter_first_raw(), and then dereferencing it.

Michael

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#5 lore raw

Re: [RFC PATCH 1/6] hv_sock: Check hv_pkt_iter_first_raw()'s return value

Quoted 29 lines · Michael Kelley (LINUX)

From: Andrea Parri <[email protected]> Sent: Thursday, April 14, 2022 11:42 PM

Quoted 28 lines · Michael Kelley (LINUX)
Quoted 1 line

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13,

2022 1:48 PM

Quoted 7 lines
Quoted 3 lines

The function returns NULL if the ring buffer has no enough space available for a packet descriptor. The ring buffer's write_index

The first sentence wording is a bit scrambled. I think you mean the ring buffer doesn't contain enough readable bytes to constitute a packet descriptor.

Indeed, replaced with your working.

Quoted 12 lines
Quoted 2 lines

is in memory which is shared with the Hyper-V host, its value is thus subject to being changed at any time.

This second sentence is true, but I'm not making the connection
with the code change below. Evidently, there is some previous
check made to ensure that enough bytes are available to be received when hvs_stream_dequeue() is called, so we assumed that NULL could never be returned? I looked but didn't find such a check, so maybe I didn't look carefully enough. But now we are assuming that Hyper-V might have invalidated that previous check by subsequently changing the write_index in a bogus way? So now, NULL could be returned when previously we assumed it couldn't.

Quoted 6 lines
Quoted 2 lines

I think you're looking for hvs_stream_has_data(). (Previous checks apart, hvs_stream_dequeue() will "dereference" the pointer so...)

Agreed. I didn't say this explicitly, but I was wondering about the risk in the current code (without these hardening patches) of getting a NULL pointer from hv_pkt_iter_first_raw(), and then dereferencing it.

Got it. Updated the changelog to:

"The ring buffer's write_index is in memory which is shared with the
Hyper-V host, an erroneous or malicious host could thus change its value and overturn the result of hvs_stream_has_data()."

Hopefully this can clarify the issue (without introducing other typos).

Thanks,
Andrea

#6 lore raw

RFC PATCH 2/6 hv_sock: Copy packets sent by Hyper-V out of the ring buffer

In reply to Andrea Parri (Microsoft)

Pointers to VMbus packets sent by Hyper-V are used by the hv_sock driver within the gues VM. Hyper-V can send packets with erroneous values or modify packet fields after they are processed by the guest. To defend against these scenarios, copy the incoming packet after validating its length and offset fields using hv_pkt_iter_{first,next}(). In this way, the packet can no longer be modified by the host.

---
 net/vmw_vsock/hyperv_transport.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
net/vmw_vsock/hyperv_transport.c +7 -2
index 943352530936e..8c37d07017fc4 100644
@@ -78,6 +78,9 @@ struct hvs_send_buf {
 					 ALIGN((payload_len), 8) + \
 					 VMBUS_PKT_TRAILER_SIZE)
 
+/* Upper bound on the size of a VMbus packet for hv_sock */
+#define HVS_MAX_PKT_SIZE	HVS_PKT_LEN(HVS_MTU_SIZE)
+
 union hvs_service_id {
 	guid_t	srv_id;
 
@@ -378,6 +381,8 @@ static void hvs_open_connection(struct vmbus_channel *chan)
 		rcvbuf = ALIGN(rcvbuf, HV_HYP_PAGE_SIZE);
 	}
 
+	chan->max_pkt_size = HVS_MAX_PKT_SIZE;
+
 	ret = vmbus_open(chan, sndbuf, rcvbuf, NULL, 0, hvs_channel_cb,
 			 conn_from_host ? new : sk);
 	if (ret != 0) {
@@ -602,7 +607,7 @@ static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk, struct msghdr *msg,
 		return -EOPNOTSUPP;
 
 	if (need_refill) {
-		hvs->recv_desc = hv_pkt_iter_first_raw(hvs->chan);
+		hvs->recv_desc = hv_pkt_iter_first(hvs->chan);
 		if (!hvs->recv_desc)
 			return -ENOBUFS;
 		ret = hvs_update_recv_data(hvs);
@@ -618,7 +623,7 @@ static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk, struct msghdr *msg,
 
 	hvs->recv_data_len -= to_read;
 	if (hvs->recv_data_len == 0) {
-		hvs->recv_desc = hv_pkt_iter_next_raw(hvs->chan, hvs->recv_desc);
+		hvs->recv_desc = hv_pkt_iter_next(hvs->chan, hvs->recv_desc);
 		if (hvs->recv_desc) {
 			ret = hvs_update_recv_data(hvs);
 			if (ret)
Signature · 1 line
2.25.1
#7 lore raw

RE: [RFC PATCH 2/6] hv_sock: Copy packets sent by Hyper-V out of the ring buffer

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 3 lines

Pointers to VMbus packets sent by Hyper-V are used by the hv_sock driver within the gues VM. Hyper-V can send packets with erroneous values or

s/gues/guest/

Quoted 52 lines

modify packet fields after they are processed by the guest. To defend against these scenarios, copy the incoming packet after validating its length and offset fields using hv_pkt_iter_{first,next}(). In this way, the packet can no longer be modified by the host.

---
 net/vmw_vsock/hyperv_transport.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
net/vmw_vsock/hyperv_transport.c +6 -1
index 943352530936e..8c37d07017fc4 100644
@@ -78,6 +78,9 @@ struct hvs_send_buf {
 					 ALIGN((payload_len), 8) + \
 					 VMBUS_PKT_TRAILER_SIZE)
 
+/* Upper bound on the size of a VMbus packet for hv_sock */
+#define HVS_MAX_PKT_SIZE	HVS_PKT_LEN(HVS_MTU_SIZE)
+
 union hvs_service_id {
 	guid_t	srv_id;
 
@@ -378,6 +381,8 @@ static void hvs_open_connection(struct vmbus_channel *chan)
 		rcvbuf = ALIGN(rcvbuf, HV_HYP_PAGE_SIZE);
 	}
 
+	chan->max_pkt_size = HVS_MAX_PKT_SIZE;
+
 	ret = vmbus_open(chan, sndbuf, rcvbuf, NULL, 0, hvs_channel_cb,
 			 conn_from_host ? new : sk);
 	if (ret != 0) {
@@ -602,7 +607,7 @@ static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk,
 struct msghdr *msg,
 		return -EOPNOTSUPP;
 
 	if (need_refill) {
-		hvs->recv_desc = hv_pkt_iter_first_raw(hvs->chan);
+		hvs->recv_desc = hv_pkt_iter_first(hvs->chan);
 		if (!hvs->recv_desc)
 			return -ENOBUFS;
net/vmw_vsock/hyperv_transport.c +1 -1
 		ret = hvs_update_recv_data(hvs);
@@ -618,7 +623,7 @@ static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk,
 struct msghdr *msg,
 
 	hvs->recv_data_len -= to_read;
 	if (hvs->recv_data_len == 0) {
-		hvs->recv_desc = hv_pkt_iter_next_raw(hvs->chan, hvs->recv_desc);
+		hvs->recv_desc = hv_pkt_iter_next(hvs->chan, hvs->recv_desc);
 		if (hvs->recv_desc) {
 			ret = hvs_update_recv_data(hvs);
Quoted 3 lines

if (ret)
--
2.25.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#8 lore raw

Re: [RFC PATCH 2/6] hv_sock: Copy packets sent by Hyper-V out of the ring buffer

Quoted 6 lines · Michael Kelley (LINUX)

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 3 lines

Pointers to VMbus packets sent by Hyper-V are used by the hv_sock driver within the gues VM. Hyper-V can send packets with erroneous values or

s/gues/guest/

Fixed.

Thanks,
Andrea

#9 lore raw

RFC PATCH 3/6 hv_sock: Add validation for untrusted Hyper-V values

In reply to Andrea Parri (Microsoft)

For additional robustness in the face of Hyper-V errors or malicious behavior, validate all values that originate from packets that Hyper-V has sent to the guest in the host-to-guest ring buffer. Ensure that invalid values cannot cause data being copied out of the bounds of the source buffer in hvs_stream_dequeue().

---
 include/linux/hyperv.h           |  5 +++++
 net/vmw_vsock/hyperv_transport.c | 11 +++++++++--
 2 files changed, 14 insertions(+), 2 deletions(-)
include/linux/hyperv.h +5 -0
index fe2e0179ed51e..55478a6810b60 100644
@@ -1663,6 +1663,11 @@ static inline u32 hv_pkt_datalen(const struct vmpacket_descriptor *desc)
 	return (desc->len8 << 3) - (desc->offset8 << 3);
 }
 
+/* Get packet length associated with descriptor */
+static inline u32 hv_pkt_len(const struct vmpacket_descriptor *desc)
+{
+	return desc->len8 << 3;
+}
 
 struct vmpacket_descriptor *
 hv_pkt_iter_first_raw(struct vmbus_channel *channel);
net/vmw_vsock/hyperv_transport.c +9 -2
index 8c37d07017fc4..092cadc2c866d 100644
@@ -577,12 +577,19 @@ static bool hvs_dgram_allow(u32 cid, u32 port)
 static int hvs_update_recv_data(struct hvsock *hvs)
 {
 	struct hvs_recv_buf *recv_buf;
-	u32 payload_len;
+	u32 pkt_len, payload_len;
+
+	pkt_len = hv_pkt_len(hvs->recv_desc);
+
+	/* Ensure the packet is big enough to read its header */
+	if (pkt_len < HVS_HEADER_LEN)
+		return -EIO;
 
 	recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1);
 	payload_len = recv_buf->hdr.data_size;
 
-	if (payload_len > HVS_MTU_SIZE)
+	/* Ensure the packet is big enough to read its payload */
+	if (payload_len > pkt_len - HVS_HEADER_LEN || payload_len > HVS_MTU_SIZE)
 		return -EIO;
 
 	if (payload_len == 0)
Signature · 1 line
2.25.1
#10 lore raw

RFC PATCH 4/6 hv_sock: Initialize send_buf in hvs_stream_enqueue()

In reply to Andrea Parri (Microsoft)

So that padding or uninitialized bytes can't leak guest memory contents.

---
 net/vmw_vsock/hyperv_transport.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/vmw_vsock/hyperv_transport.c +1 -1
index 092cadc2c866d..72ce00928c8e7 100644
@@ -655,7 +655,7 @@ static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk, struct msghdr *msg,
 
 	BUILD_BUG_ON(sizeof(*send_buf) != HV_HYP_PAGE_SIZE);
 
-	send_buf = kmalloc(sizeof(*send_buf), GFP_KERNEL);
+	send_buf = kzalloc(sizeof(*send_buf), GFP_KERNEL);
 	if (!send_buf)
 		return -ENOMEM;
 
Signature · 1 line
2.25.1
#11 lore raw

RE: [RFC PATCH 4/6] hv_sock: Initialize send_buf in hvs_stream_enqueue()

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 13 lines

So that padding or uninitialized bytes can't leak guest memory contents.

---
 net/vmw_vsock/hyperv_transport.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/vmw_vsock/hyperv_transport.c +0 -0
index 092cadc2c866d..72ce00928c8e7 100644

net/vmw_vsock/hyperv_transport.c static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk, · line 659

Quoted 6 lines
net/vmw_vsock/hyperv_transport.c +1 -1
@@ -655,7 +655,7 @@ static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk,
 struct msghdr *msg,
 
 	BUILD_BUG_ON(sizeof(*send_buf) != HV_HYP_PAGE_SIZE);
 
-	send_buf = kmalloc(sizeof(*send_buf), GFP_KERNEL);
+	send_buf = kzalloc(sizeof(*send_buf), GFP_KERNEL);

Is this change really needed? All fields are explicitly initialized, and in the data
array, only the populated bytes are copied to the ring buffer. There should not
be any uninitialized values sent to the host. Zeroing the memory ahead of
time certainly provides an extra protection (particularly against padding bytes, but there can't be any since the layout of the data is part of the protocol with Hyper-V). It is expensive protection to zero out 16K+ bytes every time we send out a small message.

Michael

Quoted 5 lines
net/vmw_vsock/hyperv_transport.c +0 -1
 	if (!send_buf)
 		return -ENOMEM;
 
--

2.25.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#12 lore raw

Re: [RFC PATCH 4/6] hv_sock: Initialize send_buf in hvs_stream_enqueue()

Quoted 9 lines

net/vmw_vsock/hyperv_transport.c static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk, · line 659

Quoted 7 lines
net/vmw_vsock/hyperv_transport.c +1 -1
@@ -655,7 +655,7 @@ static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk,
 struct msghdr *msg,
 
 	BUILD_BUG_ON(sizeof(*send_buf) != HV_HYP_PAGE_SIZE);
 
-	send_buf = kmalloc(sizeof(*send_buf), GFP_KERNEL);
+	send_buf = kzalloc(sizeof(*send_buf), GFP_KERNEL);

Is this change really needed?

The idea was...

Quoted 6 lines

All fields are explicitly initialized, and in the data
array, only the populated bytes are copied to the ring buffer. There should not
be any uninitialized values sent to the host. Zeroing the memory ahead of
time certainly provides an extra protection (particularly against padding bytes, but there can't be any since the layout of the data is part of the protocol with Hyper-V).

Rather than keeping checking that...

Quoted 2 lines

It is expensive protection to zero out 16K+ bytes every time we send out a small message.

Do this. ;-)

Will drop the patch.

Thanks,
Andrea

#13 lore raw

RE: [RFC PATCH 4/6] hv_sock: Initialize send_buf in hvs_stream_enqueue()

From: Andrea Parri <[email protected]> Sent: Thursday, April 14, 2022 11:51 PM

Quoted 14 lines
Quoted 10 lines

net/vmw_vsock/hyperv_transport.c static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk, · line 659

Quoted 8 lines
net/vmw_vsock/hyperv_transport.c +1 -1
@@ -655,7 +655,7 @@ static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk,
 struct msghdr *msg,
 
 	BUILD_BUG_ON(sizeof(*send_buf) != HV_HYP_PAGE_SIZE);
 
-	send_buf = kmalloc(sizeof(*send_buf), GFP_KERNEL);
+	send_buf = kzalloc(sizeof(*send_buf), GFP_KERNEL);

Is this change really needed?

The idea was...

Quoted 8 lines
Quoted 6 lines

All fields are explicitly initialized, and in the data
array, only the populated bytes are copied to the ring buffer. There should not
be any uninitialized values sent to the host. Zeroing the memory ahead of
time certainly provides an extra protection (particularly against padding bytes, but there can't be any since the layout of the data is part of the protocol with Hyper-V).

Rather than keeping checking that...

The extra protection might be obtained by just zero'ing the header (i.e., the
bytes up to the 16 Kbyte data array). I don't have a strong preference either
way, so up to you.

Michael

Quoted 11 lines
Quoted 4 lines

It is expensive protection to zero out 16K+ bytes every time we send out a small message.

Do this. ;-)

Will drop the patch.

Thanks,
Andrea

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#14 lore raw

Re: [RFC PATCH 4/6] hv_sock: Initialize send_buf in hvs_stream_enqueue()

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

All fields are explicitly initialized, and in the data
array, only the populated bytes are copied to the ring buffer. There should not
be any uninitialized values sent to the host. Zeroing the memory ahead of
time certainly provides an extra protection (particularly against padding bytes, but there can't be any since the layout of the data is part of the protocol with Hyper-V).

Quoted 5 lines
Quoted 1 line

Rather than keeping checking that...

The extra protection might be obtained by just zero'ing the header (i.e., the
bytes up to the 16 Kbyte data array). I don't have a strong preference either
way, so up to you.

A main reason behind this RFC is that I don't have either. IIUC, you're suggesting something like (the compiled only):

net/vmw_vsock/hyperv_transport.c +3 -1
index 092cadc2c866d..200f12c432863 100644
@@ -234,7 +234,8 @@ static int __hvs_send_data(struct vmbus_channel *chan,
 {
 	hdr->pkt_type = 1;
 	hdr->data_size = to_write;
-	return vmbus_sendpacket(chan, hdr, sizeof(*hdr) + to_write,
+	return vmbus_sendpacket(chan, hdr,
+				offsetof(struct hvs_send_buf, data) + to_write,
 				0, VM_PKT_DATA_INBAND, 0);
 }
 
@@ -658,6 +659,7 @@ static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk, struct msghdr *msg,
 	send_buf = kmalloc(sizeof(*send_buf), GFP_KERNEL);
 	if (!send_buf)
 		return -ENOMEM;
+	memset(send_buf, 0, offsetof(struct hvs_send_buf, data));
 
 	/* Reader(s) could be draining data from the channel as we write.
 	 * Maximize bandwidth, by iterating until the channel is found to be
Signature · 4 lines
Let me queue this for further testing/review...

Thanks,
  Andrea
#15 lore raw

RFC PATCH 5/6 Drivers: hv: vmbus: Accept hv_sock offers in isolated guests

In reply to Andrea Parri (Microsoft)

So that isolated guests can communicate with the host via hv_sock channels.

---
 drivers/hv/channel_mgmt.c | 9 +++++++--
 include/linux/hyperv.h    | 8 ++++++--
 2 files changed, 13 insertions(+), 4 deletions(-)
drivers/hv/channel_mgmt.c +7 -2
index 67be81208a2d9..83d7ab90b7305 100644
@@ -976,17 +976,22 @@ find_primary_channel_by_offer(const struct vmbus_channel_offer_channel *offer)
 	return channel;
 }
 
-static bool vmbus_is_valid_device(const guid_t *guid)
+static bool vmbus_is_valid_offer(const struct vmbus_channel_offer_channel *offer)
 {
+	const guid_t *guid = &offer->offer.if_type;
 	u16 i;
 
 	if (!hv_is_isolation_supported())
 		return true;
 
+	if (is_hvsock_offer(offer))
+		return true;
+
 	for (i = 0; i < ARRAY_SIZE(vmbus_devs); i++) {
 		if (guid_equal(guid, &vmbus_devs[i].guid))
 			return vmbus_devs[i].allowed_in_isolated;
 	}
+
 	return false;
 }
 
@@ -1004,7 +1009,7 @@ static void vmbus_onoffer(struct vmbus_channel_message_header *hdr)
 
 	trace_vmbus_onoffer(offer);
 
-	if (!vmbus_is_valid_device(&offer->offer.if_type)) {
+	if (!vmbus_is_valid_offer(offer)) {
 		pr_err_ratelimited("Invalid offer %d from the host supporting isolation\n",
 				   offer->child_relid);
 		atomic_dec(&vmbus_connection.offer_in_progress);
include/linux/hyperv.h +6 -2
index 55478a6810b60..1112c5cf894e6 100644
@@ -1044,10 +1044,14 @@ struct vmbus_channel {
 u64 vmbus_next_request_id(struct vmbus_channel *channel, u64 rqst_addr);
 u64 vmbus_request_addr(struct vmbus_channel *channel, u64 trans_id);
 
+static inline bool is_hvsock_offer(const struct vmbus_channel_offer_channel *o)
+{
+	return !!(o->offer.chn_flags & VMBUS_CHANNEL_TLNPI_PROVIDER_OFFER);
+}
+
 static inline bool is_hvsock_channel(const struct vmbus_channel *c)
 {
-	return !!(c->offermsg.offer.chn_flags &
-		  VMBUS_CHANNEL_TLNPI_PROVIDER_OFFER);
+	return is_hvsock_offer(&c->offermsg);
 }
 
 static inline bool is_sub_channel(const struct vmbus_channel *c)
Signature · 1 line
2.25.1
#16 lore raw

RE: [RFC PATCH 5/6] Drivers: hv: vmbus: Accept hv_sock offers in isolated guests

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 26 lines

So that isolated guests can communicate with the host via hv_sock channels.

---
 drivers/hv/channel_mgmt.c | 9 +++++++--
 include/linux/hyperv.h    | 8 ++++++--
 2 files changed, 13 insertions(+), 4 deletions(-)
drivers/hv/channel_mgmt.c +2 -1
index 67be81208a2d9..83d7ab90b7305 100644
@@ -976,17 +976,22 @@ find_primary_channel_by_offer(const struct
 vmbus_channel_offer_channel *offer)
 	return channel;
 }
 
-static bool vmbus_is_valid_device(const guid_t *guid)
+static bool vmbus_is_valid_offer(const struct vmbus_channel_offer_channel *offer)
 {
+	const guid_t *guid = &offer->offer.if_type;
 	u16 i;
 
 	if (!hv_is_isolation_supported())

drivers/hv/channel_mgmt.c find_primary_channel_by_offer(const struct · line 995

Quoted 10 lines
drivers/hv/channel_mgmt.c +4 -0
 		return true;
 
+	if (is_hvsock_offer(offer))
+		return true;
+
 	for (i = 0; i < ARRAY_SIZE(vmbus_devs); i++) {
 		if (guid_equal(guid, &vmbus_devs[i].guid))
 			return vmbus_devs[i].allowed_in_isolated;
 	}
+

Spurious newline added?

Quoted 35 lines
drivers/hv/channel_mgmt.c +1 -1
 	return false;
 }
@@ -1004,7 +1009,7 @@ static void vmbus_onoffer(struct
 vmbus_channel_message_header *hdr)
 
 	trace_vmbus_onoffer(offer);
 
-	if (!vmbus_is_valid_device(&offer->offer.if_type)) {
+	if (!vmbus_is_valid_offer(offer)) {
 		pr_err_ratelimited("Invalid offer %d from the host supporting isolation\n",
 				   offer->child_relid);
drivers/hv/channel_mgmt.c +0 -0
 		atomic_dec(&vmbus_connection.offer_in_progress);
include/linux/hyperv.h +6 -3
index 55478a6810b60..1112c5cf894e6 100644
@@ -1044,10 +1044,14 @@ struct vmbus_channel {
 u64 vmbus_next_request_id(struct vmbus_channel *channel, u64 rqst_addr);
 u64 vmbus_request_addr(struct vmbus_channel *channel, u64 trans_id);
 
+static inline bool is_hvsock_offer(const struct vmbus_channel_offer_channel *o)
+{
+	return !!(o->offer.chn_flags & VMBUS_CHANNEL_TLNPI_PROVIDER_OFFER);
+}
+
 static inline bool is_hvsock_channel(const struct vmbus_channel *c)
 {
-	return !!(c->offermsg.offer.chn_flags &
-		  VMBUS_CHANNEL_TLNPI_PROVIDER_OFFER);
+	return is_hvsock_offer(&c->offermsg);
 }
 
 static inline bool is_sub_channel(const struct vmbus_channel *c)
--
Quoted 1 line

2.25.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#17 lore raw

Re: [RFC PATCH 5/6] Drivers: hv: vmbus: Accept hv_sock offers in isolated guests

Quoted 14 lines
Quoted 14 lines
drivers/hv/channel_mgmt.c +2 -1
@@ -976,17 +976,22 @@ find_primary_channel_by_offer(const struct
 vmbus_channel_offer_channel *offer)
 	return channel;
 }
 
-static bool vmbus_is_valid_device(const guid_t *guid)
+static bool vmbus_is_valid_offer(const struct vmbus_channel_offer_channel *offer)
 {
+	const guid_t *guid = &offer->offer.if_type;
 	u16 i;
 
 	if (!hv_is_isolation_supported())
 		return true;
 
Quoted 12 lines

drivers/hv/channel_mgmt.c find_primary_channel_by_offer(const struct vmbus_channel_offer_channel *offer) · line 994

Quoted 8 lines
drivers/hv/channel_mgmt.c +4 -0
+	if (is_hvsock_offer(offer))
+		return true;
+
 	for (i = 0; i < ARRAY_SIZE(vmbus_devs); i++) {
 		if (guid_equal(guid, &vmbus_devs[i].guid))
 			return vmbus_devs[i].allowed_in_isolated;
 	}
+

Spurious newline added?

Quoted 1 line
drivers/hv/channel_mgmt.c +0 -0
 	return false;

Intentionally added to visually separate the "hvsock", "vmbus_dev" and "default" blocks, patch seemed simple enough to try to merge in "style material" without incurring in the question. ;-)

Newline removed.

Thanks,
Andrea

#18 lore raw

RFC PATCH 6/6 Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

In reply to Andrea Parri (Microsoft)

With no users of hv_pkt_iter_next_raw() and no "external" users of hv_pkt_iter_first_raw(), the iterator functions can be refactored and simplified to remove some indirection/code.

---
 drivers/hv/ring_buffer.c | 11 +++++------
 include/linux/hyperv.h   | 35 ++++-------------------------------
 2 files changed, 9 insertions(+), 37 deletions(-)
drivers/hv/ring_buffer.c +5 -6
index 3d215d9dec433..c9357dae2a2c8 100644
@@ -421,7 +421,7 @@ int hv_ringbuffer_read(struct vmbus_channel *channel,
 	memcpy(buffer, (const char *)desc + offset, packetlen);
 
 	/* Advance ring index to next packet descriptor */
-	__hv_pkt_iter_next(channel, desc, true);
+	__hv_pkt_iter_next(channel, desc);
 
 	/* Notify host of update */
 	hv_pkt_iter_close(channel);
@@ -459,7 +459,8 @@ static u32 hv_pkt_iter_avail(const struct hv_ring_buffer_info *rbi)
 /*
  * Get first vmbus packet without copying it out of the ring buffer
  */
-struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct vmbus_channel *channel)
+static struct vmpacket_descriptor *
+hv_pkt_iter_first_raw(struct vmbus_channel *channel)
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 
@@ -470,7 +471,6 @@ struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct vmbus_channel *channel)
 
 	return (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi->priv_read_index);
 }
-EXPORT_SYMBOL_GPL(hv_pkt_iter_first_raw);
 
 /*
  * Get first vmbus packet from ring buffer after read_index
@@ -534,8 +534,7 @@ EXPORT_SYMBOL_GPL(hv_pkt_iter_first);
  */
 struct vmpacket_descriptor *
 __hv_pkt_iter_next(struct vmbus_channel *channel,
-		   const struct vmpacket_descriptor *desc,
-		   bool copy)
+		   const struct vmpacket_descriptor *desc)
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 	u32 packetlen = desc->len8 << 3;
@@ -548,7 +547,7 @@ __hv_pkt_iter_next(struct vmbus_channel *channel,
 		rbi->priv_read_index -= dsize;
 
 	/* more data? */
-	return copy ? hv_pkt_iter_first(channel) : hv_pkt_iter_first_raw(channel);
+	return hv_pkt_iter_first(channel);
 }
 EXPORT_SYMBOL_GPL(__hv_pkt_iter_next);
 
include/linux/hyperv.h +4 -31
index 1112c5cf894e6..370adc9971d3e 100644
@@ -1673,55 +1673,28 @@ static inline u32 hv_pkt_len(const struct vmpacket_descriptor *desc)
 	return desc->len8 << 3;
 }
 
-struct vmpacket_descriptor *
-hv_pkt_iter_first_raw(struct vmbus_channel *channel);
-
 struct vmpacket_descriptor *
 hv_pkt_iter_first(struct vmbus_channel *channel);
 
 struct vmpacket_descriptor *
 __hv_pkt_iter_next(struct vmbus_channel *channel,
-		   const struct vmpacket_descriptor *pkt,
-		   bool copy);
+		   const struct vmpacket_descriptor *pkt);
 
 void hv_pkt_iter_close(struct vmbus_channel *channel);
 
 static inline struct vmpacket_descriptor *
-hv_pkt_iter_next_pkt(struct vmbus_channel *channel,
-		     const struct vmpacket_descriptor *pkt,
-		     bool copy)
+hv_pkt_iter_next(struct vmbus_channel *channel,
+		 const struct vmpacket_descriptor *pkt)
 {
 	struct vmpacket_descriptor *nxt;
 
-	nxt = __hv_pkt_iter_next(channel, pkt, copy);
+	nxt = __hv_pkt_iter_next(channel, pkt);
 	if (!nxt)
 		hv_pkt_iter_close(channel);
 
 	return nxt;
 }
 
-/*
- * Get next packet descriptor without copying it out of the ring buffer
- * If at end of list, return NULL and update host.
- */
-static inline struct vmpacket_descriptor *
-hv_pkt_iter_next_raw(struct vmbus_channel *channel,
-		     const struct vmpacket_descriptor *pkt)
-{
-	return hv_pkt_iter_next_pkt(channel, pkt, false);
-}
-
-/*
- * Get next packet descriptor from iterator
- * If at end of list, return NULL and update host.
- */
-static inline struct vmpacket_descriptor *
-hv_pkt_iter_next(struct vmbus_channel *channel,
-		 const struct vmpacket_descriptor *pkt)
-{
-	return hv_pkt_iter_next_pkt(channel, pkt, true);
-}
-
 #define foreach_vmbus_pkt(pkt, channel) \
 	for (pkt = hv_pkt_iter_first(channel); pkt; \
 	    pkt = hv_pkt_iter_next(channel, pkt))
Signature · 1 line
2.25.1
#19 lore raw

RE: [RFC PATCH 6/6] Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, April 13, 2022 1:48 PM

Quoted 36 lines

With no users of hv_pkt_iter_next_raw() and no "external" users of hv_pkt_iter_first_raw(), the iterator functions can be refactored and simplified to remove some indirection/code.

---
 drivers/hv/ring_buffer.c | 11 +++++------
 include/linux/hyperv.h   | 35 ++++-------------------------------
 2 files changed, 9 insertions(+), 37 deletions(-)
drivers/hv/ring_buffer.c +3 -2
index 3d215d9dec433..c9357dae2a2c8 100644
@@ -421,7 +421,7 @@ int hv_ringbuffer_read(struct vmbus_channel *channel,
 	memcpy(buffer, (const char *)desc + offset, packetlen);
 
 	/* Advance ring index to next packet descriptor */
-	__hv_pkt_iter_next(channel, desc, true);
+	__hv_pkt_iter_next(channel, desc);
 
 	/* Notify host of update */
 	hv_pkt_iter_close(channel);
@@ -459,7 +459,8 @@ static u32 hv_pkt_iter_avail(const struct hv_ring_buffer_info
 *rbi)
 /*
  * Get first vmbus packet without copying it out of the ring buffer
  */
-struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct vmbus_channel *channel)
+static struct vmpacket_descriptor *
+hv_pkt_iter_first_raw(struct vmbus_channel *channel)
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;

drivers/hv/ring_buffer.c struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct · line 475

Quoted 6 lines
drivers/hv/ring_buffer.c +0 -1
@@ -470,7 +471,6 @@ struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct
 vmbus_channel *channel)
 
 	return (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi-
 >priv_read_index);
 }
-EXPORT_SYMBOL_GPL(hv_pkt_iter_first_raw);

Does hv_pkt_iter_first_raw() need to be retained at all as a separate function? I think after these changes, the only caller is hv_pkt_iter_first(), in which case the code could just go inline in hv_pkt_iter_first(). Doing that combining would also allow the elimination of the duplicate call to hv_pkt_iter_avail().

Michael

Quoted 86 lines
drivers/hv/ring_buffer.c +0 -0
 /*

* Get first vmbus packet from ring buffer after read_index

drivers/hv/ring_buffer.c +2 -3
@@ -534,8 +534,7 @@ EXPORT_SYMBOL_GPL(hv_pkt_iter_first);
  */
 struct vmpacket_descriptor *
 __hv_pkt_iter_next(struct vmbus_channel *channel,
-		   const struct vmpacket_descriptor *desc,
-		   bool copy)
+		   const struct vmpacket_descriptor *desc)
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 	u32 packetlen = desc->len8 << 3;
@@ -548,7 +547,7 @@ __hv_pkt_iter_next(struct vmbus_channel *channel,
 		rbi->priv_read_index -= dsize;
 
 	/* more data? */
-	return copy ? hv_pkt_iter_first(channel) : hv_pkt_iter_first_raw(channel);
+	return hv_pkt_iter_first(channel);
 }
 EXPORT_SYMBOL_GPL(__hv_pkt_iter_next);
 
include/linux/hyperv.h +4 -31
index 1112c5cf894e6..370adc9971d3e 100644
@@ -1673,55 +1673,28 @@ static inline u32 hv_pkt_len(const struct
 vmpacket_descriptor *desc)
 	return desc->len8 << 3;
 }
 
-struct vmpacket_descriptor *
-hv_pkt_iter_first_raw(struct vmbus_channel *channel);
-
 struct vmpacket_descriptor *
 hv_pkt_iter_first(struct vmbus_channel *channel);
 
 struct vmpacket_descriptor *
 __hv_pkt_iter_next(struct vmbus_channel *channel,
-		   const struct vmpacket_descriptor *pkt,
-		   bool copy);
+		   const struct vmpacket_descriptor *pkt);
 
 void hv_pkt_iter_close(struct vmbus_channel *channel);
 
 static inline struct vmpacket_descriptor *
-hv_pkt_iter_next_pkt(struct vmbus_channel *channel,
-		     const struct vmpacket_descriptor *pkt,
-		     bool copy)
+hv_pkt_iter_next(struct vmbus_channel *channel,
+		 const struct vmpacket_descriptor *pkt)
 {
 	struct vmpacket_descriptor *nxt;
 
-	nxt = __hv_pkt_iter_next(channel, pkt, copy);
+	nxt = __hv_pkt_iter_next(channel, pkt);
 	if (!nxt)
 		hv_pkt_iter_close(channel);
 
 	return nxt;
 }
 
-/*
- * Get next packet descriptor without copying it out of the ring buffer
- * If at end of list, return NULL and update host.
- */
-static inline struct vmpacket_descriptor *
-hv_pkt_iter_next_raw(struct vmbus_channel *channel,
-		     const struct vmpacket_descriptor *pkt)
-{
-	return hv_pkt_iter_next_pkt(channel, pkt, false);
-}
-
-/*
- * Get next packet descriptor from iterator
- * If at end of list, return NULL and update host.
- */
-static inline struct vmpacket_descriptor *
-hv_pkt_iter_next(struct vmbus_channel *channel,
-		 const struct vmpacket_descriptor *pkt)
-{
-	return hv_pkt_iter_next_pkt(channel, pkt, true);
-}
-
 #define foreach_vmbus_pkt(pkt, channel) \
 	for (pkt = hv_pkt_iter_first(channel); pkt; \
Quoted 3 lines

pkt = hv_pkt_iter_next(channel, pkt))
--
2.25.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#20 lore raw

Re: [RFC PATCH 6/6] Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

Quoted 8 lines

drivers/hv/ring_buffer.c struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct · line 475

Quoted 7 lines
drivers/hv/ring_buffer.c +0 -1
@@ -470,7 +471,6 @@ struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct
 vmbus_channel *channel)
 
 	return (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi-
 >priv_read_index);
 }
-EXPORT_SYMBOL_GPL(hv_pkt_iter_first_raw);
Quoted 6 lines

Does hv_pkt_iter_first_raw() need to be retained at all as a separate function? I think after these changes, the only caller is hv_pkt_iter_first(), in which case the code could just go inline in hv_pkt_iter_first(). Doing that combining would also allow the elimination of the duplicate call to hv_pkt_iter_avail().

Good point. Will do.

Thanks,
Andrea

#21 lore raw

Re: [RFC PATCH 6/6] Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

Quoted 8 lines · Andrea Parri
Quoted 8 lines
Quoted 7 lines
drivers/hv/ring_buffer.c +0 -1
@@ -470,7 +471,6 @@ struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct
 vmbus_channel *channel)
 
 	return (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi-
 >priv_read_index);
 }
-EXPORT_SYMBOL_GPL(hv_pkt_iter_first_raw);
Quoted 6 lines
Quoted 6 lines

Does hv_pkt_iter_first_raw() need to be retained at all as a separate function? I think after these changes, the only caller is hv_pkt_iter_first(), in which case the code could just go inline in hv_pkt_iter_first(). Doing that combining would also allow the elimination of the duplicate call to hv_pkt_iter_avail().

Back to this, can you clarify what you mean by "the elimination of..."? After moving the function "inline", hv_pkt_iter_avail() would be called in to check for a non-NULL descriptor (in the inline function) and later in the computation of bytes_avail.

Thanks,
Andrea

Quoted 5 lines

Good point. Will do.

Thanks,
Andrea

#22 lore raw

RE: [RFC PATCH 6/6] Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

From: Andrea Parri <[email protected]> Sent: Friday, April 15, 2022 9:28 AM

Quoted 17 lines · Andrea Parri
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
drivers/hv/ring_buffer.c +0 -1
@@ -470,7 +471,6 @@ struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct
 vmbus_channel *channel)
 
 	return (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi-
 >priv_read_index);
 }
-EXPORT_SYMBOL_GPL(hv_pkt_iter_first_raw);
Quoted 6 lines
Quoted 6 lines

Does hv_pkt_iter_first_raw() need to be retained at all as a separate function? I think after these changes, the only caller is hv_pkt_iter_first(), in which case the code could just go inline in hv_pkt_iter_first(). Doing that combining would also allow the elimination of the duplicate call to hv_pkt_iter_avail().

Quoted 4 lines

Back to this, can you clarify what you mean by "the elimination of..."? After moving the function "inline", hv_pkt_iter_avail() would be called in to check for a non-NULL descriptor (in the inline function) and later in the computation of bytes_avail.

I was thinking something like this:

bytes_avail = hv_pkt_iter_avail(rbi);
if (bytes_avail < sizeof(struct vmpacket_descriptor))

	return NULL;

bytes_avail = min(rbi->pkt_buffer_size, bytes_avail);

desc = (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi->priv_read_index);

And for that matter, hv_pkt_iter_avail() is now only called in one place. It's a judgment call whether to keep it as a separate helper function vs. inlining it in hv_pkt_iter_first() as well. I'm OK either way.

Michael

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#23 lore raw

Re: [RFC PATCH 6/6] Drivers: hv: vmbus: Refactor the ring-buffer iterator functions

Quoted 23 lines · Michael Kelley (LINUX)

From: Andrea Parri <[email protected]> Sent: Friday, April 15, 2022 9:28 AM

Quoted 17 lines · Andrea Parri
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
drivers/hv/ring_buffer.c +0 -1
@@ -470,7 +471,6 @@ struct vmpacket_descriptor *hv_pkt_iter_first_raw(struct
 vmbus_channel *channel)
 
 	return (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi-
 >priv_read_index);
 }
-EXPORT_SYMBOL_GPL(hv_pkt_iter_first_raw);
Quoted 6 lines
Quoted 6 lines

Does hv_pkt_iter_first_raw() need to be retained at all as a separate function? I think after these changes, the only caller is hv_pkt_iter_first(), in which case the code could just go inline in hv_pkt_iter_first(). Doing that combining would also allow the elimination of the duplicate call to hv_pkt_iter_avail().

Quoted 4 lines

Back to this, can you clarify what you mean by "the elimination of..."? After moving the function "inline", hv_pkt_iter_avail() would be called in to check for a non-NULL descriptor (in the inline function) and later in the computation of bytes_avail.

Quoted 8 lines

I was thinking something like this:

bytes_avail = hv_pkt_iter_avail(rbi);
if (bytes_avail < sizeof(struct vmpacket_descriptor))

	return NULL;

bytes_avail = min(rbi->pkt_buffer_size, bytes_avail);

desc = (struct vmpacket_descriptor *)(hv_get_ring_buffer(rbi) + rbi->priv_read_index);

Thanks for the clarification, I've applied it.

Andrea

Quoted 8 lines

And for that matter, hv_pkt_iter_avail() is now only called in one place. It's a judgment call whether to keep it as a separate helper function vs. inlining it in hv_pkt_iter_first() as well. I'm OK either way.

Michael