Skip to contentSkip to conversation

virtio-user: fix virtio-user server mode

PATCH Superseded Adrian Moreno 6 participants 14 messages Oct 20, 2020 – Oct 23, 2020 Network drivers

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover virtio-user: fix virtio-user server mode

This series addresses some issues identified in the virtio-user driver when configured in server mode.

Firstly, properly identify the backend type in this mode.

Secondly, ensure no get/set status commands are sent before protocol feature negotiation to avoid sending them to a backend that does not support them

And finally, add get/set status as per virtio spec so that it can work with modern vhost-user backends that rely on this message to start the device.

--
v1->v2:
Added patch 2 and 3 addressing additional issues
Check errno to select vhost-user backend and log the detected backend type

Adrian Moreno (3):
virtio-user: fix backend selection if stat fails
virtio_user: don't set/get_status until FEATURES_OK
virtio-user: set status on virtio-user reconnect

 drivers/net/virtio/virtio_user/vhost_user.c   | 10 ++++++----
 .../net/virtio/virtio_user/virtio_user_dev.c  |  7 +++++++
 .../net/virtio/virtio_user/virtio_user_dev.h  |  1 +
 drivers/net/virtio/virtio_user_ethdev.c       | 20 ++++++++++++++++---
 4 files changed, 31 insertions(+), 7 deletions(-)
Signature · 1 line
2.26.2
#1 lore raw

PATCH v2 1/3 virtio-user: fix backend selection if stat fails

If stat fails because the file does not exist, it means that the backend must be vhost-user in server mode.

Also, log the detected backend type.

Bugzilla ID: 559

Fixes f908b22ea47a ("net/virtio: move backend type selection to ethdev")
---
 drivers/net/virtio/virtio_user/virtio_user_dev.c | 7 +++++++
 drivers/net/virtio/virtio_user/virtio_user_dev.h | 1 +
 drivers/net/virtio/virtio_user_ethdev.c          | 6 +++++-
 3 files changed, 13 insertions(+), 1 deletion(-)
drivers/net/virtio/virtio_user/virtio_user_dev.c +7 -0
index 3681758ef..27814eadb 100644
@@ -22,6 +22,13 @@
 
 #define VIRTIO_USER_MEM_EVENT_CLB_NAME "virtio_user_mem_event_clb"
 
+const char * const virtio_user_backend_strings[] = {
+	[VIRTIO_USER_BACKEND_UNKNOWN] = "VIRTIO_USER_BACKEND_UNKNOWN",
+	[VIRTIO_USER_BACKEND_VHOST_USER] = "VHOST_USER",
+	[VIRTIO_USER_BACKEND_VHOST_KERNEL] = "VHOST_NET",
+	[VIRTIO_USER_BACKEND_VHOST_VDPA] = "VHOST_VDPA",
+};
+
 static int
 virtio_user_create_queue(struct virtio_user_dev *dev, uint32_t queue_sel)
 {
drivers/net/virtio/virtio_user/virtio_user_dev.h +1 -0
index 3e9d1a1eb..998986875 100644
@@ -83,4 +83,5 @@ void virtio_user_handle_cq_packed(struct virtio_user_dev *dev,
 uint8_t virtio_user_handle_mq(struct virtio_user_dev *dev, uint16_t q_pairs);
 int virtio_user_send_status_update(struct virtio_user_dev *dev, uint8_t status);
 int virtio_user_update_status(struct virtio_user_dev *dev);
+extern const char * const virtio_user_backend_strings[];
 #endif
drivers/net/virtio/virtio_user_ethdev.c +5 -1
index 042665bc0..e870fb2ff 100644
@@ -560,6 +560,9 @@ virtio_user_backend_type(const char *path)
 	struct stat sb;
 
 	if (stat(path, &sb) == -1) {
+		if (errno == ENOENT)
+			return VIRTIO_USER_BACKEND_VHOST_USER;
+
 		PMD_INIT_LOG(ERR, "Stat fails: %s (%s)\n", path,
 			     strerror(errno));
 		return VIRTIO_USER_BACKEND_UNKNOWN;
@@ -697,7 +700,8 @@ virtio_user_pmd_probe(struct rte_vdev_device *dev)
 			path);
 		goto end;
 	}
-
+	PMD_INIT_LOG(INFO, "Backend type detected: %s",
+		     virtio_user_backend_strings[backend_type]);
 
 	if (rte_kvargs_count(kvlist, VIRTIO_USER_ARG_INTERFACE_NAME) == 1) {
 		if (backend_type != VIRTIO_USER_BACKEND_VHOST_KERNEL) {
Signature · 1 line
2.26.2
Reviewed-by on v2 1/3
#2 lore raw

Re: [dpdk-dev] [PATCH v2 1/3] virtio-user: fix backend selection if stat fails

Quoted 9 lines · Adrian Moreno

If stat fails because the file does not exist, it means that the backend must be vhost-user in server mode.

Also, log the detected backend type.

Bugzilla ID: 559

Fixes f908b22ea47a ("net/virtio: move backend type selection to ethdev")
Quoted 7 lines
---
 drivers/net/virtio/virtio_user/virtio_user_dev.c | 7 +++++++
 drivers/net/virtio/virtio_user/virtio_user_dev.h | 1 +
 drivers/net/virtio/virtio_user_ethdev.c          | 6 +++++-
 3 files changed, 13 insertions(+), 1 deletion(-)

Thanks,
Maxime

Acked-by on v2 1/3
#3 lore raw

Re: [dpdk-dev] [PATCH v2 1/3] virtio-user: fix backend selection if stat fails

In reply to Adrian Moreno
Quoted 9 lines · Adrian Moreno

If stat fails because the file does not exist, it means that the backend must be vhost-user in server mode.

Also, log the detected backend type.

Bugzilla ID: 559

Fixes f908b22ea47a ("net/virtio: move backend type selection to ethdev")
Quoted 6 lines
---
 drivers/net/virtio/virtio_user/virtio_user_dev.c | 7 +++++++
 drivers/net/virtio/virtio_user/virtio_user_dev.h | 1 +
 drivers/net/virtio/virtio_user_ethdev.c          | 6 +++++-
 3 files changed, 13 insertions(+), 1 deletion(-)
#4 lore raw

PATCH v2 2/3 virtio_user: don't set/get_status until FEATURES_OK

In reply to Adrian Moreno

According to the virtio spec, ACK and DRIVER status bits should be set before feature negotiation.

However, until the protocol features are negotiated, the driver does not know if the device actually supports the those vhost-user messages. Therefore, until FEATURES_OK is set, the GET/SET_STATUS messages should not be sent.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")
---
 drivers/net/virtio/virtio_user/vhost_user.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
drivers/net/virtio/virtio_user/vhost_user.c +6 -4
index ef290c357..450d77e92 100644
@@ -278,8 +278,9 @@ vhost_user_sock(struct virtio_user_dev *dev,
 
 	switch (req) {
 	case VHOST_USER_GET_STATUS:
-		if (!(dev->protocol_features &
-				(1ULL << VHOST_USER_PROTOCOL_F_STATUS)))
+		if (!(dev->status & VIRTIO_CONFIG_STATUS_FEATURES_OK) ||
+		    (!(dev->protocol_features &
+				(1ULL << VHOST_USER_PROTOCOL_F_STATUS))))
 			return 0;
 		/* Fallthrough */
 	case VHOST_USER_GET_FEATURES:
@@ -288,8 +289,9 @@ vhost_user_sock(struct virtio_user_dev *dev,
 		break;
 
 	case VHOST_USER_SET_STATUS:
-		if (!(dev->protocol_features &
-				(1ULL << VHOST_USER_PROTOCOL_F_STATUS)))
+		if (!(dev->status & VIRTIO_CONFIG_STATUS_FEATURES_OK) ||
+		    (!(dev->protocol_features &
+				(1ULL << VHOST_USER_PROTOCOL_F_STATUS))))
 			return 0;
 
 		if (has_reply_ack)
Signature · 1 line
2.26.2
Reviewed-by on v2 2/3
#5 lore raw

Re: [dpdk-dev] [PATCH v2 2/3] virtio_user: don't set/get_status until FEATURES_OK

Quoted 5 lines · Adrian Moreno

According to the virtio spec, ACK and DRIVER status bits should be set before feature negotiation.

However, until the protocol features are negotiated, the driver does not know if the device actually supports the those vhost-user messages.

s/the those/those/

Quoted 31 lines

Therefore, until FEATURES_OK is set, the GET/SET_STATUS messages should not be sent.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")
---
 drivers/net/virtio/virtio_user/vhost_user.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
drivers/net/virtio/virtio_user/vhost_user.c +3 -2
index ef290c357..450d77e92 100644
@@ -278,8 +278,9 @@ vhost_user_sock(struct virtio_user_dev *dev,
 
 	switch (req) {
 	case VHOST_USER_GET_STATUS:
-		if (!(dev->protocol_features &
-				(1ULL << VHOST_USER_PROTOCOL_F_STATUS)))
+		if (!(dev->status & VIRTIO_CONFIG_STATUS_FEATURES_OK) ||
+		    (!(dev->protocol_features &
+				(1ULL << VHOST_USER_PROTOCOL_F_STATUS))))
 			return 0;
 		/* Fallthrough */
 	case VHOST_USER_GET_FEATURES:
@@ -288,8 +289,9 @@ vhost_user_sock(struct virtio_user_dev *dev,
 		break;

drivers/net/virtio/virtio_user/vhost_user.c vhost_user_sock(struct virtio_user_dev *dev, · line 297

Quoted 10 lines
drivers/net/virtio/virtio_user/vhost_user.c +3 -2
 
 	case VHOST_USER_SET_STATUS:
-		if (!(dev->protocol_features &
-				(1ULL << VHOST_USER_PROTOCOL_F_STATUS)))
+		if (!(dev->status & VIRTIO_CONFIG_STATUS_FEATURES_OK) ||
+		    (!(dev->protocol_features &
+				(1ULL << VHOST_USER_PROTOCOL_F_STATUS))))
 			return 0;
 
 		if (has_reply_ack)

Thanks,
Maxime

#6 lore raw

PATCH v2 3/3 virtio-user: set status on virtio-user reconnect

In reply to Adrian Moreno

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
drivers/net/virtio/virtio_user_ethdev.c +12 -2
index e870fb2ff..d8bea4537 100644
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 		return -1;
 
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {
Signature · 1 line
2.26.2
Reviewed-by on v2 3/3
#7 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Quoted 42 lines · Adrian Moreno

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
drivers/net/virtio/virtio_user_ethdev.c +9 -0
index e870fb2ff..d8bea4537 100644
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 		return -1;
 
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 		virtio_user_reset_queues_packed(eth_dev);

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev *dev) · line 136

Quoted 10 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime

#8 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Maxime/ Adrian,

Thanks for the patch. we can launch vhost-user with client mode with this fix patch.
But still fail to get throughput with basic vhost/virtio-user server mode loopback test. This is another problem which introduced by 57912824615fd7787a48a7b18e40661466.
Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=541

BR,
Yinan

Quoted 73 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: 2020年10月21日 0:43
To: Adrian Moreno <[email protected]>; [email protected]
Cc: Wang, Yinan <[email protected]>; Fu, Patrick <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>; Wang, Zhihong <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Quoted 7 lines · Adrian Moreno

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")

Review

Quoted 6 lines
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
(quoted diff) +0 -0

b/drivers/net/virtio/virtio_user_ethdev.c

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 4 lines

index e870fb2ff..d8bea4537 100644

drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 14 lines
drivers/net/virtio/virtio_user_ethdev.c +7 -0
 		return -1;
 
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 9 lines
drivers/net/virtio/virtio_user_ethdev.c +2 -0
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev *dev) · line 136

Quoted 12 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime

#9 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Yinan,

Quoted 5 lines · Yinan

Hi Maxime/ Adrian,

Thanks for the patch. we can launch vhost-user with client mode with this fix patch.
But still fail to get throughput with basic vhost/virtio-user server mode loopback test. This is another problem which introduced by 57912824615fd7787a48a7b18e40661466.
Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=541

Thanks for reporting the issue, I will look at it this morning.

BTW, this virtio-user server mode is broken by design, we should really fix it.

For example, it assumes features to be supported by the backend before the negotiation took place:
https://git.dpdk.org/dpdk/tree/drivers/net/virtio/virtio_user/virtio_user_dev.c#n513

As part of my rework, I suggest we implement the same behaviour as QEMU does, which will be reliable and consistent with QEMU. It means that if server mode is enabled in device command-line, the driver waits until the socket is ready.

Chenbo, Adrian, what do you think?

Thanks,
Maxime

Quoted 76 lines

BR,
Yinan

Quoted 73 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: 2020年10月21日 0:43
To: Adrian Moreno <[email protected]>; [email protected]
Cc: Wang, Yinan <[email protected]>; Fu, Patrick <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>; Wang, Zhihong <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Quoted 7 lines · Adrian Moreno

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")

Review

Quoted 6 lines
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
(quoted diff) +0 -0

b/drivers/net/virtio/virtio_user_ethdev.c

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 4 lines

index e870fb2ff..d8bea4537 100644

drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 14 lines
drivers/net/virtio/virtio_user_ethdev.c +7 -0
 		return -1;
 
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 9 lines
drivers/net/virtio/virtio_user_ethdev.c +2 -0
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev *dev) · line 136

Quoted 12 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime

#10 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Maxime,

Quoted 32 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, October 22, 2020 3:14 PM
To: Wang, Yinan <[email protected]>; Adrian Moreno <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>
Cc: Fu, Patrick <[email protected]>; [email protected]; Wang, Zhihong <[email protected]>; Xu, Qian Q <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Yinan,

Quoted 3 lines · Yinan

Hi Maxime/ Adrian,

Thanks for the patch. we can launch vhost-user with client mode with

this fix patch.

Quoted 1 line

But still fail to get throughput with basic vhost/virtio-user server

mode loopback test. This is another problem which introduced by 57912824615fd7787a48a7b18e40661466.

Quoted 1 line

Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=541

Thanks for reporting the issue, I will look at it this morning.

BTW, this virtio-user server mode is broken by design, we should really fix it.

For example, it assumes features to be supported by the backend before the negotiation took place:
https://git.dpdk.org/dpdk/tree/drivers/net/virtio/virtio_user/virtio_user_
dev.c#n513

Quoted 6 lines

As part of my rework, I suggest we implement the same behaviour as QEMU does, which will be reliable and consistent with QEMU. It means that if server mode is enabled in device command-line, the driver waits until the socket is ready.

Chenbo, Adrian, what do you think?

Yes! I totally agree and have the same opinion for a long time. As I remember, this feature assumption has caused problems before and I notice the new STATUS feature is also affected. It's good to make this consistent with QEMU 😊.

Cheers,
Chenbo

Quoted 84 lines

Thanks,
Maxime

Quoted 7 lines

BR,
Yinan

Quoted 5 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: 2020年10月21日 0:43
To: Adrian Moreno <[email protected]>; [email protected]
Cc: Wang, Yinan <[email protected]>; Fu, Patrick

<[email protected]>;

Quoted 3 lines
Quoted 3 lines

[email protected]; Xia, Chenbo <[email protected]>; Wang, Zhihong <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user

reconnect

Quoted 25 lines
Quoted 25 lines · Adrian Moreno
Quoted 7 lines

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")

Review

Quoted 6 lines
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
(quoted diff) +0 -0

b/drivers/net/virtio/virtio_user_ethdev.c

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 4 lines

index e870fb2ff..d8bea4537 100644

drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

Quoted 2 lines
drivers/net/virtio/virtio_user_ethdev.c +0 -0
 		return -1;
 
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -0
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
Quoted 8 lines

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 8 lines
drivers/net/virtio/virtio_user_ethdev.c +4 -0
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 10 lines
Quoted 10 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 9 lines
drivers/net/virtio/virtio_user_ethdev.c +2 -0
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 19 lines
Quoted 18 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev *dev) · line 136

Quoted 12 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime

#11 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Quoted 39 lines · Chenbo

Hi Maxime,

Quoted 37 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, October 22, 2020 3:14 PM
To: Wang, Yinan <[email protected]>; Adrian Moreno <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>
Cc: Fu, Patrick <[email protected]>; [email protected]; Wang, Zhihong <[email protected]>; Xu, Qian Q <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Yinan,

Quoted 3 lines · Yinan

Hi Maxime/ Adrian,

Thanks for the patch. we can launch vhost-user with client mode with

this fix patch.

Quoted 1 line

But still fail to get throughput with basic vhost/virtio-user server

mode loopback test. This is another problem which introduced by 57912824615fd7787a48a7b18e40661466.

Quoted 1 line

Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=541

Thanks for reporting the issue, I will look at it this morning.

BTW, this virtio-user server mode is broken by design, we should really fix it.

For example, it assumes features to be supported by the backend before the negotiation took place:
https://git.dpdk.org/dpdk/tree/drivers/net/virtio/virtio_user/virtio_user_
dev.c#n513

As part of my rework, I suggest we implement the same behaviour as QEMU does, which will be reliable and consistent with QEMU. It means that if server mode is enabled in device command-line, the driver waits until the socket is ready.

Quoted 6 lines
Quoted 1 line

Chenbo, Adrian, what do you think?

Yes! I totally agree and have the same opinion for a long time. As I remember, this feature assumption has caused problems before and I notice the new STATUS feature is also affected. It's good to make this consistent with QEMU 😊.

+1

In fact, I was looking into how to get rid of some error messages originated during the initialization of the virtio-pci layer when the socket is still not ready.

Quoted 88 lines

Cheers,
Chenbo

Quoted 84 lines

Thanks,
Maxime

Quoted 7 lines

BR,
Yinan

Quoted 5 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: 2020年10月21日 0:43
To: Adrian Moreno <[email protected]>; [email protected]
Cc: Wang, Yinan <[email protected]>; Fu, Patrick

<[email protected]>;

Quoted 3 lines
Quoted 3 lines

[email protected]; Xia, Chenbo <[email protected]>; Wang, Zhihong <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user

reconnect

Quoted 25 lines
Quoted 25 lines · Adrian Moreno
Quoted 7 lines

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")

Review

Quoted 6 lines
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
(quoted diff) +0 -0

b/drivers/net/virtio/virtio_user_ethdev.c

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 4 lines

index e870fb2ff..d8bea4537 100644

drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

Quoted 2 lines
drivers/net/virtio/virtio_user_ethdev.c +0 -0
 		return -1;
 
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -0
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
Quoted 8 lines

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 8 lines
drivers/net/virtio/virtio_user_ethdev.c +4 -0
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 10 lines
Quoted 10 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 9 lines
drivers/net/virtio/virtio_user_ethdev.c +2 -0
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 19 lines
Quoted 18 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev *dev) · line 136

Quoted 12 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime

Signature · 1 line
Adrián Moreno
#12 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Quoted 48 lines · Adrian Moreno
Quoted 39 lines · Chenbo

Hi Maxime,

Quoted 37 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, October 22, 2020 3:14 PM
To: Wang, Yinan <[email protected]>; Adrian Moreno <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>
Cc: Fu, Patrick <[email protected]>; [email protected]; Wang, Zhihong <[email protected]>; Xu, Qian Q <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Yinan,

Quoted 3 lines · Yinan

Hi Maxime/ Adrian,

Thanks for the patch. we can launch vhost-user with client mode with

this fix patch.

Quoted 1 line

But still fail to get throughput with basic vhost/virtio-user server

mode loopback test. This is another problem which introduced by 57912824615fd7787a48a7b18e40661466.

Quoted 1 line

Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=541

Thanks for reporting the issue, I will look at it this morning.

BTW, this virtio-user server mode is broken by design, we should really fix it.

For example, it assumes features to be supported by the backend before the negotiation took place:
https://git.dpdk.org/dpdk/tree/drivers/net/virtio/virtio_user/virtio_user_
dev.c#n513

As part of my rework, I suggest we implement the same behaviour as QEMU does, which will be reliable and consistent with QEMU. It means that if server mode is enabled in device command-line, the driver waits until the socket is ready.

Quoted 6 lines
Quoted 1 line

Chenbo, Adrian, what do you think?

Yes! I totally agree and have the same opinion for a long time. As I remember, this feature assumption has caused problems before and I notice the new STATUS feature is also affected. It's good to make this consistent with QEMU 😊.

Quoted 5 lines

+1

In fact, I was looking into how to get rid of some error messages originated during the initialization of the virtio-pci layer when the socket is still not ready.

Is that the error messages you meant?

virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1): Success

That would be good to have them fixed

Quoted 90 lines
Quoted 89 lines

Cheers,
Chenbo

Quoted 84 lines

Thanks,
Maxime

Quoted 7 lines

BR,
Yinan

Quoted 5 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: 2020年10月21日 0:43
To: Adrian Moreno <[email protected]>; [email protected]
Cc: Wang, Yinan <[email protected]>; Fu, Patrick

<[email protected]>;

Quoted 3 lines
Quoted 3 lines

[email protected]; Xia, Chenbo <[email protected]>; Wang, Zhihong <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user

reconnect

Quoted 25 lines
Quoted 25 lines · Adrian Moreno
Quoted 7 lines

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status setting")

Review

Quoted 6 lines
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
(quoted diff) +0 -0

b/drivers/net/virtio/virtio_user_ethdev.c

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev

Quoted 4 lines

index e870fb2ff..d8bea4537 100644

drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct virtio_user_dev

*dev)

Quoted 2 lines
drivers/net/virtio/virtio_user_ethdev.c +0 -0
 		return -1;
 
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -0
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
Quoted 8 lines

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 8 lines
drivers/net/virtio/virtio_user_ethdev.c +4 -0
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev, VHOST_USER_GET_FEATURES,
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s",
@@ -111,6 +118,8 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 10 lines
Quoted 10 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 9 lines
drivers/net/virtio/virtio_user_ethdev.c +2 -0
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_FEATURES_OK);
+
 	/* For packed ring, resetting queues is required in reconnection. */
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) & VIRTIO_CONFIG_STATUS_DRIVER_OK)) {
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 19 lines
Quoted 18 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct virtio_user_dev *dev) · line 136

Quoted 12 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime

Tested-by
#13 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Best Regards
Jiang yu

Quoted 194 lines

-----Original Message-----
From: dev [mailto:[email protected]] On Behalf Of Maxime Coquelin
Sent: Thursday, October 22, 2020 4:32 PM
To: Adrian Moreno <[email protected]>; Xia, Chenbo <[email protected]>; Wang, Yinan <[email protected]>; [email protected]
Cc: Fu, Patrick <[email protected]>; [email protected]; Wang, Zhihong <[email protected]>; Xu, Qian Q <[email protected]>
Subject: Re: [dpdk-dev] [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Quoted 46 lines · Adrian Moreno
Quoted 43 lines · Chenbo

Hi Maxime,

Quoted 41 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, October 22, 2020 3:14 PM
To: Wang, Yinan <[email protected]>; Adrian Moreno <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>
Cc: Fu, Patrick <[email protected]>; [email protected]; Wang, Zhihong <[email protected]>; Xu, Qian Q <[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user reconnect

Hi Yinan,

Quoted 4 lines · Yinan

Hi Maxime/ Adrian,

Thanks for the patch. we can launch vhost-user with client mode with

this fix patch.

Quoted 2 lines

But still fail to get throughput with basic vhost/virtio-user server

mode loopback test. This is another problem which introduced by 57912824615fd7787a48a7b18e40661466.

Quoted 1 line

Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=541

Thanks for reporting the issue, I will look at it this morning.

BTW, this virtio-user server mode is broken by design, we should really fix it.

For example, it assumes features to be supported by the backend before the negotiation took place:
https://git.dpdk.org/dpdk/tree/drivers/net/virtio/virtio_user/virtio
_user_
dev.c#n513

As part of my rework, I suggest we implement the same behaviour as QEMU does, which will be reliable and consistent with QEMU. It means that if server mode is enabled in device command-line, the driver waits until the socket is ready.

Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Chenbo, Adrian, what do you think?

Yes! I totally agree and have the same opinion for a long time. As I remember, this feature assumption has caused problems before and I notice the new STATUS feature is also affected. It's good to make this

consistent with QEMU 😊.

Quoted 6 lines
Quoted 1 line

+1

In fact, I was looking into how to get rid of some error messages originated during the initialization of the virtio-pci layer when the socket is still not ready.

Is that the error messages you meant?

virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1):
Success
virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1):
Success
virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1):
Success
virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1):
Success
virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_update_status(): VHOST_USER_GET_STATUS failed (-1): Success virtio_user_send_status_update(): VHOST_USER_SET_STATUS failed (-1):
Success

That would be good to have them fixed

Quoted 10 lines
Quoted 10 lines

Cheers,
Chenbo

Quoted 6 lines

Thanks,
Maxime

Quoted 2 lines

BR,
Yinan

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 5 lines
Quoted 5 lines

-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: 2020年10月21日 0:43
To: Adrian Moreno <[email protected]>; [email protected]
Cc: Wang, Yinan <[email protected]>; Fu, Patrick

<[email protected]>;

Quoted 1 line
Quoted 1 line

[email protected]; Xia, Chenbo <[email protected]>; Wang,

Zhihong

Quoted 30 lines
Quoted 30 lines
Quoted 30 lines
Quoted 2 lines
Quoted 2 lines

<[email protected]>
Subject: Re: [PATCH v2 3/3] virtio-user: set status on virtio-user

reconnect

Quoted 27 lines
Quoted 27 lines · Adrian Moreno
Quoted 8 lines

Newer vhost-user backends will rely on SET_STATUS to start the device so this required to support them.

Fixes 57912824615f ("net/virtio-user: support vhost status

setting")
Cc: [email protected]
Cc: [email protected]

Review

Quoted 6 lines
---
 drivers/net/virtio/virtio_user_ethdev.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
(quoted diff) +0 -0

b/drivers/net/virtio/virtio_user_ethdev.c

Quoted 5 lines

index e870fb2ff..d8bea4537 100644

drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -78,6 +78,13 @@ virtio_user_server_reconnect(struct

virtio_user_dev

*dev)

Quoted 2 lines

return -1;

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

drivers/net/virtio/virtio_user_ethdev.c

Quoted 9 lines
drivers/net/virtio/virtio_user_ethdev.c +7 -0
 	dev->vhostfd = connectfd;
+
+	vtpci_reset(hw);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
+
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER);
+
 	if (dev->ops->send_request(dev,

VHOST_USER_GET_FEATURES,

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/net/virtio/virtio_user_ethdev.c

Quoted 2 lines
drivers/net/virtio/virtio_user_ethdev.c +0 -0
 				   &dev->device_features) < 0) {
 		PMD_INIT_LOG(ERR, "get_features failed: %s", @@ -

111,6 +118,8

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

@@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 5 lines
Quoted 5 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c

Quoted 4 lines
drivers/net/virtio/virtio_user_ethdev.c +1 -0
 	dev->features &= dev->device_features;
 
+	vtpci_set_status(hw,

VIRTIO_CONFIG_STATUS_FEATURES_OK);

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/net/virtio/virtio_user_ethdev.c

Quoted 2 lines
drivers/net/virtio/virtio_user_ethdev.c +1 -0
+
 	/* For packed ring, resetting queues is required in

reconnection. */

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/net/virtio/virtio_user_ethdev.c

Quoted 2 lines
drivers/net/virtio/virtio_user_ethdev.c +0 -0
 	if (vtpci_packed_queue(hw) &&
 	   (vtpci_get_status(hw) &

VIRTIO_CONFIG_STATUS_DRIVER_OK)) {

Quoted 23 lines
Quoted 22 lines
Quoted 21 lines
Quoted 1 line
Quoted 1 line

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct

Quoted 1 line
drivers/net/virtio/virtio_user_ethdev.c +0 -0
@@ -119,8 +128,9 @@ virtio_user_server_reconnect(struct

virtio_user_dev

Quoted 19 lines
Quoted 18 lines

*dev)

drivers/net/virtio/virtio_user_ethdev.c virtio_user_server_reconnect(struct · line 136

Quoted 12 lines
drivers/net/virtio/virtio_user_ethdev.c +3 -2
 		virtio_user_reset_queues_packed(eth_dev);
 	}
 
-	ret = virtio_user_start_device(dev);
-	if (ret < 0)
+	/*  Start the device */
+	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
+	if (!dev->started)
 		return -1;
 
 	if (dev->queue_pairs > 1) {

Thanks,
Maxime