freebsd-nq/sys/dev/cxgbe/t4_sched.c

907 lines
22 KiB
C
Raw Normal View History

/*-
* Copyright (c) 2017 Chelsio Communications, Inc.
* All rights reserved.
* Written by: Navdeep Parhar <np@FreeBSD.org>
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_inet.h"
#include "opt_inet6.h"
#include "opt_ratelimit.h"
#include <sys/types.h>
#include <sys/malloc.h>
#include <sys/queue.h>
#include <sys/sbuf.h>
#include <sys/taskqueue.h>
#include <sys/sysctl.h>
#include "common/common.h"
#include "common/t4_regs.h"
#include "common/t4_regs_values.h"
#include "common/t4_msg.h"
static int
in_range(int val, int lo, int hi)
{
return (val < 0 || (val <= hi && val >= lo));
}
static int
set_sched_class_config(struct adapter *sc, int minmax)
{
int rc;
if (minmax < 0)
return (EINVAL);
rc = begin_synchronized_op(sc, NULL, SLEEP_OK | INTR_OK, "t4sscc");
if (rc)
return (rc);
rc = -t4_sched_config(sc, FW_SCHED_TYPE_PKTSCHED, minmax, 1);
end_synchronized_op(sc, 0);
return (rc);
}
static int
set_sched_class_params(struct adapter *sc, struct t4_sched_class_params *p,
int sleep_ok)
{
int rc, top_speed, fw_level, fw_mode, fw_rateunit, fw_ratemode;
struct port_info *pi;
struct tx_cl_rl_params *tc, old;
bool check_pktsize = false;
if (p->level == SCHED_CLASS_LEVEL_CL_RL)
fw_level = FW_SCHED_PARAMS_LEVEL_CL_RL;
else if (p->level == SCHED_CLASS_LEVEL_CL_WRR)
fw_level = FW_SCHED_PARAMS_LEVEL_CL_WRR;
else if (p->level == SCHED_CLASS_LEVEL_CH_RL)
fw_level = FW_SCHED_PARAMS_LEVEL_CH_RL;
else
return (EINVAL);
if (p->level == SCHED_CLASS_LEVEL_CL_RL) {
if (p->mode == SCHED_CLASS_MODE_CLASS)
fw_mode = FW_SCHED_PARAMS_MODE_CLASS;
else if (p->mode == SCHED_CLASS_MODE_FLOW) {
check_pktsize = true;
fw_mode = FW_SCHED_PARAMS_MODE_FLOW;
} else
return (EINVAL);
} else
fw_mode = 0;
/* Valid channel must always be provided. */
if (p->channel < 0)
return (EINVAL);
if (!in_range(p->channel, 0, sc->chip_params->nchan - 1))
return (ERANGE);
pi = sc->port[sc->chan_map[p->channel]];
if (pi == NULL)
return (ENXIO);
MPASS(pi->tx_chan == p->channel);
top_speed = port_top_speed(pi) * 1000000; /* Gbps -> Kbps */
if (p->level == SCHED_CLASS_LEVEL_CL_RL ||
p->level == SCHED_CLASS_LEVEL_CH_RL) {
/*
* Valid rate (mode, unit and values) must be provided.
*/
if (p->minrate < 0)
p->minrate = 0;
if (p->maxrate < 0)
return (EINVAL);
if (p->rateunit == SCHED_CLASS_RATEUNIT_BITS) {
fw_rateunit = FW_SCHED_PARAMS_UNIT_BITRATE;
/* ratemode could be relative (%) or absolute. */
if (p->ratemode == SCHED_CLASS_RATEMODE_REL) {
fw_ratemode = FW_SCHED_PARAMS_RATE_REL;
/* maxrate is % of port bandwidth. */
if (!in_range(p->minrate, 0, 100) ||
!in_range(p->maxrate, 0, 100)) {
return (ERANGE);
}
} else if (p->ratemode == SCHED_CLASS_RATEMODE_ABS) {
fw_ratemode = FW_SCHED_PARAMS_RATE_ABS;
/* maxrate is absolute value in kbps. */
if (!in_range(p->minrate, 0, top_speed) ||
!in_range(p->maxrate, 0, top_speed)) {
return (ERANGE);
}
} else
return (EINVAL);
} else if (p->rateunit == SCHED_CLASS_RATEUNIT_PKTS) {
/* maxrate is the absolute value in pps. */
check_pktsize = true;
fw_rateunit = FW_SCHED_PARAMS_UNIT_PKTRATE;
} else
return (EINVAL);
} else {
MPASS(p->level == SCHED_CLASS_LEVEL_CL_WRR);
/*
* Valid weight must be provided.
*/
if (p->weight < 0)
return (EINVAL);
if (!in_range(p->weight, 1, 99))
return (ERANGE);
fw_rateunit = 0;
fw_ratemode = 0;
}
if (p->level == SCHED_CLASS_LEVEL_CL_RL ||
p->level == SCHED_CLASS_LEVEL_CL_WRR) {
/*
* Valid scheduling class must be provided.
*/
if (p->cl < 0)
return (EINVAL);
if (!in_range(p->cl, 0, sc->chip_params->nsched_cls - 1))
return (ERANGE);
}
if (check_pktsize) {
if (p->pktsize < 0)
return (EINVAL);
if (!in_range(p->pktsize, 64, pi->vi[0].ifp->if_mtu))
return (ERANGE);
}
if (p->level == SCHED_CLASS_LEVEL_CL_RL) {
tc = &pi->sched_params->cl_rl[p->cl];
mtx_lock(&sc->tc_lock);
if (tc->refcount > 0 || tc->flags & (CLRL_SYNC | CLRL_ASYNC))
rc = EBUSY;
else {
tc->flags |= CLRL_SYNC | CLRL_USER;
tc->ratemode = fw_ratemode;
tc->rateunit = fw_rateunit;
tc->mode = fw_mode;
tc->maxrate = p->maxrate;
tc->pktsize = p->pktsize;
rc = 0;
old= *tc;
}
mtx_unlock(&sc->tc_lock);
if (rc != 0)
return (rc);
}
rc = begin_synchronized_op(sc, NULL,
sleep_ok ? (SLEEP_OK | INTR_OK) : HOLD_LOCK, "t4sscp");
if (rc != 0) {
if (p->level == SCHED_CLASS_LEVEL_CL_RL) {
mtx_lock(&sc->tc_lock);
*tc = old;
mtx_unlock(&sc->tc_lock);
}
return (rc);
}
rc = -t4_sched_params(sc, FW_SCHED_TYPE_PKTSCHED, fw_level, fw_mode,
fw_rateunit, fw_ratemode, p->channel, p->cl, p->minrate, p->maxrate,
p->weight, p->pktsize, 0, sleep_ok);
end_synchronized_op(sc, sleep_ok ? 0 : LOCK_HELD);
if (p->level == SCHED_CLASS_LEVEL_CL_RL) {
mtx_lock(&sc->tc_lock);
MPASS(tc->flags & CLRL_SYNC);
MPASS(tc->flags & CLRL_USER);
MPASS(tc->refcount == 0);
tc->flags &= ~CLRL_SYNC;
if (rc == 0)
tc->flags &= ~CLRL_ERR;
else
tc->flags |= CLRL_ERR;
mtx_unlock(&sc->tc_lock);
}
return (rc);
}
static void
update_tx_sched(void *context, int pending)
{
int i, j, rc;
struct port_info *pi;
struct tx_cl_rl_params *tc;
struct adapter *sc = context;
const int n = sc->chip_params->nsched_cls;
mtx_lock(&sc->tc_lock);
for_each_port(sc, i) {
pi = sc->port[i];
tc = &pi->sched_params->cl_rl[0];
for (j = 0; j < n; j++, tc++) {
MPASS(mtx_owned(&sc->tc_lock));
if ((tc->flags & CLRL_ASYNC) == 0)
continue;
mtx_unlock(&sc->tc_lock);
if (begin_synchronized_op(sc, NULL, SLEEP_OK | INTR_OK,
"t4utxs") != 0) {
mtx_lock(&sc->tc_lock);
continue;
}
rc = -t4_sched_params(sc, FW_SCHED_TYPE_PKTSCHED,
FW_SCHED_PARAMS_LEVEL_CL_RL, tc->mode, tc->rateunit,
tc->ratemode, pi->tx_chan, j, 0, tc->maxrate, 0,
tc->pktsize, tc->burstsize, 1);
end_synchronized_op(sc, 0);
mtx_lock(&sc->tc_lock);
MPASS(tc->flags & CLRL_ASYNC);
tc->flags &= ~CLRL_ASYNC;
if (rc == 0)
tc->flags &= ~CLRL_ERR;
else
tc->flags |= CLRL_ERR;
}
}
mtx_unlock(&sc->tc_lock);
}
int
t4_set_sched_class(struct adapter *sc, struct t4_sched_params *p)
{
if (p->type != SCHED_CLASS_TYPE_PACKET)
return (EINVAL);
if (p->subcmd == SCHED_CLASS_SUBCMD_CONFIG)
return (set_sched_class_config(sc, p->u.config.minmax));
if (p->subcmd == SCHED_CLASS_SUBCMD_PARAMS)
return (set_sched_class_params(sc, &p->u.params, 1));
return (EINVAL);
}
static int
bind_txq_to_traffic_class(struct adapter *sc, struct sge_txq *txq, int idx)
{
struct tx_cl_rl_params *tc0, *tc;
int rc, old_idx;
uint32_t fw_mnem, fw_class;
if (!(txq->eq.flags & EQ_ALLOCATED))
return (EAGAIN);
mtx_lock(&sc->tc_lock);
if (txq->tc_idx == -2) {
rc = EBUSY; /* Another bind/unbind in progress already. */
goto done;
}
if (idx == txq->tc_idx) {
rc = 0; /* No change, nothing to do. */
goto done;
}
tc0 = &sc->port[txq->eq.tx_chan]->sched_params->cl_rl[0];
if (idx != -1) {
/*
* Bind to a different class at index idx.
*/
tc = &tc0[idx];
if (tc->flags & CLRL_ERR) {
rc = ENXIO;
goto done;
} else {
/*
* Ok to proceed. Place a reference on the new class
* while still holding on to the reference on the
* previous class, if any.
*/
tc->refcount++;
}
}
/* Mark as busy before letting go of the lock. */
old_idx = txq->tc_idx;
txq->tc_idx = -2;
mtx_unlock(&sc->tc_lock);
rc = begin_synchronized_op(sc, NULL, SLEEP_OK | INTR_OK, "t4btxq");
if (rc != 0)
return (rc);
fw_mnem = (V_FW_PARAMS_MNEM(FW_PARAMS_MNEM_DMAQ) |
V_FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_DMAQ_EQ_SCHEDCLASS_ETH) |
V_FW_PARAMS_PARAM_YZ(txq->eq.cntxt_id));
fw_class = idx < 0 ? 0xffffffff : idx;
rc = -t4_set_params(sc, sc->mbox, sc->pf, 0, 1, &fw_mnem, &fw_class);
end_synchronized_op(sc, 0);
mtx_lock(&sc->tc_lock);
MPASS(txq->tc_idx == -2);
if (rc == 0) {
/*
* Unbind, bind, or bind to a different class succeeded. Remove
* the reference on the old traffic class, if any.
*/
if (old_idx != -1) {
tc = &tc0[old_idx];
MPASS(tc->refcount > 0);
tc->refcount--;
}
txq->tc_idx = idx;
} else {
/*
* Unbind, bind, or bind to a different class failed. Remove
* the anticipatory reference on the new traffic class, if any.
*/
if (idx != -1) {
tc = &tc0[idx];
MPASS(tc->refcount > 0);
tc->refcount--;
}
txq->tc_idx = old_idx;
}
done:
MPASS(txq->tc_idx >= -1 && txq->tc_idx < sc->chip_params->nsched_cls);
mtx_unlock(&sc->tc_lock);
return (rc);
}
int
t4_set_sched_queue(struct adapter *sc, struct t4_sched_queue *p)
{
struct port_info *pi = NULL;
struct vi_info *vi;
struct sge_txq *txq;
int i, rc;
if (p->port >= sc->params.nports)
return (EINVAL);
/*
* XXX: cxgbetool allows the user to specify the physical port only. So
* we always operate on the main VI.
*/
pi = sc->port[p->port];
vi = &pi->vi[0];
/* Checking VI_INIT_DONE outside a synch-op is a harmless race here. */
if (!(vi->flags & VI_INIT_DONE))
return (EAGAIN);
MPASS(vi->ntxq > 0);
if (!in_range(p->queue, 0, vi->ntxq - 1) ||
!in_range(p->cl, 0, sc->chip_params->nsched_cls - 1))
return (EINVAL);
if (p->queue < 0) {
/*
* Change the scheduling on all the TX queues for the
* interface.
*/
for_each_txq(vi, i, txq) {
rc = bind_txq_to_traffic_class(sc, txq, p->cl);
if (rc != 0)
break;
}
} else {
/*
* If op.queue is non-negative, then we're only changing the
* scheduling on a single specified TX queue.
*/
txq = &sc->sge.txq[vi->first_txq + p->queue];
rc = bind_txq_to_traffic_class(sc, txq, p->cl);
}
return (rc);
}
int
t4_init_tx_sched(struct adapter *sc)
{
int i, j;
const int n = sc->chip_params->nsched_cls;
struct port_info *pi;
struct tx_cl_rl_params *tc;
mtx_init(&sc->tc_lock, "tx_sched lock", NULL, MTX_DEF);
TASK_INIT(&sc->tc_task, 0, update_tx_sched, sc);
for_each_port(sc, i) {
pi = sc->port[i];
pi->sched_params = malloc(sizeof(*pi->sched_params) +
n * sizeof(*tc), M_CXGBE, M_ZERO | M_WAITOK);
tc = &pi->sched_params->cl_rl[0];
for (j = 0; j < n; j++, tc++) {
tc->refcount = 0;
tc->ratemode = FW_SCHED_PARAMS_RATE_ABS;
tc->rateunit = FW_SCHED_PARAMS_UNIT_BITRATE;
tc->mode = FW_SCHED_PARAMS_MODE_CLASS;
tc->maxrate = 1000 * 1000; /* 1 Gbps. Arbitrary */
if (t4_sched_params_cl_rl_kbps(sc, pi->tx_chan, j,
tc->mode, tc->maxrate, tc->pktsize, 1) != 0)
tc->flags = CLRL_ERR;
}
}
return (0);
}
int
t4_free_tx_sched(struct adapter *sc)
{
int i;
taskqueue_drain(taskqueue_thread, &sc->tc_task);
for_each_port(sc, i) {
if (sc->port[i] != NULL)
free(sc->port[i]->sched_params, M_CXGBE);
}
if (mtx_initialized(&sc->tc_lock))
mtx_destroy(&sc->tc_lock);
return (0);
}
void
t4_update_tx_sched(struct adapter *sc)
{
taskqueue_enqueue(taskqueue_thread, &sc->tc_task);
}
int
t4_reserve_cl_rl_kbps(struct adapter *sc, int port_id, u_int maxrate,
int *tc_idx)
{
int rc = 0, fa = -1, i, pktsize, burstsize;
bool update;
struct tx_cl_rl_params *tc;
struct port_info *pi;
MPASS(port_id >= 0 && port_id < sc->params.nports);
pi = sc->port[port_id];
if (pi->sched_params->pktsize > 0)
pktsize = pi->sched_params->pktsize;
else
pktsize = pi->vi[0].ifp->if_mtu;
if (pi->sched_params->burstsize > 0)
burstsize = pi->sched_params->burstsize;
else
burstsize = pktsize * 4;
tc = &pi->sched_params->cl_rl[0];
update = false;
mtx_lock(&sc->tc_lock);
for (i = 0; i < sc->chip_params->nsched_cls; i++, tc++) {
if (fa < 0 && tc->refcount == 0 && !(tc->flags & CLRL_USER))
fa = i; /* first available */
if (tc->ratemode == FW_SCHED_PARAMS_RATE_ABS &&
tc->rateunit == FW_SCHED_PARAMS_UNIT_BITRATE &&
tc->mode == FW_SCHED_PARAMS_MODE_FLOW &&
tc->maxrate == maxrate && tc->pktsize == pktsize &&
tc->burstsize == burstsize) {
tc->refcount++;
*tc_idx = i;
if ((tc->flags & (CLRL_ERR | CLRL_ASYNC | CLRL_SYNC)) ==
CLRL_ERR) {
update = true;
}
goto done;
}
}
/* Not found */
MPASS(i == sc->chip_params->nsched_cls);
if (fa != -1) {
tc = &pi->sched_params->cl_rl[fa];
tc->refcount = 1;
tc->ratemode = FW_SCHED_PARAMS_RATE_ABS;
tc->rateunit = FW_SCHED_PARAMS_UNIT_BITRATE;
tc->mode = FW_SCHED_PARAMS_MODE_FLOW;
tc->maxrate = maxrate;
tc->pktsize = pktsize;
tc->burstsize = burstsize;
*tc_idx = fa;
update = true;
} else {
*tc_idx = -1;
rc = ENOSPC;
}
done:
mtx_unlock(&sc->tc_lock);
if (update) {
tc->flags |= CLRL_ASYNC;
t4_update_tx_sched(sc);
}
return (rc);
}
void
t4_release_cl_rl(struct adapter *sc, int port_id, int tc_idx)
{
struct tx_cl_rl_params *tc;
MPASS(port_id >= 0 && port_id < sc->params.nports);
MPASS(tc_idx >= 0 && tc_idx < sc->chip_params->nsched_cls);
mtx_lock(&sc->tc_lock);
tc = &sc->port[port_id]->sched_params->cl_rl[tc_idx];
MPASS(tc->refcount > 0);
tc->refcount--;
mtx_unlock(&sc->tc_lock);
}
int
sysctl_tc(SYSCTL_HANDLER_ARGS)
{
struct vi_info *vi = arg1;
struct port_info *pi;
struct adapter *sc;
struct sge_txq *txq;
int qidx = arg2, rc, tc_idx;
MPASS(qidx >= 0 && qidx < vi->ntxq);
pi = vi->pi;
sc = pi->adapter;
txq = &sc->sge.txq[vi->first_txq + qidx];
tc_idx = txq->tc_idx;
rc = sysctl_handle_int(oidp, &tc_idx, 0, req);
if (rc != 0 || req->newptr == NULL)
return (rc);
if (sc->flags & IS_VF)
return (EPERM);
if (!in_range(tc_idx, 0, sc->chip_params->nsched_cls - 1))
return (EINVAL);
return (bind_txq_to_traffic_class(sc, txq, tc_idx));
}
int
sysctl_tc_params(SYSCTL_HANDLER_ARGS)
{
struct adapter *sc = arg1;
struct tx_cl_rl_params tc;
struct sbuf *sb;
int i, rc, port_id, mbps, gbps;
rc = sysctl_wire_old_buffer(req, 0);
if (rc != 0)
return (rc);
sb = sbuf_new_for_sysctl(NULL, NULL, 4096, req);
if (sb == NULL)
return (ENOMEM);
port_id = arg2 >> 16;
MPASS(port_id < sc->params.nports);
MPASS(sc->port[port_id] != NULL);
i = arg2 & 0xffff;
MPASS(i < sc->chip_params->nsched_cls);
mtx_lock(&sc->tc_lock);
tc = sc->port[port_id]->sched_params->cl_rl[i];
mtx_unlock(&sc->tc_lock);
switch (tc.rateunit) {
case SCHED_CLASS_RATEUNIT_BITS:
switch (tc.ratemode) {
case SCHED_CLASS_RATEMODE_REL:
/* XXX: top speed or actual link speed? */
gbps = port_top_speed(sc->port[port_id]);
sbuf_printf(sb, "%u%% of %uGbps", tc.maxrate, gbps);
break;
case SCHED_CLASS_RATEMODE_ABS:
mbps = tc.maxrate / 1000;
gbps = tc.maxrate / 1000000;
if (tc.maxrate == gbps * 1000000)
sbuf_printf(sb, "%uGbps", gbps);
else if (tc.maxrate == mbps * 1000)
sbuf_printf(sb, "%uMbps", mbps);
else
sbuf_printf(sb, "%uKbps", tc.maxrate);
break;
default:
rc = ENXIO;
goto done;
}
break;
case SCHED_CLASS_RATEUNIT_PKTS:
sbuf_printf(sb, "%upps", tc.maxrate);
break;
default:
rc = ENXIO;
goto done;
}
switch (tc.mode) {
case SCHED_CLASS_MODE_CLASS:
sbuf_printf(sb, " aggregate");
break;
case SCHED_CLASS_MODE_FLOW:
sbuf_printf(sb, " per-flow");
if (tc.pktsize > 0)
sbuf_printf(sb, " pkt-size %u", tc.pktsize);
if (tc.burstsize > 0)
sbuf_printf(sb, " burst-size %u", tc.burstsize);
break;
default:
rc = ENXIO;
goto done;
}
done:
if (rc == 0)
rc = sbuf_finish(sb);
sbuf_delete(sb);
return (rc);
}
#ifdef RATELIMIT
void
t4_init_etid_table(struct adapter *sc)
{
int i;
struct tid_info *t;
if (!is_ethoffload(sc))
return;
t = &sc->tids;
MPASS(t->netids > 0);
mtx_init(&t->etid_lock, "etid lock", NULL, MTX_DEF);
t->etid_tab = malloc(sizeof(*t->etid_tab) * t->netids, M_CXGBE,
M_ZERO | M_WAITOK);
t->efree = t->etid_tab;
t->etids_in_use = 0;
for (i = 1; i < t->netids; i++)
t->etid_tab[i - 1].next = &t->etid_tab[i];
t->etid_tab[t->netids - 1].next = NULL;
}
void
t4_free_etid_table(struct adapter *sc)
{
struct tid_info *t;
if (!is_ethoffload(sc))
return;
t = &sc->tids;
MPASS(t->netids > 0);
free(t->etid_tab, M_CXGBE);
t->etid_tab = NULL;
if (mtx_initialized(&t->etid_lock))
mtx_destroy(&t->etid_lock);
}
/* etid services */
static int alloc_etid(struct adapter *, struct cxgbe_snd_tag *);
static void free_etid(struct adapter *, int);
static int
alloc_etid(struct adapter *sc, struct cxgbe_snd_tag *cst)
{
struct tid_info *t = &sc->tids;
int etid = -1;
mtx_lock(&t->etid_lock);
if (t->efree) {
union etid_entry *p = t->efree;
etid = p - t->etid_tab + t->etid_base;
t->efree = p->next;
p->cst = cst;
t->etids_in_use++;
}
mtx_unlock(&t->etid_lock);
return (etid);
}
struct cxgbe_snd_tag *
lookup_etid(struct adapter *sc, int etid)
{
struct tid_info *t = &sc->tids;
return (t->etid_tab[etid - t->etid_base].cst);
}
static void
free_etid(struct adapter *sc, int etid)
{
struct tid_info *t = &sc->tids;
union etid_entry *p = &t->etid_tab[etid - t->etid_base];
mtx_lock(&t->etid_lock);
p->next = t->efree;
t->efree = p;
t->etids_in_use--;
mtx_unlock(&t->etid_lock);
}
int
cxgbe_snd_tag_alloc(struct ifnet *ifp, union if_snd_tag_alloc_params *params,
struct m_snd_tag **pt)
{
int rc, schedcl;
struct vi_info *vi = ifp->if_softc;
struct port_info *pi = vi->pi;
struct adapter *sc = pi->adapter;
struct cxgbe_snd_tag *cst;
if (params->hdr.type != IF_SND_TAG_TYPE_RATE_LIMIT)
return (ENOTSUP);
rc = t4_reserve_cl_rl_kbps(sc, pi->port_id,
(params->rate_limit.max_rate * 8ULL / 1000), &schedcl);
if (rc != 0)
return (rc);
MPASS(schedcl >= 0 && schedcl < sc->chip_params->nsched_cls);
cst = malloc(sizeof(*cst), M_CXGBE, M_ZERO | M_NOWAIT);
if (cst == NULL) {
failed:
t4_release_cl_rl(sc, pi->port_id, schedcl);
return (ENOMEM);
}
cst->etid = alloc_etid(sc, cst);
if (cst->etid < 0) {
free(cst, M_CXGBE);
goto failed;
}
mtx_init(&cst->lock, "cst_lock", NULL, MTX_DEF);
mbufq_init(&cst->pending_tx, INT_MAX);
mbufq_init(&cst->pending_fwack, INT_MAX);
Restructure mbuf send tags to provide stronger guarantees. - Perform ifp mismatch checks (to determine if a send tag is allocated for a different ifp than the one the packet is being output on), in ip_output() and ip6_output(). This avoids sending packets with send tags to ifnet drivers that don't support send tags. Since we are now checking for ifp mismatches before invoking if_output, we can now try to allocate a new tag before invoking if_output sending the original packet on the new tag if allocation succeeds. To avoid code duplication for the fragment and unfragmented cases, add ip_output_send() and ip6_output_send() as wrappers around if_output and nd6_output_ifp, respectively. All of the logic for setting send tags and dealing with send tag-related errors is done in these wrapper functions. For pseudo interfaces that wrap other network interfaces (vlan and lagg), wrapper send tags are now allocated so that ip*_output see the wrapper ifp as the ifp in the send tag. The if_transmit routines rewrite the send tags after performing an ifp mismatch check. If an ifp mismatch is detected, the transmit routines fail with EAGAIN. - To provide clearer life cycle management of send tags, especially in the presence of vlan and lagg wrapper tags, add a reference count to send tags managed via m_snd_tag_ref() and m_snd_tag_rele(). Provide a helper function (m_snd_tag_init()) for use by drivers supporting send tags. m_snd_tag_init() takes care of the if_ref on the ifp meaning that code alloating send tags via if_snd_tag_alloc no longer has to manage that manually. Similarly, m_snd_tag_rele drops the refcount on the ifp after invoking if_snd_tag_free when the last reference to a send tag is dropped. This also closes use after free races if there are pending packets in driver tx rings after the socket is closed (e.g. from tcpdrop). In order for m_free to work reliably, add a new CSUM_SND_TAG flag in csum_flags to indicate 'snd_tag' is set (rather than 'rcvif'). Drivers now also check this flag instead of checking snd_tag against NULL. This avoids false positive matches when a forwarded packet has a non-NULL rcvif that was treated as a send tag. - cxgbe was relying on snd_tag_free being called when the inp was detached so that it could kick the firmware to flush any pending work on the flow. This is because the driver doesn't require ACK messages from the firmware for every request, but instead does a kind of manual interrupt coalescing by only setting a flag to request a completion on a subset of requests. If all of the in-flight requests don't have the flag when the tag is detached from the inp, the flow might never return the credits. The current snd_tag_free command issues a flush command to force the credits to return. However, the credit return is what also frees the mbufs, and since those mbufs now hold references on the tag, this meant that snd_tag_free would never be called. To fix, explicitly drop the mbuf's reference on the snd tag when the mbuf is queued in the firmware work queue. This means that once the inp's reference on the tag goes away and all in-flight mbufs have been queued to the firmware, tag's refcount will drop to zero and snd_tag_free will kick in and send the flush request. Note that we need to avoid doing this in the middle of ethofld_tx(), so the driver grabs a temporary reference on the tag around that loop to defer the free to the end of the function in case it sends the last mbuf to the queue after the inp has dropped its reference on the tag. - mlx5 preallocates send tags and was using the ifp pointer even when the send tag wasn't in use. Explicitly use the ifp from other data structures instead. - Sprinkle some assertions in various places to assert that received packets don't have a send tag, and that other places that overwrite rcvif (e.g. 802.11 transmit) don't clobber a send tag pointer. Reviewed by: gallatin, hselasky, rgrimes, ae Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D20117
2019-05-24 22:30:40 +00:00
m_snd_tag_init(&cst->com, ifp);
cst->flags |= EO_FLOWC_PENDING | EO_SND_TAG_REF;
cst->adapter = sc;
cst->port_id = pi->port_id;
cst->schedcl = schedcl;
cst->max_rate = params->rate_limit.max_rate;
cst->tx_credits = sc->params.eo_wr_cred;
cst->tx_total = cst->tx_credits;
cst->plen = 0;
cst->ctrl0 = htobe32(V_TXPKT_OPCODE(CPL_TX_PKT) |
V_TXPKT_INTF(pi->tx_chan) | V_TXPKT_PF(sc->pf) |
V_TXPKT_VF(vi->vin) | V_TXPKT_VF_VLD(vi->vfvld));
/*
* Queues will be selected later when the connection flowid is available.
*/
*pt = &cst->com;
return (0);
}
/*
* Change in parameters, no change in ifp.
*/
int
cxgbe_snd_tag_modify(struct m_snd_tag *mst,
union if_snd_tag_modify_params *params)
{
int rc, schedcl;
struct cxgbe_snd_tag *cst = mst_to_cst(mst);
struct adapter *sc = cst->adapter;
/* XXX: is schedcl -1 ok here? */
MPASS(cst->schedcl >= 0 && cst->schedcl < sc->chip_params->nsched_cls);
mtx_lock(&cst->lock);
MPASS(cst->flags & EO_SND_TAG_REF);
rc = t4_reserve_cl_rl_kbps(sc, cst->port_id,
(params->rate_limit.max_rate * 8ULL / 1000), &schedcl);
if (rc != 0)
return (rc);
MPASS(schedcl >= 0 && schedcl < sc->chip_params->nsched_cls);
t4_release_cl_rl(sc, cst->port_id, cst->schedcl);
cst->schedcl = schedcl;
cst->max_rate = params->rate_limit.max_rate;
mtx_unlock(&cst->lock);
return (0);
}
int
cxgbe_snd_tag_query(struct m_snd_tag *mst,
union if_snd_tag_query_params *params)
{
struct cxgbe_snd_tag *cst = mst_to_cst(mst);
params->rate_limit.max_rate = cst->max_rate;
#define CST_TO_MST_QLEVEL_SCALE (IF_SND_QUEUE_LEVEL_MAX / cst->tx_total)
params->rate_limit.queue_level =
(cst->tx_total - cst->tx_credits) * CST_TO_MST_QLEVEL_SCALE;
return (0);
}
/*
* Unlocks cst and frees it.
*/
void
cxgbe_snd_tag_free_locked(struct cxgbe_snd_tag *cst)
{
struct adapter *sc = cst->adapter;
mtx_assert(&cst->lock, MA_OWNED);
MPASS((cst->flags & EO_SND_TAG_REF) == 0);
MPASS(cst->tx_credits == cst->tx_total);
MPASS(cst->plen == 0);
MPASS(mbufq_first(&cst->pending_tx) == NULL);
MPASS(mbufq_first(&cst->pending_fwack) == NULL);
if (cst->etid >= 0)
free_etid(sc, cst->etid);
if (cst->schedcl != -1)
t4_release_cl_rl(sc, cst->port_id, cst->schedcl);
mtx_unlock(&cst->lock);
mtx_destroy(&cst->lock);
free(cst, M_CXGBE);
}
void
cxgbe_snd_tag_free(struct m_snd_tag *mst)
{
struct cxgbe_snd_tag *cst = mst_to_cst(mst);
mtx_lock(&cst->lock);
/* The kernel is done with the snd_tag. Remove its reference. */
MPASS(cst->flags & EO_SND_TAG_REF);
cst->flags &= ~EO_SND_TAG_REF;
if (cst->ncompl == 0) {
/*
* No fw4_ack in flight. Free the tag right away if there are
* no outstanding credits. Request the firmware to return all
* credits for the etid otherwise.
*/
if (cst->tx_credits == cst->tx_total) {
cxgbe_snd_tag_free_locked(cst);
return; /* cst is gone. */
}
send_etid_flush_wr(cst);
}
mtx_unlock(&cst->lock);
}
#endif