Skip to contentSkip to conversation

vdpa tool enhancements

PATCH Superseded Eli Cohen 3 participants 12 messages Feb 17, 2022 – Feb 21, 2022

Subsystems (2)

v2 of 7 Superseded Latest is v7 →

lore
Share & tools
lore

Eli Cohen

author
#0 lore raw

v2 cover vdpa tool enhancements

The following four patch series enhances vdpa to show negotiated features for a vdpa device, max features for a management device and allows to configure max number of virtqueue pairs.

v1->v2:
1. Re-arrange print functions to better reuse code.
2. Fix printing management dev features instead of negotiated features
3. Update help string with max_vqp option

Eli Cohen (4):
vdpa: Remove unsupported command line option
vdpa: Allow for printing negotiated features of a device
vdpa: Support for configuring max VQ pairs for a device
vdpa: Support reading device features

 vdpa/include/uapi/linux/vdpa.h |   4 +
 vdpa/vdpa.c                    | 151 +++++++++++++++++++++++++++++++--
 2 files changed, 148 insertions(+), 7 deletions(-)
Signature · 1 line
2.35.1
author
#1 lore raw

PATCH v2 1/4 vdpa: Remove unsupported command line option

"-v[erbose]" option is not supported.
Remove it.

---
 vdpa/vdpa.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
vdpa/vdpa.c +1 -1
index f048e470c929..4ccb564872a0 100644
@@ -711,7 +711,7 @@ static void help(void)
 	fprintf(stderr,
 		"Usage: vdpa [ OPTIONS ] OBJECT { COMMAND | help }\n"
 		"where  OBJECT := { mgmtdev | dev }\n"
-		"       OPTIONS := { -V[ersion] | -n[o-nice-names] | -j[son] | -p[retty] | -v[erbose] }\n");
+		"       OPTIONS := { -V[ersion] | -n[o-nice-names] | -j[son] | -p[retty] }\n");
 }
 
 static int vdpa_cmd(struct vdpa *vdpa, int argc, char **argv)
Signature · 1 line
2.35.1
Reviewed-by on v2 1/4
#2 lore raw

Re: [PATCH v2 1/4] vdpa: Remove unsupported command line option

Quoted 5 lines · Eli Cohen

"-v[erbose]" option is not supported.
Remove it.

Quoted 18 lines
---
  vdpa/vdpa.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
vdpa/vdpa.c +1 -1
index f048e470c929..4ccb564872a0 100644
@@ -711,7 +711,7 @@ static void help(void)
  	fprintf(stderr,
  		"Usage: vdpa [ OPTIONS ] OBJECT { COMMAND | help }\n"
  		"where  OBJECT := { mgmtdev | dev }\n"
-		"       OPTIONS := { -V[ersion] | -n[o-nice-names] | -j[son] | -p[retty] | -v[erbose] }\n");
+		"       OPTIONS := { -V[ersion] | -n[o-nice-names] | -j[son] | -p[retty] }\n");
  }
 
  static int vdpa_cmd(struct vdpa *vdpa, int argc, char **argv)
author
#3 lore raw

PATCH v2 2/4 vdpa: Allow for printing negotiated features of a device

In reply to Eli Cohen

When reading the configuration of a vdpa device, check if the VDPA_ATTR_DEV_NEGOTIATED_FEATURES is available. If it is, parse the feature bits and print a string representation of each of the feature bits.

We keep the strings in two different arrays. One for net device related devices and one for generic feature bits.

In this patch we parse only net device specific features. Support for other devices can be added later. If the device queried is not a net device, we print its bit number only.

Examples:
1. Standard presentation

$ vdpa dev config show vdpa-a
vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 2 mtu 9000
  negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM

2. json output

$ vdpa -j dev config show vdpa-a
{"config":{"vdpa-a":{"mac":"00:00:00:00:88:88","link":"up","link_announce":false,\
"max_vq_pairs":2,"mtu":9000,"negotiated_features":["CSUM","GUEST_CSUM",\
"MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ","CTRL_MAC_ADDR",\
"VERSION_1","ACCESS_PLATFORM"]}}}

3. Pretty json

$ vdpa -jp dev config show vdpa-a
{
    "config": {
        "vdpa-a": {
            "mac": "00:00:00:00:88:88",
            "link ": "up",
            "link_announce ": false,
            "max_vq_pairs": 2,
            "mtu": 9000,
            "negotiated_features": [
"CSUM","GUEST_CSUM","MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ",\
"MQ","CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
        }
    }
}
---
 vdpa/include/uapi/linux/vdpa.h |   2 +
 vdpa/vdpa.c                    | 108 ++++++++++++++++++++++++++++++++-
 2 files changed, 107 insertions(+), 3 deletions(-)
vdpa/include/uapi/linux/vdpa.h +2 -0
index b7eab069988a..748c350450b2 100644
@@ -40,6 +40,8 @@ enum vdpa_attr {
 	VDPA_ATTR_DEV_NET_CFG_MAX_VQP,		/* u16 */
 	VDPA_ATTR_DEV_NET_CFG_MTU,		/* u16 */
 
+	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
+
 	/* new attributes must be added above here */
 	VDPA_ATTR_MAX,
 };
vdpa/vdpa.c +105 -3

152 lines not on this page · raw message

Signature · 1 line
2.35.1
Reviewed-by on v2 2/4
#4 lore raw

Re: [PATCH v2 2/4] vdpa: Allow for printing negotiated features of a device

Quoted 43 lines · Eli Cohen

When reading the configuration of a vdpa device, check if the VDPA_ATTR_DEV_NEGOTIATED_FEATURES is available. If it is, parse the feature bits and print a string representation of each of the feature bits.

We keep the strings in two different arrays. One for net device related devices and one for generic feature bits.

In this patch we parse only net device specific features. Support for other devices can be added later. If the device queried is not a net device, we print its bit number only.

Examples:
1. Standard presentation

$ vdpa dev config show vdpa-a
vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 2 mtu 9000
   negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM

2. json output

$ vdpa -j dev config show vdpa-a
{"config":{"vdpa-a":{"mac":"00:00:00:00:88:88","link":"up","link_announce":false,\
"max_vq_pairs":2,"mtu":9000,"negotiated_features":["CSUM","GUEST_CSUM",\
"MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ","CTRL_MAC_ADDR",\
"VERSION_1","ACCESS_PLATFORM"]}}}

3. Pretty json

$ vdpa -jp dev config show vdpa-a
{
     "config": {
         "vdpa-a": {
             "mac": "00:00:00:00:88:88",
             "link ": "up",
             "link_announce ": false,
             "max_vq_pairs": 2,
             "mtu": 9000,
             "negotiated_features": [
"CSUM","GUEST_CSUM","MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ",\
"MQ","CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
         }
     }
}
Quoted 1 line

Other than the VIRTIO_F_NOTIFICATION_DATA definition you'd need to include from uapi header file, the changes in this patch look good to me overall. So,

Quoted 181 lines
---
  vdpa/include/uapi/linux/vdpa.h |   2 +
  vdpa/vdpa.c                    | 108 ++++++++++++++++++++++++++++++++-
  2 files changed, 107 insertions(+), 3 deletions(-)
vdpa/include/uapi/linux/vdpa.h +2 -0
index b7eab069988a..748c350450b2 100644
@@ -40,6 +40,8 @@ enum vdpa_attr {
  	VDPA_ATTR_DEV_NET_CFG_MAX_VQP,		/* u16 */
  	VDPA_ATTR_DEV_NET_CFG_MTU,		/* u16 */
 
+	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
+
  	/* new attributes must be added above here */
  	VDPA_ATTR_MAX,
  };
vdpa/vdpa.c +105 -3
index 4ccb564872a0..f60e647b8cf8 100644
@@ -10,6 +10,8 @@
  #include <linux/virtio_net.h>
  #include <linux/netlink.h>
  #include <libmnl/libmnl.h>
+#include <linux/virtio_ring.h>
+#include <linux/virtio_config.h>
  #include "mnl_utils.h"
  #include <rt_names.h>
 
@@ -78,6 +80,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
  	[VDPA_ATTR_DEV_VENDOR_ID] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQS] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
+	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
  };
 
  static int attr_cb(const struct nlattr *attr, void *data)
@@ -385,6 +388,96 @@ static const char *parse_class(int num)
  	return class ? class : "< unknown class >";
  }
 
+static const char * const net_feature_strs[64] = {
+	[VIRTIO_NET_F_CSUM] = "CSUM",
+	[VIRTIO_NET_F_GUEST_CSUM] = "GUEST_CSUM",
+	[VIRTIO_NET_F_CTRL_GUEST_OFFLOADS] = "CTRL_GUEST_OFFLOADS",
+	[VIRTIO_NET_F_MTU] = "MTU",
+	[VIRTIO_NET_F_MAC] = "MAC",
+	[VIRTIO_NET_F_GUEST_TSO4] = "GUEST_TSO4",
+	[VIRTIO_NET_F_GUEST_TSO6] = "GUEST_TSO6",
+	[VIRTIO_NET_F_GUEST_ECN] = "GUEST_ECN",
+	[VIRTIO_NET_F_GUEST_UFO] = "GUEST_UFO",
+	[VIRTIO_NET_F_HOST_TSO4] = "HOST_TSO4",
+	[VIRTIO_NET_F_HOST_TSO6] = "HOST_TSO6",
+	[VIRTIO_NET_F_HOST_ECN] = "HOST_ECN",
+	[VIRTIO_NET_F_HOST_UFO] = "HOST_UFO",
+	[VIRTIO_NET_F_MRG_RXBUF] = "MRG_RXBUF",
+	[VIRTIO_NET_F_STATUS] = "STATUS",
+	[VIRTIO_NET_F_CTRL_VQ] = "CTRL_VQ",
+	[VIRTIO_NET_F_CTRL_RX] = "CTRL_RX",
+	[VIRTIO_NET_F_CTRL_VLAN] = "CTRL_VLAN",
+	[VIRTIO_NET_F_CTRL_RX_EXTRA] = "CTRL_RX_EXTRA",
+	[VIRTIO_NET_F_GUEST_ANNOUNCE] = "GUEST_ANNOUNCE",
+	[VIRTIO_NET_F_MQ] = "MQ",
+	[VIRTIO_F_NOTIFY_ON_EMPTY] = "NOTIFY_ON_EMPTY",
+	[VIRTIO_NET_F_CTRL_MAC_ADDR] = "CTRL_MAC_ADDR",
+	[VIRTIO_F_ANY_LAYOUT] = "ANY_LAYOUT",
+	[VIRTIO_NET_F_RSC_EXT] = "RSC_EXT",
+	[VIRTIO_NET_F_HASH_REPORT] = "HASH_REPORT",
+	[VIRTIO_NET_F_RSS] = "RSS",
+	[VIRTIO_NET_F_STANDBY] = "STANDBY",
+	[VIRTIO_NET_F_SPEED_DUPLEX] = "SPEED_DUPLEX",
+};
+
+#define VIRTIO_F_IN_ORDER 35
+#define VIRTIO_F_NOTIFICATION_DATA 38
+#define VDPA_EXT_FEATURES_SZ (VIRTIO_TRANSPORT_F_END - \
+			      VIRTIO_TRANSPORT_F_START + 1)
+
+static const char * const ext_feature_strs[VDPA_EXT_FEATURES_SZ] = {
+	[VIRTIO_RING_F_INDIRECT_DESC - VIRTIO_TRANSPORT_F_START] = "RING_INDIRECT_DESC",
+	[VIRTIO_RING_F_EVENT_IDX - VIRTIO_TRANSPORT_F_START] = "RING_EVENT_IDX",
+	[VIRTIO_F_VERSION_1 - VIRTIO_TRANSPORT_F_START] = "VERSION_1",
+	[VIRTIO_F_ACCESS_PLATFORM - VIRTIO_TRANSPORT_F_START] = "ACCESS_PLATFORM",
+	[VIRTIO_F_RING_PACKED - VIRTIO_TRANSPORT_F_START] = "RING_PACKED",
+	[VIRTIO_F_IN_ORDER - VIRTIO_TRANSPORT_F_START] = "IN_ORDER",
+	[VIRTIO_F_ORDER_PLATFORM - VIRTIO_TRANSPORT_F_START] = "ORDER_PLATFORM",
+	[VIRTIO_F_SR_IOV - VIRTIO_TRANSPORT_F_START] = "SR_IOV",
+	[VIRTIO_F_NOTIFICATION_DATA - VIRTIO_TRANSPORT_F_START] = "NOTIFICATION_DATA",
+};
+
+static const char * const *dev_to_feature_str[] = {
+	[VIRTIO_ID_NET] = net_feature_strs,
+};
+
+static void print_features(struct vdpa *vdpa, uint64_t features, bool mgmtdevf,
+			   uint16_t dev_id)
+{
+	const char * const *feature_strs = NULL;
+	const char *s;
+	int i;
+
+	if (dev_id < ARRAY_SIZE(dev_to_feature_str))
+		feature_strs = dev_to_feature_str[dev_id];
+
+	if (mgmtdevf)
+		pr_out_array_start(vdpa, "dev_features");
+	else
+		pr_out_array_start(vdpa, "negotiated_features");
+
+	for (i = 0; i < 64; i++) {
+		if (!(features & (1ULL << i)))
+			continue;
+
+		if (i < VIRTIO_TRANSPORT_F_START || i > VIRTIO_TRANSPORT_F_END) {
+			if (feature_strs) {
+				s = feature_strs[i];
+			} else {
+				s = NULL;
+			}
+		} else {
+			s = ext_feature_strs[i - VIRTIO_TRANSPORT_F_START];
+		}
+		if (!s)
+			print_uint(PRINT_ANY, NULL, " bit_%d", i);
+		else
+			print_string(PRINT_ANY, NULL, " %s", s);
+	}
+
+	pr_out_array_end(vdpa);
+}
+
  static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  				struct nlattr **tb)
  {
@@ -395,7 +488,6 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
 
  	if (tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]) {
  		uint64_t classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
-
  		pr_out_array_start(vdpa, "supported_classes");
 
  		for (i = 1; i < 64; i++) {
@@ -579,9 +671,10 @@ static int cmd_dev_del(struct vdpa *vdpa,  int argc, char **argv)
  	return mnlu_gen_socket_sndrcv(&vdpa->nlg, nlh, NULL, NULL);
  }
 
-static void pr_out_dev_net_config(struct nlattr **tb)
+static void pr_out_dev_net_config(struct vdpa *vdpa, struct nlattr **tb)
  {
  	SPRINT_BUF(macaddr);
+	uint64_t val_u64;
  	uint16_t val_u16;
 
  	if (tb[VDPA_ATTR_DEV_NET_CFG_MACADDR]) {
@@ -610,6 +703,15 @@ static void pr_out_dev_net_config(struct nlattr **tb)
  		val_u16 = mnl_attr_get_u16(tb[VDPA_ATTR_DEV_NET_CFG_MTU]);
  		print_uint(PRINT_ANY, "mtu", "mtu %d ", val_u16);
  	}
+	if (tb[VDPA_ATTR_DEV_NEGOTIATED_FEATURES]) {
+		uint16_t dev_id = 0;
+
+		if (tb[VDPA_ATTR_DEV_ID])
+			dev_id = mnl_attr_get_u32(tb[VDPA_ATTR_DEV_ID]);
+
+		val_u64 = mnl_attr_get_u64(tb[VDPA_ATTR_DEV_NEGOTIATED_FEATURES]);
+		print_features(vdpa, val_u64, false, dev_id);
+	}
  }
 
  static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb)
@@ -619,7 +721,7 @@ static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb)
  	pr_out_vdev_handle_start(vdpa, tb);
  	switch (device_id) {
  	case VIRTIO_ID_NET:
-		pr_out_dev_net_config(tb);
+		pr_out_dev_net_config(vdpa, tb);
  		break;
  	default:
  		break;
#5 lore raw

Re: [PATCH v2 2/4] vdpa: Allow for printing negotiated features of a device

In reply to Eli Cohen
Quoted 181 lines · Eli Cohen

When reading the configuration of a vdpa device, check if the VDPA_ATTR_DEV_NEGOTIATED_FEATURES is available. If it is, parse the feature bits and print a string representation of each of the feature bits.

We keep the strings in two different arrays. One for net device related devices and one for generic feature bits.

In this patch we parse only net device specific features. Support for other devices can be added later. If the device queried is not a net device, we print its bit number only.

Examples:
1. Standard presentation

$ vdpa dev config show vdpa-a
vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 2 mtu 9000
   negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM

2. json output

$ vdpa -j dev config show vdpa-a
{"config":{"vdpa-a":{"mac":"00:00:00:00:88:88","link":"up","link_announce":false,\
"max_vq_pairs":2,"mtu":9000,"negotiated_features":["CSUM","GUEST_CSUM",\
"MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ","CTRL_MAC_ADDR",\
"VERSION_1","ACCESS_PLATFORM"]}}}

3. Pretty json

$ vdpa -jp dev config show vdpa-a
{
     "config": {
         "vdpa-a": {
             "mac": "00:00:00:00:88:88",
             "link ": "up",
             "link_announce ": false,
             "max_vq_pairs": 2,
             "mtu": 9000,
             "negotiated_features": [
"CSUM","GUEST_CSUM","MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ",\
"MQ","CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
         }
     }
}
---
  vdpa/include/uapi/linux/vdpa.h |   2 +
  vdpa/vdpa.c                    | 108 ++++++++++++++++++++++++++++++++-
  2 files changed, 107 insertions(+), 3 deletions(-)
vdpa/include/uapi/linux/vdpa.h +2 -0
index b7eab069988a..748c350450b2 100644
@@ -40,6 +40,8 @@ enum vdpa_attr {
  	VDPA_ATTR_DEV_NET_CFG_MAX_VQP,		/* u16 */
  	VDPA_ATTR_DEV_NET_CFG_MTU,		/* u16 */
 
+	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
+
  	/* new attributes must be added above here */
  	VDPA_ATTR_MAX,
  };
vdpa/vdpa.c +93 -0
index 4ccb564872a0..f60e647b8cf8 100644
@@ -10,6 +10,8 @@
  #include <linux/virtio_net.h>
  #include <linux/netlink.h>
  #include <libmnl/libmnl.h>
+#include <linux/virtio_ring.h>
+#include <linux/virtio_config.h>
  #include "mnl_utils.h"
  #include <rt_names.h>
 
@@ -78,6 +80,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
  	[VDPA_ATTR_DEV_VENDOR_ID] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQS] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
+	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
  };
 
  static int attr_cb(const struct nlattr *attr, void *data)
@@ -385,6 +388,96 @@ static const char *parse_class(int num)
  	return class ? class : "< unknown class >";
  }
 
+static const char * const net_feature_strs[64] = {
+	[VIRTIO_NET_F_CSUM] = "CSUM",
+	[VIRTIO_NET_F_GUEST_CSUM] = "GUEST_CSUM",
+	[VIRTIO_NET_F_CTRL_GUEST_OFFLOADS] = "CTRL_GUEST_OFFLOADS",
+	[VIRTIO_NET_F_MTU] = "MTU",
+	[VIRTIO_NET_F_MAC] = "MAC",
+	[VIRTIO_NET_F_GUEST_TSO4] = "GUEST_TSO4",
+	[VIRTIO_NET_F_GUEST_TSO6] = "GUEST_TSO6",
+	[VIRTIO_NET_F_GUEST_ECN] = "GUEST_ECN",
+	[VIRTIO_NET_F_GUEST_UFO] = "GUEST_UFO",
+	[VIRTIO_NET_F_HOST_TSO4] = "HOST_TSO4",
+	[VIRTIO_NET_F_HOST_TSO6] = "HOST_TSO6",
+	[VIRTIO_NET_F_HOST_ECN] = "HOST_ECN",
+	[VIRTIO_NET_F_HOST_UFO] = "HOST_UFO",
+	[VIRTIO_NET_F_MRG_RXBUF] = "MRG_RXBUF",
+	[VIRTIO_NET_F_STATUS] = "STATUS",
+	[VIRTIO_NET_F_CTRL_VQ] = "CTRL_VQ",
+	[VIRTIO_NET_F_CTRL_RX] = "CTRL_RX",
+	[VIRTIO_NET_F_CTRL_VLAN] = "CTRL_VLAN",
+	[VIRTIO_NET_F_CTRL_RX_EXTRA] = "CTRL_RX_EXTRA",
+	[VIRTIO_NET_F_GUEST_ANNOUNCE] = "GUEST_ANNOUNCE",
+	[VIRTIO_NET_F_MQ] = "MQ",
+	[VIRTIO_F_NOTIFY_ON_EMPTY] = "NOTIFY_ON_EMPTY",
+	[VIRTIO_NET_F_CTRL_MAC_ADDR] = "CTRL_MAC_ADDR",
+	[VIRTIO_F_ANY_LAYOUT] = "ANY_LAYOUT",
+	[VIRTIO_NET_F_RSC_EXT] = "RSC_EXT",
+	[VIRTIO_NET_F_HASH_REPORT] = "HASH_REPORT",
+	[VIRTIO_NET_F_RSS] = "RSS",
+	[VIRTIO_NET_F_STANDBY] = "STANDBY",
+	[VIRTIO_NET_F_SPEED_DUPLEX] = "SPEED_DUPLEX",
+};
+
+#define VIRTIO_F_IN_ORDER 35
+#define VIRTIO_F_NOTIFICATION_DATA 38
+#define VDPA_EXT_FEATURES_SZ (VIRTIO_TRANSPORT_F_END - \
+			      VIRTIO_TRANSPORT_F_START + 1)
+
+static const char * const ext_feature_strs[VDPA_EXT_FEATURES_SZ] = {
+	[VIRTIO_RING_F_INDIRECT_DESC - VIRTIO_TRANSPORT_F_START] = "RING_INDIRECT_DESC",
+	[VIRTIO_RING_F_EVENT_IDX - VIRTIO_TRANSPORT_F_START] = "RING_EVENT_IDX",
+	[VIRTIO_F_VERSION_1 - VIRTIO_TRANSPORT_F_START] = "VERSION_1",
+	[VIRTIO_F_ACCESS_PLATFORM - VIRTIO_TRANSPORT_F_START] = "ACCESS_PLATFORM",
+	[VIRTIO_F_RING_PACKED - VIRTIO_TRANSPORT_F_START] = "RING_PACKED",
+	[VIRTIO_F_IN_ORDER - VIRTIO_TRANSPORT_F_START] = "IN_ORDER",
+	[VIRTIO_F_ORDER_PLATFORM - VIRTIO_TRANSPORT_F_START] = "ORDER_PLATFORM",
+	[VIRTIO_F_SR_IOV - VIRTIO_TRANSPORT_F_START] = "SR_IOV",
+	[VIRTIO_F_NOTIFICATION_DATA - VIRTIO_TRANSPORT_F_START] = "NOTIFICATION_DATA",
+};
+
+static const char * const *dev_to_feature_str[] = {
+	[VIRTIO_ID_NET] = net_feature_strs,
+};
+
+static void print_features(struct vdpa *vdpa, uint64_t features, bool mgmtdevf,
+			   uint16_t dev_id)
+{
+	const char * const *feature_strs = NULL;
+	const char *s;
+	int i;
+
+	if (dev_id < ARRAY_SIZE(dev_to_feature_str))
+		feature_strs = dev_to_feature_str[dev_id];
+
+	if (mgmtdevf)
+		pr_out_array_start(vdpa, "dev_features");
+	else
+		pr_out_array_start(vdpa, "negotiated_features");
+
+	for (i = 0; i < 64; i++) {
+		if (!(features & (1ULL << i)))
+			continue;
+
+		if (i < VIRTIO_TRANSPORT_F_START || i > VIRTIO_TRANSPORT_F_END) {
+			if (feature_strs) {
+				s = feature_strs[i];
+			} else {
+				s = NULL;
+			}
+		} else {
+			s = ext_feature_strs[i - VIRTIO_TRANSPORT_F_START];
+		}
+		if (!s)
+			print_uint(PRINT_ANY, NULL, " bit_%d", i);
+		else
+			print_string(PRINT_ANY, NULL, " %s", s);
+	}
+
+	pr_out_array_end(vdpa);
+}
+
  static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  				struct nlattr **tb)
  {

vdpa/vdpa.c static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh, · line 398

Quoted 4 lines
vdpa/vdpa.c +0 -1
@@ -395,7 +488,6 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
 
  	if (tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]) {
  		uint64_t classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
-

Unnecessary changes.

Quoted 32 lines
vdpa/vdpa.c +11 -1
  		pr_out_array_start(vdpa, "supported_classes");
 
  		for (i = 1; i < 64; i++) {
@@ -579,9 +671,10 @@ static int cmd_dev_del(struct vdpa *vdpa,  int argc, char **argv)
  	return mnlu_gen_socket_sndrcv(&vdpa->nlg, nlh, NULL, NULL);
  }
 
-static void pr_out_dev_net_config(struct nlattr **tb)
+static void pr_out_dev_net_config(struct vdpa *vdpa, struct nlattr **tb)
  {
  	SPRINT_BUF(macaddr);
+	uint64_t val_u64;
  	uint16_t val_u16;
 
  	if (tb[VDPA_ATTR_DEV_NET_CFG_MACADDR]) {
@@ -610,6 +703,15 @@ static void pr_out_dev_net_config(struct nlattr **tb)
  		val_u16 = mnl_attr_get_u16(tb[VDPA_ATTR_DEV_NET_CFG_MTU]);
  		print_uint(PRINT_ANY, "mtu", "mtu %d ", val_u16);
  	}
+	if (tb[VDPA_ATTR_DEV_NEGOTIATED_FEATURES]) {
+		uint16_t dev_id = 0;
+
+		if (tb[VDPA_ATTR_DEV_ID])
+			dev_id = mnl_attr_get_u32(tb[VDPA_ATTR_DEV_ID]);
+
+		val_u64 = mnl_attr_get_u64(tb[VDPA_ATTR_DEV_NEGOTIATED_FEATURES]);
+		print_features(vdpa, val_u64, false, dev_id);
+	}
  }
 
  static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb)

vdpa/vdpa.c static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb) · line 726

Quoted 7 lines
vdpa/vdpa.c +1 -1
@@ -619,7 +721,7 @@ static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb)
  	pr_out_vdev_handle_start(vdpa, tb);
  	switch (device_id) {
  	case VIRTIO_ID_NET:
-		pr_out_dev_net_config(tb);
+		pr_out_dev_net_config(vdpa, tb);
  		break;
  	default:

I wonder if we need a warning here.

Otherwise looks good to me.

Thanks

Quoted 1 line
vdpa/vdpa.c +0 -0
  		break;
author
#6 lore raw

Re: [PATCH v2 2/4] vdpa: Allow for printing negotiated features of a device

Quoted 177 lines · Jason Wang
Quoted 175 lines · Eli Cohen

When reading the configuration of a vdpa device, check if the VDPA_ATTR_DEV_NEGOTIATED_FEATURES is available. If it is, parse the feature bits and print a string representation of each of the feature bits.

We keep the strings in two different arrays. One for net device related devices and one for generic feature bits.

In this patch we parse only net device specific features. Support for other devices can be added later. If the device queried is not a net device, we print its bit number only.

Examples:
1. Standard presentation

$ vdpa dev config show vdpa-a
vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 2 mtu 9000
   negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM

2. json output

$ vdpa -j dev config show vdpa-a
{"config":{"vdpa-a":{"mac":"00:00:00:00:88:88","link":"up","link_announce":false,\
"max_vq_pairs":2,"mtu":9000,"negotiated_features":["CSUM","GUEST_CSUM",\
"MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ","CTRL_MAC_ADDR",\
"VERSION_1","ACCESS_PLATFORM"]}}}

3. Pretty json

$ vdpa -jp dev config show vdpa-a
{
     "config": {
         "vdpa-a": {
             "mac": "00:00:00:00:88:88",
             "link ": "up",
             "link_announce ": false,
             "max_vq_pairs": 2,
             "mtu": 9000,
             "negotiated_features": [
"CSUM","GUEST_CSUM","MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ",\
"MQ","CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
         }
     }
}
---
  vdpa/include/uapi/linux/vdpa.h |   2 +
  vdpa/vdpa.c                    | 108 ++++++++++++++++++++++++++++++++-
  2 files changed, 107 insertions(+), 3 deletions(-)
vdpa/include/uapi/linux/vdpa.h +2 -0
index b7eab069988a..748c350450b2 100644
@@ -40,6 +40,8 @@ enum vdpa_attr {
  	VDPA_ATTR_DEV_NET_CFG_MAX_VQP,		/* u16 */
  	VDPA_ATTR_DEV_NET_CFG_MTU,		/* u16 */
+	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
+
  	/* new attributes must be added above here */
  	VDPA_ATTR_MAX,
  };
vdpa/vdpa.c +93 -0
index 4ccb564872a0..f60e647b8cf8 100644
@@ -10,6 +10,8 @@
  #include <linux/virtio_net.h>
  #include <linux/netlink.h>
  #include <libmnl/libmnl.h>
+#include <linux/virtio_ring.h>
+#include <linux/virtio_config.h>
  #include "mnl_utils.h"
  #include <rt_names.h>
@@ -78,6 +80,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
  	[VDPA_ATTR_DEV_VENDOR_ID] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQS] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
+	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
  };
  static int attr_cb(const struct nlattr *attr, void *data)
@@ -385,6 +388,96 @@ static const char *parse_class(int num)
  	return class ? class : "< unknown class >";
  }
+static const char * const net_feature_strs[64] = {
+	[VIRTIO_NET_F_CSUM] = "CSUM",
+	[VIRTIO_NET_F_GUEST_CSUM] = "GUEST_CSUM",
+	[VIRTIO_NET_F_CTRL_GUEST_OFFLOADS] = "CTRL_GUEST_OFFLOADS",
+	[VIRTIO_NET_F_MTU] = "MTU",
+	[VIRTIO_NET_F_MAC] = "MAC",
+	[VIRTIO_NET_F_GUEST_TSO4] = "GUEST_TSO4",
+	[VIRTIO_NET_F_GUEST_TSO6] = "GUEST_TSO6",
+	[VIRTIO_NET_F_GUEST_ECN] = "GUEST_ECN",
+	[VIRTIO_NET_F_GUEST_UFO] = "GUEST_UFO",
+	[VIRTIO_NET_F_HOST_TSO4] = "HOST_TSO4",
+	[VIRTIO_NET_F_HOST_TSO6] = "HOST_TSO6",
+	[VIRTIO_NET_F_HOST_ECN] = "HOST_ECN",
+	[VIRTIO_NET_F_HOST_UFO] = "HOST_UFO",
+	[VIRTIO_NET_F_MRG_RXBUF] = "MRG_RXBUF",
+	[VIRTIO_NET_F_STATUS] = "STATUS",
+	[VIRTIO_NET_F_CTRL_VQ] = "CTRL_VQ",
+	[VIRTIO_NET_F_CTRL_RX] = "CTRL_RX",
+	[VIRTIO_NET_F_CTRL_VLAN] = "CTRL_VLAN",
+	[VIRTIO_NET_F_CTRL_RX_EXTRA] = "CTRL_RX_EXTRA",
+	[VIRTIO_NET_F_GUEST_ANNOUNCE] = "GUEST_ANNOUNCE",
+	[VIRTIO_NET_F_MQ] = "MQ",
+	[VIRTIO_F_NOTIFY_ON_EMPTY] = "NOTIFY_ON_EMPTY",
+	[VIRTIO_NET_F_CTRL_MAC_ADDR] = "CTRL_MAC_ADDR",
+	[VIRTIO_F_ANY_LAYOUT] = "ANY_LAYOUT",
+	[VIRTIO_NET_F_RSC_EXT] = "RSC_EXT",
+	[VIRTIO_NET_F_HASH_REPORT] = "HASH_REPORT",
+	[VIRTIO_NET_F_RSS] = "RSS",
+	[VIRTIO_NET_F_STANDBY] = "STANDBY",
+	[VIRTIO_NET_F_SPEED_DUPLEX] = "SPEED_DUPLEX",
+};
+
+#define VIRTIO_F_IN_ORDER 35
+#define VIRTIO_F_NOTIFICATION_DATA 38
+#define VDPA_EXT_FEATURES_SZ (VIRTIO_TRANSPORT_F_END - \
+			      VIRTIO_TRANSPORT_F_START + 1)
+
+static const char * const ext_feature_strs[VDPA_EXT_FEATURES_SZ] = {
+	[VIRTIO_RING_F_INDIRECT_DESC - VIRTIO_TRANSPORT_F_START] = "RING_INDIRECT_DESC",
+	[VIRTIO_RING_F_EVENT_IDX - VIRTIO_TRANSPORT_F_START] = "RING_EVENT_IDX",
+	[VIRTIO_F_VERSION_1 - VIRTIO_TRANSPORT_F_START] = "VERSION_1",
+	[VIRTIO_F_ACCESS_PLATFORM - VIRTIO_TRANSPORT_F_START] = "ACCESS_PLATFORM",
+	[VIRTIO_F_RING_PACKED - VIRTIO_TRANSPORT_F_START] = "RING_PACKED",
+	[VIRTIO_F_IN_ORDER - VIRTIO_TRANSPORT_F_START] = "IN_ORDER",
+	[VIRTIO_F_ORDER_PLATFORM - VIRTIO_TRANSPORT_F_START] = "ORDER_PLATFORM",
+	[VIRTIO_F_SR_IOV - VIRTIO_TRANSPORT_F_START] = "SR_IOV",
+	[VIRTIO_F_NOTIFICATION_DATA - VIRTIO_TRANSPORT_F_START] = "NOTIFICATION_DATA",
+};
+
+static const char * const *dev_to_feature_str[] = {
+	[VIRTIO_ID_NET] = net_feature_strs,
+};
+
+static void print_features(struct vdpa *vdpa, uint64_t features, bool mgmtdevf,
+			   uint16_t dev_id)
+{
+	const char * const *feature_strs = NULL;
+	const char *s;
+	int i;
+
+	if (dev_id < ARRAY_SIZE(dev_to_feature_str))
+		feature_strs = dev_to_feature_str[dev_id];
+
+	if (mgmtdevf)
+		pr_out_array_start(vdpa, "dev_features");
+	else
+		pr_out_array_start(vdpa, "negotiated_features");
+
+	for (i = 0; i < 64; i++) {
+		if (!(features & (1ULL << i)))
+			continue;
+
+		if (i < VIRTIO_TRANSPORT_F_START || i > VIRTIO_TRANSPORT_F_END) {
+			if (feature_strs) {
+				s = feature_strs[i];
+			} else {
+				s = NULL;
+			}
+		} else {
+			s = ext_feature_strs[i - VIRTIO_TRANSPORT_F_START];
+		}
+		if (!s)
+			print_uint(PRINT_ANY, NULL, " bit_%d", i);
+		else
+			print_string(PRINT_ANY, NULL, " %s", s);
+	}
+
+	pr_out_array_end(vdpa);
+}
+
  static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  				struct nlattr **tb)
Quoted 8 lines

vdpa/vdpa.c static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh, · line 397

Quoted 5 lines
vdpa/vdpa.c +0 -1
  {
@@ -395,7 +488,6 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  	if (tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]) {
  		uint64_t classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
-

Unnecessary changes.

Will fix.

Quoted 32 lines
Quoted 32 lines
vdpa/vdpa.c +11 -1
  		pr_out_array_start(vdpa, "supported_classes");
  		for (i = 1; i < 64; i++) {
@@ -579,9 +671,10 @@ static int cmd_dev_del(struct vdpa *vdpa,  int argc, char **argv)
  	return mnlu_gen_socket_sndrcv(&vdpa->nlg, nlh, NULL, NULL);
  }
-static void pr_out_dev_net_config(struct nlattr **tb)
+static void pr_out_dev_net_config(struct vdpa *vdpa, struct nlattr **tb)
  {
  	SPRINT_BUF(macaddr);
+	uint64_t val_u64;
  	uint16_t val_u16;
  	if (tb[VDPA_ATTR_DEV_NET_CFG_MACADDR]) {
@@ -610,6 +703,15 @@ static void pr_out_dev_net_config(struct nlattr **tb)
  		val_u16 = mnl_attr_get_u16(tb[VDPA_ATTR_DEV_NET_CFG_MTU]);
  		print_uint(PRINT_ANY, "mtu", "mtu %d ", val_u16);
  	}
+	if (tb[VDPA_ATTR_DEV_NEGOTIATED_FEATURES]) {
+		uint16_t dev_id = 0;
+
+		if (tb[VDPA_ATTR_DEV_ID])
+			dev_id = mnl_attr_get_u32(tb[VDPA_ATTR_DEV_ID]);
+
+		val_u64 = mnl_attr_get_u64(tb[VDPA_ATTR_DEV_NEGOTIATED_FEATURES]);
+		print_features(vdpa, val_u64, false, dev_id);
+	}
  }
  static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb)
@@ -619,7 +721,7 @@ static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb)
  	pr_out_vdev_handle_start(vdpa, tb);
  	switch (device_id) {
Quoted 8 lines

vdpa/vdpa.c static void pr_out_dev_config(struct vdpa *vdpa, struct nlattr **tb) · line 726

Quoted 5 lines
vdpa/vdpa.c +1 -1
  	case VIRTIO_ID_NET:
-		pr_out_dev_net_config(tb);
+		pr_out_dev_net_config(vdpa, tb);
  		break;
  	default:

I wonder if we need a warning here.

I don't think. You will see the bit_xx printed instead of getting string representation of the feature.

Quoted 8 lines

Otherwise looks good to me.

Thanks

Quoted 1 line
vdpa/vdpa.c +0 -0
  		break;
author
#7 lore raw

PATCH v2 3/4 vdpa: Support for configuring max VQ pairs for a device

In reply to Eli Cohen

Use VDPA_ATTR_DEV_MGMTDEV_MAX_VQS to specify max number of virtqueue pairs to configure for a vdpa device when adding a device.

Examples:
1. Create a device with 3 virtqueue pairs:

$ vdpa dev add name vdpa-a mgmtdev auxiliary/mlx5_core.sf.1 max_vqp 3

2. Read the configuration of a vdpa device

$ vdpa dev config show vdpa-a
  vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 3 \
          mtu 1500
  negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
                      CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
---
 vdpa/include/uapi/linux/vdpa.h |  1 +
 vdpa/vdpa.c                    | 27 +++++++++++++++++++++++++--
 2 files changed, 26 insertions(+), 2 deletions(-)
vdpa/include/uapi/linux/vdpa.h +1 -0
index 748c350450b2..a3ebf4d4d9b8 100644
@@ -41,6 +41,7 @@ enum vdpa_attr {
 	VDPA_ATTR_DEV_NET_CFG_MTU,		/* u16 */
 
 	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
+	VDPA_ATTR_DEV_MGMTDEV_MAX_VQS,          /* u32 */
 
 	/* new attributes must be added above here */
 	VDPA_ATTR_MAX,
vdpa/vdpa.c +25 -2
index f60e647b8cf8..78736b1422b6 100644
@@ -25,6 +25,7 @@
 #define VDPA_OPT_VDEV_HANDLE		BIT(3)
 #define VDPA_OPT_VDEV_MAC		BIT(4)
 #define VDPA_OPT_VDEV_MTU		BIT(5)
+#define VDPA_OPT_MAX_VQP		BIT(6)
 
 struct vdpa_opts {
 	uint64_t present; /* flags of present items */
@@ -34,6 +35,7 @@ struct vdpa_opts {
 	unsigned int device_id;
 	char mac[ETH_ALEN];
 	uint16_t mtu;
+	uint16_t max_vqp;
 };
 
 struct vdpa {
@@ -81,6 +83,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
 	[VDPA_ATTR_DEV_MAX_VQS] = MNL_TYPE_U32,
 	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
 	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
+	[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS] = MNL_TYPE_U32,
 };
 
 static int attr_cb(const struct nlattr *attr, void *data)
@@ -222,6 +225,8 @@ static void vdpa_opts_put(struct nlmsghdr *nlh, struct vdpa *vdpa)
 			     sizeof(opts->mac), opts->mac);
 	if (opts->present & VDPA_OPT_VDEV_MTU)
 		mnl_attr_put_u16(nlh, VDPA_ATTR_DEV_NET_CFG_MTU, opts->mtu);
+	if (opts->present & VDPA_OPT_MAX_VQP)
+		mnl_attr_put_u16(nlh, VDPA_ATTR_DEV_NET_CFG_MAX_VQP, opts->max_vqp);
 }
 
 static int vdpa_argv_parse(struct vdpa *vdpa, int argc, char **argv,
@@ -290,6 +295,14 @@ static int vdpa_argv_parse(struct vdpa *vdpa, int argc, char **argv,
 
 			NEXT_ARG_FWD();
 			o_found |= VDPA_OPT_VDEV_MTU;
+		} else if ((matches(*argv, "max_vqp")  == 0) && (o_optional & VDPA_OPT_MAX_VQP)) {
+			NEXT_ARG_FWD();
+			err = vdpa_argv_u16(vdpa, argc, argv, &opts->max_vqp);
+			if (err)
+				return err;
+
+			NEXT_ARG_FWD();
+			o_found |= VDPA_OPT_MAX_VQP;
 		} else {
 			fprintf(stderr, "Unknown option \"%s\"\n", *argv);
 			return -EINVAL;
@@ -500,6 +513,15 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
 		pr_out_array_end(vdpa);
 	}
 
+	if (tb[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS]) {
+		uint16_t num_vqs;
+
+		if (!vdpa->json_output)
+			printf("\n");
+		num_vqs = mnl_attr_get_u16(tb[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS]);
+		print_uint(PRINT_ANY, "max_supported_vqs", "  max_supported_vqs %d", num_vqs);
+	}
+
 	pr_out_handle_end(vdpa);
 }
 
@@ -559,7 +581,7 @@ static int cmd_mgmtdev(struct vdpa *vdpa, int argc, char **argv)
 static void cmd_dev_help(void)
 {
 	fprintf(stderr, "Usage: vdpa dev show [ DEV ]\n");
-	fprintf(stderr, "       vdpa dev add name NAME mgmtdev MANAGEMENTDEV [ mac MACADDR ] [ mtu MTU ]\n");
+	fprintf(stderr, "       vdpa dev add name NAME mgmtdev MANAGEMENTDEV [ mac MACADDR ] [ mtu MTU ] [max_vqp MAX_VQ_PAIRS]\n");
 	fprintf(stderr, "       vdpa dev del DEV\n");
 	fprintf(stderr, "Usage: vdpa dev config COMMAND [ OPTIONS ]\n");
 }
@@ -649,7 +671,8 @@ static int cmd_dev_add(struct vdpa *vdpa, int argc, char **argv)
 					  NLM_F_REQUEST | NLM_F_ACK);
 	err = vdpa_argv_parse_put(nlh, vdpa, argc, argv,
 				  VDPA_OPT_VDEV_MGMTDEV_HANDLE | VDPA_OPT_VDEV_NAME,
-				  VDPA_OPT_VDEV_MAC | VDPA_OPT_VDEV_MTU);
+				  VDPA_OPT_VDEV_MAC | VDPA_OPT_VDEV_MTU |
+				  VDPA_OPT_MAX_VQP);
 	if (err)
 		return err;
 
Signature · 1 line
2.35.1
Acked-by on v2 3/4
#8 lore raw

Re: [PATCH v2 3/4] vdpa: Support for configuring max VQ pairs for a device

Quoted 8 lines · Eli Cohen

Use VDPA_ATTR_DEV_MGMTDEV_MAX_VQS to specify max number of virtqueue pairs to configure for a vdpa device when adding a device.

Examples:
1. Create a device with 3 virtqueue pairs:

$ vdpa dev add name vdpa-a mgmtdev auxiliary/mlx5_core.sf.1 max_vqp 3

2. Read the configuration of a vdpa device

Quoted 8 lines
$ vdpa dev config show vdpa-a
   vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 3 \
           mtu 1500
   negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
                       CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
Quoted 104 lines
---
  vdpa/include/uapi/linux/vdpa.h |  1 +
  vdpa/vdpa.c                    | 27 +++++++++++++++++++++++++--
  2 files changed, 26 insertions(+), 2 deletions(-)
vdpa/include/uapi/linux/vdpa.h +1 -0
index 748c350450b2..a3ebf4d4d9b8 100644
@@ -41,6 +41,7 @@ enum vdpa_attr {
  	VDPA_ATTR_DEV_NET_CFG_MTU,		/* u16 */
 
  	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
+	VDPA_ATTR_DEV_MGMTDEV_MAX_VQS,          /* u32 */
 
  	/* new attributes must be added above here */
  	VDPA_ATTR_MAX,
vdpa/vdpa.c +25 -2
index f60e647b8cf8..78736b1422b6 100644
@@ -25,6 +25,7 @@
  #define VDPA_OPT_VDEV_HANDLE		BIT(3)
  #define VDPA_OPT_VDEV_MAC		BIT(4)
  #define VDPA_OPT_VDEV_MTU		BIT(5)
+#define VDPA_OPT_MAX_VQP		BIT(6)
 
  struct vdpa_opts {
  	uint64_t present; /* flags of present items */
@@ -34,6 +35,7 @@ struct vdpa_opts {
  	unsigned int device_id;
  	char mac[ETH_ALEN];
  	uint16_t mtu;
+	uint16_t max_vqp;
  };
 
  struct vdpa {
@@ -81,6 +83,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
  	[VDPA_ATTR_DEV_MAX_VQS] = MNL_TYPE_U32,
  	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
  	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
+	[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS] = MNL_TYPE_U32,
  };
 
  static int attr_cb(const struct nlattr *attr, void *data)
@@ -222,6 +225,8 @@ static void vdpa_opts_put(struct nlmsghdr *nlh, struct vdpa *vdpa)
  			     sizeof(opts->mac), opts->mac);
  	if (opts->present & VDPA_OPT_VDEV_MTU)
  		mnl_attr_put_u16(nlh, VDPA_ATTR_DEV_NET_CFG_MTU, opts->mtu);
+	if (opts->present & VDPA_OPT_MAX_VQP)
+		mnl_attr_put_u16(nlh, VDPA_ATTR_DEV_NET_CFG_MAX_VQP, opts->max_vqp);
  }
 
  static int vdpa_argv_parse(struct vdpa *vdpa, int argc, char **argv,
@@ -290,6 +295,14 @@ static int vdpa_argv_parse(struct vdpa *vdpa, int argc, char **argv,
 
  			NEXT_ARG_FWD();
  			o_found |= VDPA_OPT_VDEV_MTU;
+		} else if ((matches(*argv, "max_vqp")  == 0) && (o_optional & VDPA_OPT_MAX_VQP)) {
+			NEXT_ARG_FWD();
+			err = vdpa_argv_u16(vdpa, argc, argv, &opts->max_vqp);
+			if (err)
+				return err;
+
+			NEXT_ARG_FWD();
+			o_found |= VDPA_OPT_MAX_VQP;
  		} else {
  			fprintf(stderr, "Unknown option \"%s\"\n", *argv);
  			return -EINVAL;
@@ -500,6 +513,15 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  		pr_out_array_end(vdpa);
  	}
 
+	if (tb[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS]) {
+		uint16_t num_vqs;
+
+		if (!vdpa->json_output)
+			printf("\n");
+		num_vqs = mnl_attr_get_u16(tb[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS]);
+		print_uint(PRINT_ANY, "max_supported_vqs", "  max_supported_vqs %d", num_vqs);
+	}
+
  	pr_out_handle_end(vdpa);
  }
 
@@ -559,7 +581,7 @@ static int cmd_mgmtdev(struct vdpa *vdpa, int argc, char **argv)
  static void cmd_dev_help(void)
  {
  	fprintf(stderr, "Usage: vdpa dev show [ DEV ]\n");
-	fprintf(stderr, "       vdpa dev add name NAME mgmtdev MANAGEMENTDEV [ mac MACADDR ] [ mtu MTU ]\n");
+	fprintf(stderr, "       vdpa dev add name NAME mgmtdev MANAGEMENTDEV [ mac MACADDR ] [ mtu MTU ] [max_vqp MAX_VQ_PAIRS]\n");
  	fprintf(stderr, "       vdpa dev del DEV\n");
  	fprintf(stderr, "Usage: vdpa dev config COMMAND [ OPTIONS ]\n");
  }
@@ -649,7 +671,8 @@ static int cmd_dev_add(struct vdpa *vdpa, int argc, char **argv)
  					  NLM_F_REQUEST | NLM_F_ACK);
  	err = vdpa_argv_parse_put(nlh, vdpa, argc, argv,
  				  VDPA_OPT_VDEV_MGMTDEV_HANDLE | VDPA_OPT_VDEV_NAME,
-				  VDPA_OPT_VDEV_MAC | VDPA_OPT_VDEV_MTU);
+				  VDPA_OPT_VDEV_MAC | VDPA_OPT_VDEV_MTU |
+				  VDPA_OPT_MAX_VQP);
  	if (err)
  		return err;
 
author
#9 lore raw

PATCH v2 4/4 vdpa: Support reading device features

In reply to Eli Cohen

When showing the available management devices, check if VDPA_ATTR_DEV_SUPPORTED_FEATURES feature is available and print the supported features for a management device.

Examples:

$ vdpa mgmtdev show
auxiliary/mlx5_core.sf.1:
  supported_classes net
  max_supported_vqs 257
  dev_features CSUM GUEST_CSUM MTU HOST_TSO4 HOST_TSO6 STATUS CTRL_VQ MQ \
               CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
$ vdpa -jp mgmtdev show
{
    "mgmtdev": {
        "auxiliary/mlx5_core.sf.1": {
            "supported_classes": [ "net" ],
            "max_supported_vqs": 257,
            "dev_features": [
"CSUM","GUEST_CSUM","MTU","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ",\
"CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
        }
    }
}
---
 vdpa/include/uapi/linux/vdpa.h |  1 +
 vdpa/vdpa.c                    | 14 +++++++++++++-
 2 files changed, 14 insertions(+), 1 deletion(-)
vdpa/include/uapi/linux/vdpa.h +1 -0
index a3ebf4d4d9b8..96ccbf305d14 100644
@@ -42,6 +42,7 @@ enum vdpa_attr {
 
 	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
 	VDPA_ATTR_DEV_MGMTDEV_MAX_VQS,          /* u32 */
+	VDPA_ATTR_DEV_SUPPORTED_FEATURES,	/* u64 */
 
 	/* new attributes must be added above here */
 	VDPA_ATTR_MAX,
vdpa/vdpa.c +13 -1
index 78736b1422b6..bdc366880ab9 100644
@@ -84,6 +84,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
 	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
 	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
 	[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS] = MNL_TYPE_U32,
+	[VDPA_ATTR_DEV_SUPPORTED_FEATURES] = MNL_TYPE_U64,
 };
 
 static int attr_cb(const struct nlattr *attr, void *data)
@@ -494,13 +495,14 @@ static void print_features(struct vdpa *vdpa, uint64_t features, bool mgmtdevf,
 static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
 				struct nlattr **tb)
 {
+	uint64_t classes = 0;
 	const char *class;
 	unsigned int i;
 
 	pr_out_handle_start(vdpa, tb);
 
 	if (tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]) {
-		uint64_t classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
+		classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
 		pr_out_array_start(vdpa, "supported_classes");
 
 		for (i = 1; i < 64; i++) {
@@ -522,6 +524,16 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
 		print_uint(PRINT_ANY, "max_supported_vqs", "  max_supported_vqs %d", num_vqs);
 	}
 
+	if (tb[VDPA_ATTR_DEV_SUPPORTED_FEATURES]) {
+		uint64_t features;
+
+		features  = mnl_attr_get_u64(tb[VDPA_ATTR_DEV_SUPPORTED_FEATURES]);
+		if (classes & BIT(VIRTIO_ID_NET))
+			print_features(vdpa, features, true, VIRTIO_ID_NET);
+		else
+			print_features(vdpa, features, true, 0);
+	}
+
 	pr_out_handle_end(vdpa);
 }
 
Signature · 1 line
2.35.1
#10 lore raw

Re: [PATCH v2 4/4] vdpa: Support reading device features

Quoted 74 lines · Eli Cohen

When showing the available management devices, check if VDPA_ATTR_DEV_SUPPORTED_FEATURES feature is available and print the supported features for a management device.

Examples:

$ vdpa mgmtdev show
auxiliary/mlx5_core.sf.1:
   supported_classes net
   max_supported_vqs 257
   dev_features CSUM GUEST_CSUM MTU HOST_TSO4 HOST_TSO6 STATUS CTRL_VQ MQ \
                CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
$ vdpa -jp mgmtdev show
{
     "mgmtdev": {
         "auxiliary/mlx5_core.sf.1": {
             "supported_classes": [ "net" ],
             "max_supported_vqs": 257,
             "dev_features": [
"CSUM","GUEST_CSUM","MTU","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ",\
"CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
         }
     }
}
---
  vdpa/include/uapi/linux/vdpa.h |  1 +
  vdpa/vdpa.c                    | 14 +++++++++++++-
  2 files changed, 14 insertions(+), 1 deletion(-)
vdpa/include/uapi/linux/vdpa.h +1 -0
index a3ebf4d4d9b8..96ccbf305d14 100644
@@ -42,6 +42,7 @@ enum vdpa_attr {
 
  	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
  	VDPA_ATTR_DEV_MGMTDEV_MAX_VQS,          /* u32 */
+	VDPA_ATTR_DEV_SUPPORTED_FEATURES,	/* u64 */
 
  	/* new attributes must be added above here */
  	VDPA_ATTR_MAX,
vdpa/vdpa.c +3 -1
index 78736b1422b6..bdc366880ab9 100644
@@ -84,6 +84,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
  	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
  	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
  	[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS] = MNL_TYPE_U32,
+	[VDPA_ATTR_DEV_SUPPORTED_FEATURES] = MNL_TYPE_U64,
  };
 
  static int attr_cb(const struct nlattr *attr, void *data)
@@ -494,13 +495,14 @@ static void print_features(struct vdpa *vdpa, uint64_t features, bool mgmtdevf,
  static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  				struct nlattr **tb)
  {
+	uint64_t classes = 0;
  	const char *class;
  	unsigned int i;
 
  	pr_out_handle_start(vdpa, tb);
 
  	if (tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]) {
-		uint64_t classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
+		classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
  		pr_out_array_start(vdpa, "supported_classes");
 
  		for (i = 1; i < 64; i++) {
@@ -522,6 +524,16 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  		print_uint(PRINT_ANY, "max_supported_vqs", "  max_supported_vqs %d", num_vqs);

vdpa/vdpa.c static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh, · line 534

Quoted 10 lines
vdpa/vdpa.c +8 -0
  	}
 
+	if (tb[VDPA_ATTR_DEV_SUPPORTED_FEATURES]) {
+		uint64_t features;
+
+		features  = mnl_attr_get_u64(tb[VDPA_ATTR_DEV_SUPPORTED_FEATURES]);
+		if (classes & BIT(VIRTIO_ID_NET))
+			print_features(vdpa, features, true, VIRTIO_ID_NET);
+		else
+			print_features(vdpa, features, true, 0);

I wonder what happens if we try to read a virtio_blk device consider:

static const char * const *dev_to_feature_str[] = { [VIRTIO_ID_NET] = net_feature_strs, };

Thanks

Quoted 5 lines
vdpa/vdpa.c +2 -0
+	}
+
  	pr_out_handle_end(vdpa);
  }
 
author
#11 lore raw

Re: [PATCH v2 4/4] vdpa: Support reading device features

Quoted 83 lines · Jason Wang
Quoted 67 lines · Eli Cohen

When showing the available management devices, check if VDPA_ATTR_DEV_SUPPORTED_FEATURES feature is available and print the supported features for a management device.

Examples:

$ vdpa mgmtdev show
auxiliary/mlx5_core.sf.1:
   supported_classes net
   max_supported_vqs 257
   dev_features CSUM GUEST_CSUM MTU HOST_TSO4 HOST_TSO6 STATUS CTRL_VQ MQ \
                CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
$ vdpa -jp mgmtdev show
{
     "mgmtdev": {
         "auxiliary/mlx5_core.sf.1": {
             "supported_classes": [ "net" ],
             "max_supported_vqs": 257,
             "dev_features": [
"CSUM","GUEST_CSUM","MTU","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ",\
"CTRL_MAC_ADDR","VERSION_1","ACCESS_PLATFORM" ]
         }
     }
}
---
  vdpa/include/uapi/linux/vdpa.h |  1 +
  vdpa/vdpa.c                    | 14 +++++++++++++-
  2 files changed, 14 insertions(+), 1 deletion(-)
vdpa/include/uapi/linux/vdpa.h +1 -0
index a3ebf4d4d9b8..96ccbf305d14 100644
@@ -42,6 +42,7 @@ enum vdpa_attr {
  	VDPA_ATTR_DEV_NEGOTIATED_FEATURES,	/* u64 */
  	VDPA_ATTR_DEV_MGMTDEV_MAX_VQS,          /* u32 */
+	VDPA_ATTR_DEV_SUPPORTED_FEATURES,	/* u64 */
  	/* new attributes must be added above here */
  	VDPA_ATTR_MAX,
vdpa/vdpa.c +3 -1
index 78736b1422b6..bdc366880ab9 100644
@@ -84,6 +84,7 @@ static const enum mnl_attr_data_type vdpa_policy[VDPA_ATTR_MAX + 1] = {
  	[VDPA_ATTR_DEV_MAX_VQ_SIZE] = MNL_TYPE_U16,
  	[VDPA_ATTR_DEV_NEGOTIATED_FEATURES] = MNL_TYPE_U64,
  	[VDPA_ATTR_DEV_MGMTDEV_MAX_VQS] = MNL_TYPE_U32,
+	[VDPA_ATTR_DEV_SUPPORTED_FEATURES] = MNL_TYPE_U64,
  };
  static int attr_cb(const struct nlattr *attr, void *data)
@@ -494,13 +495,14 @@ static void print_features(struct vdpa *vdpa, uint64_t features, bool mgmtdevf,
  static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  				struct nlattr **tb)
  {
+	uint64_t classes = 0;
  	const char *class;
  	unsigned int i;
  	pr_out_handle_start(vdpa, tb);
  	if (tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]) {
-		uint64_t classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
+		classes = mnl_attr_get_u64(tb[VDPA_ATTR_MGMTDEV_SUPPORTED_CLASSES]);
  		pr_out_array_start(vdpa, "supported_classes");
  		for (i = 1; i < 64; i++) {

vdpa/vdpa.c static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh, · line 533

Quoted 10 lines
vdpa/vdpa.c +8 -0
@@ -522,6 +524,16 @@ static void pr_out_mgmtdev_show(struct vdpa *vdpa, const struct nlmsghdr *nlh,
  		print_uint(PRINT_ANY, "max_supported_vqs", "  max_supported_vqs %d", num_vqs);
  	}
+	if (tb[VDPA_ATTR_DEV_SUPPORTED_FEATURES]) {
+		uint64_t features;
+
+		features  = mnl_attr_get_u64(tb[VDPA_ATTR_DEV_SUPPORTED_FEATURES]);
+		if (classes & BIT(VIRTIO_ID_NET))
+			print_features(vdpa, features, true, VIRTIO_ID_NET);
+		else
+			print_features(vdpa, features, true, 0);

I wonder what happens if we try to read a virtio_blk device consider:

Quoted 3 lines

static const char * const *dev_to_feature_str[] = { [VIRTIO_ID_NET] = net_feature_strs, };

In that case we will print bit_xx for each non general bit.

Quoted 9 lines

Thanks

Quoted 4 lines
vdpa/vdpa.c +2 -0
+	}
+
  	pr_out_handle_end(vdpa);
  }