From cd502236835b678738810ecd501c85a3a7a11150 Mon Sep 17 00:00:00 2001 From: Michal Wilczynski Date: Tue, 15 Nov 2022 11:48:15 +0100 Subject: devlink: Introduce new attribute 'tx_priority' to devlink-rate To fully utilize offload capabilities of Intel 100G card QoS capabilities new attribute 'tx_priority' needs to be introduced. This attribute allows for usage of strict priority arbiter among siblings. This arbitration scheme attempts to schedule nodes based on their priority as long as the nodes remain within their bandwidth limit. Introduce new attribute in devlink-rate that will allow for configuration of strict priority. New attribute is optional. Signed-off-by: Michal Wilczynski Reviewed-by: Jiri Pirko Signed-off-by: Jakub Kicinski --- include/uapi/linux/devlink.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include/uapi/linux/devlink.h') diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h index 2f24b53a87a5..1a9214d35ef5 100644 --- a/include/uapi/linux/devlink.h +++ b/include/uapi/linux/devlink.h @@ -607,6 +607,7 @@ enum devlink_attr { DEVLINK_ATTR_SELFTESTS, /* nested */ + DEVLINK_ATTR_RATE_TX_PRIORITY, /* u32 */ /* add new attributes above here, update the policy in devlink.c */ __DEVLINK_ATTR_MAX, -- cgit v1.2.3 From 6e2d7e84fcfee62d70e62aa9e469d10b8b6a7dc7 Mon Sep 17 00:00:00 2001 From: Michal Wilczynski Date: Tue, 15 Nov 2022 11:48:16 +0100 Subject: devlink: Introduce new attribute 'tx_weight' to devlink-rate To fully utilize offload capabilities of Intel 100G card QoS capabilities new attribute 'tx_weight' needs to be introduced. This attribute allows for usage of Weighted Fair Queuing arbitration scheme among siblings. This arbitration scheme can be used simultaneously with the strict priority. Introduce new attribute in devlink-rate that will allow for configuration of Weighted Fair Queueing. New attribute is optional. Signed-off-by: Michal Wilczynski Reviewed-by: Jiri Pirko Signed-off-by: Jakub Kicinski --- include/net/devlink.h | 5 +++++ include/uapi/linux/devlink.h | 2 ++ net/core/devlink.c | 33 +++++++++++++++++++++++++++++++++ 3 files changed, 40 insertions(+) (limited to 'include/uapi/linux/devlink.h') diff --git a/include/net/devlink.h b/include/net/devlink.h index 90d59d673cb1..366b23d3f973 100644 --- a/include/net/devlink.h +++ b/include/net/devlink.h @@ -116,6 +116,7 @@ struct devlink_rate { }; u32 tx_priority; + u32 tx_weight; }; struct devlink_port { @@ -1515,12 +1516,16 @@ struct devlink_ops { u64 tx_max, struct netlink_ext_ack *extack); int (*rate_leaf_tx_priority_set)(struct devlink_rate *devlink_rate, void *priv, u32 tx_priority, struct netlink_ext_ack *extack); + int (*rate_leaf_tx_weight_set)(struct devlink_rate *devlink_rate, void *priv, + u32 tx_weight, struct netlink_ext_ack *extack); int (*rate_node_tx_share_set)(struct devlink_rate *devlink_rate, void *priv, u64 tx_share, struct netlink_ext_ack *extack); int (*rate_node_tx_max_set)(struct devlink_rate *devlink_rate, void *priv, u64 tx_max, struct netlink_ext_ack *extack); int (*rate_node_tx_priority_set)(struct devlink_rate *devlink_rate, void *priv, u32 tx_priority, struct netlink_ext_ack *extack); + int (*rate_node_tx_weight_set)(struct devlink_rate *devlink_rate, void *priv, + u32 tx_weight, struct netlink_ext_ack *extack); int (*rate_node_new)(struct devlink_rate *rate_node, void **priv, struct netlink_ext_ack *extack); int (*rate_node_del)(struct devlink_rate *rate_node, void *priv, diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h index 1a9214d35ef5..498d0d5d0957 100644 --- a/include/uapi/linux/devlink.h +++ b/include/uapi/linux/devlink.h @@ -608,6 +608,8 @@ enum devlink_attr { DEVLINK_ATTR_SELFTESTS, /* nested */ DEVLINK_ATTR_RATE_TX_PRIORITY, /* u32 */ + DEVLINK_ATTR_RATE_TX_WEIGHT, /* u32 */ + /* add new attributes above here, update the policy in devlink.c */ __DEVLINK_ATTR_MAX, diff --git a/net/core/devlink.c b/net/core/devlink.c index bf6d3a3c28bb..525bdf426163 100644 --- a/net/core/devlink.c +++ b/net/core/devlink.c @@ -1206,6 +1206,11 @@ static int devlink_nl_rate_fill(struct sk_buff *msg, if (nla_put_u32(msg, DEVLINK_ATTR_RATE_TX_PRIORITY, devlink_rate->tx_priority)) goto nla_put_failure; + + if (nla_put_u32(msg, DEVLINK_ATTR_RATE_TX_WEIGHT, + devlink_rate->tx_weight)) + goto nla_put_failure; + if (devlink_rate->parent) if (nla_put_string(msg, DEVLINK_ATTR_RATE_PARENT_NODE_NAME, devlink_rate->parent->name)) @@ -1940,6 +1945,7 @@ static int devlink_nl_rate_set(struct devlink_rate *devlink_rate, struct nlattr *nla_parent, **attrs = info->attrs; int err = -EOPNOTSUPP; u32 priority; + u32 weight; u64 rate; if (attrs[DEVLINK_ATTR_RATE_TX_SHARE]) { @@ -1982,6 +1988,20 @@ static int devlink_nl_rate_set(struct devlink_rate *devlink_rate, devlink_rate->tx_priority = priority; } + if (attrs[DEVLINK_ATTR_RATE_TX_WEIGHT]) { + weight = nla_get_u32(attrs[DEVLINK_ATTR_RATE_TX_WEIGHT]); + if (devlink_rate_is_leaf(devlink_rate)) + err = ops->rate_leaf_tx_weight_set(devlink_rate, devlink_rate->priv, + weight, info->extack); + else if (devlink_rate_is_node(devlink_rate)) + err = ops->rate_node_tx_weight_set(devlink_rate, devlink_rate->priv, + weight, info->extack); + + if (err) + return err; + devlink_rate->tx_weight = weight; + } + nla_parent = attrs[DEVLINK_ATTR_RATE_PARENT_NODE_NAME]; if (nla_parent) { err = devlink_nl_rate_parent_node_set(devlink_rate, info, @@ -2019,6 +2039,12 @@ static bool devlink_rate_set_ops_supported(const struct devlink_ops *ops, "TX priority set isn't supported for the leafs"); return false; } + if (attrs[DEVLINK_ATTR_RATE_TX_WEIGHT] && !ops->rate_leaf_tx_weight_set) { + NL_SET_ERR_MSG_ATTR(info->extack, + attrs[DEVLINK_ATTR_RATE_TX_WEIGHT], + "TX weight set isn't supported for the leafs"); + return false; + } } else if (type == DEVLINK_RATE_TYPE_NODE) { if (attrs[DEVLINK_ATTR_RATE_TX_SHARE] && !ops->rate_node_tx_share_set) { NL_SET_ERR_MSG_MOD(info->extack, "TX share set isn't supported for the nodes"); @@ -2039,6 +2065,12 @@ static bool devlink_rate_set_ops_supported(const struct devlink_ops *ops, "TX priority set isn't supported for the nodes"); return false; } + if (attrs[DEVLINK_ATTR_RATE_TX_WEIGHT] && !ops->rate_node_tx_weight_set) { + NL_SET_ERR_MSG_ATTR(info->extack, + attrs[DEVLINK_ATTR_RATE_TX_WEIGHT], + "TX weight set isn't supported for the nodes"); + return false; + } } else { WARN(1, "Unknown type of rate object"); return false; @@ -9218,6 +9250,7 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = { [DEVLINK_ATTR_LINECARD_TYPE] = { .type = NLA_NUL_STRING }, [DEVLINK_ATTR_SELFTESTS] = { .type = NLA_NESTED }, [DEVLINK_ATTR_RATE_TX_PRIORITY] = { .type = NLA_U32 }, + [DEVLINK_ATTR_RATE_TX_WEIGHT] = { .type = NLA_U32 }, }; static const struct genl_small_ops devlink_nl_ops[] = { -- cgit v1.2.3 From af6397c9ee2b42988c912dcad2fca1f43d5c1c99 Mon Sep 17 00:00:00 2001 From: Jacob Keller Date: Mon, 28 Nov 2022 12:36:44 -0800 Subject: devlink: support directly reading from region memory To read from a region, user space must currently request a new snapshot of the region and then read from that snapshot. This can sometimes be overkill if user space only reads a tiny portion. They first create the snapshot, then request a read, then destroy the snapshot. For regions which have a single underlying "contents", it makes sense to allow supporting direct reading of the region data. Extend the DEVLINK_CMD_REGION_READ to allow direct reading from a region if requested via the new DEVLINK_ATTR_REGION_DIRECT. If this attribute is set, then perform a direct read instead of using a snapshot. Direct read is mutually exclusive with DEVLINK_ATTR_REGION_SNAPSHOT_ID, and care is taken to ensure that we reject commands which provide incorrect attributes. Regions must enable support for direct read by implementing the .read() callback function. If a region does not support such direct reads, a suitable extended error message is reported. Signed-off-by: Jacob Keller Reviewed-by: Jiri Pirko Signed-off-by: Jakub Kicinski --- .../networking/devlink/devlink-region.rst | 13 ++++ include/net/devlink.h | 16 +++++ include/uapi/linux/devlink.h | 2 + net/core/devlink.c | 80 +++++++++++++++++----- 4 files changed, 94 insertions(+), 17 deletions(-) (limited to 'include/uapi/linux/devlink.h') diff --git a/Documentation/networking/devlink/devlink-region.rst b/Documentation/networking/devlink/devlink-region.rst index f06dca9a1eb6..9232cd7da301 100644 --- a/Documentation/networking/devlink/devlink-region.rst +++ b/Documentation/networking/devlink/devlink-region.rst @@ -31,6 +31,15 @@ in its ``devlink_region_ops`` structure. If snapshot id is not set in the ``DEVLINK_CMD_REGION_NEW`` request kernel will allocate one and send the snapshot information to user space. +Regions may optionally allow directly reading from their contents without a +snapshot. Direct read requests are not atomic. In particular a read request +of size 256 bytes or larger will be split into multiple chunks. If atomic +access is required, use a snapshot. A driver wishing to enable this for a +region should implement the ``.read`` callback in the ``devlink_region_ops`` +structure. User space can request a direct read by using the +``DEVLINK_ATTR_REGION_DIRECT`` attribute instead of specifying a snapshot +id. + example usage ------------- @@ -65,6 +74,10 @@ example usage $ devlink region read pci/0000:00:05.0/fw-health snapshot 1 address 0 length 16 0000000000000000 0014 95dc 0014 9514 0035 1670 0034 db30 + # Read from the region without a snapshot + $ devlink region read pci/0000:00:05.0/fw-health address 16 length 16 + 0000000000000010 0000 0000 ffff ff04 0029 8c00 0028 8cc8 + As regions are likely very device or driver specific, no generic regions are defined. See the driver-specific documentation files for information on the specific regions a driver supports. diff --git a/include/net/devlink.h b/include/net/devlink.h index 074a79b8933f..02528f736f65 100644 --- a/include/net/devlink.h +++ b/include/net/devlink.h @@ -650,6 +650,10 @@ struct devlink_info_req; * the data variable must be updated to point to the snapshot data. * The function will be called while the devlink instance lock is * held. + * @read: callback to directly read a portion of the region. On success, + * the data pointer will be updated with the contents of the + * requested portion of the region. The function will be called + * while the devlink instance lock is held. * @priv: Pointer to driver private data for the region operation */ struct devlink_region_ops { @@ -659,6 +663,10 @@ struct devlink_region_ops { const struct devlink_region_ops *ops, struct netlink_ext_ack *extack, u8 **data); + int (*read)(struct devlink *devlink, + const struct devlink_region_ops *ops, + struct netlink_ext_ack *extack, + u64 offset, u32 size, u8 *data); void *priv; }; @@ -670,6 +678,10 @@ struct devlink_region_ops { * the data variable must be updated to point to the snapshot data. * The function will be called while the devlink instance lock is * held. + * @read: callback to directly read a portion of the region. On success, + * the data pointer will be updated with the contents of the + * requested portion of the region. The function will be called + * while the devlink instance lock is held. * @priv: Pointer to driver private data for the region operation */ struct devlink_port_region_ops { @@ -679,6 +691,10 @@ struct devlink_port_region_ops { const struct devlink_port_region_ops *ops, struct netlink_ext_ack *extack, u8 **data); + int (*read)(struct devlink_port *port, + const struct devlink_port_region_ops *ops, + struct netlink_ext_ack *extack, + u64 offset, u32 size, u8 *data); void *priv; }; diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h index 498d0d5d0957..70191d96af89 100644 --- a/include/uapi/linux/devlink.h +++ b/include/uapi/linux/devlink.h @@ -610,6 +610,8 @@ enum devlink_attr { DEVLINK_ATTR_RATE_TX_PRIORITY, /* u32 */ DEVLINK_ATTR_RATE_TX_WEIGHT, /* u32 */ + DEVLINK_ATTR_REGION_DIRECT, /* flag */ + /* add new attributes above here, update the policy in devlink.c */ __DEVLINK_ATTR_MAX, diff --git a/net/core/devlink.c b/net/core/devlink.c index 6c05cfaa571d..298041a44aa8 100644 --- a/net/core/devlink.c +++ b/net/core/devlink.c @@ -6515,6 +6515,26 @@ devlink_region_snapshot_fill(void *cb_priv, u8 *chunk, u32 chunk_size, return 0; } +static int +devlink_region_port_direct_fill(void *cb_priv, u8 *chunk, u32 chunk_size, + u64 curr_offset, struct netlink_ext_ack *extack) +{ + struct devlink_region *region = cb_priv; + + return region->port_ops->read(region->port, region->port_ops, extack, + curr_offset, chunk_size, chunk); +} + +static int +devlink_region_direct_fill(void *cb_priv, u8 *chunk, u32 chunk_size, + u64 curr_offset, struct netlink_ext_ack *extack) +{ + struct devlink_region *region = cb_priv; + + return region->ops->read(region->devlink, region->ops, extack, + curr_offset, chunk_size, chunk); +} + static int devlink_nl_cmd_region_read_dumpit(struct sk_buff *skb, struct netlink_callback *cb) { @@ -6523,12 +6543,12 @@ static int devlink_nl_cmd_region_read_dumpit(struct sk_buff *skb, u64 ret_offset, start_offset, end_offset = U64_MAX; struct nlattr **attrs = info->attrs; struct devlink_port *port = NULL; - struct devlink_snapshot *snapshot; + devlink_chunk_fill_t *region_cb; struct devlink_region *region; const char *region_name; struct devlink *devlink; unsigned int index; - u32 snapshot_id; + void *region_cb_priv; void *hdr; int err; @@ -6546,12 +6566,6 @@ static int devlink_nl_cmd_region_read_dumpit(struct sk_buff *skb, goto out_unlock; } - if (!attrs[DEVLINK_ATTR_REGION_SNAPSHOT_ID]) { - NL_SET_ERR_MSG(cb->extack, "No snapshot id provided"); - err = -EINVAL; - goto out_unlock; - } - if (info->attrs[DEVLINK_ATTR_PORT_INDEX]) { index = nla_get_u32(info->attrs[DEVLINK_ATTR_PORT_INDEX]); @@ -6577,12 +6591,43 @@ static int devlink_nl_cmd_region_read_dumpit(struct sk_buff *skb, } snapshot_attr = attrs[DEVLINK_ATTR_REGION_SNAPSHOT_ID]; - snapshot_id = nla_get_u32(snapshot_attr); - snapshot = devlink_region_snapshot_get_by_id(region, snapshot_id); - if (!snapshot) { - NL_SET_ERR_MSG_ATTR(cb->extack, snapshot_attr, "Requested snapshot does not exist"); - err = -EINVAL; - goto out_unlock; + if (!snapshot_attr) { + if (!nla_get_flag(attrs[DEVLINK_ATTR_REGION_DIRECT])) { + NL_SET_ERR_MSG(cb->extack, "No snapshot id provided"); + err = -EINVAL; + goto out_unlock; + } + + if (!region->ops->read) { + NL_SET_ERR_MSG(cb->extack, "Requested region does not support direct read"); + err = -EOPNOTSUPP; + goto out_unlock; + } + + if (port) + region_cb = &devlink_region_port_direct_fill; + else + region_cb = &devlink_region_direct_fill; + region_cb_priv = region; + } else { + struct devlink_snapshot *snapshot; + u32 snapshot_id; + + if (nla_get_flag(attrs[DEVLINK_ATTR_REGION_DIRECT])) { + NL_SET_ERR_MSG_ATTR(cb->extack, snapshot_attr, "Direct region read does not use snapshot"); + err = -EINVAL; + goto out_unlock; + } + + snapshot_id = nla_get_u32(snapshot_attr); + snapshot = devlink_region_snapshot_get_by_id(region, snapshot_id); + if (!snapshot) { + NL_SET_ERR_MSG_ATTR(cb->extack, snapshot_attr, "Requested snapshot does not exist"); + err = -EINVAL; + goto out_unlock; + } + region_cb = &devlink_region_snapshot_fill; + region_cb_priv = snapshot; } if (attrs[DEVLINK_ATTR_REGION_CHUNK_ADDR] && @@ -6633,9 +6678,9 @@ static int devlink_nl_cmd_region_read_dumpit(struct sk_buff *skb, goto nla_put_failure; } - err = devlink_nl_region_read_fill(skb, &devlink_region_snapshot_fill, - snapshot, start_offset, end_offset, - &ret_offset, cb->extack); + err = devlink_nl_region_read_fill(skb, region_cb, region_cb_priv, + start_offset, end_offset, &ret_offset, + cb->extack); if (err && err != -EMSGSIZE) goto nla_put_failure; @@ -9280,6 +9325,7 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = { [DEVLINK_ATTR_SELFTESTS] = { .type = NLA_NESTED }, [DEVLINK_ATTR_RATE_TX_PRIORITY] = { .type = NLA_U32 }, [DEVLINK_ATTR_RATE_TX_WEIGHT] = { .type = NLA_U32 }, + [DEVLINK_ATTR_REGION_DIRECT] = { .type = NLA_FLAG }, }; static const struct genl_small_ops devlink_nl_ops[] = { -- cgit v1.2.3 From da65e9ff3bf614d2836e38e1d405c7073e6ba3b7 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 6 Dec 2022 20:51:15 +0200 Subject: devlink: Expose port function commands to control RoCE Expose port function commands to enable / disable RoCE, this is used to control the port RoCE device capabilities. When RoCE is disabled for a function of the port, function cannot create any RoCE specific resources (e.g GID table). It also saves system memory utilization. For example disabling RoCE enable a VF/SF saves 1 Mbytes of system memory per function. Example of a PCI VF port which supports function configuration: Set RoCE of the VF's port function. $ devlink port show pci/0000:06:00.0/2 pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 function: hw_addr 00:00:00:00:00:00 roce enable $ devlink port function set pci/0000:06:00.0/2 roce disable $ devlink port show pci/0000:06:00.0/2 pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 function: hw_addr 00:00:00:00:00:00 roce disable Signed-off-by: Shay Drory Reviewed-by: Jiri Pirko Signed-off-by: Jakub Kicinski --- Documentation/networking/devlink/devlink-port.rst | 34 ++++++- include/net/devlink.h | 18 ++++ include/uapi/linux/devlink.h | 10 ++ net/core/devlink.c | 113 ++++++++++++++++++++++ 4 files changed, 174 insertions(+), 1 deletion(-) (limited to 'include/uapi/linux/devlink.h') diff --git a/Documentation/networking/devlink/devlink-port.rst b/Documentation/networking/devlink/devlink-port.rst index 2c637f4aae8e..c3302d23e480 100644 --- a/Documentation/networking/devlink/devlink-port.rst +++ b/Documentation/networking/devlink/devlink-port.rst @@ -110,7 +110,7 @@ devlink ports for both the controllers. Function configuration ====================== -A user can configure the function attribute before enumerating the PCI +Users can configure one or more function attributes before enumerating the PCI function. Usually it means, user should configure function attribute before a bus specific device for the function is created. However, when SRIOV is enabled, virtual function devices are created on the PCI bus. @@ -122,6 +122,9 @@ A user may set the hardware address of the function using `devlink port function set hw_addr` command. For Ethernet port function this means a MAC address. +Users may also set the RoCE capability of the function using +`devlink port function set roce` command. + Function attributes =================== @@ -162,6 +165,35 @@ device created for the PCI VF/SF. function: hw_addr 00:00:00:00:88:88 +RoCE capability setup +--------------------- +Not all PCI VFs/SFs require RoCE capability. + +When RoCE capability is disabled, it saves system memory per PCI VF/SF. + +When user disables RoCE capability for a VF/SF, user application cannot send or +receive any RoCE packets through this VF/SF and RoCE GID table for this PCI +will be empty. + +When RoCE capability is disabled in the device using port function attribute, +VF/SF driver cannot override it. + +- Get RoCE capability of the VF device:: + + $ devlink port show pci/0000:06:00.0/2 + pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 + function: + hw_addr 00:00:00:00:00:00 roce enable + +- Set RoCE capability of the VF device:: + + $ devlink port function set pci/0000:06:00.0/2 roce disable + + $ devlink port show pci/0000:06:00.0/2 + pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 + function: + hw_addr 00:00:00:00:00:00 roce disable + Subfunction ============ diff --git a/include/net/devlink.h b/include/net/devlink.h index 5f6eca5e4a40..ce4c65d2f2e7 100644 --- a/include/net/devlink.h +++ b/include/net/devlink.h @@ -1451,6 +1451,24 @@ struct devlink_ops { int (*port_function_hw_addr_set)(struct devlink_port *port, const u8 *hw_addr, int hw_addr_len, struct netlink_ext_ack *extack); + /** + * @port_fn_roce_get: Port function's roce get function. + * + * Query RoCE state of a function managed by the devlink port. + * Return -EOPNOTSUPP if port function RoCE handling is not supported. + */ + int (*port_fn_roce_get)(struct devlink_port *devlink_port, + bool *is_enable, + struct netlink_ext_ack *extack); + /** + * @port_fn_roce_set: Port function's roce set function. + * + * Enable/Disable the RoCE state of a function managed by the devlink + * port. + * Return -EOPNOTSUPP if port function RoCE handling is not supported. + */ + int (*port_fn_roce_set)(struct devlink_port *devlink_port, + bool enable, struct netlink_ext_ack *extack); /** * port_new() - Add a new port function of a specified flavor * @devlink: Devlink instance diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h index 70191d96af89..6cc2925bd478 100644 --- a/include/uapi/linux/devlink.h +++ b/include/uapi/linux/devlink.h @@ -658,11 +658,21 @@ enum devlink_resource_unit { DEVLINK_RESOURCE_UNIT_ENTRY, }; +enum devlink_port_fn_attr_cap { + DEVLINK_PORT_FN_ATTR_CAP_ROCE_BIT, + + /* Add new caps above */ + __DEVLINK_PORT_FN_ATTR_CAPS_MAX, +}; + +#define DEVLINK_PORT_FN_CAP_ROCE _BITUL(DEVLINK_PORT_FN_ATTR_CAP_ROCE_BIT) + enum devlink_port_function_attr { DEVLINK_PORT_FUNCTION_ATTR_UNSPEC, DEVLINK_PORT_FUNCTION_ATTR_HW_ADDR, /* binary */ DEVLINK_PORT_FN_ATTR_STATE, /* u8 */ DEVLINK_PORT_FN_ATTR_OPSTATE, /* u8 */ + DEVLINK_PORT_FN_ATTR_CAPS, /* bitfield32 */ __DEVLINK_PORT_FUNCTION_ATTR_MAX, DEVLINK_PORT_FUNCTION_ATTR_MAX = __DEVLINK_PORT_FUNCTION_ATTR_MAX - 1 diff --git a/net/core/devlink.c b/net/core/devlink.c index 035249c5dd17..8c0ad52431c5 100644 --- a/net/core/devlink.c +++ b/net/core/devlink.c @@ -195,11 +195,16 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(devlink_hwmsg); EXPORT_TRACEPOINT_SYMBOL_GPL(devlink_hwerr); EXPORT_TRACEPOINT_SYMBOL_GPL(devlink_trap_report); +#define DEVLINK_PORT_FN_CAPS_VALID_MASK \ + (_BITUL(__DEVLINK_PORT_FN_ATTR_CAPS_MAX) - 1) + static const struct nla_policy devlink_function_nl_policy[DEVLINK_PORT_FUNCTION_ATTR_MAX + 1] = { [DEVLINK_PORT_FUNCTION_ATTR_HW_ADDR] = { .type = NLA_BINARY }, [DEVLINK_PORT_FN_ATTR_STATE] = NLA_POLICY_RANGE(NLA_U8, DEVLINK_PORT_FN_STATE_INACTIVE, DEVLINK_PORT_FN_STATE_ACTIVE), + [DEVLINK_PORT_FN_ATTR_CAPS] = + NLA_POLICY_BITFIELD32(DEVLINK_PORT_FN_CAPS_VALID_MASK), }; static const struct nla_policy devlink_selftest_nl_policy[DEVLINK_ATTR_SELFTEST_ID_MAX + 1] = { @@ -680,6 +685,60 @@ devlink_sb_tc_index_get_from_attrs(struct devlink_sb *devlink_sb, return 0; } +static void devlink_port_fn_cap_fill(struct nla_bitfield32 *caps, + u32 cap, bool is_enable) +{ + caps->selector |= cap; + if (is_enable) + caps->value |= cap; +} + +static int devlink_port_fn_roce_fill(const struct devlink_ops *ops, + struct devlink_port *devlink_port, + struct nla_bitfield32 *caps, + struct netlink_ext_ack *extack) +{ + bool is_enable; + int err; + + if (!ops->port_fn_roce_get) + return 0; + + err = ops->port_fn_roce_get(devlink_port, &is_enable, extack); + if (err) { + if (err == -EOPNOTSUPP) + return 0; + return err; + } + + devlink_port_fn_cap_fill(caps, DEVLINK_PORT_FN_CAP_ROCE, is_enable); + return 0; +} + +static int devlink_port_fn_caps_fill(const struct devlink_ops *ops, + struct devlink_port *devlink_port, + struct sk_buff *msg, + struct netlink_ext_ack *extack, + bool *msg_updated) +{ + struct nla_bitfield32 caps = {}; + int err; + + err = devlink_port_fn_roce_fill(ops, devlink_port, &caps, extack); + if (err) + return err; + + if (!caps.selector) + return 0; + err = nla_put_bitfield32(msg, DEVLINK_PORT_FN_ATTR_CAPS, caps.value, + caps.selector); + if (err) + return err; + + *msg_updated = true; + return 0; +} + static int devlink_sb_tc_index_get_from_info(struct devlink_sb *devlink_sb, struct genl_info *info, @@ -1263,6 +1322,35 @@ static int devlink_port_fn_state_fill(const struct devlink_ops *ops, return 0; } +static int +devlink_port_fn_roce_set(struct devlink_port *devlink_port, bool enable, + struct netlink_ext_ack *extack) +{ + const struct devlink_ops *ops = devlink_port->devlink->ops; + + return ops->port_fn_roce_set(devlink_port, enable, extack); +} + +static int devlink_port_fn_caps_set(struct devlink_port *devlink_port, + const struct nlattr *attr, + struct netlink_ext_ack *extack) +{ + struct nla_bitfield32 caps; + u32 caps_value; + int err; + + caps = nla_get_bitfield32(attr); + caps_value = caps.value & caps.selector; + if (caps.selector & DEVLINK_PORT_FN_CAP_ROCE) { + err = devlink_port_fn_roce_set(devlink_port, + caps_value & DEVLINK_PORT_FN_CAP_ROCE, + extack); + if (err) + return err; + } + return 0; +} + static int devlink_nl_port_function_attrs_put(struct sk_buff *msg, struct devlink_port *port, struct netlink_ext_ack *extack) @@ -1281,6 +1369,10 @@ devlink_nl_port_function_attrs_put(struct sk_buff *msg, struct devlink_port *por &msg_updated); if (err) goto out; + err = devlink_port_fn_caps_fill(ops, port, msg, extack, + &msg_updated); + if (err) + goto out; err = devlink_port_fn_state_fill(ops, port, msg, extack, &msg_updated); out: if (err || !msg_updated) @@ -1653,6 +1745,7 @@ static int devlink_port_function_validate(struct devlink_port *devlink_port, struct netlink_ext_ack *extack) { const struct devlink_ops *ops = devlink_port->devlink->ops; + struct nlattr *attr; if (tb[DEVLINK_PORT_FUNCTION_ATTR_HW_ADDR] && !ops->port_function_hw_addr_set) { @@ -1665,6 +1758,18 @@ static int devlink_port_function_validate(struct devlink_port *devlink_port, "Function does not support state setting"); return -EOPNOTSUPP; } + attr = tb[DEVLINK_PORT_FN_ATTR_CAPS]; + if (attr) { + struct nla_bitfield32 caps; + + caps = nla_get_bitfield32(attr); + if (caps.selector & DEVLINK_PORT_FN_CAP_ROCE && + !ops->port_fn_roce_set) { + NL_SET_ERR_MSG_ATTR(extack, attr, + "Port doesn't support RoCE function attribute"); + return -EOPNOTSUPP; + } + } return 0; } @@ -1692,6 +1797,14 @@ static int devlink_port_function_set(struct devlink_port *port, if (err) return err; } + + attr = tb[DEVLINK_PORT_FN_ATTR_CAPS]; + if (attr) { + err = devlink_port_fn_caps_set(port, attr, extack); + if (err) + return err; + } + /* Keep this as the last function attribute set, so that when * multiple port function attributes are set along with state, * Those can be applied first before activating the state. -- cgit v1.2.3 From a8ce7b26a51efc4d7753b23d639ae092878a6193 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 6 Dec 2022 20:51:18 +0200 Subject: devlink: Expose port function commands to control migratable Expose port function commands to enable / disable migratable capability, this is used to set the port function as migratable. Live migration is the process of transferring a live virtual machine from one physical host to another without disrupting its normal operation. In order for a VM to be able to perform LM, all the VM components must be able to perform migration. e.g.: to be migratable. In order for VF to be migratable, VF must be bound to VFIO driver with migration support. When migratable capability is enabled for a function of the port, the device is making the necessary preparations for the function to be migratable, which might include disabling features which cannot be migrated. Example of LM with migratable function configuration: Set migratable of the VF's port function. $ devlink port show pci/0000:06:00.0/2 pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 function: hw_addr 00:00:00:00:00:00 migratable disable $ devlink port function set pci/0000:06:00.0/2 migratable enable $ devlink port show pci/0000:06:00.0/2 pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 function: hw_addr 00:00:00:00:00:00 migratable enable Bind VF to VFIO driver with migration support: $ echo > /sys/bus/pci/devices/0000:08:00.0/driver/unbind $ echo mlx5_vfio_pci > /sys/bus/pci/devices/0000:08:00.0/driver_override $ echo > /sys/bus/pci/devices/0000:08:00.0/driver/bind Attach VF to the VM. Start the VM. Perform LM. Signed-off-by: Shay Drory Reviewed-by: Jiri Pirko Acked-by: Shannon Nelson Signed-off-by: Jakub Kicinski --- Documentation/networking/devlink/devlink-port.rst | 46 +++++++++++++++++++ include/net/devlink.h | 21 +++++++++ include/uapi/linux/devlink.h | 3 ++ net/core/devlink.c | 55 +++++++++++++++++++++++ 4 files changed, 125 insertions(+) (limited to 'include/uapi/linux/devlink.h') diff --git a/Documentation/networking/devlink/devlink-port.rst b/Documentation/networking/devlink/devlink-port.rst index c3302d23e480..3da590953ce8 100644 --- a/Documentation/networking/devlink/devlink-port.rst +++ b/Documentation/networking/devlink/devlink-port.rst @@ -125,6 +125,9 @@ this means a MAC address. Users may also set the RoCE capability of the function using `devlink port function set roce` command. +Users may also set the function as migratable using +'devlink port function set migratable' command. + Function attributes =================== @@ -194,6 +197,49 @@ VF/SF driver cannot override it. function: hw_addr 00:00:00:00:00:00 roce disable +migratable capability setup +--------------------------- +Live migration is the process of transferring a live virtual machine +from one physical host to another without disrupting its normal +operation. + +User who want PCI VFs to be able to perform live migration need to +explicitly enable the VF migratable capability. + +When user enables migratable capability for a VF, and the HV binds the VF to VFIO driver +with migration support, the user can migrate the VM with this VF from one HV to a +different one. + +However, when migratable capability is enable, device will disable features which cannot +be migrated. Thus migratable cap can impose limitations on a VF so let the user decide. + +Example of LM with migratable function configuration: +- Get migratable capability of the VF device:: + + $ devlink port show pci/0000:06:00.0/2 + pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 + function: + hw_addr 00:00:00:00:00:00 migratable disable + +- Set migratable capability of the VF device:: + + $ devlink port function set pci/0000:06:00.0/2 migratable enable + + $ devlink port show pci/0000:06:00.0/2 + pci/0000:06:00.0/2: type eth netdev enp6s0pf0vf1 flavour pcivf pfnum 0 vfnum 1 + function: + hw_addr 00:00:00:00:00:00 migratable enable + +- Bind VF to VFIO driver with migration support:: + + $ echo > /sys/bus/pci/devices/0000:08:00.0/driver/unbind + $ echo mlx5_vfio_pci > /sys/bus/pci/devices/0000:08:00.0/driver_override + $ echo > /sys/bus/pci/devices/0000:08:00.0/driver/bind + +Attach VF to the VM. +Start the VM. +Perform live migration. + Subfunction ============ diff --git a/include/net/devlink.h b/include/net/devlink.h index ce4c65d2f2e7..0f376a28b9c4 100644 --- a/include/net/devlink.h +++ b/include/net/devlink.h @@ -1469,6 +1469,27 @@ struct devlink_ops { */ int (*port_fn_roce_set)(struct devlink_port *devlink_port, bool enable, struct netlink_ext_ack *extack); + /** + * @port_fn_migratable_get: Port function's migratable get function. + * + * Query migratable state of a function managed by the devlink port. + * Return -EOPNOTSUPP if port function migratable handling is not + * supported. + */ + int (*port_fn_migratable_get)(struct devlink_port *devlink_port, + bool *is_enable, + struct netlink_ext_ack *extack); + /** + * @port_fn_migratable_set: Port function's migratable set function. + * + * Enable/Disable migratable state of a function managed by the devlink + * port. + * Return -EOPNOTSUPP if port function migratable handling is not + * supported. + */ + int (*port_fn_migratable_set)(struct devlink_port *devlink_port, + bool enable, + struct netlink_ext_ack *extack); /** * port_new() - Add a new port function of a specified flavor * @devlink: Devlink instance diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h index 6cc2925bd478..3782d4219ac9 100644 --- a/include/uapi/linux/devlink.h +++ b/include/uapi/linux/devlink.h @@ -660,12 +660,15 @@ enum devlink_resource_unit { enum devlink_port_fn_attr_cap { DEVLINK_PORT_FN_ATTR_CAP_ROCE_BIT, + DEVLINK_PORT_FN_ATTR_CAP_MIGRATABLE_BIT, /* Add new caps above */ __DEVLINK_PORT_FN_ATTR_CAPS_MAX, }; #define DEVLINK_PORT_FN_CAP_ROCE _BITUL(DEVLINK_PORT_FN_ATTR_CAP_ROCE_BIT) +#define DEVLINK_PORT_FN_CAP_MIGRATABLE \ + _BITUL(DEVLINK_PORT_FN_ATTR_CAP_MIGRATABLE_BIT) enum devlink_port_function_attr { DEVLINK_PORT_FUNCTION_ATTR_UNSPEC, diff --git a/net/core/devlink.c b/net/core/devlink.c index 8c0ad52431c5..ab40ebcb4aea 100644 --- a/net/core/devlink.c +++ b/net/core/devlink.c @@ -715,6 +715,29 @@ static int devlink_port_fn_roce_fill(const struct devlink_ops *ops, return 0; } +static int devlink_port_fn_migratable_fill(const struct devlink_ops *ops, + struct devlink_port *devlink_port, + struct nla_bitfield32 *caps, + struct netlink_ext_ack *extack) +{ + bool is_enable; + int err; + + if (!ops->port_fn_migratable_get || + devlink_port->attrs.flavour != DEVLINK_PORT_FLAVOUR_PCI_VF) + return 0; + + err = ops->port_fn_migratable_get(devlink_port, &is_enable, extack); + if (err) { + if (err == -EOPNOTSUPP) + return 0; + return err; + } + + devlink_port_fn_cap_fill(caps, DEVLINK_PORT_FN_CAP_MIGRATABLE, is_enable); + return 0; +} + static int devlink_port_fn_caps_fill(const struct devlink_ops *ops, struct devlink_port *devlink_port, struct sk_buff *msg, @@ -728,6 +751,10 @@ static int devlink_port_fn_caps_fill(const struct devlink_ops *ops, if (err) return err; + err = devlink_port_fn_migratable_fill(ops, devlink_port, &caps, extack); + if (err) + return err; + if (!caps.selector) return 0; err = nla_put_bitfield32(msg, DEVLINK_PORT_FN_ATTR_CAPS, caps.value, @@ -1322,6 +1349,15 @@ static int devlink_port_fn_state_fill(const struct devlink_ops *ops, return 0; } +static int +devlink_port_fn_mig_set(struct devlink_port *devlink_port, bool enable, + struct netlink_ext_ack *extack) +{ + const struct devlink_ops *ops = devlink_port->devlink->ops; + + return ops->port_fn_migratable_set(devlink_port, enable, extack); +} + static int devlink_port_fn_roce_set(struct devlink_port *devlink_port, bool enable, struct netlink_ext_ack *extack) @@ -1348,6 +1384,13 @@ static int devlink_port_fn_caps_set(struct devlink_port *devlink_port, if (err) return err; } + if (caps.selector & DEVLINK_PORT_FN_CAP_MIGRATABLE) { + err = devlink_port_fn_mig_set(devlink_port, caps_value & + DEVLINK_PORT_FN_CAP_MIGRATABLE, + extack); + if (err) + return err; + } return 0; } @@ -1769,6 +1812,18 @@ static int devlink_port_function_validate(struct devlink_port *devlink_port, "Port doesn't support RoCE function attribute"); return -EOPNOTSUPP; } + if (caps.selector & DEVLINK_PORT_FN_CAP_MIGRATABLE) { + if (!ops->port_fn_migratable_set) { + NL_SET_ERR_MSG_ATTR(extack, attr, + "Port doesn't support migratable function attribute"); + return -EOPNOTSUPP; + } + if (devlink_port->attrs.flavour != DEVLINK_PORT_FLAVOUR_PCI_VF) { + NL_SET_ERR_MSG_ATTR(extack, attr, + "migratable function attribute supported for VFs only"); + return -EOPNOTSUPP; + } + } } return 0; } -- cgit v1.2.3