Skip to content

Commit

Permalink
octeontx2-pf: TC_MATCHALL egress ratelimiting offload
Browse files Browse the repository at this point in the history
Add TC_MATCHALL egress ratelimiting offload support with POLICE
action for entire traffic going out of the interface.

Eg: To ratelimit egress traffic to 100Mbps

$ ethtool -K eth0 hw-tc-offload on
$ tc qdisc add dev eth0 clsact
$ tc filter add dev eth0 egress matchall skip_sw \
                action police rate 100Mbit burst 16Kbit

HW supports a max burst size of ~128KB.
Only one ratelimiting filter can be installed at a time.

Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: Naveen Mamindlapalli <naveenm@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
  • Loading branch information
Sunil Goutham authored and David S. Miller committed Mar 18, 2021
1 parent d8ce30e commit e638a83
Show file tree
Hide file tree
Showing 3 changed files with 236 additions and 4 deletions.
1 change: 1 addition & 0 deletions drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
Original file line number Diff line number Diff line change
Expand Up @@ -318,6 +318,7 @@ struct otx2_nic {
#define OTX2_FLAG_RX_PAUSE_ENABLED BIT_ULL(9)
#define OTX2_FLAG_TX_PAUSE_ENABLED BIT_ULL(10)
#define OTX2_FLAG_TC_FLOWER_SUPPORT BIT_ULL(11)
#define OTX2_FLAG_TC_MATCHALL_EGRESS_ENABLED BIT_ULL(12)
u64 flags;

struct otx2_qset qset;
Expand Down
1 change: 1 addition & 0 deletions drivers/net/ethernet/marvell/octeontx2/nic/otx2_reg.h
Original file line number Diff line number Diff line change
Expand Up @@ -152,6 +152,7 @@
#define NIX_AF_TL3X_SCHEDULE(a) (0x1000 | (a) << 16)
#define NIX_AF_TL4X_PARENT(a) (0x1288 | (a) << 16)
#define NIX_AF_TL4X_SCHEDULE(a) (0x1200 | (a) << 16)
#define NIX_AF_TL4X_PIR(a) (0x1230 | (a) << 16)
#define NIX_AF_MDQX_SCHEDULE(a) (0x1400 | (a) << 16)
#define NIX_AF_MDQX_PARENT(a) (0x1480 | (a) << 16)
#define NIX_AF_TL3_TL2X_LINKX_CFG(a, b) (0x1700 | (a) << 16 | (b) << 3)
Expand Down
238 changes: 234 additions & 4 deletions drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@
#include <linux/etherdevice.h>
#include <linux/inetdevice.h>
#include <linux/rhashtable.h>
#include <linux/bitfield.h>
#include <net/flow_dissector.h>
#include <net/pkt_cls.h>
#include <net/tc_act/tc_gact.h>
Expand All @@ -16,6 +17,21 @@

#include "otx2_common.h"

/* Egress rate limiting definitions */
#define MAX_BURST_EXPONENT 0x0FULL
#define MAX_BURST_MANTISSA 0xFFULL
#define MAX_BURST_SIZE 130816ULL
#define MAX_RATE_DIVIDER_EXPONENT 12ULL
#define MAX_RATE_EXPONENT 0x0FULL
#define MAX_RATE_MANTISSA 0xFFULL

/* Bitfields in NIX_TLX_PIR register */
#define TLX_RATE_MANTISSA GENMASK_ULL(8, 1)
#define TLX_RATE_EXPONENT GENMASK_ULL(12, 9)
#define TLX_RATE_DIVIDER_EXPONENT GENMASK_ULL(16, 13)
#define TLX_BURST_MANTISSA GENMASK_ULL(36, 29)
#define TLX_BURST_EXPONENT GENMASK_ULL(40, 37)

struct otx2_tc_flow_stats {
u64 bytes;
u64 pkts;
Expand All @@ -32,6 +48,178 @@ struct otx2_tc_flow {
spinlock_t lock; /* lock for stats */
};

static void otx2_get_egress_burst_cfg(u32 burst, u32 *burst_exp,
u32 *burst_mantissa)
{
unsigned int tmp;

/* Burst is calculated as
* ((256 + BURST_MANTISSA) << (1 + BURST_EXPONENT)) / 256
* Max supported burst size is 130,816 bytes.
*/
burst = min_t(u32, burst, MAX_BURST_SIZE);
if (burst) {
*burst_exp = ilog2(burst) ? ilog2(burst) - 1 : 0;
tmp = burst - rounddown_pow_of_two(burst);
if (burst < MAX_BURST_MANTISSA)
*burst_mantissa = tmp * 2;
else
*burst_mantissa = tmp / (1ULL << (*burst_exp - 7));
} else {
*burst_exp = MAX_BURST_EXPONENT;
*burst_mantissa = MAX_BURST_MANTISSA;
}
}

static void otx2_get_egress_rate_cfg(u32 maxrate, u32 *exp,
u32 *mantissa, u32 *div_exp)
{
unsigned int tmp;

/* Rate calculation by hardware
*
* PIR_ADD = ((256 + mantissa) << exp) / 256
* rate = (2 * PIR_ADD) / ( 1 << div_exp)
* The resultant rate is in Mbps.
*/

/* 2Mbps to 100Gbps can be expressed with div_exp = 0.
* Setting this to '0' will ease the calculation of
* exponent and mantissa.
*/
*div_exp = 0;

if (maxrate) {
*exp = ilog2(maxrate) ? ilog2(maxrate) - 1 : 0;
tmp = maxrate - rounddown_pow_of_two(maxrate);
if (maxrate < MAX_RATE_MANTISSA)
*mantissa = tmp * 2;
else
*mantissa = tmp / (1ULL << (*exp - 7));
} else {
/* Instead of disabling rate limiting, set all values to max */
*exp = MAX_RATE_EXPONENT;
*mantissa = MAX_RATE_MANTISSA;
}
}

static int otx2_set_matchall_egress_rate(struct otx2_nic *nic, u32 burst, u32 maxrate)
{
struct otx2_hw *hw = &nic->hw;
struct nix_txschq_config *req;
u32 burst_exp, burst_mantissa;
u32 exp, mantissa, div_exp;
int txschq, err;

/* All SQs share the same TL4, so pick the first scheduler */
txschq = hw->txschq_list[NIX_TXSCH_LVL_TL4][0];

/* Get exponent and mantissa values from the desired rate */
otx2_get_egress_burst_cfg(burst, &burst_exp, &burst_mantissa);
otx2_get_egress_rate_cfg(maxrate, &exp, &mantissa, &div_exp);

mutex_lock(&nic->mbox.lock);
req = otx2_mbox_alloc_msg_nix_txschq_cfg(&nic->mbox);
if (!req) {
mutex_unlock(&nic->mbox.lock);
return -ENOMEM;
}

req->lvl = NIX_TXSCH_LVL_TL4;
req->num_regs = 1;
req->reg[0] = NIX_AF_TL4X_PIR(txschq);
req->regval[0] = FIELD_PREP(TLX_BURST_EXPONENT, burst_exp) |
FIELD_PREP(TLX_BURST_MANTISSA, burst_mantissa) |
FIELD_PREP(TLX_RATE_DIVIDER_EXPONENT, div_exp) |
FIELD_PREP(TLX_RATE_EXPONENT, exp) |
FIELD_PREP(TLX_RATE_MANTISSA, mantissa) | BIT_ULL(0);

err = otx2_sync_mbox_msg(&nic->mbox);
mutex_unlock(&nic->mbox.lock);
return err;
}

static int otx2_tc_validate_flow(struct otx2_nic *nic,
struct flow_action *actions,
struct netlink_ext_ack *extack)
{
if (nic->flags & OTX2_FLAG_INTF_DOWN) {
NL_SET_ERR_MSG_MOD(extack, "Interface not initialized");
return -EINVAL;
}

if (!flow_action_has_entries(actions)) {
NL_SET_ERR_MSG_MOD(extack, "MATCHALL offload called with no action");
return -EINVAL;
}

if (!flow_offload_has_one_action(actions)) {
NL_SET_ERR_MSG_MOD(extack,
"Egress MATCHALL offload supports only 1 policing action");
return -EINVAL;
}
return 0;
}

static int otx2_tc_egress_matchall_install(struct otx2_nic *nic,
struct tc_cls_matchall_offload *cls)
{
struct netlink_ext_ack *extack = cls->common.extack;
struct flow_action *actions = &cls->rule->action;
struct flow_action_entry *entry;
u32 rate;
int err;

err = otx2_tc_validate_flow(nic, actions, extack);
if (err)
return err;

if (nic->flags & OTX2_FLAG_TC_MATCHALL_EGRESS_ENABLED) {
NL_SET_ERR_MSG_MOD(extack,
"Only one Egress MATCHALL ratelimitter can be offloaded");
return -ENOMEM;
}

entry = &cls->rule->action.entries[0];
switch (entry->id) {
case FLOW_ACTION_POLICE:
if (entry->police.rate_pkt_ps) {
NL_SET_ERR_MSG_MOD(extack, "QoS offload not support packets per second");
return -EOPNOTSUPP;
}
/* Convert bytes per second to Mbps */
rate = entry->police.rate_bytes_ps * 8;
rate = max_t(u32, rate / 1000000, 1);
err = otx2_set_matchall_egress_rate(nic, entry->police.burst, rate);
if (err)
return err;
nic->flags |= OTX2_FLAG_TC_MATCHALL_EGRESS_ENABLED;
break;
default:
NL_SET_ERR_MSG_MOD(extack,
"Only police action is supported with Egress MATCHALL offload");
return -EOPNOTSUPP;
}

return 0;
}

static int otx2_tc_egress_matchall_delete(struct otx2_nic *nic,
struct tc_cls_matchall_offload *cls)
{
struct netlink_ext_ack *extack = cls->common.extack;
int err;

if (nic->flags & OTX2_FLAG_INTF_DOWN) {
NL_SET_ERR_MSG_MOD(extack, "Interface not initialized");
return -EINVAL;
}

err = otx2_set_matchall_egress_rate(nic, 0, 0);
nic->flags &= ~OTX2_FLAG_TC_MATCHALL_EGRESS_ENABLED;
return err;
}

static int otx2_tc_parse_actions(struct otx2_nic *nic,
struct flow_action *flow_action,
struct npc_install_flow_req *req)
Expand Down Expand Up @@ -504,22 +692,64 @@ static int otx2_setup_tc_block_ingress_cb(enum tc_setup_type type,
return -EOPNOTSUPP;
}

static int otx2_setup_tc_egress_matchall(struct otx2_nic *nic,
struct tc_cls_matchall_offload *cls_matchall)
{
switch (cls_matchall->command) {
case TC_CLSMATCHALL_REPLACE:
return otx2_tc_egress_matchall_install(nic, cls_matchall);
case TC_CLSMATCHALL_DESTROY:
return otx2_tc_egress_matchall_delete(nic, cls_matchall);
case TC_CLSMATCHALL_STATS:
default:
break;
}

return -EOPNOTSUPP;
}

static int otx2_setup_tc_block_egress_cb(enum tc_setup_type type,
void *type_data, void *cb_priv)
{
struct otx2_nic *nic = cb_priv;

if (!tc_cls_can_offload_and_chain0(nic->netdev, type_data))
return -EOPNOTSUPP;

switch (type) {
case TC_SETUP_CLSMATCHALL:
return otx2_setup_tc_egress_matchall(nic, type_data);
default:
break;
}

return -EOPNOTSUPP;
}

static LIST_HEAD(otx2_block_cb_list);

static int otx2_setup_tc_block(struct net_device *netdev,
struct flow_block_offload *f)
{
struct otx2_nic *nic = netdev_priv(netdev);
flow_setup_cb_t *cb;
bool ingress;

if (f->block_shared)
return -EOPNOTSUPP;

if (f->binder_type != FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS)
if (f->binder_type == FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS) {
cb = otx2_setup_tc_block_ingress_cb;
ingress = true;
} else if (f->binder_type == FLOW_BLOCK_BINDER_TYPE_CLSACT_EGRESS) {
cb = otx2_setup_tc_block_egress_cb;
ingress = false;
} else {
return -EOPNOTSUPP;
}

return flow_block_cb_setup_simple(f, &otx2_block_cb_list,
otx2_setup_tc_block_ingress_cb,
nic, nic, true);
return flow_block_cb_setup_simple(f, &otx2_block_cb_list, cb,
nic, nic, ingress);
}

int otx2_setup_tc(struct net_device *netdev, enum tc_setup_type type,
Expand Down

0 comments on commit e638a83

Please sign in to comment.