Currently, the header split feature (putting headers in one smaller
buffer and then the data in a separate bigger one) is always enabled
in idpf when supported.
One may want to not have fragmented frames per each packet, for example,
to avoid XDP frags. To better optimize setups for particular workloads,
add ability to switch the header split state on and off via Ethtool's
ringparams, as well as to query the current status.
There's currently only GET in the Ethtool Netlink interface for now,
so add SET first. I suspect idpf is not the only one supporting this.
Alexander Lobakin (1):
ethtool: add SET for TCP_DATA_SPLIT ringparam
Michal Kubiak (1):
idpf: add get/set for Ethtool's header split ringparam
drivers/net/ethernet/intel/idpf/idpf.h | 7 +-
.../net/ethernet/intel/idpf/idpf_ethtool.c | 11 ++++
drivers/net/ethernet/intel/idpf/idpf_lib.c | 65 +++++++++++++++++++
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 12 ++--
.../net/ethernet/intel/idpf/idpf_virtchnl.c | 2 +
include/linux/ethtool.h | 2 +
net/ethtool/rings.c | 12 ++++
7 files changed, 104 insertions(+), 7 deletions(-)
--
2.43.0
Follow up commit 9690ae604290 ("ethtool: add header/data split
indication") and add the set part of Ethtool's header split, i.e.
ability to enable/disable header split via the Ethtool Netlink
interface. This might be helpful to optimize the setup for particular
workloads, for example, to avoid XDP frags, and so on.
A driver should advertise ``ETHTOOL_RING_USE_TCP_DATA_SPLIT`` in its
ops->supported_ring_params to allow doing that. "Unknown" passed from
the userspace when the header split is supported means the driver is
free to choose the preferred state.
Reviewed-by: Przemek Kitszel <[email protected]>
Signed-off-by: Alexander Lobakin <[email protected]>
---
include/linux/ethtool.h | 2 ++
net/ethtool/rings.c | 12 ++++++++++++
2 files changed, 14 insertions(+)
diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h
index deb683d3360f..67b30940234b 100644
--- a/include/linux/ethtool.h
+++ b/include/linux/ethtool.h
@@ -95,6 +95,7 @@ struct kernel_ethtool_ringparam {
* @ETHTOOL_RING_USE_TX_PUSH: capture for setting tx_push
* @ETHTOOL_RING_USE_RX_PUSH: capture for setting rx_push
* @ETHTOOL_RING_USE_TX_PUSH_BUF_LEN: capture for setting tx_push_buf_len
+ * @ETHTOOL_RING_USE_TCP_DATA_SPLIT: capture for setting tcp_data_split
*/
enum ethtool_supported_ring_param {
ETHTOOL_RING_USE_RX_BUF_LEN = BIT(0),
@@ -102,6 +103,7 @@ enum ethtool_supported_ring_param {
ETHTOOL_RING_USE_TX_PUSH = BIT(2),
ETHTOOL_RING_USE_RX_PUSH = BIT(3),
ETHTOOL_RING_USE_TX_PUSH_BUF_LEN = BIT(4),
+ ETHTOOL_RING_USE_TCP_DATA_SPLIT = BIT(5),
};
#define __ETH_RSS_HASH_BIT(bit) ((u32)1 << (bit))
diff --git a/net/ethtool/rings.c b/net/ethtool/rings.c
index fb09f774ea01..b7865a14fdf8 100644
--- a/net/ethtool/rings.c
+++ b/net/ethtool/rings.c
@@ -124,6 +124,8 @@ const struct nla_policy ethnl_rings_set_policy[] = {
[ETHTOOL_A_RINGS_RX_JUMBO] = { .type = NLA_U32 },
[ETHTOOL_A_RINGS_TX] = { .type = NLA_U32 },
[ETHTOOL_A_RINGS_RX_BUF_LEN] = NLA_POLICY_MIN(NLA_U32, 1),
+ [ETHTOOL_A_RINGS_TCP_DATA_SPLIT] =
+ NLA_POLICY_MAX(NLA_U8, ETHTOOL_TCP_DATA_SPLIT_ENABLED),
[ETHTOOL_A_RINGS_CQE_SIZE] = NLA_POLICY_MIN(NLA_U32, 1),
[ETHTOOL_A_RINGS_TX_PUSH] = NLA_POLICY_MAX(NLA_U8, 1),
[ETHTOOL_A_RINGS_RX_PUSH] = NLA_POLICY_MAX(NLA_U8, 1),
@@ -145,6 +147,14 @@ ethnl_set_rings_validate(struct ethnl_req_info *req_info,
return -EOPNOTSUPP;
}
+ if (tb[ETHTOOL_A_RINGS_TCP_DATA_SPLIT] &&
+ !(ops->supported_ring_params & ETHTOOL_RING_USE_TCP_DATA_SPLIT)) {
+ NL_SET_ERR_MSG_ATTR(info->extack,
+ tb[ETHTOOL_A_RINGS_TCP_DATA_SPLIT],
+ "setting TCP data split is not supported");
+ return -EOPNOTSUPP;
+ }
+
if (tb[ETHTOOL_A_RINGS_CQE_SIZE] &&
!(ops->supported_ring_params & ETHTOOL_RING_USE_CQE_SIZE)) {
NL_SET_ERR_MSG_ATTR(info->extack,
@@ -202,6 +212,8 @@ ethnl_set_rings(struct ethnl_req_info *req_info, struct genl_info *info)
ethnl_update_u32(&ringparam.tx_pending, tb[ETHTOOL_A_RINGS_TX], &mod);
ethnl_update_u32(&kernel_ringparam.rx_buf_len,
tb[ETHTOOL_A_RINGS_RX_BUF_LEN], &mod);
+ ethnl_update_u8(&kernel_ringparam.tcp_data_split,
+ tb[ETHTOOL_A_RINGS_TCP_DATA_SPLIT], &mod);
ethnl_update_u32(&kernel_ringparam.cqe_size,
tb[ETHTOOL_A_RINGS_CQE_SIZE], &mod);
ethnl_update_u8(&kernel_ringparam.tx_push,
--
2.43.0
From: Michal Kubiak <[email protected]>
idpf supports the header split feature and that feature is always
enabled by default.
However, for flexibility reasons and to simplify some scenarios, it
would be useful to have the support for switching the header split
off (and on) from the userspace.
Address that need by adding the user config parameter, the functions
for disabling (or enabling) the header split feature, and calls to
them from the Ethtool ringparam callbacks.
It still is enabled by default if supported by the hardware.
Reviewed-by: Przemek Kitszel <[email protected]>
Signed-off-by: Michal Kubiak <[email protected]>
Co-developed-by: Alexander Lobakin <[email protected]>
Signed-off-by: Alexander Lobakin <[email protected]>
---
drivers/net/ethernet/intel/idpf/idpf.h | 7 +-
.../net/ethernet/intel/idpf/idpf_ethtool.c | 11 ++++
drivers/net/ethernet/intel/idpf/idpf_lib.c | 65 +++++++++++++++++++
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 12 ++--
.../net/ethernet/intel/idpf/idpf_virtchnl.c | 2 +
5 files changed, 90 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/intel/idpf/idpf.h b/drivers/net/ethernet/intel/idpf/idpf.h
index bee73353b56a..0acc125decb3 100644
--- a/drivers/net/ethernet/intel/idpf/idpf.h
+++ b/drivers/net/ethernet/intel/idpf/idpf.h
@@ -15,7 +15,7 @@ struct idpf_vport_max_q;
#include <linux/pci.h>
#include <linux/bitfield.h>
#include <linux/sctp.h>
-#include <linux/ethtool.h>
+#include <linux/ethtool_netlink.h>
#include <net/gro.h>
#include <linux/dim.h>
@@ -418,11 +418,13 @@ struct idpf_vport {
/**
* enum idpf_user_flags
+ * @__IDPF_USER_FLAG_HSPLIT: header split state
* @__IDPF_PROMISC_UC: Unicast promiscuous mode
* @__IDPF_PROMISC_MC: Multicast promiscuous mode
* @__IDPF_USER_FLAGS_NBITS: Must be last
*/
enum idpf_user_flags {
+ __IDPF_USER_FLAG_HSPLIT = 0U,
__IDPF_PROMISC_UC = 32,
__IDPF_PROMISC_MC,
@@ -965,4 +967,7 @@ int idpf_send_map_unmap_queue_vector_msg(struct idpf_vport *vport, bool map);
int idpf_send_set_sriov_vfs_msg(struct idpf_adapter *adapter, u16 num_vfs);
int idpf_sriov_configure(struct pci_dev *pdev, int num_vfs);
+u8 idpf_vport_get_hsplit(const struct idpf_vport *vport);
+bool idpf_vport_set_hsplit(const struct idpf_vport *vport, u8 val);
+
#endif /* !_IDPF_H_ */
diff --git a/drivers/net/ethernet/intel/idpf/idpf_ethtool.c b/drivers/net/ethernet/intel/idpf/idpf_ethtool.c
index bf58989a573e..d549d3c5910d 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_ethtool.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_ethtool.c
@@ -320,6 +320,8 @@ static void idpf_get_ringparam(struct net_device *netdev,
ring->rx_pending = vport->rxq_desc_count;
ring->tx_pending = vport->txq_desc_count;
+ kring->tcp_data_split = idpf_vport_get_hsplit(vport);
+
idpf_vport_ctrl_unlock(netdev);
}
@@ -379,6 +381,14 @@ static int idpf_set_ringparam(struct net_device *netdev,
new_rx_count == vport->rxq_desc_count)
goto unlock_mutex;
+ if (!idpf_vport_set_hsplit(vport, kring->tcp_data_split)) {
+ NL_SET_ERR_MSG_MOD(ext_ack,
+ "setting TCP data split is not supported");
+ err = -EOPNOTSUPP;
+
+ goto unlock_mutex;
+ }
+
config_data = &vport->adapter->vport_config[idx]->user_config;
config_data->num_req_txq_desc = new_tx_count;
config_data->num_req_rxq_desc = new_rx_count;
@@ -1334,6 +1344,7 @@ static int idpf_get_link_ksettings(struct net_device *netdev,
static const struct ethtool_ops idpf_ethtool_ops = {
.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
ETHTOOL_COALESCE_USE_ADAPTIVE,
+ .supported_ring_params = ETHTOOL_RING_USE_TCP_DATA_SPLIT,
.get_msglevel = idpf_get_msglevel,
.set_msglevel = idpf_set_msglevel,
.get_link = ethtool_op_get_link,
diff --git a/drivers/net/ethernet/intel/idpf/idpf_lib.c b/drivers/net/ethernet/intel/idpf/idpf_lib.c
index 19809b0ddcd9..5fea2fd957eb 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_lib.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_lib.c
@@ -1057,6 +1057,71 @@ static void idpf_vport_dealloc(struct idpf_vport *vport)
adapter->next_vport = idpf_get_free_slot(adapter);
}
+/**
+ * idpf_is_hsplit_supported - check whether the header split is supported
+ * @vport: virtual port to check the capability for
+ *
+ * Return: true if it's supported by the HW/FW, false if not.
+ */
+static bool idpf_is_hsplit_supported(const struct idpf_vport *vport)
+{
+ return idpf_is_queue_model_split(vport->rxq_model) &&
+ idpf_is_cap_ena_all(vport->adapter, IDPF_HSPLIT_CAPS,
+ IDPF_CAP_HSPLIT);
+}
+
+/**
+ * idpf_vport_get_hsplit - get the current header split feature state
+ * @vport: virtual port to query the state for
+ *
+ * Return: ``ETHTOOL_TCP_DATA_SPLIT_UNKNOWN`` if not supported,
+ * ``ETHTOOL_TCP_DATA_SPLIT_DISABLED`` if disabled,
+ * ``ETHTOOL_TCP_DATA_SPLIT_ENABLED`` if active.
+ */
+u8 idpf_vport_get_hsplit(const struct idpf_vport *vport)
+{
+ const struct idpf_vport_user_config_data *config;
+
+ if (!idpf_is_hsplit_supported(vport))
+ return ETHTOOL_TCP_DATA_SPLIT_UNKNOWN;
+
+ config = &vport->adapter->vport_config[vport->idx]->user_config;
+
+ return test_bit(__IDPF_USER_FLAG_HSPLIT, config->user_flags) ?
+ ETHTOOL_TCP_DATA_SPLIT_ENABLED :
+ ETHTOOL_TCP_DATA_SPLIT_DISABLED;
+}
+
+/**
+ * idpf_vport_set_hsplit - enable or disable header split on a given vport
+ * @vport: virtual port to configure
+ * @val: Ethtool flag controlling the header split state
+ *
+ * Return: true on success, false if not supported by the HW.
+ */
+bool idpf_vport_set_hsplit(const struct idpf_vport *vport, u8 val)
+{
+ struct idpf_vport_user_config_data *config;
+
+ if (!idpf_is_hsplit_supported(vport))
+ return val == ETHTOOL_TCP_DATA_SPLIT_UNKNOWN;
+
+ config = &vport->adapter->vport_config[vport->idx]->user_config;
+
+ switch (val) {
+ case ETHTOOL_TCP_DATA_SPLIT_UNKNOWN:
+ /* Default is to enable */
+ case ETHTOOL_TCP_DATA_SPLIT_ENABLED:
+ __set_bit(__IDPF_USER_FLAG_HSPLIT, config->user_flags);
+ return true;
+ case ETHTOOL_TCP_DATA_SPLIT_DISABLED:
+ __clear_bit(__IDPF_USER_FLAG_HSPLIT, config->user_flags);
+ return true;
+ default:
+ return false;
+ }
+}
+
/**
* idpf_vport_alloc - Allocates the next available struct vport in the adapter
* @adapter: board private structure
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.c b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
index 29d2191737d5..c01fdfd41d98 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_txrx.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
@@ -1240,12 +1240,15 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq)
struct idpf_adapter *adapter = vport->adapter;
struct idpf_queue *q;
int i, k, err = 0;
+ bool hs;
vport->rxq_grps = kcalloc(vport->num_rxq_grp,
sizeof(struct idpf_rxq_group), GFP_KERNEL);
if (!vport->rxq_grps)
return -ENOMEM;
+ hs = idpf_vport_get_hsplit(vport) == ETHTOOL_TCP_DATA_SPLIT_ENABLED;
+
for (i = 0; i < vport->num_rxq_grp; i++) {
struct idpf_rxq_group *rx_qgrp = &vport->rxq_grps[i];
int j;
@@ -1298,9 +1301,8 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq)
q->rx_buf_size = vport->bufq_size[j];
q->rx_buffer_low_watermark = IDPF_LOW_WATERMARK;
q->rx_buf_stride = IDPF_RX_BUF_STRIDE;
- if (idpf_is_cap_ena_all(adapter, IDPF_HSPLIT_CAPS,
- IDPF_CAP_HSPLIT) &&
- idpf_is_queue_model_split(vport->rxq_model)) {
+
+ if (hs) {
q->rx_hsplit_en = true;
q->rx_hbuf_size = IDPF_HDR_BUF_SIZE;
}
@@ -1344,9 +1346,7 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq)
rx_qgrp->splitq.rxq_sets[j]->refillq1 =
&rx_qgrp->splitq.bufq_sets[1].refillqs[j];
- if (idpf_is_cap_ena_all(adapter, IDPF_HSPLIT_CAPS,
- IDPF_CAP_HSPLIT) &&
- idpf_is_queue_model_split(vport->rxq_model)) {
+ if (hs) {
q->rx_hsplit_en = true;
q->rx_hbuf_size = IDPF_HDR_BUF_SIZE;
}
diff --git a/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c b/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c
index 2c1b051fdc0d..d0cdd63b3d5b 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c
@@ -3285,6 +3285,8 @@ void idpf_vport_init(struct idpf_vport *vport, struct idpf_vport_max_q *max_q)
memcpy(vport->rx_itr_profile, rx_itr, IDPF_DIM_PROFILE_SLOTS);
memcpy(vport->tx_itr_profile, tx_itr, IDPF_DIM_PROFILE_SLOTS);
+ idpf_vport_set_hsplit(vport, ETHTOOL_TCP_DATA_SPLIT_ENABLED);
+
idpf_vport_init_num_qs(vport, vport_msg);
idpf_vport_calc_num_q_desc(vport);
idpf_vport_calc_num_q_groups(vport);
--
2.43.0
On Tue, 12 Dec 2023 15:27:50 +0100 Alexander Lobakin wrote:
> Currently, the header split feature (putting headers in one smaller
> buffer and then the data in a separate bigger one) is always enabled
> in idpf when supported.
> One may want to not have fragmented frames per each packet, for example,
> to avoid XDP frags. To better optimize setups for particular workloads,
> add ability to switch the header split state on and off via Ethtool's
> ringparams, as well as to query the current status.
> There's currently only GET in the Ethtool Netlink interface for now,
> so add SET first. I suspect idpf is not the only one supporting this.
Reviewed-by: Jakub Kicinski <[email protected]>
Hello:
This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <[email protected]>:
On Tue, 12 Dec 2023 15:27:50 +0100 you wrote:
> Currently, the header split feature (putting headers in one smaller
> buffer and then the data in a separate bigger one) is always enabled
> in idpf when supported.
> One may want to not have fragmented frames per each packet, for example,
> to avoid XDP frags. To better optimize setups for particular workloads,
> add ability to switch the header split state on and off via Ethtool's
> ringparams, as well as to query the current status.
> There's currently only GET in the Ethtool Netlink interface for now,
> so add SET first. I suspect idpf is not the only one supporting this.
>
> [...]
Here is the summary with links:
- [net-next,1/2] ethtool: add SET for TCP_DATA_SPLIT ringparam
https://git.kernel.org/netdev/net-next/c/50d73710715d
- [net-next,2/2] idpf: add get/set for Ethtool's header split ringparam
https://git.kernel.org/netdev/net-next/c/9b1aa3ef2328
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html