linux/drivers/net/ethernet/intel/ice/ice_dcb_nl.c
Dave Ertman 741b7b743b ice: remove DCBNL_DEVRESET bit from PF state
The original purpose of the ICE_DCBNL_DEVRESET was to protect
the driver during DCBNL device resets.  But, the flow for
DCBNL device resets now consists of only calls up the stack
such as dev_close() and dev_open() that will result in NDO calls
to the driver.  These will be handled with state changes from the
stack.  Also, there is a problem of the dev_close and dev_open
being blocked by checks for reset in progress also using the
ICE_DCBNL_DEVRESET bit.

Since the ICE_DCBNL_DEVRESET bit is not necessary for protecting
the driver from DCBNL device resets and it is actually blocking
changes coming from the DCBNL interface, remove the bit from the
PF state and don't block driver function based on DCBNL reset in
progress.

Fixes: b94b013eb6 ("ice: Implement DCBNL support")
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Tested-by: Tony Brelinski <tonyx.brelinski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-03-29 10:37:19 -07:00

975 lines
25 KiB
C

// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2019, Intel Corporation. */
#include "ice.h"
#include "ice_dcb.h"
#include "ice_dcb_lib.h"
#include "ice_dcb_nl.h"
#include <net/dcbnl.h>
/**
* ice_dcbnl_devreset - perform enough of a ifdown/ifup to sync DCBNL info
* @netdev: device associated with interface that needs reset
*/
static void ice_dcbnl_devreset(struct net_device *netdev)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
while (ice_is_reset_in_progress(pf->state))
usleep_range(1000, 2000);
dev_close(netdev);
netdev_state_change(netdev);
dev_open(netdev, NULL);
netdev_state_change(netdev);
}
/**
* ice_dcbnl_getets - retrieve local ETS configuration
* @netdev: the relevant netdev
* @ets: struct to hold ETS configuration
*/
static int ice_dcbnl_getets(struct net_device *netdev, struct ieee_ets *ets)
{
struct ice_dcbx_cfg *dcbxcfg;
struct ice_pf *pf;
pf = ice_netdev_to_pf(netdev);
dcbxcfg = &pf->hw.port_info->qos_cfg.local_dcbx_cfg;
ets->willing = dcbxcfg->etscfg.willing;
ets->ets_cap = dcbxcfg->etscfg.maxtcs;
ets->cbs = dcbxcfg->etscfg.cbs;
memcpy(ets->tc_tx_bw, dcbxcfg->etscfg.tcbwtable, sizeof(ets->tc_tx_bw));
memcpy(ets->tc_rx_bw, dcbxcfg->etscfg.tcbwtable, sizeof(ets->tc_rx_bw));
memcpy(ets->tc_tsa, dcbxcfg->etscfg.tsatable, sizeof(ets->tc_tsa));
memcpy(ets->prio_tc, dcbxcfg->etscfg.prio_table, sizeof(ets->prio_tc));
memcpy(ets->tc_reco_bw, dcbxcfg->etsrec.tcbwtable,
sizeof(ets->tc_reco_bw));
memcpy(ets->tc_reco_tsa, dcbxcfg->etsrec.tsatable,
sizeof(ets->tc_reco_tsa));
memcpy(ets->reco_prio_tc, dcbxcfg->etscfg.prio_table,
sizeof(ets->reco_prio_tc));
return 0;
}
/**
* ice_dcbnl_setets - set IEEE ETS configuration
* @netdev: pointer to relevant netdev
* @ets: struct to hold ETS configuration
*/
static int ice_dcbnl_setets(struct net_device *netdev, struct ieee_ets *ets)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *new_cfg;
int bwcfg = 0, bwrec = 0;
int err, i, max_tc = 0;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_IEEE))
return -EINVAL;
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
mutex_lock(&pf->tc_mutex);
new_cfg->etscfg.willing = ets->willing;
new_cfg->etscfg.cbs = ets->cbs;
ice_for_each_traffic_class(i) {
new_cfg->etscfg.tcbwtable[i] = ets->tc_tx_bw[i];
bwcfg += ets->tc_tx_bw[i];
new_cfg->etscfg.tsatable[i] = ets->tc_tsa[i];
new_cfg->etscfg.prio_table[i] = ets->prio_tc[i];
if (ets->prio_tc[i] > max_tc)
max_tc = ets->prio_tc[i];
new_cfg->etsrec.tcbwtable[i] = ets->tc_reco_bw[i];
bwrec += ets->tc_reco_bw[i];
new_cfg->etsrec.tsatable[i] = ets->tc_reco_tsa[i];
new_cfg->etsrec.prio_table[i] = ets->reco_prio_tc[i];
}
if (ice_dcb_bwchk(pf, new_cfg)) {
err = -EINVAL;
goto ets_out;
}
max_tc = pf->hw.func_caps.common_cap.maxtc;
new_cfg->etscfg.maxtcs = max_tc;
if (!bwcfg)
new_cfg->etscfg.tcbwtable[0] = 100;
if (!bwrec)
new_cfg->etsrec.tcbwtable[0] = 100;
err = ice_pf_dcb_cfg(pf, new_cfg, true);
/* return of zero indicates new cfg applied */
if (err == ICE_DCB_HW_CHG_RST)
ice_dcbnl_devreset(netdev);
if (err == ICE_DCB_NO_HW_CHG)
err = ICE_DCB_HW_CHG_RST;
ets_out:
mutex_unlock(&pf->tc_mutex);
return err;
}
/**
* ice_dcbnl_getnumtcs - Get max number of traffic classes supported
* @dev: pointer to netdev struct
* @tcid: TC ID
* @num: total number of TCs supported by the adapter
*
* Return the total number of TCs supported
*/
static int
ice_dcbnl_getnumtcs(struct net_device *dev, int __always_unused tcid, u8 *num)
{
struct ice_pf *pf = ice_netdev_to_pf(dev);
if (!test_bit(ICE_FLAG_DCB_CAPABLE, pf->flags))
return -EINVAL;
*num = pf->hw.func_caps.common_cap.maxtc;
return 0;
}
/**
* ice_dcbnl_getdcbx - retrieve current DCBX capability
* @netdev: pointer to the netdev struct
*/
static u8 ice_dcbnl_getdcbx(struct net_device *netdev)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
return pf->dcbx_cap;
}
/**
* ice_dcbnl_setdcbx - set required DCBX capability
* @netdev: the corresponding netdev
* @mode: required mode
*/
static u8 ice_dcbnl_setdcbx(struct net_device *netdev, u8 mode)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_qos_cfg *qos_cfg;
/* if FW LLDP agent is running, DCBNL not allowed to change mode */
if (test_bit(ICE_FLAG_FW_LLDP_AGENT, pf->flags))
return ICE_DCB_NO_HW_CHG;
/* No support for LLD_MANAGED modes or CEE+IEEE */
if ((mode & DCB_CAP_DCBX_LLD_MANAGED) ||
((mode & DCB_CAP_DCBX_VER_IEEE) && (mode & DCB_CAP_DCBX_VER_CEE)) ||
!(mode & DCB_CAP_DCBX_HOST))
return ICE_DCB_NO_HW_CHG;
/* Already set to the given mode no change */
if (mode == pf->dcbx_cap)
return ICE_DCB_NO_HW_CHG;
pf->dcbx_cap = mode;
qos_cfg = &pf->hw.port_info->qos_cfg;
if (mode & DCB_CAP_DCBX_VER_CEE)
qos_cfg->local_dcbx_cfg.dcbx_mode = ICE_DCBX_MODE_CEE;
else
qos_cfg->local_dcbx_cfg.dcbx_mode = ICE_DCBX_MODE_IEEE;
dev_info(ice_pf_to_dev(pf), "DCBx mode = 0x%x\n", mode);
return ICE_DCB_HW_CHG_RST;
}
/**
* ice_dcbnl_get_perm_hw_addr - MAC address used by DCBX
* @netdev: pointer to netdev struct
* @perm_addr: buffer to return permanent MAC address
*/
static void ice_dcbnl_get_perm_hw_addr(struct net_device *netdev, u8 *perm_addr)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
int i, j;
memset(perm_addr, 0xff, MAX_ADDR_LEN);
for (i = 0; i < netdev->addr_len; i++)
perm_addr[i] = pi->mac.perm_addr[i];
for (j = 0; j < netdev->addr_len; j++, i++)
perm_addr[i] = pi->mac.perm_addr[j];
}
/**
* ice_get_pfc_delay - Retrieve PFC Link Delay
* @hw: pointer to HW struct
* @delay: holds the PFC Link Delay value
*/
static void ice_get_pfc_delay(struct ice_hw *hw, u16 *delay)
{
u32 val;
val = rd32(hw, PRTDCB_GENC);
*delay = (u16)((val & PRTDCB_GENC_PFCLDA_M) >> PRTDCB_GENC_PFCLDA_S);
}
/**
* ice_dcbnl_getpfc - retrieve local IEEE PFC config
* @netdev: pointer to netdev struct
* @pfc: struct to hold PFC info
*/
static int ice_dcbnl_getpfc(struct net_device *netdev, struct ieee_pfc *pfc)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
struct ice_dcbx_cfg *dcbxcfg;
int i;
dcbxcfg = &pi->qos_cfg.local_dcbx_cfg;
pfc->pfc_cap = dcbxcfg->pfc.pfccap;
pfc->pfc_en = dcbxcfg->pfc.pfcena;
pfc->mbc = dcbxcfg->pfc.mbc;
ice_get_pfc_delay(&pf->hw, &pfc->delay);
ice_for_each_traffic_class(i) {
pfc->requests[i] = pf->stats.priority_xoff_tx[i];
pfc->indications[i] = pf->stats.priority_xoff_rx[i];
}
return 0;
}
/**
* ice_dcbnl_setpfc - set local IEEE PFC config
* @netdev: pointer to relevant netdev
* @pfc: pointer to struct holding PFC config
*/
static int ice_dcbnl_setpfc(struct net_device *netdev, struct ieee_pfc *pfc)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *new_cfg;
int err;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_IEEE))
return -EINVAL;
mutex_lock(&pf->tc_mutex);
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
if (pfc->pfc_cap)
new_cfg->pfc.pfccap = pfc->pfc_cap;
else
new_cfg->pfc.pfccap = pf->hw.func_caps.common_cap.maxtc;
new_cfg->pfc.pfcena = pfc->pfc_en;
err = ice_pf_dcb_cfg(pf, new_cfg, true);
if (err == ICE_DCB_HW_CHG_RST)
ice_dcbnl_devreset(netdev);
if (err == ICE_DCB_NO_HW_CHG)
err = ICE_DCB_HW_CHG_RST;
mutex_unlock(&pf->tc_mutex);
return err;
}
/**
* ice_dcbnl_get_pfc_cfg - Get CEE PFC config
* @netdev: pointer to netdev struct
* @prio: corresponding user priority
* @setting: the PFC setting for given priority
*/
static void
ice_dcbnl_get_pfc_cfg(struct net_device *netdev, int prio, u8 *setting)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (prio >= ICE_MAX_USER_PRIORITY)
return;
*setting = (pi->qos_cfg.local_dcbx_cfg.pfc.pfcena >> prio) & 0x1;
dev_dbg(ice_pf_to_dev(pf), "Get PFC Config up=%d, setting=%d, pfcenable=0x%x\n",
prio, *setting, pi->qos_cfg.local_dcbx_cfg.pfc.pfcena);
}
/**
* ice_dcbnl_set_pfc_cfg - Set CEE PFC config
* @netdev: the corresponding netdev
* @prio: User Priority
* @set: PFC setting to apply
*/
static void ice_dcbnl_set_pfc_cfg(struct net_device *netdev, int prio, u8 set)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *new_cfg;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (prio >= ICE_MAX_USER_PRIORITY)
return;
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
new_cfg->pfc.pfccap = pf->hw.func_caps.common_cap.maxtc;
if (set)
new_cfg->pfc.pfcena |= BIT(prio);
else
new_cfg->pfc.pfcena &= ~BIT(prio);
dev_dbg(ice_pf_to_dev(pf), "Set PFC config UP:%d set:%d pfcena:0x%x\n",
prio, set, new_cfg->pfc.pfcena);
}
/**
* ice_dcbnl_getpfcstate - get CEE PFC mode
* @netdev: pointer to netdev struct
*/
static u8 ice_dcbnl_getpfcstate(struct net_device *netdev)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
/* Return enabled if any UP enabled for PFC */
if (pi->qos_cfg.local_dcbx_cfg.pfc.pfcena)
return 1;
return 0;
}
/**
* ice_dcbnl_getstate - get DCB enabled state
* @netdev: pointer to netdev struct
*/
static u8 ice_dcbnl_getstate(struct net_device *netdev)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
u8 state = 0;
state = test_bit(ICE_FLAG_DCB_CAPABLE, pf->flags);
dev_dbg(ice_pf_to_dev(pf), "DCB enabled state = %d\n", state);
return state;
}
/**
* ice_dcbnl_setstate - Set CEE DCB state
* @netdev: pointer to relevant netdev
* @state: state value to set
*/
static u8 ice_dcbnl_setstate(struct net_device *netdev, u8 state)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return ICE_DCB_NO_HW_CHG;
/* Nothing to do */
if (!!state == test_bit(ICE_FLAG_DCB_ENA, pf->flags))
return ICE_DCB_NO_HW_CHG;
if (state) {
set_bit(ICE_FLAG_DCB_ENA, pf->flags);
memcpy(&pf->hw.port_info->qos_cfg.desired_dcbx_cfg,
&pf->hw.port_info->qos_cfg.local_dcbx_cfg,
sizeof(struct ice_dcbx_cfg));
} else {
clear_bit(ICE_FLAG_DCB_ENA, pf->flags);
}
return ICE_DCB_HW_CHG;
}
/**
* ice_dcbnl_get_pg_tc_cfg_tx - get CEE PG Tx config
* @netdev: pointer to netdev struct
* @prio: the corresponding user priority
* @prio_type: traffic priority type
* @pgid: the BW group ID the traffic class belongs to
* @bw_pct: BW percentage for the corresponding BWG
* @up_map: prio mapped to corresponding TC
*/
static void
ice_dcbnl_get_pg_tc_cfg_tx(struct net_device *netdev, int prio,
u8 __always_unused *prio_type, u8 *pgid,
u8 __always_unused *bw_pct,
u8 __always_unused *up_map)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (prio >= ICE_MAX_USER_PRIORITY)
return;
*pgid = pi->qos_cfg.local_dcbx_cfg.etscfg.prio_table[prio];
dev_dbg(ice_pf_to_dev(pf), "Get PG config prio=%d tc=%d\n", prio,
*pgid);
}
/**
* ice_dcbnl_set_pg_tc_cfg_tx - set CEE PG Tx config
* @netdev: pointer to relevant netdev
* @tc: the corresponding traffic class
* @prio_type: the traffic priority type
* @bwg_id: the BW group ID the TC belongs to
* @bw_pct: the BW perventage for the BWG
* @up_map: prio mapped to corresponding TC
*/
static void
ice_dcbnl_set_pg_tc_cfg_tx(struct net_device *netdev, int tc,
u8 __always_unused prio_type,
u8 __always_unused bwg_id,
u8 __always_unused bw_pct, u8 up_map)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *new_cfg;
int i;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (tc >= ICE_MAX_TRAFFIC_CLASS)
return;
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
/* prio_type, bwg_id and bw_pct per UP are not supported */
ice_for_each_traffic_class(i) {
if (up_map & BIT(i))
new_cfg->etscfg.prio_table[i] = tc;
}
new_cfg->etscfg.tsatable[tc] = ICE_IEEE_TSA_ETS;
}
/**
* ice_dcbnl_get_pg_bwg_cfg_tx - Get CEE PGBW config
* @netdev: pointer to the netdev struct
* @pgid: corresponding traffic class
* @bw_pct: the BW percentage for the corresponding TC
*/
static void
ice_dcbnl_get_pg_bwg_cfg_tx(struct net_device *netdev, int pgid, u8 *bw_pct)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (pgid >= ICE_MAX_TRAFFIC_CLASS)
return;
*bw_pct = pi->qos_cfg.local_dcbx_cfg.etscfg.tcbwtable[pgid];
dev_dbg(ice_pf_to_dev(pf), "Get PG BW config tc=%d bw_pct=%d\n",
pgid, *bw_pct);
}
/**
* ice_dcbnl_set_pg_bwg_cfg_tx - set CEE PG Tx BW config
* @netdev: the corresponding netdev
* @pgid: Correspongind traffic class
* @bw_pct: the BW percentage for the specified TC
*/
static void
ice_dcbnl_set_pg_bwg_cfg_tx(struct net_device *netdev, int pgid, u8 bw_pct)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *new_cfg;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (pgid >= ICE_MAX_TRAFFIC_CLASS)
return;
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
new_cfg->etscfg.tcbwtable[pgid] = bw_pct;
}
/**
* ice_dcbnl_get_pg_tc_cfg_rx - Get CEE PG Rx config
* @netdev: pointer to netdev struct
* @prio: the corresponding user priority
* @prio_type: the traffic priority type
* @pgid: the PG ID
* @bw_pct: the BW percentage for the corresponding BWG
* @up_map: prio mapped to corresponding TC
*/
static void
ice_dcbnl_get_pg_tc_cfg_rx(struct net_device *netdev, int prio,
u8 __always_unused *prio_type, u8 *pgid,
u8 __always_unused *bw_pct,
u8 __always_unused *up_map)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_port_info *pi = pf->hw.port_info;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
if (prio >= ICE_MAX_USER_PRIORITY)
return;
*pgid = pi->qos_cfg.local_dcbx_cfg.etscfg.prio_table[prio];
}
/**
* ice_dcbnl_set_pg_tc_cfg_rx
* @netdev: relevant netdev struct
* @prio: corresponding user priority
* @prio_type: the traffic priority type
* @pgid: the PG ID
* @bw_pct: BW percentage for corresponding BWG
* @up_map: prio mapped to corresponding TC
*
* lldpad requires this function pointer to be non-NULL to complete CEE config.
*/
static void
ice_dcbnl_set_pg_tc_cfg_rx(struct net_device *netdev,
int __always_unused prio,
u8 __always_unused prio_type,
u8 __always_unused pgid,
u8 __always_unused bw_pct,
u8 __always_unused up_map)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
dev_dbg(ice_pf_to_dev(pf), "Rx TC PG Config Not Supported.\n");
}
/**
* ice_dcbnl_get_pg_bwg_cfg_rx - Get CEE PG BW Rx config
* @netdev: pointer to netdev struct
* @pgid: the corresponding traffic class
* @bw_pct: the BW percentage for the corresponding TC
*/
static void
ice_dcbnl_get_pg_bwg_cfg_rx(struct net_device *netdev, int __always_unused pgid,
u8 *bw_pct)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return;
*bw_pct = 0;
}
/**
* ice_dcbnl_set_pg_bwg_cfg_rx
* @netdev: the corresponding netdev
* @pgid: corresponding TC
* @bw_pct: BW percentage for given TC
*
* lldpad requires this function pointer to be non-NULL to complete CEE config.
*/
static void
ice_dcbnl_set_pg_bwg_cfg_rx(struct net_device *netdev, int __always_unused pgid,
u8 __always_unused bw_pct)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
dev_dbg(ice_pf_to_dev(pf), "Rx BWG PG Config Not Supported.\n");
}
/**
* ice_dcbnl_get_cap - Get DCBX capabilities of adapter
* @netdev: pointer to netdev struct
* @capid: the capability type
* @cap: the capability value
*/
static u8 ice_dcbnl_get_cap(struct net_device *netdev, int capid, u8 *cap)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
if (!(test_bit(ICE_FLAG_DCB_CAPABLE, pf->flags)))
return ICE_DCB_NO_HW_CHG;
switch (capid) {
case DCB_CAP_ATTR_PG:
*cap = true;
break;
case DCB_CAP_ATTR_PFC:
*cap = true;
break;
case DCB_CAP_ATTR_UP2TC:
*cap = false;
break;
case DCB_CAP_ATTR_PG_TCS:
*cap = 0x80;
break;
case DCB_CAP_ATTR_PFC_TCS:
*cap = 0x80;
break;
case DCB_CAP_ATTR_GSP:
*cap = false;
break;
case DCB_CAP_ATTR_BCN:
*cap = false;
break;
case DCB_CAP_ATTR_DCBX:
*cap = pf->dcbx_cap;
break;
default:
*cap = false;
break;
}
dev_dbg(ice_pf_to_dev(pf), "DCBX Get Capability cap=%d capval=0x%x\n",
capid, *cap);
return 0;
}
/**
* ice_dcbnl_getapp - get CEE APP
* @netdev: pointer to netdev struct
* @idtype: the App selector
* @id: the App ethtype or port number
*/
static int ice_dcbnl_getapp(struct net_device *netdev, u8 idtype, u16 id)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct dcb_app app = {
.selector = idtype,
.protocol = id,
};
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return -EINVAL;
return dcb_getapp(netdev, &app);
}
/**
* ice_dcbnl_find_app - Search for APP in given DCB config
* @cfg: struct to hold DCBX config
* @app: struct to hold app data to look for
*/
static bool
ice_dcbnl_find_app(struct ice_dcbx_cfg *cfg,
struct ice_dcb_app_priority_table *app)
{
unsigned int i;
for (i = 0; i < cfg->numapps; i++) {
if (app->selector == cfg->app[i].selector &&
app->prot_id == cfg->app[i].prot_id &&
app->priority == cfg->app[i].priority)
return true;
}
return false;
}
/**
* ice_dcbnl_setapp - set local IEEE App config
* @netdev: relevant netdev struct
* @app: struct to hold app config info
*/
static int ice_dcbnl_setapp(struct net_device *netdev, struct dcb_app *app)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcb_app_priority_table new_app;
struct ice_dcbx_cfg *old_cfg, *new_cfg;
int ret;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_IEEE))
return -EINVAL;
mutex_lock(&pf->tc_mutex);
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
old_cfg = &pf->hw.port_info->qos_cfg.local_dcbx_cfg;
if (old_cfg->numapps == ICE_DCBX_MAX_APPS) {
ret = -EINVAL;
goto setapp_out;
}
ret = dcb_ieee_setapp(netdev, app);
if (ret)
goto setapp_out;
new_app.selector = app->selector;
new_app.prot_id = app->protocol;
new_app.priority = app->priority;
if (ice_dcbnl_find_app(old_cfg, &new_app)) {
ret = 0;
goto setapp_out;
}
new_cfg->app[new_cfg->numapps++] = new_app;
ret = ice_pf_dcb_cfg(pf, new_cfg, true);
/* return of zero indicates new cfg applied */
if (ret == ICE_DCB_HW_CHG_RST)
ice_dcbnl_devreset(netdev);
if (ret == ICE_DCB_NO_HW_CHG)
ret = ICE_DCB_HW_CHG_RST;
setapp_out:
mutex_unlock(&pf->tc_mutex);
return ret;
}
/**
* ice_dcbnl_delapp - Delete local IEEE App config
* @netdev: relevant netdev
* @app: struct to hold app too delete
*
* Will not delete first application required by the FW
*/
static int ice_dcbnl_delapp(struct net_device *netdev, struct dcb_app *app)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *old_cfg, *new_cfg;
unsigned int i, j;
int ret = 0;
if (pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED)
return -EINVAL;
mutex_lock(&pf->tc_mutex);
old_cfg = &pf->hw.port_info->qos_cfg.local_dcbx_cfg;
if (old_cfg->numapps <= 1)
goto delapp_out;
ret = dcb_ieee_delapp(netdev, app);
if (ret)
goto delapp_out;
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
for (i = 1; i < new_cfg->numapps; i++) {
if (app->selector == new_cfg->app[i].selector &&
app->protocol == new_cfg->app[i].prot_id &&
app->priority == new_cfg->app[i].priority) {
new_cfg->app[i].selector = 0;
new_cfg->app[i].prot_id = 0;
new_cfg->app[i].priority = 0;
break;
}
}
/* Did not find DCB App */
if (i == new_cfg->numapps) {
ret = -EINVAL;
goto delapp_out;
}
new_cfg->numapps--;
for (j = i; j < new_cfg->numapps; j++) {
new_cfg->app[i].selector = old_cfg->app[j + 1].selector;
new_cfg->app[i].prot_id = old_cfg->app[j + 1].prot_id;
new_cfg->app[i].priority = old_cfg->app[j + 1].priority;
}
ret = ice_pf_dcb_cfg(pf, new_cfg, true);
/* return of zero indicates new cfg applied */
if (ret == ICE_DCB_HW_CHG_RST)
ice_dcbnl_devreset(netdev);
if (ret == ICE_DCB_NO_HW_CHG)
ret = ICE_DCB_HW_CHG_RST;
delapp_out:
mutex_unlock(&pf->tc_mutex);
return ret;
}
/**
* ice_dcbnl_cee_set_all - Commit CEE DCB settings to HW
* @netdev: the corresponding netdev
*/
static u8 ice_dcbnl_cee_set_all(struct net_device *netdev)
{
struct ice_pf *pf = ice_netdev_to_pf(netdev);
struct ice_dcbx_cfg *new_cfg;
int err;
if ((pf->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED) ||
!(pf->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
return ICE_DCB_NO_HW_CHG;
new_cfg = &pf->hw.port_info->qos_cfg.desired_dcbx_cfg;
mutex_lock(&pf->tc_mutex);
err = ice_pf_dcb_cfg(pf, new_cfg, true);
mutex_unlock(&pf->tc_mutex);
return (err != ICE_DCB_HW_CHG_RST) ? ICE_DCB_NO_HW_CHG : err;
}
static const struct dcbnl_rtnl_ops dcbnl_ops = {
/* IEEE 802.1Qaz std */
.ieee_getets = ice_dcbnl_getets,
.ieee_setets = ice_dcbnl_setets,
.ieee_getpfc = ice_dcbnl_getpfc,
.ieee_setpfc = ice_dcbnl_setpfc,
.ieee_setapp = ice_dcbnl_setapp,
.ieee_delapp = ice_dcbnl_delapp,
/* CEE std */
.getstate = ice_dcbnl_getstate,
.setstate = ice_dcbnl_setstate,
.getpermhwaddr = ice_dcbnl_get_perm_hw_addr,
.setpgtccfgtx = ice_dcbnl_set_pg_tc_cfg_tx,
.setpgbwgcfgtx = ice_dcbnl_set_pg_bwg_cfg_tx,
.setpgtccfgrx = ice_dcbnl_set_pg_tc_cfg_rx,
.setpgbwgcfgrx = ice_dcbnl_set_pg_bwg_cfg_rx,
.getpgtccfgtx = ice_dcbnl_get_pg_tc_cfg_tx,
.getpgbwgcfgtx = ice_dcbnl_get_pg_bwg_cfg_tx,
.getpgtccfgrx = ice_dcbnl_get_pg_tc_cfg_rx,
.getpgbwgcfgrx = ice_dcbnl_get_pg_bwg_cfg_rx,
.setpfccfg = ice_dcbnl_set_pfc_cfg,
.getpfccfg = ice_dcbnl_get_pfc_cfg,
.setall = ice_dcbnl_cee_set_all,
.getcap = ice_dcbnl_get_cap,
.getnumtcs = ice_dcbnl_getnumtcs,
.getpfcstate = ice_dcbnl_getpfcstate,
.getapp = ice_dcbnl_getapp,
/* DCBX configuration */
.getdcbx = ice_dcbnl_getdcbx,
.setdcbx = ice_dcbnl_setdcbx,
};
/**
* ice_dcbnl_set_all - set all the apps and ieee data from DCBX config
* @vsi: pointer to VSI struct
*/
void ice_dcbnl_set_all(struct ice_vsi *vsi)
{
struct net_device *netdev = vsi->netdev;
struct ice_dcbx_cfg *dcbxcfg;
struct ice_port_info *pi;
struct dcb_app sapp;
struct ice_pf *pf;
unsigned int i;
if (!netdev)
return;
pf = ice_netdev_to_pf(netdev);
pi = pf->hw.port_info;
/* SW DCB taken care of by SW Default Config */
if (pf->dcbx_cap & DCB_CAP_DCBX_HOST)
return;
/* DCB not enabled */
if (!test_bit(ICE_FLAG_DCB_ENA, pf->flags))
return;
dcbxcfg = &pi->qos_cfg.local_dcbx_cfg;
for (i = 0; i < dcbxcfg->numapps; i++) {
u8 prio, tc_map;
prio = dcbxcfg->app[i].priority;
tc_map = BIT(dcbxcfg->etscfg.prio_table[prio]);
/* Add APP only if the TC is enabled for this VSI */
if (tc_map & vsi->tc_cfg.ena_tc) {
sapp.selector = dcbxcfg->app[i].selector;
sapp.protocol = dcbxcfg->app[i].prot_id;
sapp.priority = prio;
dcb_ieee_setapp(netdev, &sapp);
}
}
/* Notify user-space of the changes */
dcbnl_ieee_notify(netdev, RTM_SETDCB, DCB_CMD_IEEE_SET, 0, 0);
}
/**
* ice_dcbnl_vsi_del_app - Delete APP on all VSIs
* @vsi: pointer to the main VSI
* @app: APP to delete
*
* Delete given APP from all the VSIs for given PF
*/
static void
ice_dcbnl_vsi_del_app(struct ice_vsi *vsi,
struct ice_dcb_app_priority_table *app)
{
struct dcb_app sapp;
int err;
sapp.selector = app->selector;
sapp.protocol = app->prot_id;
sapp.priority = app->priority;
err = ice_dcbnl_delapp(vsi->netdev, &sapp);
dev_dbg(ice_pf_to_dev(vsi->back), "Deleting app for VSI idx=%d err=%d sel=%d proto=0x%x, prio=%d\n",
vsi->idx, err, app->selector, app->prot_id, app->priority);
}
/**
* ice_dcbnl_flush_apps - Delete all removed APPs
* @pf: the corresponding PF
* @old_cfg: old DCBX configuration data
* @new_cfg: new DCBX configuration data
*
* Find and delete all APPS that are not present in the passed
* DCB configuration
*/
void
ice_dcbnl_flush_apps(struct ice_pf *pf, struct ice_dcbx_cfg *old_cfg,
struct ice_dcbx_cfg *new_cfg)
{
struct ice_vsi *main_vsi = ice_get_main_vsi(pf);
unsigned int i;
if (!main_vsi)
return;
for (i = 0; i < old_cfg->numapps; i++) {
struct ice_dcb_app_priority_table app = old_cfg->app[i];
/* The APP is not available anymore delete it */
if (!ice_dcbnl_find_app(new_cfg, &app))
ice_dcbnl_vsi_del_app(main_vsi, &app);
}
}
/**
* ice_dcbnl_setup - setup DCBNL
* @vsi: VSI to get associated netdev from
*/
void ice_dcbnl_setup(struct ice_vsi *vsi)
{
struct net_device *netdev = vsi->netdev;
struct ice_pf *pf;
pf = ice_netdev_to_pf(netdev);
if (!test_bit(ICE_FLAG_DCB_CAPABLE, pf->flags))
return;
netdev->dcbnl_ops = &dcbnl_ops;
ice_dcbnl_set_all(vsi);
}