| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <201102061850.40458.vladz@broadcom.com>
Date: Sun, 6 Feb 2011 18:50:40 +0200
From: "Vlad Zolotarov" <vladz@...adcom.com>
To: "Dave Miller" <davem@...emloft.net>
cc: "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"Eilon Greenstein" <eilong@...adcom.com>
Subject: [PATCH net-next] bnx2x: Proper netdev->ndo_set_rx_mode()
implementation.
Completed the bnx2x_set_rx_mode() to a proper netdev->ndo_set_rx_mode
implementation:
- Added a missing configuration of a unicast MAC addresses list.
- Changed bp->dma_lock from being a mutex to a spinlock as long as it's taken
under netdev->addr_list_lock now.
Signed-off-by: Vladislav Zolotarov <vladz@...adcom.com>
Signed-off-by: Eilon Greenstein <eilong@...adcom.com>
---
drivers/net/bnx2x/bnx2x.h | 13 +-
drivers/net/bnx2x/bnx2x_cmn.c | 17 +-
drivers/net/bnx2x/bnx2x_main.c | 444 +++++++++++++++++++++++++++++++---------
3 files changed, 369 insertions(+), 105 deletions(-)
diff --git a/drivers/net/bnx2x/bnx2x.h b/drivers/net/bnx2x/bnx2x.h
index 8e41837..9821bd9 100644
--- a/drivers/net/bnx2x/bnx2x.h
+++ b/drivers/net/bnx2x/bnx2x.h
@@ -129,6 +129,7 @@ void bnx2x_panic_dump(struct bnx2x *bp);
#endif
#define bnx2x_mc_addr(ha) ((ha)->addr)
+#define bnx2x_uc_addr(ha) ((ha)->addr)
#define U64_LO(x) (u32)(((u64)(x)) & 0xffffffff)
#define U64_HI(x) (u32)(((u64)(x)) >> 32)
@@ -810,6 +811,7 @@ struct bnx2x_slowpath {
struct eth_stats_query fw_stats;
struct mac_configuration_cmd mac_config;
struct mac_configuration_cmd mcast_config;
+ struct mac_configuration_cmd uc_mac_config;
struct client_init_ramrod_data client_init_data;
/* used by dmae command executer */
@@ -939,7 +941,7 @@ struct bnx2x {
struct eth_spe *spq_prod_bd;
struct eth_spe *spq_last_bd;
__le16 *dsb_sp_prod;
- atomic_t spq_left; /* serialize spq */
+ atomic_t cq_spq_left; /* ETH_XXX ramrods credit */
/* used to synchronize spq accesses */
spinlock_t spq_lock;
@@ -949,6 +951,7 @@ struct bnx2x {
u16 eq_prod;
u16 eq_cons;
__le16 *eq_cons_sb;
+ atomic_t eq_spq_left; /* COMMON_XXX ramrods credit */
/* Flags for marking that there is a STAT_QUERY or
SET_MAC ramrod pending */
@@ -1131,7 +1134,7 @@ struct bnx2x {
int dmae_ready;
/* used to synchronize dmae accesses */
- struct mutex dmae_mutex;
+ spinlock_t dmae_lock;
/* used to protect the FW mail box */
struct mutex fw_mb_mutex;
@@ -1447,6 +1450,12 @@ u32 bnx2x_fw_command(struct bnx2x *bp, u32 command, u32 param);
void bnx2x_calc_fc_adv(struct bnx2x *bp);
int bnx2x_sp_post(struct bnx2x *bp, int command, int cid,
u32 data_hi, u32 data_lo, int common);
+
+/* Clears multicast and unicast list configuration in the chip. */
+void bnx2x_invalidate_e1_mc_list(struct bnx2x *bp);
+void bnx2x_invalidate_e1h_mc_list(struct bnx2x *bp);
+void bnx2x_invalidate_uc_list(struct bnx2x *bp);
+
void bnx2x_update_coalesce(struct bnx2x *bp);
int bnx2x_get_link_cfg_idx(struct bnx2x *bp);
diff --git a/drivers/net/bnx2x/bnx2x_cmn.c b/drivers/net/bnx2x/bnx2x_cmn.c
index 710ce5d..ae98f7d 100644
--- a/drivers/net/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/bnx2x/bnx2x_cmn.c
@@ -1427,28 +1427,35 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
bnx2x_set_eth_mac(bp, 1);
+ /* Clear MC configuration */
+ if (CHIP_IS_E1(bp))
+ bnx2x_invalidate_e1_mc_list(bp);
+ else
+ bnx2x_invalidate_e1h_mc_list(bp);
+
+ /* Clear UC lists configuration */
+ bnx2x_invalidate_uc_list(bp);
+
if (bp->port.pmf)
bnx2x_initial_phy_init(bp, load_mode);
+ /* Initialize Rx filtering */
+ bnx2x_set_rx_mode(bp->dev);
+
/* Start fast path */
switch (load_mode) {
case LOAD_NORMAL:
/* Tx queue should be only reenabled */
netif_tx_wake_all_queues(bp->dev);
/* Initialize the receive filter. */
- bnx2x_set_rx_mode(bp->dev);
break;
case LOAD_OPEN:
netif_tx_start_all_queues(bp->dev);
smp_mb__after_clear_bit();
- /* Initialize the receive filter. */
- bnx2x_set_rx_mode(bp->dev);
break;
case LOAD_DIAG:
- /* Initialize the receive filter. */
- bnx2x_set_rx_mode(bp->dev);
bp->state = BNX2X_STATE_DIAG;
break;
diff --git a/drivers/net/bnx2x/bnx2x_main.c b/drivers/net/bnx2x/bnx2x_main.c
index 8cdcf5b..ea10b26 100644
--- a/drivers/net/bnx2x/bnx2x_main.c
+++ b/drivers/net/bnx2x/bnx2x_main.c
@@ -586,7 +586,7 @@ static int bnx2x_issue_dmae_with_comp(struct bnx2x *bp,
bp->slowpath->wb_data[2], bp->slowpath->wb_data[3]);
/* lock the dmae channel */
- mutex_lock(&bp->dmae_mutex);
+ spin_lock_bh(&bp->dmae_lock);
/* reset completion */
*wb_comp = 0;
@@ -617,7 +617,7 @@ static int bnx2x_issue_dmae_with_comp(struct bnx2x *bp,
bp->slowpath->wb_data[2], bp->slowpath->wb_data[3]);
unlock:
- mutex_unlock(&bp->dmae_mutex);
+ spin_unlock_bh(&bp->dmae_lock);
return rc;
}
@@ -1397,7 +1397,7 @@ void bnx2x_sp_event(struct bnx2x_fastpath *fp,
}
smp_mb__before_atomic_inc();
- atomic_inc(&bp->spq_left);
+ atomic_inc(&bp->cq_spq_left);
/* push the change in fp->state and towards the memory */
smp_wmb();
@@ -2731,11 +2731,18 @@ int bnx2x_sp_post(struct bnx2x *bp, int command, int cid,
spin_lock_bh(&bp->spq_lock);
- if (!atomic_read(&bp->spq_left)) {
- BNX2X_ERR("BUG! SPQ ring full!\n");
- spin_unlock_bh(&bp->spq_lock);
- bnx2x_panic();
- return -EBUSY;
+ if (common) {
+ if (!atomic_read(&bp->eq_spq_left)) {
+ BNX2X_ERR("BUG! EQ ring full!\n");
+ spin_unlock_bh(&bp->spq_lock);
+ bnx2x_panic();
+ return -EBUSY;
+ }
+ } else if (!atomic_read(&bp->cq_spq_left)) {
+ BNX2X_ERR("BUG! SPQ ring full!\n");
+ spin_unlock_bh(&bp->spq_lock);
+ bnx2x_panic();
+ return -EBUSY;
}
spe = bnx2x_sp_get_next(bp);
@@ -2766,20 +2773,26 @@ int bnx2x_sp_post(struct bnx2x *bp, int command, int cid,
spe->data.update_data_addr.lo = cpu_to_le32(data_lo);
/* stats ramrod has it's own slot on the spq */
- if (command != RAMROD_CMD_ID_COMMON_STAT_QUERY)
+ if (command != RAMROD_CMD_ID_COMMON_STAT_QUERY) {
/* It's ok if the actual decrement is issued towards the memory
* somewhere between the spin_lock and spin_unlock. Thus no
* more explict memory barrier is needed.
*/
- atomic_dec(&bp->spq_left);
+ if (common)
+ atomic_dec(&bp->eq_spq_left);
+ else
+ atomic_dec(&bp->cq_spq_left);
+ }
+
DP(BNX2X_MSG_SP/*NETIF_MSG_TIMER*/,
"SPQE[%x] (%x:%x) command %d hw_cid %x data (%x:%x) "
- "type(0x%x) left %x\n",
+ "type(0x%x) left (ETH, COMMON) (%x,%x)\n",
bp->spq_prod_idx, (u32)U64_HI(bp->spq_mapping),
(u32)(U64_LO(bp->spq_mapping) +
(void *)bp->spq_prod_bd - (void *)bp->spq), command,
- HW_CID(bp, cid), data_hi, data_lo, type, atomic_read(&bp->spq_left));
+ HW_CID(bp, cid), data_hi, data_lo, type,
+ atomic_read(&bp->cq_spq_left), atomic_read(&bp->eq_spq_left));
bnx2x_sp_prod_update(bp);
spin_unlock_bh(&bp->spq_lock);
@@ -3691,8 +3704,8 @@ static void bnx2x_eq_int(struct bnx2x *bp)
sw_cons = bp->eq_cons;
sw_prod = bp->eq_prod;
- DP(BNX2X_MSG_SP, "EQ: hw_cons %u sw_cons %u bp->spq_left %u\n",
- hw_cons, sw_cons, atomic_read(&bp->spq_left));
+ DP(BNX2X_MSG_SP, "EQ: hw_cons %u sw_cons %u bp->cq_spq_left %u\n",
+ hw_cons, sw_cons, atomic_read(&bp->eq_spq_left));
for (; sw_cons != hw_cons;
sw_prod = NEXT_EQ_IDX(sw_prod), sw_cons = NEXT_EQ_IDX(sw_cons)) {
@@ -3757,13 +3770,15 @@ static void bnx2x_eq_int(struct bnx2x *bp)
case (EVENT_RING_OPCODE_SET_MAC | BNX2X_STATE_OPEN):
case (EVENT_RING_OPCODE_SET_MAC | BNX2X_STATE_DIAG):
DP(NETIF_MSG_IFUP, "got set mac ramrod\n");
- bp->set_mac_pending = 0;
+ if (elem->message.data.set_mac_event.echo)
+ bp->set_mac_pending = 0;
break;
case (EVENT_RING_OPCODE_SET_MAC |
BNX2X_STATE_CLOSING_WAIT4_HALT):
DP(NETIF_MSG_IFDOWN, "got (un)set mac ramrod\n");
- bp->set_mac_pending = 0;
+ if (elem->message.data.set_mac_event.echo)
+ bp->set_mac_pending = 0;
break;
default:
/* unknown event log error and continue */
@@ -3775,7 +3790,7 @@ next_spqe:
} /* for */
smp_mb__before_atomic_inc();
- atomic_add(spqe_cnt, &bp->spq_left);
+ atomic_add(spqe_cnt, &bp->eq_spq_left);
bp->eq_cons = sw_cons;
bp->eq_prod = sw_prod;
@@ -4208,7 +4223,7 @@ void bnx2x_update_coalesce(struct bnx2x *bp)
static void bnx2x_init_sp_ring(struct bnx2x *bp)
{
spin_lock_init(&bp->spq_lock);
- atomic_set(&bp->spq_left, MAX_SPQ_PENDING);
+ atomic_set(&bp->cq_spq_left, MAX_SPQ_PENDING);
bp->spq_prod_idx = 0;
bp->dsb_sp_prod = BNX2X_SP_DSB_INDEX;
@@ -4233,6 +4248,9 @@ static void bnx2x_init_eq_ring(struct bnx2x *bp)
bp->eq_cons = 0;
bp->eq_prod = NUM_EQ_DESC;
bp->eq_cons_sb = BNX2X_EQ_INDEX;
+ /* we want a warning message before it gets rought... */
+ atomic_set(&bp->eq_spq_left,
+ min_t(int, MAX_SP_DESC_CNT - MAX_SPQ_PENDING, NUM_EQ_DESC) - 1);
}
static void bnx2x_init_ind_table(struct bnx2x *bp)
@@ -5838,7 +5856,7 @@ int bnx2x_init_hw(struct bnx2x *bp, u32 load_code)
BP_ABS_FUNC(bp), load_code);
bp->dmae_ready = 0;
- mutex_init(&bp->dmae_mutex);
+ spin_lock_init(&bp->dmae_lock);
rc = bnx2x_gunzip_init(bp);
if (rc)
return rc;
@@ -6173,12 +6191,14 @@ static void bnx2x_set_mac_addr_gen(struct bnx2x *bp, int set, const u8 *mac,
int ramrod_flags = WAIT_RAMROD_COMMON;
bp->set_mac_pending = 1;
- smp_wmb();
config->hdr.length = 1;
config->hdr.offset = cam_offset;
config->hdr.client_id = 0xff;
- config->hdr.reserved1 = 0;
+ /* Mark the single MAC configuration ramrod as opposed to a
+ * UC/MC list configuration).
+ */
+ config->hdr.echo = 1;
/* primary MAC */
config->config_table[0].msb_mac_addr =
@@ -6210,6 +6230,8 @@ static void bnx2x_set_mac_addr_gen(struct bnx2x *bp, int set, const u8 *mac,
config->config_table[0].middle_mac_addr,
config->config_table[0].lsb_mac_addr, BP_FUNC(bp), cl_bit_vec);
+ mb();
+
bnx2x_sp_post(bp, RAMROD_CMD_ID_COMMON_SET_MAC, 0,
U64_HI(bnx2x_sp_mapping(bp, mac_config)),
U64_LO(bnx2x_sp_mapping(bp, mac_config)), 1);
@@ -6274,20 +6296,15 @@ static u8 bnx2x_e1h_cam_offset(struct bnx2x *bp, u8 rel_offset)
if (CHIP_IS_E1H(bp))
return E1H_FUNC_MAX * rel_offset + BP_FUNC(bp);
else if (CHIP_MODE_IS_4_PORT(bp))
- return BP_FUNC(bp) * 32 + rel_offset;
+ return E2_FUNC_MAX * rel_offset + BP_FUNC(bp);
else
- return BP_VN(bp) * 32 + rel_offset;
+ return E2_FUNC_MAX * rel_offset + BP_VN(bp);
}
/**
* LLH CAM line allocations: currently only iSCSI and ETH macs are
* relevant. In addition, current implementation is tuned for a
* single ETH MAC.
- *
- * When multiple unicast ETH MACs PF configuration in switch
- * independent mode is required (NetQ, multiple netdev MACs,
- * etc.), consider better utilisation of 16 per function MAC
- * entries in the LLH memory.
*/
enum {
LLH_CAM_ISCSI_ETH_LINE = 0,
@@ -6362,14 +6379,37 @@ void bnx2x_set_eth_mac(struct bnx2x *bp, int set)
bnx2x_set_mac_addr_gen(bp, set, bcast, 0, cam_offset + 1, 1);
}
}
-static void bnx2x_set_e1_mc_list(struct bnx2x *bp, u8 offset)
+
+static inline u8 bnx2x_e1_cam_mc_offset(struct bnx2x *bp)
+{
+ return CHIP_REV_IS_SLOW(bp) ?
+ (BNX2X_MAX_EMUL_MULTI * (1 + BP_PORT(bp))) :
+ (BNX2X_MAX_MULTICAST * (1 + BP_PORT(bp)));
+}
+
+/* set mc list, do not wait as wait implies sleep and
+ * set_rx_mode can be invoked from non-sleepable context.
+ *
+ * Instead we use the same ramrod data buffer each time we need
+ * to configure a list of addresses, and use the fact that the
+ * list of MACs is changed in an incremental way and that the
+ * function is called under the netif_addr_lock. A temporary
+ * inconsistent CAM configuration (possible in case of a very fast
+ * sequence of add/del/add on the host side) will shortly be
+ * restored by the handler of the last ramrod.
+ */
+static int bnx2x_set_e1_mc_list(struct bnx2x *bp)
{
int i = 0, old;
struct net_device *dev = bp->dev;
+ u8 offset = bnx2x_e1_cam_mc_offset(bp);
struct netdev_hw_addr *ha;
struct mac_configuration_cmd *config_cmd = bnx2x_sp(bp, mcast_config);
dma_addr_t config_cmd_map = bnx2x_sp_mapping(bp, mcast_config);
+ if (netdev_mc_count(dev) > BNX2X_MAX_MULTICAST)
+ return -EINVAL;
+
netdev_for_each_mc_addr(ha, dev) {
/* copy mac */
config_cmd->config_table[i].msb_mac_addr =
@@ -6410,32 +6450,47 @@ static void bnx2x_set_e1_mc_list(struct bnx2x *bp, u8 offset)
}
}
+ wmb();
+
config_cmd->hdr.length = i;
config_cmd->hdr.offset = offset;
config_cmd->hdr.client_id = 0xff;
- config_cmd->hdr.reserved1 = 0;
+ /* Mark that this ramrod doesn't use bp->set_mac_pending for
+ * synchronization.
+ */
+ config_cmd->hdr.echo = 0;
- bp->set_mac_pending = 1;
- smp_wmb();
+ mb();
- bnx2x_sp_post(bp, RAMROD_CMD_ID_COMMON_SET_MAC, 0,
+ return bnx2x_sp_post(bp, RAMROD_CMD_ID_COMMON_SET_MAC, 0,
U64_HI(config_cmd_map), U64_LO(config_cmd_map), 1);
}
-static void bnx2x_invlidate_e1_mc_list(struct bnx2x *bp)
+
+void bnx2x_invalidate_e1_mc_list(struct bnx2x *bp)
{
int i;
struct mac_configuration_cmd *config_cmd = bnx2x_sp(bp, mcast_config);
dma_addr_t config_cmd_map = bnx2x_sp_mapping(bp, mcast_config);
int ramrod_flags = WAIT_RAMROD_COMMON;
+ u8 offset = bnx2x_e1_cam_mc_offset(bp);
- bp->set_mac_pending = 1;
- smp_wmb();
-
- for (i = 0; i < config_cmd->hdr.length; i++)
+ for (i = 0; i < BNX2X_MAX_MULTICAST; i++)
SET_FLAG(config_cmd->config_table[i].flags,
MAC_CONFIGURATION_ENTRY_ACTION_TYPE,
T_ETH_MAC_COMMAND_INVALIDATE);
+ wmb();
+
+ config_cmd->hdr.length = BNX2X_MAX_MULTICAST;
+ config_cmd->hdr.offset = offset;
+ config_cmd->hdr.client_id = 0xff;
+ /* We'll wait for a completion this time... */
+ config_cmd->hdr.echo = 1;
+
+ bp->set_mac_pending = 1;
+
+ mb();
+
bnx2x_sp_post(bp, RAMROD_CMD_ID_COMMON_SET_MAC, 0,
U64_HI(config_cmd_map), U64_LO(config_cmd_map), 1);
@@ -6445,6 +6500,44 @@ static void bnx2x_invlidate_e1_mc_list(struct bnx2x *bp)
}
+/* Accept one or more multicasts */
+static int bnx2x_set_e1h_mc_list(struct bnx2x *bp)
+{
+ struct net_device *dev = bp->dev;
+ struct netdev_hw_addr *ha;
+ u32 mc_filter[MC_HASH_SIZE];
+ u32 crc, bit, regidx;
+ int i;
+
+ memset(mc_filter, 0, 4 * MC_HASH_SIZE);
+
+ netdev_for_each_mc_addr(ha, dev) {
+ DP(NETIF_MSG_IFUP, "Adding mcast MAC: %pM\n",
+ bnx2x_mc_addr(ha));
+
+ crc = crc32c_le(0, bnx2x_mc_addr(ha),
+ ETH_ALEN);
+ bit = (crc >> 24) & 0xff;
+ regidx = bit >> 5;
+ bit &= 0x1f;
+ mc_filter[regidx] |= (1 << bit);
+ }
+
+ for (i = 0; i < MC_HASH_SIZE; i++)
+ REG_WR(bp, MC_HASH_OFFSET(bp, i),
+ mc_filter[i]);
+
+ return 0;
+}
+
+void bnx2x_invalidate_e1h_mc_list(struct bnx2x *bp)
+{
+ int i;
+
+ for (i = 0; i < MC_HASH_SIZE; i++)
+ REG_WR(bp, MC_HASH_OFFSET(bp, i), 0);
+}
+
#ifdef BCM_CNIC
/**
* Set iSCSI MAC(s) at the next enties in the CAM after the ETH
@@ -7110,20 +7203,15 @@ void bnx2x_chip_cleanup(struct bnx2x *bp, int unload_mode)
/* Give HW time to discard old tx messages */
msleep(1);
- if (CHIP_IS_E1(bp)) {
- /* invalidate mc list,
- * wait and poll (interrupts are off)
- */
- bnx2x_invlidate_e1_mc_list(bp);
- bnx2x_set_eth_mac(bp, 0);
-
- } else {
- REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port*8, 0);
+ bnx2x_set_eth_mac(bp, 0);
- bnx2x_set_eth_mac(bp, 0);
+ bnx2x_invalidate_uc_list(bp);
- for (i = 0; i < MC_HASH_SIZE; i++)
- REG_WR(bp, MC_HASH_OFFSET(bp, i), 0);
+ if (CHIP_IS_E1(bp))
+ bnx2x_invalidate_e1_mc_list(bp);
+ else {
+ bnx2x_invalidate_e1h_mc_list(bp);
+ REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port*8, 0);
}
#ifdef BCM_CNIC
@@ -8813,12 +8901,197 @@ static int bnx2x_close(struct net_device *dev)
return 0;
}
+#define E1_MAX_UC_LIST 29
+#define E1H_MAX_UC_LIST 30
+#define E2_MAX_UC_LIST 14
+static inline u8 bnx2x_max_uc_list(struct bnx2x *bp)
+{
+ if (CHIP_IS_E1(bp))
+ return E1_MAX_UC_LIST;
+ else if (CHIP_IS_E1H(bp))
+ return E1H_MAX_UC_LIST;
+ else
+ return E2_MAX_UC_LIST;
+}
+
+
+static inline u8 bnx2x_uc_list_cam_offset(struct bnx2x *bp)
+{
+ if (CHIP_IS_E1(bp))
+ /* CAM Entries for Port0:
+ * 0 - prim ETH MAC
+ * 1 - BCAST MAC
+ * 2 - iSCSI L2 ring ETH MAC
+ * 3-31 - UC MACs
+ *
+ * Port1 entries are allocated the same way starting from
+ * entry 32.
+ */
+ return 3 + 32 * BP_PORT(bp);
+ else if (CHIP_IS_E1H(bp)) {
+ /* CAM Entries:
+ * 0-7 - prim ETH MAC for each function
+ * 8-15 - iSCSI L2 ring ETH MAC for each function
+ * 16 till 255 UC MAC lists for each function
+ *
+ * Remark: There is no FCoE support for E1H, thus FCoE related
+ * MACs are not considered.
+ */
+ return E1H_FUNC_MAX * (CAM_ISCSI_ETH_LINE + 1) +
+ bnx2x_max_uc_list(bp) * BP_FUNC(bp);
+ } else {
+ /* CAM Entries (there is a separate CAM per engine):
+ * 0-4 - prim ETH MAC for each function
+ * 4-7 - iSCSI L2 ring ETH MAC for each function
+ * 8-11 - FIP ucast L2 MAC for each function
+ * 12-15 - ALL_ENODE_MACS mcast MAC for each function
+ * 16 till 71 UC MAC lists for each function
+ */
+ u8 func_idx =
+ (CHIP_MODE_IS_4_PORT(bp) ? BP_FUNC(bp) : BP_VN(bp));
+
+ return E2_FUNC_MAX * (CAM_MAX_PF_LINE + 1) +
+ bnx2x_max_uc_list(bp) * func_idx;
+ }
+}
+
+/* set uc list, do not wait as wait implies sleep and
+ * set_rx_mode can be invoked from non-sleepable context.
+ *
+ * Instead we use the same ramrod data buffer each time we need
+ * to configure a list of addresses, and use the fact that the
+ * list of MACs is changed in an incremental way and that the
+ * function is called under the netif_addr_lock. A temporary
+ * inconsistent CAM configuration (possible in case of very fast
+ * sequence of add/del/add on the host side) will shortly be
+ * restored by the handler of the last ramrod.
+ */
+static int bnx2x_set_uc_list(struct bnx2x *bp)
+{
+ int i = 0, old;
+ struct net_device *dev = bp->dev;
+ u8 offset = bnx2x_uc_list_cam_offset(bp);
+ struct netdev_hw_addr *ha;
+ struct mac_configuration_cmd *config_cmd = bnx2x_sp(bp, uc_mac_config);
+ dma_addr_t config_cmd_map = bnx2x_sp_mapping(bp, uc_mac_config);
+
+ if (netdev_uc_count(dev) > bnx2x_max_uc_list(bp))
+ return -EINVAL;
+
+ netdev_for_each_uc_addr(ha, dev) {
+ /* copy mac */
+ config_cmd->config_table[i].msb_mac_addr =
+ swab16(*(u16 *)&bnx2x_uc_addr(ha)[0]);
+ config_cmd->config_table[i].middle_mac_addr =
+ swab16(*(u16 *)&bnx2x_uc_addr(ha)[2]);
+ config_cmd->config_table[i].lsb_mac_addr =
+ swab16(*(u16 *)&bnx2x_uc_addr(ha)[4]);
+
+ config_cmd->config_table[i].vlan_id = 0;
+ config_cmd->config_table[i].pf_id = BP_FUNC(bp);
+ config_cmd->config_table[i].clients_bit_vector =
+ cpu_to_le32(1 << BP_L_ID(bp));
+
+ SET_FLAG(config_cmd->config_table[i].flags,
+ MAC_CONFIGURATION_ENTRY_ACTION_TYPE,
+ T_ETH_MAC_COMMAND_SET);
+
+ DP(NETIF_MSG_IFUP,
+ "setting UCAST[%d] (%04x:%04x:%04x)\n", i,
+ config_cmd->config_table[i].msb_mac_addr,
+ config_cmd->config_table[i].middle_mac_addr,
+ config_cmd->config_table[i].lsb_mac_addr);
+
+ i++;
+
+ /* Set uc MAC in NIG */
+ bnx2x_set_mac_in_nig(bp, 1, bnx2x_uc_addr(ha),
+ LLH_CAM_ETH_LINE + i);
+ }
+ old = config_cmd->hdr.length;
+ if (old > i) {
+ for (; i < old; i++) {
+ if (CAM_IS_INVALID(config_cmd->
+ config_table[i])) {
+ /* already invalidated */
+ break;
+ }
+ /* invalidate */
+ SET_FLAG(config_cmd->config_table[i].flags,
+ MAC_CONFIGURATION_ENTRY_ACTION_TYPE,
+ T_ETH_MAC_COMMAND_INVALIDATE);
+ }
+ }
+
+ wmb();
+
+ config_cmd->hdr.length = i;
+ config_cmd->hdr.offset = offset;
+ config_cmd->hdr.client_id = 0xff;
+ /* Mark that this ramrod doesn't use bp->set_mac_pending for
+ * synchronization.
+ */
+ config_cmd->hdr.echo = 0;
+
+ mb();
+
+ return bnx2x_sp_post(bp, RAMROD_CMD_ID_COMMON_SET_MAC, 0,
+ U64_HI(config_cmd_map), U64_LO(config_cmd_map), 1);
+
+}
+
+void bnx2x_invalidate_uc_list(struct bnx2x *bp)
+{
+ int i;
+ struct mac_configuration_cmd *config_cmd = bnx2x_sp(bp, uc_mac_config);
+ dma_addr_t config_cmd_map = bnx2x_sp_mapping(bp, uc_mac_config);
+ int ramrod_flags = WAIT_RAMROD_COMMON;
+ u8 offset = bnx2x_uc_list_cam_offset(bp);
+ u8 max_list_size = bnx2x_max_uc_list(bp);
+
+ for (i = 0; i < max_list_size; i++) {
+ SET_FLAG(config_cmd->config_table[i].flags,
+ MAC_CONFIGURATION_ENTRY_ACTION_TYPE,
+ T_ETH_MAC_COMMAND_INVALIDATE);
+ bnx2x_set_mac_in_nig(bp, 0, NULL, LLH_CAM_ETH_LINE + 1 + i);
+ }
+
+ wmb();
+
+ config_cmd->hdr.length = max_list_size;
+ config_cmd->hdr.offset = offset;
+ config_cmd->hdr.client_id = 0xff;
+ /* We'll wait for a completion this time... */
+ config_cmd->hdr.echo = 1;
+
+ bp->set_mac_pending = 1;
+
+ mb();
+
+ bnx2x_sp_post(bp, RAMROD_CMD_ID_COMMON_SET_MAC, 0,
+ U64_HI(config_cmd_map), U64_LO(config_cmd_map), 1);
+
+ /* Wait for a completion */
+ bnx2x_wait_ramrod(bp, 0, 0, &bp->set_mac_pending,
+ ramrod_flags);
+
+}
+
+static inline int bnx2x_set_mc_list(struct bnx2x *bp)
+{
+ /* some multicasts */
+ if (CHIP_IS_E1(bp)) {
+ return bnx2x_set_e1_mc_list(bp);
+ } else { /* E1H and newer */
+ return bnx2x_set_e1h_mc_list(bp);
+ }
+}
+
/* called with netif_tx_lock from dev_mcast.c */
void bnx2x_set_rx_mode(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
u32 rx_mode = BNX2X_RX_MODE_NORMAL;
- int port = BP_PORT(bp);
if (bp->state != BNX2X_STATE_OPEN) {
DP(NETIF_MSG_IFUP, "state is %x, returning\n", bp->state);
@@ -8829,47 +9102,16 @@ void bnx2x_set_rx_mode(struct net_device *dev)
if (dev->flags & IFF_PROMISC)
rx_mode = BNX2X_RX_MODE_PROMISC;
- else if ((dev->flags & IFF_ALLMULTI) ||
- ((netdev_mc_count(dev) > BNX2X_MAX_MULTICAST) &&
- CHIP_IS_E1(bp)))
+ else if (dev->flags & IFF_ALLMULTI)
rx_mode = BNX2X_RX_MODE_ALLMULTI;
- else { /* some multicasts */
- if (CHIP_IS_E1(bp)) {
- /*
- * set mc list, do not wait as wait implies sleep
- * and set_rx_mode can be invoked from non-sleepable
- * context
- */
- u8 offset = (CHIP_REV_IS_SLOW(bp) ?
- BNX2X_MAX_EMUL_MULTI*(1 + port) :
- BNX2X_MAX_MULTICAST*(1 + port));
-
- bnx2x_set_e1_mc_list(bp, offset);
- } else { /* E1H */
- /* Accept one or more multicasts */
- struct netdev_hw_addr *ha;
- u32 mc_filter[MC_HASH_SIZE];
- u32 crc, bit, regidx;
- int i;
-
- memset(mc_filter, 0, 4 * MC_HASH_SIZE);
-
- netdev_for_each_mc_addr(ha, dev) {
- DP(NETIF_MSG_IFUP, "Adding mcast MAC: %pM\n",
- bnx2x_mc_addr(ha));
-
- crc = crc32c_le(0, bnx2x_mc_addr(ha),
- ETH_ALEN);
- bit = (crc >> 24) & 0xff;
- regidx = bit >> 5;
- bit &= 0x1f;
- mc_filter[regidx] |= (1 << bit);
- }
+ else {
+ /* some multicasts */
+ if (bnx2x_set_mc_list(bp))
+ rx_mode = BNX2X_RX_MODE_ALLMULTI;
- for (i = 0; i < MC_HASH_SIZE; i++)
- REG_WR(bp, MC_HASH_OFFSET(bp, i),
- mc_filter[i]);
- }
+ /* some unicasts */
+ if (bnx2x_set_uc_list(bp))
+ rx_mode = BNX2X_RX_MODE_PROMISC;
}
bp->rx_mode = rx_mode;
@@ -8950,7 +9192,7 @@ static const struct net_device_ops bnx2x_netdev_ops = {
.ndo_stop = bnx2x_close,
.ndo_start_xmit = bnx2x_start_xmit,
.ndo_select_queue = bnx2x_select_queue,
- .ndo_set_multicast_list = bnx2x_set_rx_mode,
+ .ndo_set_rx_mode = bnx2x_set_rx_mode,
.ndo_set_mac_address = bnx2x_change_mac_addr,
.ndo_validate_addr = eth_validate_addr,
.ndo_do_ioctl = bnx2x_ioctl,
@@ -9776,15 +10018,21 @@ static void bnx2x_cnic_sp_post(struct bnx2x *bp, int count)
HW_CID(bp, BNX2X_ISCSI_ETH_CID));
}
- /* There may be not more than 8 L2 and COMMON SPEs and not more
- * than 8 L5 SPEs in the air.
+ /* There may be not more than 8 L2 and not more than 8 L5 SPEs
+ * We also check that the number of outstanding
+ * COMMON ramrods is not more than the EQ and SPQ can
+ * accommodate.
*/
- if ((type == NONE_CONNECTION_TYPE) ||
- (type == ETH_CONNECTION_TYPE)) {
- if (!atomic_read(&bp->spq_left))
+ if (type == ETH_CONNECTION_TYPE) {
+ if (!atomic_read(&bp->cq_spq_left))
+ break;
+ else
+ atomic_dec(&bp->cq_spq_left);
+ } else if (type == NONE_CONNECTION_TYPE) {
+ if (!atomic_read(&bp->eq_spq_left))
break;
else
- atomic_dec(&bp->spq_left);
+ atomic_dec(&bp->eq_spq_left);
} else if ((type == ISCSI_CONNECTION_TYPE) ||
(type == FCOE_CONNECTION_TYPE)) {
if (bp->cnic_spq_pending >=
@@ -9976,7 +10224,7 @@ static int bnx2x_drv_ctl(struct net_device *dev, struct drv_ctl_info *ctl)
int count = ctl->data.credit.credit_count;
smp_mb__before_atomic_inc();
- atomic_add(count, &bp->spq_left);
+ atomic_add(count, &bp->cq_spq_left);
smp_mb__after_atomic_inc();
break;
}
--
1.7.0.4
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists