* [PATCH net-next v3 07/22] bnx2x: Support ndo_set_rxmode in VF driver
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
The VF driver uses the 'q_filter' request in the VF <-> PF channel to
have the PF configure the requested rxmode to device. ndo_set_rxmode
is called under bottom half lock, so sleeping until the response
arrives over the VF <-> PF channel is out of the question. For this reason
the VF driver returns from the ndo after scheduling a work item, which
in turn processes the rx mode request and adds the classification
information through the VF <-> PF channel accordingly.
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 5 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 174 +++++++++++++++++++++-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h | 7 +
3 files changed, 180 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 4f1e06e..88cbdb9 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -1191,6 +1191,8 @@ enum {
BNX2X_SP_RTNL_TX_TIMEOUT,
BNX2X_SP_RTNL_AFEX_F_UPDATE,
BNX2X_SP_RTNL_FAN_FAILURE,
+ BNX2X_SP_RTNL_VFPF_MCAST,
+ BNX2X_SP_RTNL_VFPF_STORM_RX_MODE,
};
@@ -2229,6 +2231,9 @@ void bnx2x_vfpf_close_vf(struct bnx2x *bp);
int bnx2x_vfpf_setup_q(struct bnx2x *bp, int fp_idx);
int bnx2x_vfpf_teardown_queue(struct bnx2x *bp, int qidx);
int bnx2x_vfpf_set_mac(struct bnx2x *bp);
+int bnx2x_vfpf_set_mcast(struct net_device *dev);
+int bnx2x_vfpf_storm_rx_mode(struct bnx2x *bp);
+
int bnx2x_nic_load_analyze_req(struct bnx2x *bp, u32 load_code);
/* Congestion management fairness mode */
#define CMNG_FNS_NONE 0
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 1317813..6ac79ba 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -9406,6 +9406,19 @@ sp_rtnl_not_reset:
bnx2x_close(bp->dev);
}
+ if (test_and_clear_bit(BNX2X_SP_RTNL_VFPF_MCAST, &bp->sp_rtnl_state)) {
+ DP(BNX2X_MSG_SP,
+ "sending set mcast vf pf channel message from rtnl sp-task\n");
+ bnx2x_vfpf_set_mcast(bp->dev);
+ }
+
+ if (test_and_clear_bit(BNX2X_SP_RTNL_VFPF_STORM_RX_MODE,
+ &bp->sp_rtnl_state)) {
+ DP(BNX2X_MSG_SP,
+ "sending set storm rx mode vf pf channel message from rtnl sp-task\n");
+ bnx2x_vfpf_storm_rx_mode(bp);
+ }
+
sp_rtnl_exit:
rtnl_unlock();
}
@@ -11485,12 +11498,25 @@ void bnx2x_set_rx_mode(struct net_device *dev)
CHIP_IS_E1(bp)))
rx_mode = BNX2X_RX_MODE_ALLMULTI;
else {
- /* some multicasts */
- if (bnx2x_set_mc_list(bp) < 0)
- rx_mode = BNX2X_RX_MODE_ALLMULTI;
+ if (IS_PF(bp)) {
+ /* some multicasts */
+ if (bnx2x_set_mc_list(bp) < 0)
+ rx_mode = BNX2X_RX_MODE_ALLMULTI;
- if (bnx2x_set_uc_list(bp) < 0)
- rx_mode = BNX2X_RX_MODE_PROMISC;
+ if (bnx2x_set_uc_list(bp) < 0)
+ rx_mode = BNX2X_RX_MODE_PROMISC;
+ } else {
+ /* configuring mcast to a vf involves sleeping (when we
+ * wait for the pf's response). Since this function is
+ * called from non sleepable context we must schedule
+ * a work item for this purpose
+ */
+ smp_mb__before_clear_bit();
+ set_bit(BNX2X_SP_RTNL_VFPF_MCAST,
+ &bp->sp_rtnl_state);
+ smp_mb__after_clear_bit();
+ schedule_delayed_work(&bp->sp_rtnl_task, 0);
+ }
}
bp->rx_mode = rx_mode;
@@ -11504,7 +11530,20 @@ void bnx2x_set_rx_mode(struct net_device *dev)
return;
}
- bnx2x_set_storm_rx_mode(bp);
+ if (IS_PF(bp)) {
+ bnx2x_set_storm_rx_mode(bp);
+ } else {
+ /* configuring rx mode to storms in a vf involves sleeping (when
+ * we wait for the pf's response). Since this function is
+ * called from non sleepable context we must schedule
+ * a work item for this purpose
+ */
+ smp_mb__before_clear_bit();
+ set_bit(BNX2X_SP_RTNL_VFPF_STORM_RX_MODE,
+ &bp->sp_rtnl_state);
+ smp_mb__after_clear_bit();
+ schedule_delayed_work(&bp->sp_rtnl_task, 0);
+ }
}
/* called with rtnl_lock */
@@ -13690,3 +13729,126 @@ int bnx2x_vfpf_set_mac(struct bnx2x *bp)
return 0;
}
+
+int bnx2x_vfpf_set_mcast(struct net_device *dev)
+{
+ struct bnx2x *bp = netdev_priv(dev);
+ struct vfpf_set_q_filters_tlv *req = &bp->vf2pf_mbox->req.set_q_filters;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ int rc, i = 0;
+ struct netdev_hw_addr *ha;
+
+ if (bp->state != BNX2X_STATE_OPEN) {
+ DP(NETIF_MSG_IFUP, "state is %x, returning\n", bp->state);
+ return -EINVAL;
+ }
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_SET_Q_FILTERS,
+ sizeof(*req));
+
+ /* Get Rx mode requested */
+ DP(NETIF_MSG_IFUP, "dev->flags = %x\n", dev->flags);
+
+ netdev_for_each_mc_addr(ha, dev) {
+ DP(NETIF_MSG_IFUP, "Adding mcast MAC: %pM\n",
+ bnx2x_mc_addr(ha));
+ memcpy(req->multicast[i], bnx2x_mc_addr(ha), ETH_ALEN);
+ i++;
+ }
+
+ /* We support four PFVF_MAX_MULTICAST_PER_VF mcast
+ * addresses tops
+ */
+ if (i >= PFVF_MAX_MULTICAST_PER_VF) {
+ DP(NETIF_MSG_IFUP,
+ "VF supports not more than %d multicast MAC addresses\n",
+ PFVF_MAX_MULTICAST_PER_VF);
+ return -EINVAL;
+ }
+
+ req->n_multicast = i;
+ req->flags |= VFPF_SET_Q_FILTERS_MULTICAST_CHANGED;
+
+ req->vf_qid = 0;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+ if (rc) {
+ BNX2X_ERR("Sending a message failed: %d\n", rc);
+ return rc;
+ }
+
+ if (resp->hdr.status != PFVF_STATUS_SUCCESS) {
+ BNX2X_ERR("Set Rx mode/multicast failed: %d\n",
+ resp->hdr.status);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+int bnx2x_vfpf_storm_rx_mode(struct bnx2x *bp)
+{
+ int mode = bp->rx_mode;
+ struct vfpf_set_q_filters_tlv *req = &bp->vf2pf_mbox->req.set_q_filters;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ int rc;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_SET_Q_FILTERS,
+ sizeof(*req));
+
+ DP(NETIF_MSG_IFUP, "Rx mode is %d\n", mode);
+
+ switch (mode) {
+ case BNX2X_RX_MODE_NONE: /* no Rx */
+ req->rx_mask = VFPF_RX_MASK_ACCEPT_NONE;
+ break;
+ case BNX2X_RX_MODE_NORMAL:
+ req->rx_mask = VFPF_RX_MASK_ACCEPT_MATCHED_MULTICAST;
+ req->rx_mask |= VFPF_RX_MASK_ACCEPT_MATCHED_UNICAST;
+ req->rx_mask |= VFPF_RX_MASK_ACCEPT_BROADCAST;
+ break;
+ case BNX2X_RX_MODE_ALLMULTI:
+ req->rx_mask = VFPF_RX_MASK_ACCEPT_ALL_MULTICAST;
+ req->rx_mask |= VFPF_RX_MASK_ACCEPT_MATCHED_UNICAST;
+ req->rx_mask |= VFPF_RX_MASK_ACCEPT_BROADCAST;
+ break;
+ case BNX2X_RX_MODE_PROMISC:
+ req->rx_mask = VFPF_RX_MASK_ACCEPT_ALL_UNICAST;
+ req->rx_mask |= VFPF_RX_MASK_ACCEPT_ALL_MULTICAST;
+ req->rx_mask |= VFPF_RX_MASK_ACCEPT_BROADCAST;
+ break;
+ default:
+ BNX2X_ERR("BAD rx mode (%d)\n", mode);
+ return -EINVAL;
+ }
+
+ req->flags |= VFPF_SET_Q_FILTERS_RX_MASK_CHANGED;
+ req->vf_qid = 0;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+ if (rc)
+ BNX2X_ERR("Sending a message failed: %d\n", rc);
+
+ if (resp->hdr.status != PFVF_STATUS_SUCCESS) {
+ BNX2X_ERR("Set Rx mode failed: %d\n", resp->hdr.status);
+ return -EINVAL;
+ }
+
+ return rc;
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index 116a17a..a289e59 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -54,6 +54,13 @@ struct hw_sb_info {
#define VFPF_QUEUE_DROP_TTL0 (1 << 2)
#define VFPF_QUEUE_DROP_UDP_CS_ERR (1 << 3)
+#define VFPF_RX_MASK_ACCEPT_NONE 0x00000000
+#define VFPF_RX_MASK_ACCEPT_MATCHED_UNICAST 0x00000001
+#define VFPF_RX_MASK_ACCEPT_MATCHED_MULTICAST 0x00000002
+#define VFPF_RX_MASK_ACCEPT_ALL_UNICAST 0x00000004
+#define VFPF_RX_MASK_ACCEPT_ALL_MULTICAST 0x00000008
+#define VFPF_RX_MASK_ACCEPT_BROADCAST 0x00000010
+
enum {
PFVF_STATUS_WAITING = 0,
PFVF_STATUS_SUCCESS,
--
1.7.9.GIT
^ permalink raw reply related
* [PATCH net-next v3 06/22] bnx2x: Add teardown_q and close to VF <-> PF channel
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
When a VF is being closed its queues are released via
the 'teardown_q' and the VF itself is closed with
'close'. These are essentially the unload counterparts of
'init' and 'setup_q' from the load flow.
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 2 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 12 +++-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 85 ++++++++++++++++++++++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h | 18 +++++
4 files changed, 115 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 5c71002..4f1e06e 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -2225,7 +2225,9 @@ int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping);
int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count);
int bnx2x_vfpf_release(struct bnx2x *bp);
int bnx2x_vfpf_init(struct bnx2x *bp);
+void bnx2x_vfpf_close_vf(struct bnx2x *bp);
int bnx2x_vfpf_setup_q(struct bnx2x *bp, int fp_idx);
+int bnx2x_vfpf_teardown_queue(struct bnx2x *bp, int qidx);
int bnx2x_vfpf_set_mac(struct bnx2x *bp);
int bnx2x_nic_load_analyze_req(struct bnx2x *bp, u32 load_code);
/* Congestion management fairness mode */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 092cbee..4e87eec 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -2775,9 +2775,17 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
/* wait till consumers catch up with producers in all queues */
bnx2x_drain_tx_queues(bp);
- /* Cleanup the chip if needed */
- if (unload_mode != UNLOAD_RECOVERY)
+ /* if VF indicate to PF this function is going down (PF will delete sp
+ * elsements and clear initializations
+ */
+ if (IS_VF(bp))
+ bnx2x_vfpf_close_vf(bp);
+
+ /* if this is a normal/close unload need to clean up chip*/
+ else if (unload_mode != UNLOAD_RECOVERY)
bnx2x_chip_cleanup(bp, unload_mode, keep_link);
+
+ /* recovery */
else {
/* Send the UNLOAD_REQUEST to the MCP */
bnx2x_send_unload_req(bp, unload_mode);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index b0fa707..1317813 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -13480,6 +13480,55 @@ int bnx2x_vfpf_init(struct bnx2x *bp)
return 0;
}
+/* CLOSE VF - opposite to INIT_VF */
+void bnx2x_vfpf_close_vf(struct bnx2x *bp)
+{
+ struct vfpf_close_tlv *req = &bp->vf2pf_mbox->req.close;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ int i, rc;
+ u32 vf_id;
+
+ /* If we haven't got a valid VF id, there is no sense to
+ * continue with sending messages
+ */
+ if (bnx2x_get_vf_id(bp, &vf_id))
+ goto free_irq;
+
+ /* Close the queues */
+ for_each_queue(bp, i)
+ bnx2x_vfpf_teardown_queue(bp, i);
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_CLOSE, sizeof(*req));
+
+ req->vf_id = vf_id;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+
+ if (rc)
+ BNX2X_ERR("Sending CLOSE failed. rc was: %d\n", rc);
+
+ else if (resp->hdr.status != PFVF_STATUS_SUCCESS)
+ BNX2X_ERR("Sending CLOSE failed: pf response was %d\n",
+ resp->hdr.status);
+
+free_irq:
+ /* Disable HW interrupts, NAPI */
+ bnx2x_netif_stop(bp, 0);
+ /* Delete all NAPI objects */
+ bnx2x_del_all_napi(bp);
+
+ /* Release IRQs */
+ bnx2x_free_irq(bp);
+}
+
/* ask the pf to open a queue for the vf */
int bnx2x_vfpf_setup_q(struct bnx2x *bp, int fp_idx)
{
@@ -13563,6 +13612,42 @@ int bnx2x_vfpf_setup_q(struct bnx2x *bp, int fp_idx)
return rc;
}
+int bnx2x_vfpf_teardown_queue(struct bnx2x *bp, int qidx)
+{
+ struct vfpf_q_op_tlv *req = &bp->vf2pf_mbox->req.q_op;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ int rc;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_TEARDOWN_Q,
+ sizeof(*req));
+
+ req->vf_qid = qidx;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+
+ if (rc) {
+ BNX2X_ERR("Sending TEARDOWN for queue %d failed: %d\n", qidx,
+ rc);
+ return rc;
+ }
+
+ if (resp->hdr.status != PFVF_STATUS_SUCCESS) {
+ BNX2X_ERR("TEARDOWN for queue %d failed: %d\n", qidx,
+ resp->hdr.status);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
/* request pf to add a mac for the vf */
int bnx2x_vfpf_set_mac(struct bnx2x *bp)
{
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index ded37bc..116a17a 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -110,6 +110,13 @@ struct vfpf_acquire_tlv {
aligned_u64 bulletin_addr;
};
+/* simple operation request on queue */
+struct vfpf_q_op_tlv {
+ struct vfpf_first_tlv first_tlv;
+ u8 vf_qid;
+ u8 padding[3];
+};
+
/* acquire response tlv - carries the allocated resources */
struct pfvf_acquire_resp_tlv {
struct pfvf_tlv hdr;
@@ -246,6 +253,13 @@ struct vfpf_set_q_filters_tlv {
u32 rx_mask; /* see mask constants at the top of the file */
};
+/* close VF (disable VF) */
+struct vfpf_close_tlv {
+ struct vfpf_first_tlv first_tlv;
+ u16 vf_id; /* for debug */
+ u8 padding[2];
+};
+
/* release the VF's acquired resources */
struct vfpf_release_tlv {
struct vfpf_first_tlv first_tlv;
@@ -261,6 +275,8 @@ union vfpf_tlvs {
struct vfpf_first_tlv first_tlv;
struct vfpf_acquire_tlv acquire;
struct vfpf_init_tlv init;
+ struct vfpf_close_tlv close;
+ struct vfpf_q_op_tlv q_op;
struct vfpf_setup_q_tlv setup_q;
struct vfpf_set_q_filters_tlv set_q_filters;
struct vfpf_release_tlv release;
@@ -283,6 +299,8 @@ enum channel_tlvs {
CHANNEL_TLV_INIT,
CHANNEL_TLV_SETUP_Q,
CHANNEL_TLV_SET_Q_FILTERS,
+ CHANNEL_TLV_TEARDOWN_Q,
+ CHANNEL_TLV_CLOSE,
CHANNEL_TLV_RELEASE,
CHANNEL_TLV_LIST_END,
CHANNEL_TLV_MAX
--
1.7.9.GIT
^ permalink raw reply related
* [PATCH net-next v3 05/22] bnx2x: Add init, setup_q, set_mac to VF <-> PF channel
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
'init' - init an acquired VF. Supply allocation GPAs to PF.
'setup_q' - PF to allocate a queue in device on behalf of the VF.
'set_mac' - PF to configure a mac in device on behalf of the VF.
VF driver uses these requests in the VF <-> PF channel in nic_load
flow.
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 6 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 17 +++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 167 ++++++++++++++++++++++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h | 123 ++++++++++++++++
4 files changed, 313 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 17536ff..5c71002 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -341,6 +341,9 @@ union db_prod {
#define SGE_PAGE_SIZE PAGE_SIZE
#define SGE_PAGE_SHIFT PAGE_SHIFT
#define SGE_PAGE_ALIGN(addr) PAGE_ALIGN((typeof(PAGE_SIZE))(addr))
+#define SGE_PAGES (SGE_PAGE_SIZE * PAGES_PER_SGE)
+#define TPA_AGG_SIZE min_t(u32, (min_t(u32, 8, MAX_SKB_FRAGS) * \
+ SGE_PAGES), 0xffff)
/* SGE ring related macros */
#define NUM_RX_SGE_PAGES 2
@@ -2221,6 +2224,9 @@ int bnx2x_get_vf_id(struct bnx2x *bp, u32 *vf_id);
int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping);
int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count);
int bnx2x_vfpf_release(struct bnx2x *bp);
+int bnx2x_vfpf_init(struct bnx2x *bp);
+int bnx2x_vfpf_setup_q(struct bnx2x *bp, int fp_idx);
+int bnx2x_vfpf_set_mac(struct bnx2x *bp);
int bnx2x_nic_load_analyze_req(struct bnx2x *bp, u32 load_code);
/* Congestion management fairness mode */
#define CMNG_FNS_NONE 0
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 0f6bb1f..092cbee 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -2432,6 +2432,13 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
LOAD_ERROR_EXIT(bp, load_error0);
}
+ /* request pf to initialize status blocks */
+ if (IS_VF(bp)) {
+ rc = bnx2x_vfpf_init(bp);
+ if (rc)
+ LOAD_ERROR_EXIT(bp, load_error0);
+ }
+
/* As long as bnx2x_alloc_mem() may possibly update
* bp->num_queues, bnx2x_set_real_num_queues() should always
* come after it. At this stage cnic queues are not counted.
@@ -2549,6 +2556,14 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
LOAD_ERROR_EXIT(bp, load_error3);
}
+ } else { /* vf */
+ for_each_eth_queue(bp, i) {
+ rc = bnx2x_vfpf_setup_q(bp, i);
+ if (rc) {
+ BNX2X_ERR("Queue setup failed\n");
+ LOAD_ERROR_EXIT(bp, load_error3);
+ }
+ }
}
/* Now when Clients are configured we are ready to work */
@@ -2557,6 +2572,8 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
/* Configure a ucast MAC */
if (IS_PF(bp))
rc = bnx2x_set_eth_mac(bp, true);
+ else /* vf */
+ rc = bnx2x_vfpf_set_mac(bp);
if (rc) {
BNX2X_ERR("Setting Ethernet MAC failed\n");
LOAD_ERROR_EXIT(bp, load_error3);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 158dcec..b0fa707 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -13438,3 +13438,170 @@ int bnx2x_vfpf_release(struct bnx2x *bp)
return 0;
}
+
+/* Tell PF about SB addresses */
+int bnx2x_vfpf_init(struct bnx2x *bp)
+{
+ struct vfpf_init_tlv *req = &bp->vf2pf_mbox->req.init;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ int rc, i;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_INIT, sizeof(*req));
+
+ /* status blocks */
+ for_each_eth_queue(bp, i)
+ req->sb_addr[i] = (dma_addr_t)bnx2x_fp(bp, i,
+ status_blk_mapping);
+
+ /* statistics - requests only supports single queue for now */
+ req->stats_addr = bp->fw_stats_data_mapping +
+ offsetof(struct bnx2x_fw_stats_data, queue_stats);
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+ if (rc)
+ return rc;
+
+ if (resp->hdr.status != PFVF_STATUS_SUCCESS) {
+ BNX2X_ERR("INIT VF failed: %d. Breaking...\n",
+ resp->hdr.status);
+ return -EAGAIN;
+ }
+
+ DP(BNX2X_MSG_SP, "INIT VF Succeeded\n");
+
+ return 0;
+}
+
+/* ask the pf to open a queue for the vf */
+int bnx2x_vfpf_setup_q(struct bnx2x *bp, int fp_idx)
+{
+ struct vfpf_setup_q_tlv *req = &bp->vf2pf_mbox->req.setup_q;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ struct bnx2x_fastpath *fp = &bp->fp[fp_idx];
+ u16 tpa_agg_size = 0, flags = 0;
+ int rc;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_SETUP_Q, sizeof(*req));
+
+ /* select tpa mode to request */
+ if (!fp->disable_tpa) {
+ flags |= VFPF_QUEUE_FLG_TPA;
+ flags |= VFPF_QUEUE_FLG_TPA_IPV6;
+ if (fp->mode == TPA_MODE_GRO)
+ flags |= VFPF_QUEUE_FLG_TPA_GRO;
+ tpa_agg_size = TPA_AGG_SIZE;
+ }
+
+ /* calculate queue flags */
+ flags |= VFPF_QUEUE_FLG_STATS;
+ flags |= VFPF_QUEUE_FLG_CACHE_ALIGN;
+ flags |= IS_MF_SD(bp) ? VFPF_QUEUE_FLG_OV : 0;
+
+ flags |= VFPF_QUEUE_FLG_VLAN;
+ DP(NETIF_MSG_IFUP, "vlan removal enabled\n");
+
+ /* Common */
+ req->vf_qid = fp_idx;
+ req->param_valid = VFPF_RXQ_VALID | VFPF_TXQ_VALID;
+
+ /* Rx */
+ req->rxq.rcq_addr = fp->rx_comp_mapping;
+ req->rxq.rcq_np_addr = fp->rx_comp_mapping + BCM_PAGE_SIZE;
+ req->rxq.rxq_addr = fp->rx_desc_mapping;
+ req->rxq.sge_addr = fp->rx_sge_mapping;
+
+ req->rxq.vf_sb = fp_idx;
+ req->rxq.sb_index = HC_INDEX_ETH_RX_CQ_CONS;
+ req->rxq.hc_rate = bp->rx_ticks ? 1000000/bp->rx_ticks : 0;
+
+ req->rxq.mtu = bp->dev->mtu;
+ req->rxq.buf_sz = fp->rx_buf_size;
+ req->rxq.sge_buf_sz = BCM_PAGE_SIZE * PAGES_PER_SGE;
+ req->rxq.tpa_agg_sz = tpa_agg_size;
+ req->rxq.max_sge_pkt = SGE_PAGE_ALIGN(bp->dev->mtu) >> SGE_PAGE_SHIFT;
+ req->rxq.max_sge_pkt = ((req->rxq.max_sge_pkt + PAGES_PER_SGE - 1) &
+ (~(PAGES_PER_SGE-1))) >> PAGES_PER_SGE_SHIFT;
+ req->rxq.flags = flags;
+ req->rxq.drop_flags = 0;
+ req->rxq.cache_line_log = BNX2X_RX_ALIGN_SHIFT;
+ req->rxq.stat_id = -1; /* No stats at the moment */
+
+ /* Tx */
+ req->txq.txq_addr = fp->txdata_ptr[FIRST_TX_COS_INDEX]->tx_desc_mapping;
+ req->txq.vf_sb = fp_idx;
+ req->txq.sb_index = HC_INDEX_ETH_TX_CQ_CONS_COS0;
+ req->txq.hc_rate = bp->tx_ticks ? 1000000/bp->tx_ticks : 0;
+ req->txq.flags = flags;
+ req->txq.traffic_type = LLFC_TRAFFIC_TYPE_NW;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+ if (rc)
+ BNX2X_ERR("Sending SETUP_Q message for queue[%d] failed!\n",
+ fp_idx);
+
+ if (resp->hdr.status != PFVF_STATUS_SUCCESS) {
+ BNX2X_ERR("Status of SETUP_Q for queue[%d] is %d\n",
+ fp_idx, resp->hdr.status);
+ return -EINVAL;
+ }
+ return rc;
+}
+
+/* request pf to add a mac for the vf */
+int bnx2x_vfpf_set_mac(struct bnx2x *bp)
+{
+ struct vfpf_set_q_filters_tlv *req = &bp->vf2pf_mbox->req.set_q_filters;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ int rc;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_SET_Q_FILTERS,
+ sizeof(*req));
+
+ req->flags = VFPF_SET_Q_FILTERS_MAC_VLAN_CHANGED;
+ req->vf_qid = 0;
+ req->n_mac_vlan_filters = 1;
+
+ req->filters[0].flags =
+ VFPF_Q_FILTER_DEST_MAC_VALID | VFPF_Q_FILTER_SET_MAC;
+
+ /* copy mac from device to request */
+ memcpy(req->filters[0].mac, bp->dev->dev_addr, ETH_ALEN);
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ /* send message to pf */
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+ if (rc) {
+ BNX2X_ERR("failed to send message to pf. rc was %d\n", rc);
+ return rc;
+ }
+
+ if (resp->hdr.status != PFVF_STATUS_SUCCESS) {
+ BNX2X_ERR("vfpf SET MAC failed: %d\n", resp->hdr.status);
+ return -EINVAL;
+ }
+
+ return 0;
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index 763d1c2..ded37bc 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -38,6 +38,22 @@ struct hw_sb_info {
*/
#define TLV_BUFFER_SIZE 1024
+#define VFPF_QUEUE_FLG_TPA 0x0001
+#define VFPF_QUEUE_FLG_TPA_IPV6 0x0002
+#define VFPF_QUEUE_FLG_TPA_GRO 0x0004
+#define VFPF_QUEUE_FLG_CACHE_ALIGN 0x0008
+#define VFPF_QUEUE_FLG_STATS 0x0010
+#define VFPF_QUEUE_FLG_OV 0x0020
+#define VFPF_QUEUE_FLG_VLAN 0x0040
+#define VFPF_QUEUE_FLG_COS 0x0080
+#define VFPF_QUEUE_FLG_HC 0x0100
+#define VFPF_QUEUE_FLG_DHC 0x0200
+
+#define VFPF_QUEUE_DROP_IP_CS_ERR (1 << 0)
+#define VFPF_QUEUE_DROP_TCP_CS_ERR (1 << 1)
+#define VFPF_QUEUE_DROP_TTL0 (1 << 2)
+#define VFPF_QUEUE_DROP_UDP_CS_ERR (1 << 3)
+
enum {
PFVF_STATUS_WAITING = 0,
PFVF_STATUS_SUCCESS,
@@ -129,6 +145,107 @@ struct pfvf_acquire_resp_tlv {
} resc;
};
+/* Init VF */
+struct vfpf_init_tlv {
+ struct vfpf_first_tlv first_tlv;
+ aligned_u64 sb_addr[PFVF_MAX_SBS_PER_VF]; /* vf_sb based */
+ aligned_u64 spq_addr;
+ aligned_u64 stats_addr;
+};
+
+/* Setup Queue */
+struct vfpf_setup_q_tlv {
+ struct vfpf_first_tlv first_tlv;
+
+ struct vf_pf_rxq_params {
+ /* physical addresses */
+ aligned_u64 rcq_addr;
+ aligned_u64 rcq_np_addr;
+ aligned_u64 rxq_addr;
+ aligned_u64 sge_addr;
+
+ /* sb + hc info */
+ u8 vf_sb; /* index in hw_sbs[] */
+ u8 sb_index; /* Index in the SB */
+ u16 hc_rate; /* desired interrupts per sec. */
+ /* valid iff VFPF_QUEUE_FLG_HC */
+ /* rx buffer info */
+ u16 mtu;
+ u16 buf_sz;
+ u16 flags; /* VFPF_QUEUE_FLG_X flags */
+ u16 stat_id; /* valid iff VFPF_QUEUE_FLG_STATS */
+
+ /* valid iff VFPF_QUEUE_FLG_TPA */
+ u16 sge_buf_sz;
+ u16 tpa_agg_sz;
+ u8 max_sge_pkt;
+
+ u8 drop_flags; /* VFPF_QUEUE_DROP_X, for Linux VMs
+ * all the flags are turned off
+ */
+
+ u8 cache_line_log; /* VFPF_QUEUE_FLG_CACHE_ALIGN */
+ u8 padding;
+ } rxq;
+
+ struct vf_pf_txq_params {
+ /* physical addresses */
+ aligned_u64 txq_addr;
+
+ /* sb + hc info */
+ u8 vf_sb; /* index in hw_sbs[] */
+ u8 sb_index; /* Index in the SB */
+ u16 hc_rate; /* desired interrupts per sec. */
+ /* valid iff VFPF_QUEUE_FLG_HC */
+ u32 flags; /* VFPF_QUEUE_FLG_X flags */
+ u16 stat_id; /* valid iff VFPF_QUEUE_FLG_STATS */
+ u8 traffic_type; /* see in setup_context() */
+ u8 padding;
+ } txq;
+
+ u8 vf_qid; /* index in hw_qid[] */
+ u8 param_valid;
+ #define VFPF_RXQ_VALID 0x01
+ #define VFPF_TXQ_VALID 0x02
+ u8 padding[2];
+};
+
+/* Set Queue Filters */
+struct vfpf_q_mac_vlan_filter {
+ u32 flags;
+ #define VFPF_Q_FILTER_DEST_MAC_VALID 0x01
+ #define VFPF_Q_FILTER_VLAN_TAG_VALID 0x02
+ #define VFPF_Q_FILTER_SET_MAC 0x100 /* set/clear */
+ u8 mac[ETH_ALEN];
+ u16 vlan_tag;
+};
+
+/* configure queue filters */
+struct vfpf_set_q_filters_tlv {
+ struct vfpf_first_tlv first_tlv;
+
+ u32 flags;
+ #define VFPF_SET_Q_FILTERS_MAC_VLAN_CHANGED 0x01
+ #define VFPF_SET_Q_FILTERS_MULTICAST_CHANGED 0x02
+ #define VFPF_SET_Q_FILTERS_RX_MASK_CHANGED 0x04
+
+ u8 vf_qid; /* index in hw_qid[] */
+ u8 n_mac_vlan_filters;
+ u8 n_multicast;
+ u8 padding;
+
+ #define PFVF_MAX_MAC_FILTERS 16
+ #define PFVF_MAX_VLAN_FILTERS 16
+ #define PFVF_MAX_FILTERS (PFVF_MAX_MAC_FILTERS +\
+ PFVF_MAX_VLAN_FILTERS)
+ struct vfpf_q_mac_vlan_filter filters[PFVF_MAX_FILTERS];
+
+#define PFVF_MAX_MULTICAST_PER_VF 32
+ u8 multicast[PFVF_MAX_MULTICAST_PER_VF][ETH_ALEN];
+
+ u32 rx_mask; /* see mask constants at the top of the file */
+};
+
/* release the VF's acquired resources */
struct vfpf_release_tlv {
struct vfpf_first_tlv first_tlv;
@@ -143,6 +260,9 @@ struct tlv_buffer_size {
union vfpf_tlvs {
struct vfpf_first_tlv first_tlv;
struct vfpf_acquire_tlv acquire;
+ struct vfpf_init_tlv init;
+ struct vfpf_setup_q_tlv setup_q;
+ struct vfpf_set_q_filters_tlv set_q_filters;
struct vfpf_release_tlv release;
struct channel_list_end_tlv list_end;
struct tlv_buffer_size tlv_buf_size;
@@ -160,6 +280,9 @@ union pfvf_tlvs {
enum channel_tlvs {
CHANNEL_TLV_NONE,
CHANNEL_TLV_ACQUIRE,
+ CHANNEL_TLV_INIT,
+ CHANNEL_TLV_SETUP_Q,
+ CHANNEL_TLV_SET_Q_FILTERS,
CHANNEL_TLV_RELEASE,
CHANNEL_TLV_LIST_END,
CHANNEL_TLV_MAX
--
1.7.9.GIT
^ permalink raw reply related
* [PATCH net-next v3 04/22] bnx2x: Separate VF and PF logic
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
Generally, the VF driver cannot access the chip, except by the
narrow window its BAR allows. Care had to be taken so the VF driver
will not reach code which accesses the chip elsewhere.
Refactor the nic_load flow into parts so it would be
easier to separate the VF-only logic from the PF-only logic.
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 1 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 611 +++++++++++++-------
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h | 15 +-
.../net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c | 2 +-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 98 +++-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h | 6 +
6 files changed, 500 insertions(+), 233 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 6cafa4e..17536ff 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -2221,6 +2221,7 @@ int bnx2x_get_vf_id(struct bnx2x *bp, u32 *vf_id);
int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping);
int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count);
int bnx2x_vfpf_release(struct bnx2x *bp);
+int bnx2x_nic_load_analyze_req(struct bnx2x *bp, u32 load_code);
/* Congestion management fairness mode */
#define CMNG_FNS_NONE 0
#define CMNG_FNS_MINMAX 1
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 0a493f4..0f6bb1f 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -1048,7 +1048,7 @@ void __bnx2x_link_report(struct bnx2x *bp)
struct bnx2x_link_report_data cur_data;
/* reread mf_cfg */
- if (!CHIP_IS_E1(bp))
+ if (IS_PF(bp) && !CHIP_IS_E1(bp))
bnx2x_read_mf_cfg(bp);
/* Read the current link report info */
@@ -1391,10 +1391,14 @@ static void bnx2x_free_msix_irqs(struct bnx2x *bp, int nvecs)
if (nvecs == offset)
return;
- free_irq(bp->msix_table[offset].vector, bp->dev);
- DP(NETIF_MSG_IFDOWN, "released sp irq (%d)\n",
- bp->msix_table[offset].vector);
- offset++;
+
+ /* VFs don't have a default SB */
+ if (IS_PF(bp)) {
+ free_irq(bp->msix_table[offset].vector, bp->dev);
+ DP(NETIF_MSG_IFDOWN, "released sp irq (%d)\n",
+ bp->msix_table[offset].vector);
+ offset++;
+ }
if (CNIC_SUPPORT(bp)) {
if (nvecs == offset)
@@ -1415,11 +1419,17 @@ static void bnx2x_free_msix_irqs(struct bnx2x *bp, int nvecs)
void bnx2x_free_irq(struct bnx2x *bp)
{
if (bp->flags & USING_MSIX_FLAG &&
- !(bp->flags & USING_SINGLE_MSIX_FLAG))
- bnx2x_free_msix_irqs(bp, BNX2X_NUM_ETH_QUEUES(bp) +
- CNIC_SUPPORT(bp) + 1);
- else
+ !(bp->flags & USING_SINGLE_MSIX_FLAG)) {
+ int nvecs = BNX2X_NUM_ETH_QUEUES(bp) + CNIC_SUPPORT(bp);
+
+ /* vfs don't have a default status block */
+ if (IS_PF(bp))
+ nvecs++;
+
+ bnx2x_free_msix_irqs(bp, nvecs);
+ } else {
free_irq(bp->dev->irq, bp->dev);
+ }
}
int bnx2x_enable_msix(struct bnx2x *bp)
@@ -1515,12 +1525,15 @@ static int bnx2x_req_msix_irqs(struct bnx2x *bp)
{
int i, rc, offset = 0;
- rc = request_irq(bp->msix_table[offset++].vector,
- bnx2x_msix_sp_int, 0,
- bp->dev->name, bp->dev);
- if (rc) {
- BNX2X_ERR("request sp irq failed\n");
- return -EBUSY;
+ /* no default status block for vf */
+ if (IS_PF(bp)) {
+ rc = request_irq(bp->msix_table[offset++].vector,
+ bnx2x_msix_sp_int, 0,
+ bp->dev->name, bp->dev);
+ if (rc) {
+ BNX2X_ERR("request sp irq failed\n");
+ return -EBUSY;
+ }
}
if (CNIC_SUPPORT(bp))
@@ -1544,12 +1557,20 @@ static int bnx2x_req_msix_irqs(struct bnx2x *bp)
}
i = BNX2X_NUM_ETH_QUEUES(bp);
- offset = 1 + CNIC_SUPPORT(bp);
- netdev_info(bp->dev, "using MSI-X IRQs: sp %d fp[%d] %d ... fp[%d] %d\n",
- bp->msix_table[0].vector,
- 0, bp->msix_table[offset].vector,
- i - 1, bp->msix_table[offset + i - 1].vector);
-
+ if (IS_PF(bp)) {
+ offset = 1 + CNIC_SUPPORT(bp);
+ netdev_info(bp->dev,
+ "using MSI-X IRQs: sp %d fp[%d] %d ... fp[%d] %d\n",
+ bp->msix_table[0].vector,
+ 0, bp->msix_table[offset].vector,
+ i - 1, bp->msix_table[offset + i - 1].vector);
+ } else {
+ offset = CNIC_SUPPORT(bp);
+ netdev_info(bp->dev,
+ "using MSI-X IRQs: fp[%d] %d ... fp[%d] %d\n",
+ 0, bp->msix_table[offset].vector,
+ i - 1, bp->msix_table[offset + i - 1].vector);
+ }
return 0;
}
@@ -1956,27 +1977,206 @@ static void bnx2x_squeeze_objects(struct bnx2x *bp)
} while (0)
#endif /*BNX2X_STOP_ON_ERROR*/
-bool bnx2x_test_firmware_version(struct bnx2x *bp, bool is_err)
+static void bnx2x_free_fw_stats_mem(struct bnx2x *bp)
{
- /* build FW version dword */
- u32 my_fw = (BCM_5710_FW_MAJOR_VERSION) +
- (BCM_5710_FW_MINOR_VERSION << 8) +
- (BCM_5710_FW_REVISION_VERSION << 16) +
- (BCM_5710_FW_ENGINEERING_VERSION << 24);
+ BNX2X_PCI_FREE(bp->fw_stats, bp->fw_stats_mapping,
+ bp->fw_stats_data_sz + bp->fw_stats_req_sz);
+ return;
+}
+
+static int bnx2x_alloc_fw_stats_mem(struct bnx2x *bp)
+{
+ int num_groups;
+ int is_fcoe_stats = NO_FCOE(bp) ? 0 : 1;
+
+ /* number of queues for statistics is number of eth queues + FCoE */
+ u8 num_queue_stats = BNX2X_NUM_ETH_QUEUES(bp) + is_fcoe_stats;
+
+ /* Total number of FW statistics requests =
+ * 1 for port stats + 1 for PF stats + potential 2 for FCoE (fcoe proper
+ * and fcoe l2 queue) stats + num of queues (which includes another 1
+ * for fcoe l2 queue if applicable)
+ */
+ bp->fw_stats_num = 2 + is_fcoe_stats + num_queue_stats;
+ /* Request is built from stats_query_header and an array of
+ * stats_query_cmd_group each of which contains
+ * STATS_QUERY_CMD_COUNT rules. The real number or requests is
+ * configured in the stats_query_header.
+ */
+ num_groups =
+ (((bp->fw_stats_num) / STATS_QUERY_CMD_COUNT) +
+ (((bp->fw_stats_num) % STATS_QUERY_CMD_COUNT) ?
+ 1 : 0));
+
+ DP(BNX2X_MSG_SP, "stats fw_stats_num %d, num_groups %d",
+ bp->fw_stats_num, num_groups);
+
+ bp->fw_stats_req_sz = sizeof(struct stats_query_header) +
+ num_groups * sizeof(struct stats_query_cmd_group);
+
+ /* Data for statistics requests + stats_counter
+ *
+ * stats_counter holds per-STORM counters that are incremented
+ * when STORM has finished with the current request.
+ * memory for FCoE offloaded statistics are counted anyway,
+ * even if they will not be sent.
+ * VF stats are not accounted for here as the data of VF stats is stored
+ * in memory allocated by the VF, not here.
+ */
+ bp->fw_stats_data_sz = sizeof(struct per_port_stats) +
+ sizeof(struct per_pf_stats) +
+ sizeof(struct fcoe_statistics_params) +
+ sizeof(struct per_queue_stats) * num_queue_stats +
+ sizeof(struct stats_counter);
+
+ BNX2X_PCI_ALLOC(bp->fw_stats, &bp->fw_stats_mapping,
+ bp->fw_stats_data_sz + bp->fw_stats_req_sz);
+
+ /* Set shortcuts */
+ bp->fw_stats_req = (struct bnx2x_fw_stats_req *)bp->fw_stats;
+ bp->fw_stats_req_mapping = bp->fw_stats_mapping;
+
+ bp->fw_stats_data = (struct bnx2x_fw_stats_data *)
+ ((u8 *)bp->fw_stats + bp->fw_stats_req_sz);
+
+ bp->fw_stats_data_mapping = bp->fw_stats_mapping +
+ bp->fw_stats_req_sz;
+
+ DP(BNX2X_MSG_SP, "statistics request base address set to %x %x",
+ U64_HI(bp->fw_stats_req_mapping),
+ U64_LO(bp->fw_stats_req_mapping));
+
+ DP(BNX2X_MSG_SP, "statistics data base address set to %x %x",
+ U64_HI(bp->fw_stats_data_mapping),
+ U64_LO(bp->fw_stats_data_mapping));
+ return 0;
- /* read loaded FW from chip */
- u32 loaded_fw = REG_RD(bp, XSEM_REG_PRAM);
+alloc_mem_err:
+ bnx2x_free_fw_stats_mem(bp);
+ BNX2X_ERR("Can't allocate FW stats memory\n");
+ return -ENOMEM;
+}
- DP(NETIF_MSG_IFUP, "loaded fw %x, my fw %x\n", loaded_fw, my_fw);
+/* send load request to mcp and analyze response */
+static int bnx2x_nic_load_request(struct bnx2x *bp, u32 *load_code)
+{
+ /* init fw_seq */
+ bp->fw_seq =
+ (SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
+ DRV_MSG_SEQ_NUMBER_MASK);
+ BNX2X_DEV_INFO("fw_seq 0x%08x\n", bp->fw_seq);
+
+ /* Get current FW pulse sequence */
+ bp->fw_drv_pulse_wr_seq =
+ (SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_pulse_mb) &
+ DRV_PULSE_SEQ_MASK);
+ BNX2X_DEV_INFO("drv_pulse 0x%x\n", bp->fw_drv_pulse_wr_seq);
+ /* load request */
+ (*load_code) = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_REQ,
+ DRV_MSG_CODE_LOAD_REQ_WITH_LFA);
+
+ /* if mcp fails to respond we must abort */
+ if (!(*load_code)) {
+ BNX2X_ERR("MCP response failure, aborting\n");
+ return -EBUSY;
+ }
- if (loaded_fw != my_fw) {
- if (is_err)
- BNX2X_ERR("bnx2x with FW %x was already loaded, which mismatches my %x FW. aborting\n",
+ /* If mcp refused (e.g. other port is in diagnostic mode) we
+ * must abort
+ */
+ if ((*load_code) == FW_MSG_CODE_DRV_LOAD_REFUSED) {
+ BNX2X_ERR("MCP refused load request, aborting\n");
+ return -EBUSY;
+ }
+ return 0;
+}
+
+/* check whether another PF has already loaded FW to chip. In
+ * virtualized environments a pf from another VM may have already
+ * initialized the device including loading FW
+ */
+int bnx2x_nic_load_analyze_req(struct bnx2x *bp, u32 load_code)
+{
+ /* is another pf loaded on this engine? */
+ if (load_code != FW_MSG_CODE_DRV_LOAD_COMMON_CHIP &&
+ load_code != FW_MSG_CODE_DRV_LOAD_COMMON) {
+
+ /* build my FW version dword */
+ u32 my_fw = (BCM_5710_FW_MAJOR_VERSION) +
+ (BCM_5710_FW_MINOR_VERSION << 8) +
+ (BCM_5710_FW_REVISION_VERSION << 16) +
+ (BCM_5710_FW_ENGINEERING_VERSION << 24);
+
+ /* read loaded FW from chip */
+ u32 loaded_fw = REG_RD(bp, XSEM_REG_PRAM);
+ DP(BNX2X_MSG_SP, "loaded fw %x, my fw %x",
+ loaded_fw, my_fw);
+
+ /* abort nic load if version mismatch */
+ if (my_fw != loaded_fw) {
+ BNX2X_ERR("bnx2x with FW %x was already loaded which mismatches my %x FW. aborting",
loaded_fw, my_fw);
- return false;
+ return -EBUSY;
+ }
+ }
+ return 0;
+}
+
+/* returns the "mcp load_code" according to global load_count array */
+static int bnx2x_nic_load_no_mcp(struct bnx2x *bp, int port)
+{
+ int path = BP_PATH(bp);
+ DP(NETIF_MSG_IFUP, "NO MCP - load counts[%d] %d, %d, %d\n",
+ path, load_count[path][0], load_count[path][1],
+ load_count[path][2]);
+ load_count[path][0]++;
+ load_count[path][1 + port]++;
+ DP(NETIF_MSG_IFUP, "NO MCP - new load counts[%d] %d, %d, %d\n",
+ path, load_count[path][0], load_count[path][1],
+ load_count[path][2]);
+ if (load_count[path][0] == 1)
+ return FW_MSG_CODE_DRV_LOAD_COMMON;
+ else if (load_count[path][1 + port] == 1)
+ return FW_MSG_CODE_DRV_LOAD_PORT;
+ else
+ return FW_MSG_CODE_DRV_LOAD_FUNCTION;
+}
+
+/* mark PMF if applicable */
+static void bnx2x_nic_load_pmf(struct bnx2x *bp, u32 load_code)
+{
+ if ((load_code == FW_MSG_CODE_DRV_LOAD_COMMON) ||
+ (load_code == FW_MSG_CODE_DRV_LOAD_COMMON_CHIP) ||
+ (load_code == FW_MSG_CODE_DRV_LOAD_PORT)) {
+ bp->port.pmf = 1;
+ /* We need the barrier to ensure the ordering between the
+ * writing to bp->port.pmf here and reading it from the
+ * bnx2x_periodic_task().
+ */
+ smp_mb();
+ } else {
+ bp->port.pmf = 0;
+ }
+
+ DP(NETIF_MSG_LINK, "pmf %d\n", bp->port.pmf);
+}
+
+static void bnx2x_nic_load_afex_dcc(struct bnx2x *bp, int load_code)
+{
+ if (((load_code == FW_MSG_CODE_DRV_LOAD_COMMON) ||
+ (load_code == FW_MSG_CODE_DRV_LOAD_COMMON_CHIP)) &&
+ (bp->common.shmem2_base)) {
+ if (SHMEM2_HAS(bp, dcc_support))
+ SHMEM2_WR(bp, dcc_support,
+ (SHMEM_DCC_SUPPORT_DISABLE_ENABLE_PF_TLV |
+ SHMEM_DCC_SUPPORT_BANDWIDTH_ALLOCATION_TLV));
+ if (SHMEM2_HAS(bp, afex_driver_support))
+ SHMEM2_WR(bp, afex_driver_support,
+ SHMEM_AFEX_SUPPORTED_VERSION_ONE);
}
- return true;
+ /* Set AFEX default VLAN tag to an invalid value */
+ bp->afex_def_vlan_tag = -1;
}
/**
@@ -2079,10 +2279,12 @@ int bnx2x_load_cnic(struct bnx2x *bp)
mutex_init(&bp->cnic_mutex);
- rc = bnx2x_alloc_mem_cnic(bp);
- if (rc) {
- BNX2X_ERR("Unable to allocate bp memory for cnic\n");
- LOAD_ERROR_EXIT_CNIC(bp, load_error_cnic0);
+ if (IS_PF(bp)) {
+ rc = bnx2x_alloc_mem_cnic(bp);
+ if (rc) {
+ BNX2X_ERR("Unable to allocate bp memory for cnic\n");
+ LOAD_ERROR_EXIT_CNIC(bp, load_error_cnic0);
+ }
}
rc = bnx2x_alloc_fp_mem_cnic(bp);
@@ -2109,14 +2311,16 @@ int bnx2x_load_cnic(struct bnx2x *bp)
bnx2x_nic_init_cnic(bp);
- /* Enable Timer scan */
- REG_WR(bp, TM_REG_EN_LINEAR0_TIMER + port*4, 1);
-
- for_each_cnic_queue(bp, i) {
- rc = bnx2x_setup_queue(bp, &bp->fp[i], 0);
- if (rc) {
- BNX2X_ERR("Queue setup failed\n");
- LOAD_ERROR_EXIT(bp, load_error_cnic2);
+ if (IS_PF(bp)) {
+ /* Enable Timer scan */
+ REG_WR(bp, TM_REG_EN_LINEAR0_TIMER + port*4, 1);
+
+ for_each_cnic_queue(bp, i) {
+ rc = bnx2x_setup_queue(bp, &bp->fp[i], 0);
+ if (rc) {
+ BNX2X_ERR("Queue setup failed\n");
+ LOAD_ERROR_EXIT(bp, load_error_cnic2);
+ }
}
}
@@ -2162,8 +2366,7 @@ load_error_cnic0:
int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
{
int port = BP_PORT(bp);
- u32 load_code;
- int i, rc;
+ int i, rc = 0, load_code = 0;
DP(NETIF_MSG_IFUP, "Starting NIC load\n");
DP(NETIF_MSG_IFUP,
@@ -2185,8 +2388,9 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
&bp->last_reported_link.link_report_flags);
bnx2x_release_phy_lock(bp);
- /* must be called before memory allocation and HW init */
- bnx2x_ilt_set_info(bp);
+ if (IS_PF(bp))
+ /* must be called before memory allocation and HW init */
+ bnx2x_ilt_set_info(bp);
/*
* Zero fastpath structures preserving invariants like napi, which are
@@ -2205,8 +2409,28 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
/* Set the receive queues buffer size */
bnx2x_set_rx_buf_size(bp);
- if (bnx2x_alloc_mem(bp))
- return -ENOMEM;
+ if (IS_PF(bp)) {
+ rc = bnx2x_alloc_mem(bp);
+ if (rc) {
+ BNX2X_ERR("Unable to allocate bp memory\n");
+ return rc;
+ }
+ }
+
+ /* Allocated memory for FW statistics */
+ if (bnx2x_alloc_fw_stats_mem(bp))
+ LOAD_ERROR_EXIT(bp, load_error0);
+
+ /* fastpath: */
+
+ /* need to be done after alloc mem, since it's self adjusting to amount
+ * of memory available for RSS queues
+ */
+ rc = bnx2x_alloc_fp_mem(bp);
+ if (rc) {
+ BNX2X_ERR("Unable to allocate memory for fps\n");
+ LOAD_ERROR_EXIT(bp, load_error0);
+ }
/* As long as bnx2x_alloc_mem() may possibly update
* bp->num_queues, bnx2x_set_real_num_queues() should always
@@ -2229,98 +2453,49 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
DP(NETIF_MSG_IFUP, "napi added\n");
bnx2x_napi_enable(bp);
- /* set pf load just before approaching the MCP */
- bnx2x_set_pf_load(bp);
+ if (IS_PF(bp)) {
+ /* set pf load just before approaching the MCP */
+ bnx2x_set_pf_load(bp);
- /* Send LOAD_REQUEST command to MCP
- * Returns the type of LOAD command:
- * if it is the first port to be initialized
- * common blocks should be initialized, otherwise - not
- */
- if (!BP_NOMCP(bp)) {
- /* init fw_seq */
- bp->fw_seq =
- (SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
- DRV_MSG_SEQ_NUMBER_MASK);
- BNX2X_DEV_INFO("fw_seq 0x%08x\n", bp->fw_seq);
-
- /* Get current FW pulse sequence */
- bp->fw_drv_pulse_wr_seq =
- (SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_pulse_mb) &
- DRV_PULSE_SEQ_MASK);
- BNX2X_DEV_INFO("drv_pulse 0x%x\n", bp->fw_drv_pulse_wr_seq);
-
- load_code = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_REQ,
- DRV_MSG_CODE_LOAD_REQ_WITH_LFA);
- if (!load_code) {
- BNX2X_ERR("MCP response failure, aborting\n");
- rc = -EBUSY;
- LOAD_ERROR_EXIT(bp, load_error1);
- }
- if (load_code == FW_MSG_CODE_DRV_LOAD_REFUSED) {
- BNX2X_ERR("Driver load refused\n");
- rc = -EBUSY; /* other port in diagnostic mode */
- LOAD_ERROR_EXIT(bp, load_error1);
- }
- if (load_code != FW_MSG_CODE_DRV_LOAD_COMMON_CHIP &&
- load_code != FW_MSG_CODE_DRV_LOAD_COMMON) {
- /* abort nic load if version mismatch */
- if (!bnx2x_test_firmware_version(bp, true)) {
- rc = -EBUSY;
+ /* if mcp exists send load request and analyze response */
+ if (!BP_NOMCP(bp)) {
+
+ /* attempt to load pf */
+ rc = bnx2x_nic_load_request(bp, &load_code);
+ if (rc)
+ LOAD_ERROR_EXIT(bp, load_error1);
+
+ /* what did mcp say? */
+ rc = bnx2x_nic_load_analyze_req(bp, load_code);
+ if (rc) {
+ bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
LOAD_ERROR_EXIT(bp, load_error2);
}
+ } else {
+ load_code = bnx2x_nic_load_no_mcp(bp, port);
}
- } else {
- int path = BP_PATH(bp);
-
- DP(NETIF_MSG_IFUP, "NO MCP - load counts[%d] %d, %d, %d\n",
- path, load_count[path][0], load_count[path][1],
- load_count[path][2]);
- load_count[path][0]++;
- load_count[path][1 + port]++;
- DP(NETIF_MSG_IFUP, "NO MCP - new load counts[%d] %d, %d, %d\n",
- path, load_count[path][0], load_count[path][1],
- load_count[path][2]);
- if (load_count[path][0] == 1)
- load_code = FW_MSG_CODE_DRV_LOAD_COMMON;
- else if (load_count[path][1 + port] == 1)
- load_code = FW_MSG_CODE_DRV_LOAD_PORT;
- else
- load_code = FW_MSG_CODE_DRV_LOAD_FUNCTION;
- }
-
- if ((load_code == FW_MSG_CODE_DRV_LOAD_COMMON) ||
- (load_code == FW_MSG_CODE_DRV_LOAD_COMMON_CHIP) ||
- (load_code == FW_MSG_CODE_DRV_LOAD_PORT)) {
- bp->port.pmf = 1;
- /*
- * We need the barrier to ensure the ordering between the
- * writing to bp->port.pmf here and reading it from the
- * bnx2x_periodic_task().
- */
- smp_mb();
- } else
- bp->port.pmf = 0;
-
- DP(NETIF_MSG_IFUP, "pmf %d\n", bp->port.pmf);
+ /* mark pmf if applicable */
+ bnx2x_nic_load_pmf(bp, load_code);
- /* Init Function state controlling object */
- bnx2x__init_func_obj(bp);
+ /* Init Function state controlling object */
+ bnx2x__init_func_obj(bp);
- /* Initialize HW */
- rc = bnx2x_init_hw(bp, load_code);
- if (rc) {
- BNX2X_ERR("HW init failed, aborting\n");
- bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
- LOAD_ERROR_EXIT(bp, load_error2);
+ /* Initialize HW */
+ rc = bnx2x_init_hw(bp, load_code);
+ if (rc) {
+ BNX2X_ERR("HW init failed, aborting\n");
+ bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
+ LOAD_ERROR_EXIT(bp, load_error2);
+ }
}
/* Connect to IRQs */
rc = bnx2x_setup_irqs(bp);
if (rc) {
- BNX2X_ERR("IRQs setup failed\n");
- bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
+ BNX2X_ERR("setup irqs failed\n");
+ if (IS_PF(bp))
+ bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
LOAD_ERROR_EXIT(bp, load_error2);
}
@@ -2328,79 +2503,75 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
bnx2x_nic_init(bp, load_code);
/* Init per-function objects */
- bnx2x_init_bp_objs(bp);
-
- if (((load_code == FW_MSG_CODE_DRV_LOAD_COMMON) ||
- (load_code == FW_MSG_CODE_DRV_LOAD_COMMON_CHIP)) &&
- (bp->common.shmem2_base)) {
- if (SHMEM2_HAS(bp, dcc_support))
- SHMEM2_WR(bp, dcc_support,
- (SHMEM_DCC_SUPPORT_DISABLE_ENABLE_PF_TLV |
- SHMEM_DCC_SUPPORT_BANDWIDTH_ALLOCATION_TLV));
- if (SHMEM2_HAS(bp, afex_driver_support))
- SHMEM2_WR(bp, afex_driver_support,
- SHMEM_AFEX_SUPPORTED_VERSION_ONE);
- }
+ if (IS_PF(bp)) {
+ bnx2x_init_bp_objs(bp);
- /* Set AFEX default VLAN tag to an invalid value */
- bp->afex_def_vlan_tag = -1;
- bp->state = BNX2X_STATE_OPENING_WAIT4_PORT;
- rc = bnx2x_func_start(bp);
- if (rc) {
- BNX2X_ERR("Function start failed!\n");
- bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
- LOAD_ERROR_EXIT(bp, load_error3);
- }
+ /* Set AFEX default VLAN tag to an invalid value */
+ bp->afex_def_vlan_tag = -1;
+ bnx2x_nic_load_afex_dcc(bp, load_code);
+ bp->state = BNX2X_STATE_OPENING_WAIT4_PORT;
+ rc = bnx2x_func_start(bp);
+ if (rc) {
+ BNX2X_ERR("Function start failed!\n");
+ bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
- /* Send LOAD_DONE command to MCP */
- if (!BP_NOMCP(bp)) {
- load_code = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
- if (!load_code) {
- BNX2X_ERR("MCP response failure, aborting\n");
- rc = -EBUSY;
LOAD_ERROR_EXIT(bp, load_error3);
}
- }
- rc = bnx2x_setup_leading(bp);
- if (rc) {
- BNX2X_ERR("Setup leading failed!\n");
- LOAD_ERROR_EXIT(bp, load_error3);
- }
+ /* Send LOAD_DONE command to MCP */
+ if (!BP_NOMCP(bp)) {
+ load_code = bnx2x_fw_command(bp,
+ DRV_MSG_CODE_LOAD_DONE, 0);
+ if (!load_code) {
+ BNX2X_ERR("MCP response failure, aborting\n");
+ rc = -EBUSY;
+ LOAD_ERROR_EXIT(bp, load_error3);
+ }
+ }
- for_each_nondefault_eth_queue(bp, i) {
- rc = bnx2x_setup_queue(bp, &bp->fp[i], 0);
+ rc = bnx2x_setup_leading(bp);
if (rc) {
- BNX2X_ERR("Queue setup failed\n");
+ BNX2X_ERR("Setup leading failed!\n");
+ LOAD_ERROR_EXIT(bp, load_error3);
+ }
+
+ for_each_nondefault_eth_queue(bp, i) {
+ rc = bnx2x_setup_queue(bp, &bp->fp[i], 0);
+ if (rc) {
+ BNX2X_ERR("Queue setup failed\n");
+ LOAD_ERROR_EXIT(bp, load_error3);
+ }
+ }
+ rc = bnx2x_init_rss_pf(bp);
+ if (rc) {
+ BNX2X_ERR("PF RSS init failed\n");
LOAD_ERROR_EXIT(bp, load_error3);
}
- }
- rc = bnx2x_init_rss_pf(bp);
- if (rc) {
- BNX2X_ERR("PF RSS init failed\n");
- LOAD_ERROR_EXIT(bp, load_error3);
}
/* Now when Clients are configured we are ready to work */
bp->state = BNX2X_STATE_OPEN;
/* Configure a ucast MAC */
- rc = bnx2x_set_eth_mac(bp, true);
+ if (IS_PF(bp))
+ rc = bnx2x_set_eth_mac(bp, true);
if (rc) {
BNX2X_ERR("Setting Ethernet MAC failed\n");
LOAD_ERROR_EXIT(bp, load_error3);
}
- if (bp->pending_max) {
+ if (IS_PF(bp) && bp->pending_max) {
bnx2x_update_max_mf_config(bp, bp->pending_max);
bp->pending_max = 0;
}
- if (bp->port.pmf)
- bnx2x_initial_phy_init(bp, load_mode);
- bp->link_params.feature_config_flags &= ~FEATURE_CONFIG_BOOT_FROM_SAN;
+ if (bp->port.pmf) {
+ rc = bnx2x_initial_phy_init(bp, load_mode);
+ if (rc)
+ LOAD_ERROR_EXIT(bp, load_error3);
+ }
/* Start fast path */
@@ -2441,8 +2612,8 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
if (CNIC_ENABLED(bp))
bnx2x_load_cnic(bp);
- /* mark driver is loaded in shmem2 */
- if (SHMEM2_HAS(bp, drv_capabilities_flag)) {
+ if (IS_PF(bp) && SHMEM2_HAS(bp, drv_capabilities_flag)) {
+ /* mark driver is loaded in shmem2 */
u32 val;
val = SHMEM2_RD(bp, drv_capabilities_flag[BP_FW_MB_IDX(bp)]);
SHMEM2_WR(bp, drv_capabilities_flag[BP_FW_MB_IDX(bp)],
@@ -2451,7 +2622,7 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
}
/* Wait for all pending SP commands to complete */
- if (!bnx2x_wait_sp_comp(bp, ~0x0UL)) {
+ if (IS_PF(bp) && !bnx2x_wait_sp_comp(bp, ~0x0UL)) {
BNX2X_ERR("Timeout waiting for SP elements to complete\n");
bnx2x_nic_unload(bp, UNLOAD_CLOSE, false);
return -EBUSY;
@@ -2467,10 +2638,12 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
#ifndef BNX2X_STOP_ON_ERROR
load_error3:
- bnx2x_int_disable_sync(bp, 1);
+ if (IS_PF(bp)) {
+ bnx2x_int_disable_sync(bp, 1);
- /* Clean queueable objects */
- bnx2x_squeeze_objects(bp);
+ /* Clean queueable objects */
+ bnx2x_squeeze_objects(bp);
+ }
/* Free SKBs, SGEs, TPA pool and driver internals */
bnx2x_free_skbs(bp);
@@ -2480,7 +2653,7 @@ load_error3:
/* Release IRQs */
bnx2x_free_irq(bp);
load_error2:
- if (!BP_NOMCP(bp)) {
+ if (IS_PF(bp) && !BP_NOMCP(bp)) {
bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_REQ_WOL_MCP, 0);
bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_DONE, 0);
}
@@ -2488,15 +2661,34 @@ load_error2:
bp->port.pmf = 0;
load_error1:
bnx2x_napi_disable(bp);
+
/* clear pf_load status, as it was already set */
- bnx2x_clear_pf_load(bp);
+ if (IS_PF(bp))
+ bnx2x_clear_pf_load(bp);
load_error0:
+ bnx2x_free_fp_mem(bp);
+ bnx2x_free_fw_stats_mem(bp);
bnx2x_free_mem(bp);
return rc;
#endif /* ! BNX2X_STOP_ON_ERROR */
}
+static int bnx2x_drain_tx_queues(struct bnx2x *bp)
+{
+ u8 rc = 0, cos, i;
+
+ /* Wait until tx fastpath tasks complete */
+ for_each_tx_queue(bp, i) {
+ struct bnx2x_fastpath *fp = &bp->fp[i];
+ for_each_cos_in_tx_queue(fp, cos)
+ rc = bnx2x_clean_tx_queue(bp, fp->txdata_ptr[cos]);
+ if (rc)
+ return rc;
+ }
+ return 0;
+}
+
/* must be called with rtnl_lock */
int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
{
@@ -2506,15 +2698,16 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
DP(NETIF_MSG_IFUP, "Starting NIC unload\n");
/* mark driver is unloaded in shmem2 */
- if (SHMEM2_HAS(bp, drv_capabilities_flag)) {
+ if (IS_PF(bp) && SHMEM2_HAS(bp, drv_capabilities_flag)) {
u32 val;
val = SHMEM2_RD(bp, drv_capabilities_flag[BP_FW_MB_IDX(bp)]);
SHMEM2_WR(bp, drv_capabilities_flag[BP_FW_MB_IDX(bp)],
val & ~DRV_FLAGS_CAPABILITIES_LOADED_L2);
}
- if ((bp->state == BNX2X_STATE_CLOSED) ||
- (bp->state == BNX2X_STATE_ERROR)) {
+ if (IS_PF(bp) &&
+ (bp->state == BNX2X_STATE_CLOSED ||
+ bp->state == BNX2X_STATE_ERROR)) {
/* We can get here if the driver has been unloaded
* during parity error recovery and is either waiting for a
* leader to complete or for other functions to unload and
@@ -2551,13 +2744,19 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
del_timer_sync(&bp->timer);
- /* Set ALWAYS_ALIVE bit in shmem */
- bp->fw_drv_pulse_wr_seq |= DRV_PULSE_ALWAYS_ALIVE;
+ if (IS_PF(bp)) {
+
+ /* Set ALWAYS_ALIVE bit in shmem */
+ bp->fw_drv_pulse_wr_seq |= DRV_PULSE_ALWAYS_ALIVE;
- bnx2x_drv_pulse(bp);
+ bnx2x_drv_pulse(bp);
- bnx2x_stats_handle(bp, STATS_EVENT_STOP);
- bnx2x_save_statistics(bp);
+ bnx2x_stats_handle(bp, STATS_EVENT_STOP);
+ bnx2x_save_statistics(bp);
+ }
+
+ /* wait till consumers catch up with producers in all queues */
+ bnx2x_drain_tx_queues(bp);
/* Cleanup the chip if needed */
if (unload_mode != UNLOAD_RECOVERY)
@@ -2593,7 +2792,8 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
* At this stage no more interrupts will arrive so we may safly clean
* the queueable objects here in case they failed to get cleaned so far.
*/
- bnx2x_squeeze_objects(bp);
+ if (IS_PF(bp))
+ bnx2x_squeeze_objects(bp);
/* There should be no more pending SP commands at this stage */
bp->sp_state = 0;
@@ -2607,19 +2807,22 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
for_each_rx_queue(bp, i)
bnx2x_free_rx_sge_range(bp, bp->fp + i, NUM_RX_SGE);
- if (CNIC_LOADED(bp)) {
+ bnx2x_free_fp_mem(bp);
+ if (CNIC_LOADED(bp))
bnx2x_free_fp_mem_cnic(bp);
- bnx2x_free_mem_cnic(bp);
- }
- bnx2x_free_mem(bp);
+ if (IS_PF(bp)) {
+ bnx2x_free_mem(bp);
+ if (CNIC_LOADED(bp))
+ bnx2x_free_mem_cnic(bp);
+ }
bp->state = BNX2X_STATE_CLOSED;
bp->cnic_loaded = false;
/* Check if there are pending parity attentions. If there are - set
* RECOVERY_IN_PROGRESS.
*/
- if (bnx2x_chk_parity_attn(bp, &global, false)) {
+ if (IS_PF(bp) && bnx2x_chk_parity_attn(bp, &global, false)) {
bnx2x_set_reset_in_progress(bp);
/* Set RESET_IS_GLOBAL if needed */
@@ -2631,7 +2834,9 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
/* The last driver must disable a "close the gate" if there is no
* parity attention or "process kill" pending.
*/
- if (!bnx2x_clear_pf_load(bp) && bnx2x_reset_is_done(bp, BP_PATH(bp)))
+ if (IS_PF(bp) &&
+ !bnx2x_clear_pf_load(bp) &&
+ bnx2x_reset_is_done(bp, BP_PATH(bp)))
bnx2x_disable_close_the_gate(bp);
DP(NETIF_MSG_IFUP, "Ending NIC unload\n");
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
index bca371e..91e432d 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
@@ -1128,11 +1128,18 @@ static inline u8 bnx2x_fp_qzone_id(struct bnx2x_fastpath *fp)
static inline u32 bnx2x_rx_ustorm_prods_offset(struct bnx2x_fastpath *fp)
{
struct bnx2x *bp = fp->bp;
-
- if (!CHIP_IS_E1x(bp))
- return USTORM_RX_PRODS_E2_OFFSET(fp->cl_qzone_id);
+ u32 offset = BAR_USTRORM_INTMEM;
+
+ if (IS_VF(bp))
+ return PXP_VF_ADDR_USDM_QUEUES_START +
+ bp->acquire_resp.resc.hw_qid[fp->index] *
+ sizeof(struct ustorm_queue_zone_data);
+ else if (!CHIP_IS_E1x(bp))
+ offset += USTORM_RX_PRODS_E2_OFFSET(fp->cl_qzone_id);
else
- return USTORM_RX_PRODS_E1X_OFFSET(BP_PORT(bp), fp->cl_id);
+ offset += USTORM_RX_PRODS_E1X_OFFSET(BP_PORT(bp), fp->cl_id);
+
+ return offset;
}
static inline void bnx2x_init_txdata(struct bnx2x *bp,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
index 277f17e..b7c82f9 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
@@ -890,7 +890,7 @@ static void bnx2x_set_msglevel(struct net_device *dev, u32 level)
if (capable(CAP_NET_ADMIN)) {
/* dump MCP trace */
- if (level & BNX2X_MSG_MCP)
+ if (IS_PF(bp) && (level & BNX2X_MSG_MCP))
bnx2x_fw_dump_lvl(bp, KERN_INFO);
bp->msg_enable = level;
}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 2a8c4c2..158dcec 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -2460,17 +2460,49 @@ void bnx2x__link_status_update(struct bnx2x *bp)
return;
/* read updated dcb configuration */
- bnx2x_dcbx_pmf_update(bp);
-
- bnx2x_link_status_update(&bp->link_params, &bp->link_vars);
+ if (IS_PF(bp)) {
+ bnx2x_dcbx_pmf_update(bp);
+ bnx2x_link_status_update(&bp->link_params, &bp->link_vars);
+ if (bp->link_vars.link_up)
+ bnx2x_stats_handle(bp, STATS_EVENT_LINK_UP);
+ else
+ bnx2x_stats_handle(bp, STATS_EVENT_STOP);
+ /* indicate link status */
+ bnx2x_link_report(bp);
- if (bp->link_vars.link_up)
+ } else { /* VF */
+ bp->port.supported[0] |= (SUPPORTED_10baseT_Half |
+ SUPPORTED_10baseT_Full |
+ SUPPORTED_100baseT_Half |
+ SUPPORTED_100baseT_Full |
+ SUPPORTED_1000baseT_Full |
+ SUPPORTED_2500baseX_Full |
+ SUPPORTED_10000baseT_Full |
+ SUPPORTED_TP |
+ SUPPORTED_FIBRE |
+ SUPPORTED_Autoneg |
+ SUPPORTED_Pause |
+ SUPPORTED_Asym_Pause);
+ bp->port.advertising[0] = bp->port.supported[0];
+
+ bp->link_params.bp = bp;
+ bp->link_params.port = BP_PORT(bp);
+ bp->link_params.req_duplex[0] = DUPLEX_FULL;
+ bp->link_params.req_flow_ctrl[0] = BNX2X_FLOW_CTRL_NONE;
+ bp->link_params.req_line_speed[0] = SPEED_10000;
+ bp->link_params.speed_cap_mask[0] = 0x7f0000;
+ bp->link_params.switch_cfg = SWITCH_CFG_10G;
+ bp->link_vars.mac_type = MAC_TYPE_BMAC;
+ bp->link_vars.line_speed = SPEED_10000;
+ bp->link_vars.link_status =
+ (LINK_STATUS_LINK_UP |
+ LINK_STATUS_SPEED_AND_DUPLEX_10GTFD);
+ bp->link_vars.link_up = 1;
+ bp->link_vars.duplex = DUPLEX_FULL;
+ bp->link_vars.flow_ctrl = BNX2X_FLOW_CTRL_NONE;
+ __bnx2x_link_report(bp);
bnx2x_stats_handle(bp, STATS_EVENT_LINK_UP);
- else
- bnx2x_stats_handle(bp, STATS_EVENT_STOP);
-
- /* indicate link status */
- bnx2x_link_report(bp);
+ }
}
static int bnx2x_afex_func_update(struct bnx2x *bp, u16 vifid,
@@ -5700,6 +5732,15 @@ static void bnx2x_init_eth_fp(struct bnx2x *bp, int fp_idx)
cids[cos] = fp->txdata_ptr[cos]->cid;
}
+ /* nothing more for vf to do here */
+ if (IS_VF(bp))
+ return;
+
+ bnx2x_init_sb(bp, fp->status_blk_mapping, BNX2X_VF_ID_INVALID, false,
+ fp->fw_sb_id, fp->igu_sb_id);
+
+ bnx2x_update_fpsb_idx(fp);
+
bnx2x_init_queue_obj(bp, &bnx2x_sp_obj(bp, fp).q_obj, fp->cl_id, cids,
fp->max_cos, BP_FUNC(bp), bnx2x_sp(bp, q_rdata),
bnx2x_sp_mapping(bp, q_rdata), q_type);
@@ -5709,13 +5750,10 @@ static void bnx2x_init_eth_fp(struct bnx2x *bp, int fp_idx)
*/
bnx2x_init_vlan_mac_fp_objs(fp, BNX2X_OBJ_TYPE_RX_TX);
- DP(NETIF_MSG_IFUP, "queue[%d]: bnx2x_init_sb(%p,%p) cl_id %d fw_sb %d igu_sb %d\n",
- fp_idx, bp, fp->status_blk.e2_sb, fp->cl_id, fp->fw_sb_id,
- fp->igu_sb_id);
- bnx2x_init_sb(bp, fp->status_blk_mapping, BNX2X_VF_ID_INVALID, false,
- fp->fw_sb_id, fp->igu_sb_id);
-
- bnx2x_update_fpsb_idx(fp);
+ DP(NETIF_MSG_IFUP,
+ "queue[%d]: bnx2x_init_sb(%p,%p) cl_id %d fw_sb %d igu_sb %d\n",
+ fp_idx, bp, fp->status_blk.e2_sb, fp->cl_id, fp->fw_sb_id,
+ fp->igu_sb_id);
}
static void bnx2x_init_tx_ring_one(struct bnx2x_fp_txdata *txdata)
@@ -5752,6 +5790,7 @@ static void bnx2x_init_tx_rings_cnic(struct bnx2x *bp)
for_each_tx_queue_cnic(bp, i)
bnx2x_init_tx_ring_one(bp->fp[i].txdata_ptr[0]);
}
+
static void bnx2x_init_tx_rings(struct bnx2x *bp)
{
int i;
@@ -5787,17 +5826,22 @@ void bnx2x_nic_init(struct bnx2x *bp, u32 load_code)
for_each_eth_queue(bp, i)
bnx2x_init_eth_fp(bp, i);
+
+ /* ensure status block indices were read */
+ rmb();
+ bnx2x_init_rx_rings(bp);
+ bnx2x_init_tx_rings(bp);
+
+ if (IS_VF(bp))
+ return;
+
/* Initialize MOD_ABS interrupts */
bnx2x_init_mod_abs_int(bp, &bp->link_vars, bp->common.chip_id,
bp->common.shmem_base, bp->common.shmem2_base,
BP_PORT(bp));
- /* ensure status block indices were read */
- rmb();
bnx2x_init_def_sb(bp);
bnx2x_update_dsb_idx(bp);
- bnx2x_init_rx_rings(bp);
- bnx2x_init_tx_rings(bp);
bnx2x_init_sp_ring(bp);
bnx2x_init_eq_ring(bp);
bnx2x_init_internal(bp, load_code);
@@ -9656,7 +9700,7 @@ static int bnx2x_prev_unload_uncommon(struct bnx2x *bp)
* the one required, then FLR will be sufficient to clean any residue
* left by previous driver
*/
- rc = bnx2x_test_firmware_version(bp, false);
+ rc = bnx2x_nic_load_analyze_req(bp, FW_MSG_CODE_DRV_LOAD_FUNCTION);
if (!rc) {
/* fw version is good */
@@ -11236,17 +11280,21 @@ static int bnx2x_open(struct net_device *dev)
bnx2x_set_power_state(bp, PCI_D0);
- other_load_status = bnx2x_get_load_status(bp, other_engine);
- load_status = bnx2x_get_load_status(bp, BP_PATH(bp));
+ if (IS_PF(bp)) {
+ other_load_status = bnx2x_get_load_status(bp, other_engine);
+ load_status = bnx2x_get_load_status(bp, BP_PATH(bp));
+ }
/*
* If parity had happen during the unload, then attentions
* and/or RECOVERY_IN_PROGRES may still be set. In this case we
* want the first function loaded on the current engine to
* complete the recovery.
+ * Parity recovery is only relevant for PF driver.
*/
- if (!bnx2x_reset_is_done(bp, BP_PATH(bp)) ||
- bnx2x_chk_parity_attn(bp, &global, true))
+ if (IS_PF(bp) &&
+ (!bnx2x_reset_is_done(bp, BP_PATH(bp)) ||
+ bnx2x_chk_parity_attn(bp, &global, true)))
do {
/*
* If there are attentions and they are in a global
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
index c302de4..3f01526 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
@@ -6559,6 +6559,12 @@
#define PXP_VF_ADDR_IGU_END\
((PXP_VF_ADDR_IGU_START) + (PXP_VF_ADDR_IGU_SIZE) - 1)
+#define PXP_VF_ADDR_USDM_QUEUES_START 0x3000
+#define PXP_VF_ADDR_USDM_QUEUES_SIZE\
+ (PXP_VF_ADRR_NUM_QUEUES * PXP_ADDR_QUEUE_SIZE)
+#define PXP_VF_ADDR_USDM_QUEUES_END\
+ ((PXP_VF_ADDR_USDM_QUEUES_START) + (PXP_VF_ADDR_USDM_QUEUES_SIZE) - 1)
+
#define PXP_VF_ADDR_CSDM_GLOBAL_START 0x7600
#define PXP_VF_ADDR_CSDM_GLOBAL_SIZE (PXP_ADDR_REG_SIZE)
#define PXP_VF_ADDR_CSDM_GLOBAL_END\
--
1.7.9.GIT
^ permalink raw reply related
* [PATCH net-next v3 02/22] bnx2x: VF <-> PF channel 'acquire' at vf probe
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
Add the 'acquire' request to VF <-> PF channel and use it at
VF probe. In the acquire request the VF driver lists the resources
it would like to have. In the response the PF either ratifies the
request, or denies it and supplies the maximum values supported.
The VF may then attempt another acquire request.
This patch adds the bnx2x_vfpf.c file which contains the
implementation of the VF to PF hardware channel.
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/Makefile | 2 +-
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 13 ++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 186 +++++++++++++++++++++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h | 6 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h | 5 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c | 79 +++++++++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h | 117 +++++++++++++
7 files changed, 407 insertions(+), 1 deletions(-)
create mode 100644 drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c
diff --git a/drivers/net/ethernet/broadcom/bnx2x/Makefile b/drivers/net/ethernet/broadcom/bnx2x/Makefile
index 48fbdd4..d862ea6 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/Makefile
+++ b/drivers/net/ethernet/broadcom/bnx2x/Makefile
@@ -4,4 +4,4 @@
obj-$(CONFIG_BNX2X) += bnx2x.o
-bnx2x-objs := bnx2x_main.o bnx2x_link.o bnx2x_cmn.o bnx2x_ethtool.o bnx2x_stats.o bnx2x_dcb.o bnx2x_sp.o
+bnx2x-objs := bnx2x_main.o bnx2x_link.o bnx2x_cmn.o bnx2x_ethtool.o bnx2x_stats.o bnx2x_dcb.o bnx2x_sp.o bnx2x_vfpf.o
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index b2f7425..a201ab2 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -48,6 +48,7 @@
#include "bnx2x_sp.h"
#include "bnx2x_dcb.h"
#include "bnx2x_stats.h"
+#include "bnx2x_vfpf.h"
enum bnx2x_int_mode {
BNX2X_INT_MODE_MSIX,
@@ -1244,6 +1245,9 @@ struct bnx2x {
struct bnx2x_vf_mbx_msg *vf2pf_mbox;
dma_addr_t vf2pf_mbox_mapping;
+ /* we set aside a copy of the acquire response */
+ struct pfvf_acquire_resp_tlv acquire_resp;
+
struct net_device *dev;
struct pci_dev *pdev;
@@ -2207,6 +2211,15 @@ static inline u32 reg_poll(struct bnx2x *bp, u32 reg, u32 expected, int ms,
#define BNX2X_VPD_LEN 128
#define VENDOR_ID_LEN 4
+#define VF_ACQUIRE_THRESH 3
+#define VF_ACQUIRE_MAC_FILTERS 1
+#define VF_ACQUIRE_MC_FILTERS 10
+
+#define GOOD_ME_REG(me_reg) (((me_reg) & ME_REG_VF_VALID) && \
+ (!((me_reg) & ME_REG_VF_ERR)))
+int bnx2x_get_vf_id(struct bnx2x *bp, u32 *vf_id);
+int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping);
+int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count);
/* Congestion management fairness mode */
#define CMNG_FNS_NONE 0
#define CMNG_FNS_MINMAX 1
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index b9bc677..061c25b 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -12235,6 +12235,12 @@ static int bnx2x_init_one(struct pci_dev *pdev,
goto init_one_exit;
}
+ if (IS_VF(bp)) {
+ rc = bnx2x_vfpf_acquire(bp, tx_count, rx_count);
+ if (rc)
+ goto init_one_exit;
+ }
+
/* calc qm_cid_count */
bp->qm_cid_count = bnx2x_set_qm_cid_count(bp);
BNX2X_DEV_INFO("qm_cid_count %d\n", bp->qm_cid_count);
@@ -13158,4 +13164,184 @@ struct cnic_eth_dev *bnx2x_cnic_probe(struct net_device *dev)
return cp;
}
+int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping)
+{
+ struct cstorm_vf_zone_data __iomem *zone_data =
+ REG_ADDR(bp, PXP_VF_ADDR_CSDM_GLOBAL_START);
+ int tout = 600, interval = 100; /* wait for 60 seconds */
+
+ if (*done) {
+ BNX2X_ERR("done was non zero before message to pf was sent");
+ WARN_ON(true);
+ return -EINVAL;
+ }
+
+ /* Write message address */
+ writel(U64_LO(msg_mapping),
+ &zone_data->non_trigger.vf_pf_channel.msg_addr_lo);
+ writel(U64_HI(msg_mapping),
+ &zone_data->non_trigger.vf_pf_channel.msg_addr_hi);
+
+ /* make sure the address is written before FW accesses it */
+ wmb();
+
+ /* Trigger the PF FW */
+ writeb(1, &zone_data->trigger.vf_pf_channel.addr_valid);
+
+ /* Wait for PF to complete */
+ while ((tout >= 0) && (!*done)) {
+ msleep(interval);
+ tout -= 1;
+
+ /* progress indicator - HV can take its own sweet time in
+ * answering VFs...
+ */
+ DP_CONT(BNX2X_MSG_IOV, ".");
+ }
+
+ if (!*done) {
+ BNX2X_ERR("PF response has timed out\n");
+ return -EAGAIN;
+ }
+ DP(BNX2X_MSG_SP, "Got a response from PF\n");
+ return 0;
+}
+
+int bnx2x_get_vf_id(struct bnx2x *bp, u32 *vf_id)
+{
+ u32 me_reg;
+ int tout = 10, interval = 100; /* Wait for 1 sec */
+
+ do {
+ /* pxp traps vf read of doorbells and returns me reg value */
+ me_reg = readl(bp->doorbells);
+ if (GOOD_ME_REG(me_reg))
+ break;
+
+ msleep(interval);
+
+ BNX2X_ERR("Invalid ME register value: 0x%08x\n. Is pf driver up?",
+ me_reg);
+ } while (tout-- > 0);
+
+ if (!GOOD_ME_REG(me_reg)) {
+ BNX2X_ERR("Invalid ME register value: 0x%08x\n", me_reg);
+ return -EINVAL;
+ }
+
+ BNX2X_ERR("valid ME register value: 0x%08x\n", me_reg);
+
+ *vf_id = (me_reg & ME_REG_VF_NUM_MASK) >> ME_REG_VF_NUM_SHIFT;
+ return 0;
+}
+
+int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count)
+{
+ int rc = 0, attempts = 0;
+ struct vfpf_acquire_tlv *req = &bp->vf2pf_mbox->req.acquire;
+ struct pfvf_acquire_resp_tlv *resp = &bp->vf2pf_mbox->resp.acquire_resp;
+ u32 vf_id;
+ bool resources_acquired = false;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_ACQUIRE, sizeof(*req));
+
+ if (bnx2x_get_vf_id(bp, &vf_id))
+ return -EAGAIN;
+
+ req->vfdev_info.vf_id = vf_id;
+ req->vfdev_info.vf_os = 0;
+
+ req->resc_request.num_rxqs = rx_count;
+ req->resc_request.num_txqs = tx_count;
+ req->resc_request.num_sbs = bp->igu_sb_cnt;
+ req->resc_request.num_mac_filters = VF_ACQUIRE_MAC_FILTERS;
+ req->resc_request.num_mc_filters = VF_ACQUIRE_MC_FILTERS;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ while (!resources_acquired) {
+ DP(BNX2X_MSG_SP, "attempting to acquire resources");
+
+ /* send acquire request */
+ rc = bnx2x_send_msg2pf(bp,
+ &resp->hdr.status,
+ bp->vf2pf_mbox_mapping);
+
+ /* PF timeout */
+ if (rc)
+ return rc;
+
+ /* copy acquire response from buffer to bp */
+ memcpy(&bp->acquire_resp, resp, sizeof(bp->acquire_resp));
+
+ attempts++;
+
+ /* PF agrees to allocate our resources */
+ if (bp->acquire_resp.hdr.status == PFVF_STATUS_SUCCESS) {
+ DP(BNX2X_MSG_SP, "resources acquired");
+ resources_acquired = true;
+
+ /* PF refuses to allocate our resources */
+ } else if (bp->acquire_resp.hdr.status ==
+ PFVF_STATUS_NO_RESOURCE &&
+ attempts < VF_ACQUIRE_THRESH) {
+ DP(BNX2X_MSG_SP,
+ "PF unwilling to fulfill resource request. Try PF recommended amount");
+
+ /* humble our request */
+ req->resc_request.num_txqs =
+ bp->acquire_resp.resc.num_txqs;
+ req->resc_request.num_rxqs =
+ bp->acquire_resp.resc.num_rxqs;
+ req->resc_request.num_sbs =
+ bp->acquire_resp.resc.num_sbs;
+ req->resc_request.num_mac_filters =
+ bp->acquire_resp.resc.num_mac_filters;
+ req->resc_request.num_vlan_filters =
+ bp->acquire_resp.resc.num_vlan_filters;
+ req->resc_request.num_mc_filters =
+ bp->acquire_resp.resc.num_mc_filters;
+
+ /* Clear response buffer */
+ memset(&bp->vf2pf_mbox->resp, 0,
+ sizeof(union pfvf_tlvs));
+
+ /* PF reports error */
+ } else {
+ BNX2X_ERR("Failed to get the requested amount of resources: %d. Breaking...\n",
+ bp->acquire_resp.hdr.status);
+ return -EAGAIN;
+ }
+ }
+
+ /* get HW info */
+ bp->common.chip_id |= (bp->acquire_resp.pfdev_info.chip_num & 0xffff);
+ bp->link_params.chip_id = bp->common.chip_id;
+ bp->db_size = bp->acquire_resp.pfdev_info.db_size;
+ bp->common.int_block = INT_BLOCK_IGU;
+ bp->common.chip_port_mode = CHIP_2_PORT_MODE;
+ bp->igu_dsb_id = -1;
+ bp->mf_ov = 0;
+ bp->mf_mode = 0;
+ bp->common.flash_size = 0;
+ bp->flags |=
+ NO_WOL_FLAG | NO_ISCSI_OOO_FLAG | NO_ISCSI_FLAG | NO_FCOE_FLAG;
+ bp->igu_sb_cnt = 1;
+ bp->igu_base_sb = bp->acquire_resp.resc.hw_sbs[0].hw_sb_id;
+ strlcpy(bp->fw_ver, bp->acquire_resp.pfdev_info.fw_ver,
+ sizeof(bp->fw_ver));
+
+ if (is_valid_ether_addr(bp->acquire_resp.resc.current_mac_addr))
+ memcpy(bp->dev->dev_addr,
+ bp->acquire_resp.resc.current_mac_addr,
+ ETH_ALEN);
+
+ return 0;
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
index 463a984..c302de4 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
@@ -6558,6 +6558,12 @@
#define PXP_VF_ADDR_IGU_SIZE 0x3000
#define PXP_VF_ADDR_IGU_END\
((PXP_VF_ADDR_IGU_START) + (PXP_VF_ADDR_IGU_SIZE) - 1)
+
+#define PXP_VF_ADDR_CSDM_GLOBAL_START 0x7600
+#define PXP_VF_ADDR_CSDM_GLOBAL_SIZE (PXP_ADDR_REG_SIZE)
+#define PXP_VF_ADDR_CSDM_GLOBAL_END\
+ ((PXP_VF_ADDR_CSDM_GLOBAL_START) + (PXP_VF_ADDR_CSDM_GLOBAL_SIZE) - 1)
+
#define PXP_VF_ADDR_DB_START 0x7c00
#define PXP_VF_ADDR_DB_SIZE 0x200
#define PXP_VF_ADDR_DB_END\
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
index 1b14745..6d0df33 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
@@ -24,4 +24,9 @@ struct bnx2x_vf_mbx_msg {
union pfvf_tlvs resp;
};
+void bnx2x_add_tlv(struct bnx2x *bp, void *tlvs_list, u16 offset, u16 type,
+ u16 length);
+void bnx2x_vfpf_prep(struct bnx2x *bp, struct vfpf_first_tlv *first_tlv,
+ u16 type, u16 length);
+void bnx2x_dp_tlv_list(struct bnx2x *bp, void *tlvs_list);
#endif /* bnx2x_sriov.h */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c
new file mode 100644
index 0000000..cc9b83f
--- /dev/null
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c
@@ -0,0 +1,79 @@
+/* bnx2x_vfpf.c: Broadcom Everest network driver.
+ *
+ * Copyright 2009-2012 Broadcom Corporation
+ *
+ * Unless you and Broadcom execute a separate written software license
+ * agreement governing use of this software, this software is licensed to you
+ * under the terms of the GNU General Public License version 2, available
+ * at http://www.gnu.org/licenses/old-licenses/gpl-2.0.html (the "GPL").
+ *
+ * Notwithstanding the above, under no circumstances may you combine this
+ * software in any way with any other Broadcom software provided under a
+ * license other than the GPL, without Broadcom's express prior written
+ * consent.
+ *
+ * Maintained by: Eilon Greenstein <eilong@broadcom.com>
+ * Written by: Shmulik Ravid <shmulikr@broadcom.com>
+ * Ariel Elior <ariele@broadcom.com>
+ */
+
+/* VF MBX (aka vf-pf channel) */
+#include "bnx2x.h"
+#include "bnx2x_sriov.h"
+/* place a given tlv on the tlv buffer at a given offset */
+void bnx2x_add_tlv(struct bnx2x *bp, void *tlvs_list, u16 offset, u16 type,
+ u16 length)
+{
+ struct channel_tlv *tl =
+ (struct channel_tlv *)(tlvs_list + offset);
+ tl->type = type;
+ tl->length = length;
+}
+
+/* Clear the mailbox and init the header of the first tlv */
+void bnx2x_vfpf_prep(struct bnx2x *bp, struct vfpf_first_tlv *first_tlv,
+ u16 type, u16 length)
+{
+ DP(BNX2X_MSG_IOV, "preparing to send %d tlv over vf pf channel",
+ type);
+
+ /* Clear mailbox */
+ memset(bp->vf2pf_mbox, 0, sizeof(struct bnx2x_vf_mbx_msg));
+
+ /* init type and length */
+ bnx2x_add_tlv(bp, &first_tlv->tl, 0, type, length);
+
+ /* init first tlv header */
+ first_tlv->resp_msg_offset = sizeof(bp->vf2pf_mbox->req);
+}
+
+/* list the types and lengths of the tlvs on the buffer */
+void bnx2x_dp_tlv_list(struct bnx2x *bp, void *tlvs_list)
+{
+ int i = 1;
+ struct channel_tlv *tlv = (struct channel_tlv *)tlvs_list;
+
+ while (tlv->type != CHANNEL_TLV_LIST_END) {
+
+ /* output tlv */
+ DP(BNX2X_MSG_IOV, "TLV number %d: type %d, length %d\n", i,
+ tlv->type, tlv->length);
+
+ /* advance to next tlv */
+ tlvs_list += tlv->length;
+
+ /* cast general tlv list pointer to channel tlv header*/
+ tlv = (struct channel_tlv *)tlvs_list;
+
+ i++;
+
+ if (i > MAX_TLVS_IN_LIST) {
+ WARN(true, "corrupt tlvs");
+ return;
+ }
+ }
+
+ /* output last tlv */
+ DP(BNX2X_MSG_IOV, "TLV number %d: type %d, length %d\n", i,
+ tlv->type, tlv->length);
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index bb37675..f201a53 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -18,20 +18,137 @@
#ifndef VF_PF_IF_H
#define VF_PF_IF_H
+/* Common definitions for all HVs */
+struct vf_pf_resc_request {
+ u8 num_rxqs;
+ u8 num_txqs;
+ u8 num_sbs;
+ u8 num_mac_filters;
+ u8 num_vlan_filters;
+ u8 num_mc_filters; /* No limit so superfluous */
+};
+
+struct hw_sb_info {
+ u8 hw_sb_id; /* aka absolute igu id, used to ack the sb */
+ u8 sb_qid; /* used to update DHC for sb */
+};
+
/* HW VF-PF channel definitions
* A.K.A VF-PF mailbox
*/
#define TLV_BUFFER_SIZE 1024
+enum {
+ PFVF_STATUS_WAITING = 0,
+ PFVF_STATUS_SUCCESS,
+ PFVF_STATUS_FAILURE,
+ PFVF_STATUS_NOT_SUPPORTED,
+ PFVF_STATUS_NO_RESOURCE
+};
+
+/* vf pf channel tlvs */
+/* general tlv header (used for both vf->pf request and pf->vf response) */
+struct channel_tlv {
+ u16 type;
+ u16 length;
+};
+
+/* header of first vf->pf tlv carries the offset used to calculate response
+ * buffer address
+ */
+struct vfpf_first_tlv {
+ struct channel_tlv tl;
+ u32 resp_msg_offset;
+};
+
+/* header of pf->vf tlvs, carries the status of handling the request */
+struct pfvf_tlv {
+ struct channel_tlv tl;
+ u8 status;
+ u8 padding[3];
+};
+
+/* used to terminate and pad a tlv list */
+struct channel_list_end_tlv {
+ struct channel_tlv tl;
+ u8 padding[4];
+};
+
+/* Acquire */
+struct vfpf_acquire_tlv {
+ struct vfpf_first_tlv first_tlv;
+
+ struct vf_pf_vfdev_info {
+ /* the following fields are for debug purposes */
+ u8 vf_id; /* ME register value */
+ u8 vf_os; /* e.g. Linux, W2K8 */
+ u8 padding[2];
+ } vfdev_info;
+
+ struct vf_pf_resc_request resc_request;
+
+ aligned_u64 bulletin_addr;
+};
+
+/* acquire response tlv - carries the allocated resources */
+struct pfvf_acquire_resp_tlv {
+ struct pfvf_tlv hdr;
+ struct pf_vf_pfdev_info {
+ u32 chip_num;
+ u32 pf_cap;
+ #define PFVF_CAP_RSS 0x00000001
+ #define PFVF_CAP_DHC 0x00000002
+ #define PFVF_CAP_TPA 0x00000004
+ char fw_ver[32];
+ u16 db_size;
+ u8 indices_per_sb;
+ u8 padding;
+ } pfdev_info;
+ struct pf_vf_resc {
+ /* in case of status NO_RESOURCE in message hdr, pf will fill
+ * this struct with suggested amount of resources for next
+ * acquire request
+ */
+ #define PFVF_MAX_QUEUES_PER_VF 16
+ #define PFVF_MAX_SBS_PER_VF 16
+ struct hw_sb_info hw_sbs[PFVF_MAX_SBS_PER_VF];
+ u8 hw_qid[PFVF_MAX_QUEUES_PER_VF];
+ u8 num_rxqs;
+ u8 num_txqs;
+ u8 num_sbs;
+ u8 num_mac_filters;
+ u8 num_vlan_filters;
+ u8 num_mc_filters;
+ u8 permanent_mac_addr[ETH_ALEN];
+ u8 current_mac_addr[ETH_ALEN];
+ u8 padding[2];
+ } resc;
+};
+
struct tlv_buffer_size {
u8 tlv_buffer[TLV_BUFFER_SIZE];
};
union vfpf_tlvs {
+ struct vfpf_first_tlv first_tlv;
+ struct vfpf_acquire_tlv acquire;
+ struct channel_list_end_tlv list_end;
struct tlv_buffer_size tlv_buf_size;
};
union pfvf_tlvs {
+ struct pfvf_acquire_resp_tlv acquire_resp;
+ struct channel_list_end_tlv list_end;
struct tlv_buffer_size tlv_buf_size;
};
+
+#define MAX_TLVS_IN_LIST 50
+
+enum channel_tlvs {
+ CHANNEL_TLV_NONE, /* ends tlv sequence */
+ CHANNEL_TLV_ACQUIRE,
+ CHANNEL_TLV_LIST_END,
+ CHANNEL_TLV_MAX
+};
+
#endif /* VF_PF_IF_H */
--
1.7.9.GIT
^ permalink raw reply related
* [PATCH net-next v3 03/22] bnx2x: Add to VF <-> PF channel the release request
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
VF driver uses this request when removed. The PF driver
reclaims all resources allocated for that VF at this
time.
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 1 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 45 ++++++++++++++++++++++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h | 16 +++++++-
3 files changed, 61 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index a201ab2..6cafa4e 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -2220,6 +2220,7 @@ static inline u32 reg_poll(struct bnx2x *bp, u32 reg, u32 expected, int ms,
int bnx2x_get_vf_id(struct bnx2x *bp, u32 *vf_id);
int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping);
int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count);
+int bnx2x_vfpf_release(struct bnx2x *bp);
/* Congestion management fairness mode */
#define CMNG_FNS_NONE 0
#define CMNG_FNS_MINMAX 1
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 061c25b..2a8c4c2 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -12364,6 +12364,9 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
/* Make sure RESET task is not scheduled before continuing */
cancel_delayed_work_sync(&bp->sp_rtnl_task);
+ /* send message via vfpf channel to release the resources of this vf */
+ if (IS_VF(bp))
+ bnx2x_vfpf_release(bp);
if (bp->regview)
iounmap(bp->regview);
@@ -13345,3 +13348,45 @@ int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count)
return 0;
}
+
+int bnx2x_vfpf_release(struct bnx2x *bp)
+{
+ struct vfpf_release_tlv *req = &bp->vf2pf_mbox->req.release;
+ struct pfvf_general_resp_tlv *resp = &bp->vf2pf_mbox->resp.general_resp;
+ u32 rc = 0, vf_id;
+
+ /* clear mailbox and prep first tlv */
+ bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_RELEASE, sizeof(*req));
+
+ if (bnx2x_get_vf_id(bp, &vf_id))
+ return -EAGAIN;
+
+ req->vf_id = vf_id;
+
+ /* add list termination tlv */
+ bnx2x_add_tlv(bp, req, req->first_tlv.tl.length, CHANNEL_TLV_LIST_END,
+ sizeof(struct channel_list_end_tlv));
+
+ /* output tlvs list */
+ bnx2x_dp_tlv_list(bp, req);
+
+ /* send release request */
+ rc = bnx2x_send_msg2pf(bp, &resp->hdr.status, bp->vf2pf_mbox_mapping);
+
+ /* PF timeout */
+ if (rc)
+ return rc;
+
+ /* PF released us */
+ if (resp->hdr.status == PFVF_STATUS_SUCCESS) {
+ DP(BNX2X_MSG_SP, "vf released\n");
+
+ /* PF reports error */
+ } else {
+ BNX2X_ERR("PF failed our release request - are we out of sync? response status: %d\n",
+ resp->hdr.status);
+ return -EAGAIN;
+ }
+
+ return 0;
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index f201a53..763d1c2 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -68,6 +68,10 @@ struct pfvf_tlv {
u8 padding[3];
};
+/* response tlv used for most tlvs */
+struct pfvf_general_resp_tlv {
+ struct pfvf_tlv hdr;
+};
/* used to terminate and pad a tlv list */
struct channel_list_end_tlv {
struct channel_tlv tl;
@@ -125,6 +129,13 @@ struct pfvf_acquire_resp_tlv {
} resc;
};
+/* release the VF's acquired resources */
+struct vfpf_release_tlv {
+ struct vfpf_first_tlv first_tlv;
+ u16 vf_id;
+ u8 padding[2];
+};
+
struct tlv_buffer_size {
u8 tlv_buffer[TLV_BUFFER_SIZE];
};
@@ -132,11 +143,13 @@ struct tlv_buffer_size {
union vfpf_tlvs {
struct vfpf_first_tlv first_tlv;
struct vfpf_acquire_tlv acquire;
+ struct vfpf_release_tlv release;
struct channel_list_end_tlv list_end;
struct tlv_buffer_size tlv_buf_size;
};
union pfvf_tlvs {
+ struct pfvf_general_resp_tlv general_resp;
struct pfvf_acquire_resp_tlv acquire_resp;
struct channel_list_end_tlv list_end;
struct tlv_buffer_size tlv_buf_size;
@@ -145,8 +158,9 @@ union pfvf_tlvs {
#define MAX_TLVS_IN_LIST 50
enum channel_tlvs {
- CHANNEL_TLV_NONE, /* ends tlv sequence */
+ CHANNEL_TLV_NONE,
CHANNEL_TLV_ACQUIRE,
+ CHANNEL_TLV_RELEASE,
CHANNEL_TLV_LIST_END,
CHANNEL_TLV_MAX
};
--
1.7.9.GIT
^ permalink raw reply related
* [PATCH v3 net-next 00/22] bnx2x: support SR-IOV
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior
Hi Dave,
Changes for v2:
-Remove redundant empty lines
-Remove redundant 'inline'
Changes for v3:
-Really remove all redundant empty lines
-Remove __dev* attributes from series
This patch series adds support for SR-IOV in the bnx2x driver.
In bnx2x SR-IOV scheme the same bnx2x driver drives both VFs and PFs (single
binary).
The bulk of the communication between the VF drivers and the PF driver is done
via the VF <-> PF channel, a hardware-based communications channel with TLV
messages. The TLVs are designed to support different versions of VF drivers
(from multiple VMs) communicating with the same PF driver.
Patches:
01-03 - Probing and removing a VF driver
Includes sending the 'acquire' and 'release' messages on the VF PF channel.
Here the VF PF channel infrastructure is added, including the allocations of
the mailboxes and the definition of thin API structures (which will fill out
with more content as the series progresses).
04-07 - Loading/Unloading a VF driver
Includes refactoring of the driver load code and differentiating the VF flow
from the PF flow. This also includes requests from the VF for the PF to open
a queue in the HW on its behalf, configure the device with macs/vlans/rxmode
data, etc. Likewise the unload flow has been modified for the PF to undo these
configurations when the VF indicates it is going down.
08 - Modify fastpath flows for VFs
VFs have almost identical behavior in fastpath to PFs. In this patch the
VFs prepare transmit transaction for tx-switching, and the code for acking a
fastpath interrupt has been reorganized to allow a VF or PF to preconfigure the
offset of the interrupt's location in the BAR (as they have BARs with different
mapping) so as not to do so in fastpath.
09-10 - Probe and Load a PF driver with SR-IOV
The PF driver allocates and initializes the VF database to manage and keep track
of its VFs, their resources, queues, etc.
11-19 - The PF side of the VF <-> PF channel requests
Here is the implementation on the PF's side of the requests submitted in patches
01 through 07 by the VF. Patch 14 adds support for statistics collection by the
PF for all of its VFs (stats are DMAEed directly to VM GPA memory).
20 - Support for VF function level reset
When FLR indication is received for VFs, the PF reclaims all the resources
allocated for these VFs (interrupts, queues) and releases allocation it
performed for the FLRed VFs (it does so by consulting the VF database).
21 - Bulletin Board interface
This patch adds the PF <-> VF Bulletin Board interface. This interface is a
simple interface where the PF can be the initiator, and indicate to a VF that
it has a new MAC for it to use. In this interface each post "covers" any
previous posts (hence the name).
22 - Add the VF device ids and enable feature
In this patch we add the VF device ids of the various devices driven by bnx2x.
Here we also add the calls to "pci_enable_sriov" and "pci_disable_sriov".
Important: In this patch series we have laid the ground work for interfacing
with the infrastructure submitted to the PCI tree for dynamically controlling
the number of VFs of a physical device.
Meanwhile, in patch 9 this code is added to bnx2x_init_one():
--snip snip--
rc = bnx2x_iov_init_one(bp, int_mode, 0/*num vfs*/);
--snip snip--
recompiling the bnx2x with a number of vfs which is not 0 will enable SR-IOV
with that number of VFs.
Please consider applying these patches.
^ permalink raw reply
* [PATCH net-next v3 01/22] bnx2x: Support probing and removing of VF device
From: Ariel Elior @ 2012-12-10 15:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Ariel Elior, Eilon Greenstein
In-Reply-To: <1355154406-10855-1-git-send-email-ariele@broadcom.com>
To support probing and removing of a bnx2x virtual function
the following were added:
1. add bnx2x_vfpf.h: defines the VF to PF channel
2. add bnx2x_sriov.h: header for bnx2x SR-IOV functionality
3. enumerate VF hw types (identify VFs)
4. if driving a VF, map VF bar
5. if driving a VF, allocate Vf to PF channel
6. refactor interrupt flows to include VF
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 21 +-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 25 +-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h | 2 +-
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 391 +++++++++++++--------
drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h | 9 +
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h | 27 ++
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h | 37 ++
7 files changed, 354 insertions(+), 158 deletions(-)
create mode 100644 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
create mode 100644 drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 9a3b81e..b2f7425 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -49,6 +49,12 @@
#include "bnx2x_dcb.h"
#include "bnx2x_stats.h"
+enum bnx2x_int_mode {
+ BNX2X_INT_MODE_MSIX,
+ BNX2X_INT_MODE_INTX,
+ BNX2X_INT_MODE_MSI
+};
+
/* error/debug prints */
#define DRV_MODULE_NAME "bnx2x"
@@ -954,6 +960,9 @@ struct bnx2x_port {
extern struct workqueue_struct *bnx2x_wq;
#define BNX2X_MAX_NUM_OF_VFS 64
+#define BNX2X_VF_CID_WND 0
+#define BNX2X_CIDS_PER_VF (1 << BNX2X_VF_CID_WND)
+#define BNX2X_VF_CIDS (BNX2X_MAX_NUM_OF_VFS * BNX2X_CIDS_PER_VF)
#define BNX2X_VF_ID_INVALID 0xFF
/*
@@ -1231,6 +1240,10 @@ struct bnx2x {
(vn) * ((CHIP_IS_E1x(bp) || (CHIP_MODE_IS_4_PORT(bp))) ? 2 : 1))
#define BP_FW_MB_IDX(bp) BP_FW_MB_IDX_VN(bp, BP_VN(bp))
+ /* vf pf channel mailbox contains request and response buffers */
+ struct bnx2x_vf_mbx_msg *vf2pf_mbox;
+ dma_addr_t vf2pf_mbox_mapping;
+
struct net_device *dev;
struct pci_dev *pdev;
@@ -1318,8 +1331,6 @@ struct bnx2x {
#define DISABLE_MSI_FLAG (1 << 7)
#define TPA_ENABLE_FLAG (1 << 8)
#define NO_MCP_FLAG (1 << 9)
-
-#define BP_NOMCP(bp) (bp->flags & NO_MCP_FLAG)
#define GRO_ENABLE_FLAG (1 << 10)
#define MF_FUNC_DIS (1 << 11)
#define OWN_CNIC_IRQ (1 << 12)
@@ -1330,6 +1341,11 @@ struct bnx2x {
#define BC_SUPPORTS_FCOE_FEATURES (1 << 19)
#define USING_SINGLE_MSIX_FLAG (1 << 20)
#define BC_SUPPORTS_DCBX_MSG_NON_PMF (1 << 21)
+#define IS_VF_FLAG (1 << 22)
+
+#define BP_NOMCP(bp) ((bp)->flags & NO_MCP_FLAG)
+#define IS_VF(bp) ((bp)->flags & IS_VF_FLAG)
+#define IS_PF(bp) (!((bp)->flags & IS_VF_FLAG))
#define NO_ISCSI(bp) ((bp)->flags & NO_ISCSI_FLAG)
#define NO_ISCSI_OOO(bp) ((bp)->flags & NO_ISCSI_OOO_FLAG)
@@ -1432,6 +1448,7 @@ struct bnx2x {
u8 igu_sb_cnt;
u8 min_msix_vec_cnt;
+ u32 igu_base_addr;
dma_addr_t def_status_blk_mapping;
struct bnx2x_slowpath *slowpath;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 67baddd..0a493f4 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -1424,12 +1424,15 @@ void bnx2x_free_irq(struct bnx2x *bp)
int bnx2x_enable_msix(struct bnx2x *bp)
{
- int msix_vec = 0, i, rc, req_cnt;
+ int msix_vec = 0, i, rc;
- bp->msix_table[msix_vec].entry = msix_vec;
- BNX2X_DEV_INFO("msix_table[0].entry = %d (slowpath)\n",
- bp->msix_table[0].entry);
- msix_vec++;
+ /* VFs don't have a default status block */
+ if (IS_PF(bp)) {
+ bp->msix_table[msix_vec].entry = msix_vec;
+ BNX2X_DEV_INFO("msix_table[0].entry = %d (slowpath)\n",
+ bp->msix_table[0].entry);
+ msix_vec++;
+ }
/* Cnic requires an msix vector for itself */
if (CNIC_SUPPORT(bp)) {
@@ -1447,9 +1450,10 @@ int bnx2x_enable_msix(struct bnx2x *bp)
msix_vec++;
}
- req_cnt = BNX2X_NUM_ETH_QUEUES(bp) + CNIC_SUPPORT(bp) + 1;
+ DP(BNX2X_MSG_SP, "about to request enable msix with %d vectors",
+ msix_vec);
- rc = pci_enable_msix(bp->pdev, &bp->msix_table[0], req_cnt);
+ rc = pci_enable_msix(bp->pdev, &bp->msix_table[0], msix_vec);
/*
* reconfigure number of tx/rx queues according to available
@@ -1457,7 +1461,7 @@ int bnx2x_enable_msix(struct bnx2x *bp)
*/
if (rc >= BNX2X_MIN_MSIX_VEC_CNT(bp)) {
/* how less vectors we will have? */
- int diff = req_cnt - rc;
+ int diff = msix_vec - rc;
BNX2X_DEV_INFO("Trying to use less MSI-X vectors: %d\n", rc);
@@ -3889,7 +3893,10 @@ int bnx2x_alloc_mem_bp(struct bnx2x *bp)
* The biggest MSI-X table we might need is as a maximum number of fast
* path IGU SBs plus default SB (for PF).
*/
- msix_table_size = bp->igu_sb_cnt + 1;
+ msix_table_size = bp->igu_sb_cnt;
+ if (IS_PF(bp))
+ msix_table_size++;
+ BNX2X_DEV_INFO("msix_table_size %d", msix_table_size);
/* fp array: RSS plus CNIC related L2 queues */
fp_array_size = BNX2X_MAX_RSS_COUNT(bp) + CNIC_SUPPORT(bp);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
index 0991534..bca371e 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
@@ -863,7 +863,7 @@ static inline void bnx2x_del_all_napi(struct bnx2x *bp)
netif_napi_del(&bnx2x_fp(bp, i, napi));
}
-void bnx2x_set_int_mode(struct bnx2x *bp);
+int bnx2x_set_int_mode(struct bnx2x *bp);
static inline void bnx2x_disable_msi(struct bnx2x *bp)
{
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 940ef85..b9bc677 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -59,6 +59,8 @@
#include "bnx2x_init.h"
#include "bnx2x_init_ops.h"
#include "bnx2x_cmn.h"
+#include "bnx2x_vfpf.h"
+#include "bnx2x_sriov.h"
#include "bnx2x_dcb.h"
#include "bnx2x_sp.h"
@@ -133,39 +135,49 @@ enum bnx2x_board_type {
BCM57711E,
BCM57712,
BCM57712_MF,
+ BCM57712_VF,
BCM57800,
BCM57800_MF,
+ BCM57800_VF,
BCM57810,
BCM57810_MF,
- BCM57840_O,
+ BCM57810_VF,
BCM57840_4_10,
BCM57840_2_20,
- BCM57840_MFO,
BCM57840_MF,
+ BCM57840_VF,
BCM57811,
- BCM57811_MF
+ BCM57811_MF,
+ BCM57840_O,
+ BCM57840_MFO,
+ BCM57811_VF
};
/* indexed by board_type, above */
static struct {
char *name;
} board_info[] = {
- { "Broadcom NetXtreme II BCM57710 10 Gigabit PCIe [Everest]" },
- { "Broadcom NetXtreme II BCM57711 10 Gigabit PCIe" },
- { "Broadcom NetXtreme II BCM57711E 10 Gigabit PCIe" },
- { "Broadcom NetXtreme II BCM57712 10 Gigabit Ethernet" },
- { "Broadcom NetXtreme II BCM57712 10 Gigabit Ethernet Multi Function" },
- { "Broadcom NetXtreme II BCM57800 10 Gigabit Ethernet" },
- { "Broadcom NetXtreme II BCM57800 10 Gigabit Ethernet Multi Function" },
- { "Broadcom NetXtreme II BCM57810 10 Gigabit Ethernet" },
- { "Broadcom NetXtreme II BCM57810 10 Gigabit Ethernet Multi Function" },
- { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet" },
- { "Broadcom NetXtreme II BCM57840 10 Gigabit Ethernet" },
- { "Broadcom NetXtreme II BCM57840 20 Gigabit Ethernet" },
- { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet Multi Function"},
- { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet Multi Function"},
- { "Broadcom NetXtreme II BCM57811 10 Gigabit Ethernet"},
- { "Broadcom NetXtreme II BCM57811 10 Gigabit Ethernet Multi Function"},
+ [BCM57710] = { "Broadcom NetXtreme II BCM57710 10 Gigabit PCIe [Everest]" },
+ [BCM57711] = { "Broadcom NetXtreme II BCM57711 10 Gigabit PCIe" },
+ [BCM57711E] = { "Broadcom NetXtreme II BCM57711E 10 Gigabit PCIe" },
+ [BCM57712] = { "Broadcom NetXtreme II BCM57712 10 Gigabit Ethernet" },
+ [BCM57712_MF] = { "Broadcom NetXtreme II BCM57712 10 Gigabit Ethernet Multi Function" },
+ [BCM57712_VF] = { "Broadcom NetXtreme II BCM57712 10 Gigabit Ethernet Virtual Function" },
+ [BCM57800] = { "Broadcom NetXtreme II BCM57800 10 Gigabit Ethernet" },
+ [BCM57800_MF] = { "Broadcom NetXtreme II BCM57800 10 Gigabit Ethernet Multi Function" },
+ [BCM57800_VF] = { "Broadcom NetXtreme II BCM57800 10 Gigabit Ethernet Virtual Function" },
+ [BCM57810] = { "Broadcom NetXtreme II BCM57810 10 Gigabit Ethernet" },
+ [BCM57810_MF] = { "Broadcom NetXtreme II BCM57810 10 Gigabit Ethernet Multi Function" },
+ [BCM57810_VF] = { "Broadcom NetXtreme II BCM57810 10 Gigabit Ethernet Virtual Function" },
+ [BCM57840_4_10] = { "Broadcom NetXtreme II BCM57840 10 Gigabit Ethernet" },
+ [BCM57840_2_20] = { "Broadcom NetXtreme II BCM57840 20 Gigabit Ethernet" },
+ [BCM57840_MF] = { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet Multi Function" },
+ [BCM57840_VF] = { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet Virtual Function" },
+ [BCM57811] = { "Broadcom NetXtreme II BCM57811 10 Gigabit Ethernet" },
+ [BCM57811_MF] = { "Broadcom NetXtreme II BCM57811 10 Gigabit Ethernet Multi Function" },
+ [BCM57840_O] = { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet" },
+ [BCM57840_MFO] = { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet Multi Function" },
+ [BCM57811_VF] = { "Broadcom NetXtreme II BCM57840 10/20 Gigabit Ethernet Virtual Function" }
};
#ifndef PCI_DEVICE_ID_NX2_57710
@@ -7792,41 +7804,49 @@ int bnx2x_setup_leading(struct bnx2x *bp)
*
* In case of MSI-X it will also try to enable MSI-X.
*/
-void bnx2x_set_int_mode(struct bnx2x *bp)
+int bnx2x_set_int_mode(struct bnx2x *bp)
{
+ int rc = 0;
+
+ if (IS_VF(bp) && int_mode != BNX2X_INT_MODE_MSIX)
+ return -EINVAL;
+
switch (int_mode) {
- case INT_MODE_MSI:
+ case BNX2X_INT_MODE_MSIX:
+ /* attempt to enable msix */
+ rc = bnx2x_enable_msix(bp);
+
+ /* msix attained */
+ if (!rc)
+ return 0;
+
+ /* vfs use only msix */
+ if (rc && IS_VF(bp))
+ return rc;
+
+ /* failed to enable multiple MSI-X */
+ BNX2X_DEV_INFO("Failed to enable multiple MSI-X (%d), set number of queues to %d\n",
+ bp->num_queues,
+ 1 + bp->num_cnic_queues);
+
+ /* falling through... */
+ case BNX2X_INT_MODE_MSI:
bnx2x_enable_msi(bp);
+
/* falling through... */
- case INT_MODE_INTx:
+ case BNX2X_INT_MODE_INTX:
bp->num_ethernet_queues = 1;
bp->num_queues = bp->num_ethernet_queues + bp->num_cnic_queues;
BNX2X_DEV_INFO("set number of queues to 1\n");
break;
default:
- /* if we can't use MSI-X we only need one fp,
- * so try to enable MSI-X with the requested number of fp's
- * and fallback to MSI or legacy INTx with one fp
- */
- if (bnx2x_enable_msix(bp) ||
- bp->flags & USING_SINGLE_MSIX_FLAG) {
- /* failed to enable multiple MSI-X */
- BNX2X_DEV_INFO("Failed to enable multiple MSI-X (%d), set number of queues to %d\n",
- bp->num_queues,
- 1 + bp->num_cnic_queues);
-
- bp->num_queues = 1 + bp->num_cnic_queues;
-
- /* Try to enable MSI */
- if (!(bp->flags & USING_SINGLE_MSIX_FLAG) &&
- !(bp->flags & DISABLE_MSI_FLAG))
- bnx2x_enable_msi(bp);
- }
- break;
+ BNX2X_DEV_INFO("unknown value in int_mode module parameter\n");
+ return -EINVAL;
}
+ return 0;
}
-/* must be called prioir to any HW initializations */
+/* must be called prior to any HW initializations */
static inline u16 bnx2x_cid_ilt_lines(struct bnx2x *bp)
{
return L2_ILT_LINES(bp);
@@ -11081,9 +11101,13 @@ static int bnx2x_init_bp(struct bnx2x *bp)
INIT_DELAYED_WORK(&bp->sp_task, bnx2x_sp_task);
INIT_DELAYED_WORK(&bp->sp_rtnl_task, bnx2x_sp_rtnl_task);
INIT_DELAYED_WORK(&bp->period_task, bnx2x_period_task);
- rc = bnx2x_get_hwinfo(bp);
- if (rc)
- return rc;
+ if (IS_PF(bp)) {
+ rc = bnx2x_get_hwinfo(bp);
+ if (rc)
+ return rc;
+ } else {
+ random_ether_addr(bp->dev->dev_addr);
+ }
bnx2x_set_modes_bitmap(bp);
@@ -11096,7 +11120,7 @@ static int bnx2x_init_bp(struct bnx2x *bp)
func = BP_FUNC(bp);
/* need to reset chip if undi was active */
- if (!BP_NOMCP(bp)) {
+ if (IS_PF(bp) && !BP_NOMCP(bp)) {
/* init fw_seq */
bp->fw_seq =
SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
@@ -11133,6 +11157,8 @@ static int bnx2x_init_bp(struct bnx2x *bp)
bp->mrrs = mrrs;
bp->tx_ring_size = IS_MF_FCOE_AFEX(bp) ? 0 : MAX_TX_AVAIL;
+ if (IS_VF(bp))
+ bp->rx_ring_size = MAX_RX_AVAIL;
/* make sure that the numbers are in the right granularity */
bp->tx_ticks = (50 / BNX2X_BTR) * BNX2X_BTR;
@@ -11161,12 +11187,18 @@ static int bnx2x_init_bp(struct bnx2x *bp)
bp->cnic_base_cl_id = FP_SB_MAX_E2;
/* multiple tx priority */
- if (CHIP_IS_E1x(bp))
+ if (IS_VF(bp))
+ bp->max_cos = 1;
+ else if (CHIP_IS_E1x(bp))
bp->max_cos = BNX2X_MULTI_TX_COS_E1X;
- if (CHIP_IS_E2(bp) || CHIP_IS_E3A0(bp))
+ else if (CHIP_IS_E2(bp) || CHIP_IS_E3A0(bp))
bp->max_cos = BNX2X_MULTI_TX_COS_E2_E3A0;
- if (CHIP_IS_E3B0(bp))
+ else if (CHIP_IS_E3B0(bp))
bp->max_cos = BNX2X_MULTI_TX_COS_E3B0;
+ else
+ BNX2X_ERR("unknown chip %x revision %x\n",
+ CHIP_NUM(bp), CHIP_REV(bp));
+ pr_info("set bp->max_cos to %d\n", bp->max_cos);
/* We need at least one default status block for slow-path events,
* second status block for the L2 queue, and a third status block for
@@ -11551,10 +11583,9 @@ static int bnx2x_set_coherency_mask(struct bnx2x *bp)
return 0;
}
-static int bnx2x_init_dev(struct pci_dev *pdev, struct net_device *dev,
- unsigned long board_type)
+static int bnx2x_init_dev(struct bnx2x *bp, struct pci_dev *pdev,
+ struct net_device *dev, unsigned long board_type)
{
- struct bnx2x *bp;
int rc;
u32 pci_cfg_dword;
bool chip_is_e1x = (board_type == BCM57710 ||
@@ -11562,11 +11593,9 @@ static int bnx2x_init_dev(struct pci_dev *pdev, struct net_device *dev,
board_type == BCM57711E);
SET_NETDEV_DEV(dev, &pdev->dev);
- bp = netdev_priv(dev);
bp->dev = dev;
bp->pdev = pdev;
- bp->flags = 0;
rc = pci_enable_device(pdev);
if (rc) {
@@ -11582,9 +11611,8 @@ static int bnx2x_init_dev(struct pci_dev *pdev, struct net_device *dev,
goto err_out_disable;
}
- if (!(pci_resource_flags(pdev, 2) & IORESOURCE_MEM)) {
- dev_err(&bp->pdev->dev, "Cannot find second PCI device"
- " base address, aborting\n");
+ if (IS_PF(bp) && !(pci_resource_flags(pdev, 2) & IORESOURCE_MEM)) {
+ dev_err(&bp->pdev->dev, "Cannot find second PCI device base address, aborting\n");
rc = -ENODEV;
goto err_out_disable;
}
@@ -11609,12 +11637,14 @@ static int bnx2x_init_dev(struct pci_dev *pdev, struct net_device *dev,
pci_save_state(pdev);
}
- bp->pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM);
- if (bp->pm_cap == 0) {
- dev_err(&bp->pdev->dev,
- "Cannot find power management capability, aborting\n");
- rc = -EIO;
- goto err_out_release;
+ if (IS_PF(bp)) {
+ bp->pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM);
+ if (bp->pm_cap == 0) {
+ dev_err(&bp->pdev->dev,
+ "Cannot find power management capability, aborting\n");
+ rc = -EIO;
+ goto err_out_release;
+ }
}
if (!pci_is_pcie(pdev)) {
@@ -11665,25 +11695,28 @@ static int bnx2x_init_dev(struct pci_dev *pdev, struct net_device *dev,
* Clean the following indirect addresses for all functions since it
* is not used by the driver.
*/
- REG_WR(bp, PXP2_REG_PGL_ADDR_88_F0, 0);
- REG_WR(bp, PXP2_REG_PGL_ADDR_8C_F0, 0);
- REG_WR(bp, PXP2_REG_PGL_ADDR_90_F0, 0);
- REG_WR(bp, PXP2_REG_PGL_ADDR_94_F0, 0);
+ if (IS_PF(bp)) {
+ REG_WR(bp, PXP2_REG_PGL_ADDR_88_F0, 0);
+ REG_WR(bp, PXP2_REG_PGL_ADDR_8C_F0, 0);
+ REG_WR(bp, PXP2_REG_PGL_ADDR_90_F0, 0);
+ REG_WR(bp, PXP2_REG_PGL_ADDR_94_F0, 0);
+
+ if (chip_is_e1x) {
+ REG_WR(bp, PXP2_REG_PGL_ADDR_88_F1, 0);
+ REG_WR(bp, PXP2_REG_PGL_ADDR_8C_F1, 0);
+ REG_WR(bp, PXP2_REG_PGL_ADDR_90_F1, 0);
+ REG_WR(bp, PXP2_REG_PGL_ADDR_94_F1, 0);
+ }
- if (chip_is_e1x) {
- REG_WR(bp, PXP2_REG_PGL_ADDR_88_F1, 0);
- REG_WR(bp, PXP2_REG_PGL_ADDR_8C_F1, 0);
- REG_WR(bp, PXP2_REG_PGL_ADDR_90_F1, 0);
- REG_WR(bp, PXP2_REG_PGL_ADDR_94_F1, 0);
+ /* Enable internal target-read (in case we are probed after PF
+ * FLR). Must be done prior to any BAR read access. Only for
+ * 57712 and up
+ */
+ if (!chip_is_e1x)
+ REG_WR(bp,
+ PGLUE_B_REG_INTERNAL_PFID_ENABLE_TARGET_READ, 1);
}
- /*
- * Enable internal target-read (in case we are probed after PF FLR).
- * Must be done prior to any BAR read access. Only for 57712 and up
- */
- if (!chip_is_e1x)
- REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_TARGET_READ, 1);
-
dev->watchdog_timeo = TX_TIMEOUT;
dev->netdev_ops = &bnx2x_netdev_ops;
@@ -11734,7 +11767,8 @@ err_out:
static void bnx2x_get_pcie_width_speed(struct bnx2x *bp, int *width, int *speed)
{
- u32 val = REG_RD(bp, PCICFG_OFFSET + PCICFG_LINK_CONTROL);
+ u32 val = 0;
+ pci_read_config_dword(bp->pdev, PCICFG_LINK_CONTROL, &val);
*width = (val & PCICFG_LINK_WIDTH) >> PCICFG_LINK_WIDTH_SHIFT;
@@ -12012,10 +12046,10 @@ static int bnx2x_set_qm_cid_count(struct bnx2x *bp)
*
*/
static int bnx2x_get_num_non_def_sbs(struct pci_dev *pdev,
- int cnic_cnt)
+ int cnic_cnt, bool is_vf)
{
- int pos;
- u16 control;
+ int pos, index;
+ u16 control = 0;
pos = pci_find_capability(pdev, PCI_CAP_ID_MSIX);
@@ -12023,85 +12057,115 @@ static int bnx2x_get_num_non_def_sbs(struct pci_dev *pdev,
* If MSI-X is not supported - return number of SBs needed to support
* one fast path queue: one FP queue + SB for CNIC
*/
- if (!pos)
+ if (!pos) {
+ pr_info("no msix capability found");
return 1 + cnic_cnt;
+ }
+
+ pr_info("msix capability found");
/*
* The value in the PCI configuration space is the index of the last
* entry, namely one less than the actual size of the table, which is
* exactly what we want to return from this function: number of all SBs
* without the default SB.
+ * For VFs there is no default SB, then we return (index+1).
*/
pci_read_config_word(pdev, pos + PCI_MSI_FLAGS, &control);
- return control & PCI_MSIX_FLAGS_QSIZE;
-}
-struct cnic_eth_dev *bnx2x_cnic_probe(struct net_device *);
+ index = control & PCI_MSIX_FLAGS_QSIZE;
-static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
-{
- struct net_device *dev = NULL;
- struct bnx2x *bp;
- int pcie_width, pcie_speed;
- int rc, max_non_def_sbs;
- int rx_count, tx_count, rss_count, doorbell_size;
- int cnic_cnt;
- /*
- * An estimated maximum supported CoS number according to the chip
- * version.
- * We will try to roughly estimate the maximum number of CoSes this chip
- * may support in order to minimize the memory allocated for Tx
- * netdev_queue's. This number will be accurately calculated during the
- * initialization of bp->max_cos based on the chip versions AND chip
- * revision in the bnx2x_init_bp().
- */
- u8 max_cos_est = 0;
+ return is_vf ? index + 1 : index;
+}
- switch (ent->driver_data) {
+static int set_max_cos_est(int chip_id)
+{
+ switch (chip_id) {
case BCM57710:
case BCM57711:
case BCM57711E:
- max_cos_est = BNX2X_MULTI_TX_COS_E1X;
- break;
-
+ return BNX2X_MULTI_TX_COS_E1X;
case BCM57712:
case BCM57712_MF:
- max_cos_est = BNX2X_MULTI_TX_COS_E2_E3A0;
- break;
-
+ case BCM57712_VF:
+ return BNX2X_MULTI_TX_COS_E2_E3A0;
case BCM57800:
case BCM57800_MF:
+ case BCM57800_VF:
case BCM57810:
case BCM57810_MF:
- case BCM57840_O:
case BCM57840_4_10:
case BCM57840_2_20:
+ case BCM57840_O:
case BCM57840_MFO:
+ case BCM57810_VF:
case BCM57840_MF:
+ case BCM57840_VF:
case BCM57811:
case BCM57811_MF:
- max_cos_est = BNX2X_MULTI_TX_COS_E3B0;
- break;
-
+ case BCM57811_VF:
+ return BNX2X_MULTI_TX_COS_E3B0;
+ return 1;
default:
- pr_err("Unknown board_type (%ld), aborting\n",
- ent->driver_data);
+ pr_err("Unknown board_type (%d), aborting\n", chip_id);
return -ENODEV;
}
+}
- cnic_cnt = 1;
- max_non_def_sbs = bnx2x_get_num_non_def_sbs(pdev, cnic_cnt);
+static int set_is_vf(int chip_id)
+{
+ switch (chip_id) {
+ case BCM57712_VF:
+ case BCM57800_VF:
+ case BCM57810_VF:
+ case BCM57840_VF:
+ case BCM57811_VF:
+ return true;
+ default:
+ return false;
+ }
+}
+
+struct cnic_eth_dev *bnx2x_cnic_probe(struct net_device *dev);
+
+static int bnx2x_init_one(struct pci_dev *pdev,
+ const struct pci_device_id *ent)
+{
+ struct net_device *dev = NULL;
+ struct bnx2x *bp;
+ int pcie_width, pcie_speed;
+ int rc, max_non_def_sbs;
+ int rx_count, tx_count, rss_count, doorbell_size;
+ int max_cos_est;
+ bool is_vf;
+ int cnic_cnt;
+
+ /* An estimated maximum supported CoS number according to the chip
+ * version.
+ * We will try to roughly estimate the maximum number of CoSes this chip
+ * may support in order to minimize the memory allocated for Tx
+ * netdev_queue's. This number will be accurately calculated during the
+ * initialization of bp->max_cos based on the chip versions AND chip
+ * revision in the bnx2x_init_bp().
+ */
+ max_cos_est = set_max_cos_est(ent->driver_data);
+ if (max_cos_est < 0)
+ return max_cos_est;
+ is_vf = set_is_vf(ent->driver_data);
+ cnic_cnt = is_vf ? 0 : 1;
- WARN_ON(!max_non_def_sbs);
+ max_non_def_sbs = bnx2x_get_num_non_def_sbs(pdev, cnic_cnt, is_vf);
/* Maximum number of RSS queues: one IGU SB goes to CNIC */
- rss_count = max_non_def_sbs - cnic_cnt;
+ rss_count = is_vf ? 1 : max_non_def_sbs - cnic_cnt;
+
+ if (rss_count < 1)
+ return -EINVAL;
/* Maximum number of netdev Rx queues: RSS + FCoE L2 */
rx_count = rss_count + cnic_cnt;
- /*
- * Maximum number of netdev Tx queues:
+ /* Maximum number of netdev Tx queues:
* Maximum TSS queues * Maximum supported number of CoS + FCoE L2
*/
tx_count = rss_count * max_cos_est + cnic_cnt;
@@ -12113,22 +12177,28 @@ static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
bp = netdev_priv(dev);
+ bp->flags = 0;
+ if (is_vf)
+ bp->flags |= IS_VF_FLAG;
+
bp->igu_sb_cnt = max_non_def_sbs;
+ bp->igu_base_addr = IS_VF(bp) ? PXP_VF_ADDR_IGU_START : BAR_IGU_INTMEM;
bp->msg_enable = debug;
bp->cnic_support = cnic_cnt;
bp->cnic_probe = bnx2x_cnic_probe;
pci_set_drvdata(pdev, dev);
- rc = bnx2x_init_dev(pdev, dev, ent->driver_data);
+ rc = bnx2x_init_dev(bp, pdev, dev, ent->driver_data);
if (rc < 0) {
free_netdev(dev);
return rc;
}
+ BNX2X_DEV_INFO("This is a %s function\n",
+ IS_PF(bp) ? "physical" : "virtual");
BNX2X_DEV_INFO("Cnic support is %s\n", CNIC_SUPPORT(bp) ? "on" : "off");
- BNX2X_DEV_INFO("max_non_def_sbs %d\n", max_non_def_sbs);
-
+ BNX2X_DEV_INFO("Max num of status blocks %d\n", max_non_def_sbs);
BNX2X_DEV_INFO("Allocated netdev with %d tx and %d rx queues\n",
tx_count, rx_count);
@@ -12136,19 +12206,28 @@ static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
if (rc)
goto init_one_exit;
- /*
- * Map doorbels here as we need the real value of bp->max_cos which
- * is initialized in bnx2x_init_bp().
+ /* Map doorbells here as we need the real value of bp->max_cos which
+ * is initialized in bnx2x_init_bp() to determine the number of
+ * l2 connections.
*/
- doorbell_size = BNX2X_L2_MAX_CID(bp) * (1 << BNX2X_DB_SHIFT);
- if (doorbell_size > pci_resource_len(pdev, 2)) {
- dev_err(&bp->pdev->dev,
- "Cannot map doorbells, bar size too small, aborting\n");
- rc = -ENOMEM;
- goto init_one_exit;
+ if (IS_VF(bp)) {
+ /* vf doorbells are embedded within the regview */
+ bp->doorbells = bp->regview + PXP_VF_ADDR_DB_START;
+
+ /* allocate vf2pf mailbox for vf to pf channel */
+ BNX2X_PCI_ALLOC(bp->vf2pf_mbox, &bp->vf2pf_mbox_mapping,
+ sizeof(struct bnx2x_vf_mbx_msg));
+ } else {
+ doorbell_size = BNX2X_L2_MAX_CID(bp) * (1 << BNX2X_DB_SHIFT);
+ if (doorbell_size > pci_resource_len(pdev, 2)) {
+ dev_err(&bp->pdev->dev,
+ "Cannot map doorbells, bar size too small, aborting\n");
+ rc = -ENOMEM;
+ goto init_one_exit;
+ }
+ bp->doorbells = ioremap_nocache(pci_resource_start(pdev, 2),
+ doorbell_size);
}
- bp->doorbells = ioremap_nocache(pci_resource_start(pdev, 2),
- doorbell_size);
if (!bp->doorbells) {
dev_err(&bp->pdev->dev,
"Cannot map doorbell space, aborting\n");
@@ -12158,6 +12237,7 @@ static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
/* calc qm_cid_count */
bp->qm_cid_count = bnx2x_set_qm_cid_count(bp);
+ BNX2X_DEV_INFO("qm_cid_count %d\n", bp->qm_cid_count);
/* disable FCOE L2 queue for E1x*/
if (CHIP_IS_E1x(bp))
@@ -12179,13 +12259,19 @@ static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
/* Configure interrupt mode: try to enable MSI-X/MSI if
* needed.
*/
- bnx2x_set_int_mode(bp);
+ rc = bnx2x_set_int_mode(bp);
+ if (rc) {
+ dev_err(&pdev->dev, "Cannot set interrupts\n");
+ goto init_one_exit;
+ }
+ /* register the net device */
rc = register_netdev(dev);
if (rc) {
dev_err(&pdev->dev, "Cannot register net device\n");
goto init_one_exit;
}
+ BNX2X_DEV_INFO("device name after netdev register %s\n", dev->name);
if (!NO_FCOE(bp)) {
@@ -12196,6 +12282,8 @@ static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
}
bnx2x_get_pcie_width_speed(bp, &pcie_width, &pcie_speed);
+ BNX2X_DEV_INFO("got pcie width %d and speed %d\n",
+ pcie_width, pcie_speed);
BNX2X_DEV_INFO(
"%s (%c%d) PCI-E x%d %s found at mem %lx, IRQ %d, node addr %pM\n",
@@ -12209,11 +12297,16 @@ static int bnx2x_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
return 0;
+alloc_mem_err:
+ BNX2X_PCI_FREE(bp->vf2pf_mbox, bp->vf2pf_mbox_mapping,
+ sizeof(struct bnx2x_vf_mbx_msg));
+ rc = -ENOMEM;
+
init_one_exit:
if (bp->regview)
iounmap(bp->regview);
- if (bp->doorbells)
+ if (IS_PF(bp) && bp->doorbells)
iounmap(bp->doorbells);
free_netdev(dev);
@@ -12253,13 +12346,15 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
unregister_netdev(dev);
/* Power on: we can't let PCI layer write to us while we are in D3 */
- bnx2x_set_power_state(bp, PCI_D0);
+ if (IS_PF(bp))
+ bnx2x_set_power_state(bp, PCI_D0);
/* Disable MSI/MSI-X */
bnx2x_disable_msi(bp);
/* Power off */
- bnx2x_set_power_state(bp, PCI_D3hot);
+ if (IS_PF(bp))
+ bnx2x_set_power_state(bp, PCI_D3hot);
/* Make sure RESET task is not scheduled before continuing */
cancel_delayed_work_sync(&bp->sp_rtnl_task);
@@ -12267,11 +12362,15 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
if (bp->regview)
iounmap(bp->regview);
- if (bp->doorbells)
- iounmap(bp->doorbells);
-
- bnx2x_release_firmware(bp);
+ /* for vf doorbells are part of the regview and were unmapped along with
+ * it. FW is only loaded by PF.
+ */
+ if (IS_PF(bp)) {
+ if (bp->doorbells)
+ iounmap(bp->doorbells);
+ bnx2x_release_firmware(bp);
+ }
bnx2x_free_mem_bp(bp);
free_netdev(dev);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
index bc2f65b..463a984 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
@@ -6554,6 +6554,15 @@
(7L<<ME_REG_ABS_PF_NUM_SHIFT) /* Absolute PF Num */
+#define PXP_VF_ADDR_IGU_START 0
+#define PXP_VF_ADDR_IGU_SIZE 0x3000
+#define PXP_VF_ADDR_IGU_END\
+ ((PXP_VF_ADDR_IGU_START) + (PXP_VF_ADDR_IGU_SIZE) - 1)
+#define PXP_VF_ADDR_DB_START 0x7c00
+#define PXP_VF_ADDR_DB_SIZE 0x200
+#define PXP_VF_ADDR_DB_END\
+ ((PXP_VF_ADDR_DB_START) + (PXP_VF_ADDR_DB_SIZE) - 1)
+
#define MDIO_REG_BANK_CL73_IEEEB0 0x0
#define MDIO_CL73_IEEEB0_CL73_AN_CONTROL 0x0
#define MDIO_CL73_IEEEB0_CL73_AN_CONTROL_RESTART_AN 0x0200
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
new file mode 100644
index 0000000..1b14745
--- /dev/null
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
@@ -0,0 +1,27 @@
+/* bnx2x_sriov.h: Broadcom Everest network driver.
+ *
+ * Copyright 2009-2012 Broadcom Corporation
+ *
+ * Unless you and Broadcom execute a separate written software license
+ * agreement governing use of this software, this software is licensed to you
+ * under the terms of the GNU General Public License version 2, available
+ * at http://www.gnu.org/licenses/old-licenses/gpl-2.0.html (the "GPL").
+ *
+ * Notwithstanding the above, under no circumstances may you combine this
+ * software in any way with any other Broadcom software provided under a
+ * license other than the GPL, without Broadcom's express prior written
+ * consent.
+ *
+ * Maintained by: Eilon Greenstein <eilong@broadcom.com>
+ * Written by: Shmulik Ravid <shmulikr@broadcom.com>
+ * Ariel Elior <ariele@broadcom.com>
+ */
+#ifndef BNX2X_SRIOV_H
+#define BNX2X_SRIOV_H
+
+struct bnx2x_vf_mbx_msg {
+ union vfpf_tlvs req;
+ union pfvf_tlvs resp;
+};
+
+#endif /* bnx2x_sriov.h */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
new file mode 100644
index 0000000..bb37675
--- /dev/null
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -0,0 +1,37 @@
+/* bnx2x_vfpf.h: Broadcom Everest network driver.
+ *
+ * Copyright (c) 2011-2012 Broadcom Corporation
+ *
+ * Unless you and Broadcom execute a separate written software license
+ * agreement governing use of this software, this software is licensed to you
+ * under the terms of the GNU General Public License version 2, available
+ * at http://www.gnu.org/licenses/old-licenses/gpl-2.0.html (the "GPL").
+ *
+ * Notwithstanding the above, under no circumstances may you combine this
+ * software in any way with any other Broadcom software provided under a
+ * license other than the GPL, without Broadcom's express prior written
+ * consent.
+ *
+ * Maintained by: Eilon Greenstein <eilong@broadcom.com>
+ * Written by: Ariel Elior <ariele@broadcom.com>
+ */
+#ifndef VF_PF_IF_H
+#define VF_PF_IF_H
+
+/* HW VF-PF channel definitions
+ * A.K.A VF-PF mailbox
+ */
+#define TLV_BUFFER_SIZE 1024
+
+struct tlv_buffer_size {
+ u8 tlv_buffer[TLV_BUFFER_SIZE];
+};
+
+union vfpf_tlvs {
+ struct tlv_buffer_size tlv_buf_size;
+};
+
+union pfvf_tlvs {
+ struct tlv_buffer_size tlv_buf_size;
+};
+#endif /* VF_PF_IF_H */
--
1.7.9.GIT
^ permalink raw reply related
* RE: ipgre rss is broken since gro
From: Eric Dumazet @ 2012-12-10 15:17 UTC (permalink / raw)
To: Dmitry Kravkov; +Cc: Eric Dumazet, netdev@vger.kernel.org
In-Reply-To: <504C9EFCA2D0054393414C9CB605C37F1BFC104B@SJEXCHMB06.corp.ad.broadcom.com>
On Mon, 2012-12-10 at 11:32 +0000, Dmitry Kravkov wrote:
> Current bnx2x do not apply RSS for GRE, non GRE RSS is working w/o problem.
What about you post the changes you did ?
^ permalink raw reply
* Re: [PATCH RESEND] net: remove obsolete simple_strto<foo>
From: Neil Horman @ 2012-12-10 15:03 UTC (permalink / raw)
To: Abhijit Pawar
Cc: David S. Miller, Pablo Neira Ayuso, Patrick McHardy,
Alexey Kuznetsov, James Morris, Hideaki YOSHIFUJI,
John W. Linville, Johannes Berg, Cong Wang, Eric Dumazet,
Joe Perches, netdev, linux-kernel, netfilter-devel, netfilter,
coreteam, linux-wireless
In-Reply-To: <1355130748-7828-1-git-send-email-abhi.c.pawar@gmail.com>
On Mon, Dec 10, 2012 at 02:42:28PM +0530, Abhijit Pawar wrote:
> This patch replace the obsolete simple_strto<foo> with kstrto<foo>
>
> Signed-off-by: Abhijit Pawar <abhi.c.pawar@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
^ permalink raw reply
* Re: [PATCH net-next v3] tipc: sk_recv_queue size check only for connectionless sockets
From: Neil Horman @ 2012-12-10 14:51 UTC (permalink / raw)
To: Ying Xue; +Cc: Paul.Gortmaker, jon.maloy, erik.hugne, netdev, tipc-discussion
In-Reply-To: <1355131380-8542-1-git-send-email-ying.xue@windriver.com>
On Mon, Dec 10, 2012 at 05:23:00PM +0800, Ying Xue wrote:
> The sk_receive_queue limit control is currently performed for all
> arriving messages, disregarding socket and message type. But for
> connectionless sockets this check is redundant, since the protocol
> flow already makes queue overflow impossible.
>
> We move the sk_receive_queue limit control so that it's only performed
> for connectionless sockets, i.e. SOCK_RDM and SOCK_DGRAM type sockets.
>
> However, as Neil Horman specified, we cannot simply force the socket
> receive queue limit against connectionless sockets as it may create a
> DoS vulnerability. For example, if a sender floods a receiver with
> messages containing an invalid set of message importance bits or
> CRITICAL importance, we will queue messages indefinitely.
>
> To avoid DoS attack, socket receive queue will be marked as overflow
> if we receive messages with invalid message importances, meanwhile,
> we also set one new threshold for CRITICAL importance messages.
>
> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
> Cc: Neil Horman <nhorman@tuxdriver.com>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> ---
> v3 changes:
> - set new threshold for CRITICAL message
> - defined an importance factor table to avoid multiplication and
> division operations in rx_queue_full().
> - changed return value of rx_queue_full() from integer to boolean.
>
> net/tipc/socket.c | 44 +++++++++++++++++++-------------------------
> 1 files changed, 19 insertions(+), 25 deletions(-)
>
> diff --git a/net/tipc/socket.c b/net/tipc/socket.c
> index 9b4e483..a18a757 100644
> --- a/net/tipc/socket.c
> +++ b/net/tipc/socket.c
> @@ -43,7 +43,7 @@
> #define SS_LISTENING -1 /* socket is listening */
> #define SS_READY -2 /* socket is connectionless */
>
> -#define OVERLOAD_LIMIT_BASE 10000
> +#define OVERLOAD_LIMIT_BASE 5000
> #define CONN_TIMEOUT_DEFAULT 8000 /* default connect timeout = 8s */
>
> struct tipc_sock {
> @@ -73,6 +73,13 @@ static struct proto tipc_proto;
>
> static int sockets_enabled;
>
> +static const u32 msg_importance_factor[] = {
> + OVERLOAD_LIMIT_BASE, /* TIPC_LOW_IMPORTANCE limit */
> + OVERLOAD_LIMIT_BASE * 2, /* TIPC_MEDIUM_IMPORTANCE limit */
> + OVERLOAD_LIMIT_BASE * 100, /* TIPC_HIGH_IMPORTANCE limit */
> + OVERLOAD_LIMIT_BASE * 200 /* TIPC_CRITICAL_IMPORTANCE limit */
> + };
> +
> /*
> * Revised TIPC socket locking policy:
> *
> @@ -1158,28 +1165,17 @@ static void tipc_data_ready(struct sock *sk, int len)
> * rx_queue_full - determine if receive queue can accept another message
> * @msg: message to be added to queue
> * @queue_size: current size of queue
> - * @base: nominal maximum size of queue
> *
> - * Returns 1 if queue is unable to accept message, 0 otherwise
> + * Returns true if queue is unable to accept message, false otherwise
> */
> -static int rx_queue_full(struct tipc_msg *msg, u32 queue_size, u32 base)
> +static bool rx_queue_full(struct tipc_msg *msg, u32 queue_size)
> {
> - u32 threshold;
> u32 imp = msg_importance(msg);
>
> - if (imp == TIPC_LOW_IMPORTANCE)
> - threshold = base;
> - else if (imp == TIPC_MEDIUM_IMPORTANCE)
> - threshold = base * 2;
> - else if (imp == TIPC_HIGH_IMPORTANCE)
> - threshold = base * 100;
> - else
> - return 0;
> + if (unlikely(imp > TIPC_CRITICAL_IMPORTANCE))
> + return true;
>
> - if (msg_connected(msg))
> - threshold *= 4;
> -
> - return queue_size >= threshold;
> + return queue_size >= msg_importance_factor[imp];
> }
>
> /**
> @@ -1275,7 +1271,6 @@ static u32 filter_rcv(struct sock *sk, struct sk_buff *buf)
> {
> struct socket *sock = sk->sk_socket;
> struct tipc_msg *msg = buf_msg(buf);
> - u32 recv_q_len;
> u32 res = TIPC_OK;
>
> /* Reject message if it is wrong sort of message for socket */
> @@ -1285,19 +1280,18 @@ static u32 filter_rcv(struct sock *sk, struct sk_buff *buf)
> if (sock->state == SS_READY) {
> if (msg_connected(msg))
> return TIPC_ERR_NO_PORT;
> + /* Reject SOCK_DGRAM and SOCK_RDM message if there isn't room
> + * to queue it
> + */
> + if (unlikely(rx_queue_full(msg,
> + skb_queue_len(&sk->sk_receive_queue))))
> + return TIPC_ERR_OVERLOAD;
> } else {
> res = filter_connect(tipc_sk(sk), &buf);
> if (res != TIPC_OK || buf == NULL)
> return res;
> }
>
> - /* Reject message if there isn't room to queue it */
> - recv_q_len = skb_queue_len(&sk->sk_receive_queue);
> - if (unlikely(recv_q_len >= (OVERLOAD_LIMIT_BASE / 2))) {
> - if (rx_queue_full(msg, recv_q_len, OVERLOAD_LIMIT_BASE / 2))
> - return TIPC_ERR_OVERLOAD;
> - }
> -
> /* Enqueue message (finally!) */
> TIPC_SKB_CB(buf)->handle = 0;
> __skb_queue_tail(&sk->sk_receive_queue, buf);
> --
> 1.7.1
>
>
That looks more reasonable, thanks.
Acked-by: Neil Horman <nhorman@tuxdriver.com>
^ permalink raw reply
* Re: [PATCH net-next 03/10] tipc: sk_recv_queue size check only for connectionless sockets
From: Neil Horman @ 2012-12-10 14:22 UTC (permalink / raw)
To: Ying Xue; +Cc: Jon Maloy, Paul Gortmaker, David Miller, netdev
In-Reply-To: <50C580E6.7030905@windriver.com>
On Mon, Dec 10, 2012 at 02:27:50PM +0800, Ying Xue wrote:
> Neil Horman wrote:
> >On Fri, Dec 07, 2012 at 05:30:11PM -0500, Jon Maloy wrote:
> >>On 12/07/2012 02:20 PM, Neil Horman wrote:
> >>>On Fri, Dec 07, 2012 at 09:28:11AM -0500, Paul Gortmaker wrote:
> >>>>From: Ying Xue <ying.xue@windriver.com>
> >>>>
> >>>>The sk_receive_queue limit control is currently performed for
> >>>>all arriving messages, disregarding socket and message type.
> >>>>But for connected sockets this check is redundant, since the protocol
> >>>>flow control already makes queue overflow impossible.
> >>>>
> >>>Can you explain where that occurs?
> >>It happens in the functions port_dispatcher_sigh() and
> >>tipc_send(), among other places. Both are to be found in the
> >>file port.c, which was supposed to contain the 'generic' (i.e.,
> >>API independent) part of the send/receive code.
> >>Now that we have only one API left, the socket API, we are
> >>planning to merge the code in socket.c and port.c, and get rid
> >>of some code overhead.
> >>
> >>The flow control in TIPC is message based, where the sender
> >>requires to receive an explicit acknowledge message for each 512
> >>message the receiver reads to user space.
> >>If the sender has more than 1024 messages outstanding without having
> >>received an acknowledge he will be suspended or receive EAGAIN
> >>until he does.
> >>The plan going forward is to replace this mechanism with a more
> >>standard looking byte based flow control, while maintaining
> >>backwards compatibility.
> >>
> >Ok, That makes more sense, thank you. Although I still don't think this is
> >safe (but the problem may not be solely introduced by this patch). Using a
> >global limit that assumes the sender will block when the congestion window is
> >reached just doesn't seem sane to me. It clearly works with the Linux
> >implementation, as it conforms to your expectations, but an alternate
> >implementation could create a DOS situation by simply ignoring the window limit,
> >and continuing to send. I see that we drop frames over the global limit in
> >filter_rcv, but the check in rx_queue_full bumps up that limit based on the
> >value of msg_importance(msg), but that threshold is ignored if the value of
> >msg_importance is invalid. All a sender needs to do is flood a receiver with
> >frames containing an invalid set of message importance bits, and you will queue
> >frames indefinately. In fact that will also happen if you send message of
> >CRITICAL importance as well, so you don't even need to supply an invalid value
> >here.
> >
>
> You are absolutely right. I will correct these drawbacks in next version.
>
> >>>I see where the tipc dispatch function calls
> >>>sk_add_backlog, which checks the per socket recieve queue (regardless of weather
> >>>the receiving socket is connection oriented or connectionless), but if the
> >>>receiver doesn't call receive very often, This just adds a check against your
> >>>global limit, doing nothing for your per-socket limits.
> >>OVERLOAD_LIMIT_BASE is tested against a per-socket message counter, so it _is_
> >>our per-socket limit. In fact, TIPC connectionless overflow
> >>control currently is a kind of a hybrid, based on a message
> >>counter when the socket is not locked, and based on
> >>sk_rcv_queue's byte limit when a message has to be added to the
> >>backlog.
> >>We are planning to fix this inconsistency too.
> >Good, thank you, that was seeming quite wrong to me.
> >
> >> In fact it seems to
> >>>repeat the same check twice, as in the worst case of the incomming message being
> >>>TIPC_LOW_IMPORTANCE, its just going to check that the global limit is exactly
> >>>OVERLOAD_LIMIT_BASE/2 again.
> >>Yes, you are right. The intention is that only the first test,
> >>if (unlikely(recv_q_len >= (OVERLOAD_LIMIT_BASE / 2)){..}
> >>will be run for the vast majority of messages, since we must assume
> >>that there is no overload most of the time.
> >>An inelegant optimization, perhaps, but not logically wrong.
> >No, not logically wrong, but not an optimization either. With this change,
> >your only use of rx_queue_full passes OVERLOAD_LIMIT_BASE/2 as the base value to
> >rx_queue_full, and then you do some multiplication based on that. If you really
> >want to optimize this, leave OVERLOAD_LIMIT_BASE where it is (rather than
> >doubling it like this patch series does), mark rx_queue_full as inline, and just
> >pass OVERLOAD_LIMIT_BASE as the argument, it will save you a division opration,
> >the conditional branch and a call instruction. If you add a multiplication
> >factor table, you can eliminate the if/else clauses in rx_queue_full as well.
> >
>
> Good suggestion with a factor table. Maybe it's unnecessary to
> explicitly mark rx_queue_full as inline. Currently it sounds like we
> let complier decide whether a function is defined as inline or not.
>
Thats correct, the compiler usually decides if something should be inlined
(unless you use the __always_inline) attribute. In this case, given a single
call site, it most like will just inline anyway. But if you're interested in
optimizing here, it might be worth taking the extra steps to make sure. In
fact, since this is your only call site, it may be worthwhile to just remove the
function entirely, and manually inline the check.
Neil
> Regards,
> Ying
>
> >Neil
> >
> >>///jon
> >>
> >>>Neil
> >>>
> >>>--
> >>>To unsubscribe from this list: send the line "unsubscribe netdev" in
> >>>the body of a message to majordomo@vger.kernel.org
> >>>More majordomo info at http://vger.kernel.org/majordomo-info.html
> >>>
> >--
> >To unsubscribe from this list: send the line "unsubscribe netdev" in
> >the body of a message to majordomo@vger.kernel.org
> >More majordomo info at http://vger.kernel.org/majordomo-info.html
> >
>
>
^ permalink raw reply
* Re: [PATCH][RFC] smsc95xx: enable dynamic autosuspend (RFC)
From: Steve Glendinning @ 2012-12-10 14:18 UTC (permalink / raw)
To: Oliver Neukum
Cc: Steve Glendinning, netdev, Ming Lei, linux-usb,
Greg Kroah-Hartman
In-Reply-To: <1404210.ufpq7a3oeI@linux-lqwf.site>
On 10 December 2012 12:09, Oliver Neukum <oliver@neukum.org> wrote:
> So this is a problem with remote wakeup on older hardware?
Exactly, the older hardware revisions can't reliably do it.
>> Unfortunately we don't know if the connected device supports
>> this feature until we query its ID register at runtime.
>>
>> Suggestions on how best to indicate this capability at runtime
>> instead of compile-time would be appreciated, so we don't have
>> to repeatedly fail if accidentally enabled. Or maybe this is
>> actually the best way?
>
> If this is a problem with remote wakeup, you should up the
> pm counter (usb_autopm_get_noresume()) in .manage_power
> That was the reason I implemented this is a callback and not as
> a helper in usbnet.
Thanks, so something like this should do the job?
static int smsc95xx_manage_power(struct usbnet *dev, int on)
{
struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
dev->intf->needs_remote_wakeup = on;
if (pdata->features & FEATURE_AUTOSUSPEND)
return 0;
/* this chip revision doesn't support autosuspend */
netdev_info(dev->net, "hardware doesn't support USB autosuspend\n");
if (on)
usb_autopm_get_interface_no_resume(dev->intf);
else
usb_autopm_put_interface_no_suspend(dev->intf);
return 0;
}
^ permalink raw reply
* Re: [PATCH][RFC] smsc95xx: enable dynamic autosuspend (RFC)
From: Ming Lei @ 2012-12-10 13:59 UTC (permalink / raw)
To: Steve Glendinning
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, Oliver Neukum,
linux-usb-u79uwXL29TY76Z2rM5mHXA,
gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r
In-Reply-To: <1355140301-10518-1-git-send-email-steve.glendinning-nksJyM/082jR7s880joybQ@public.gmane.org>
On Mon, Dec 10, 2012 at 7:51 PM, Steve Glendinning
<steve.glendinning-nksJyM/082jR7s880joybQ@public.gmane.org> wrote:
> This is a work in-progress patch. It's not yet complete but
> I thought I'd share it for comments, feedback and testing.
>
> This patch enables dynamic autosuspend for all devices
> supported by the driver, but it will only actually work on
> LAN9500A (as this has a new SUSPEND3 mode for this purpose).
> Unfortunately we don't know if the connected device supports
> this feature until we query its ID register at runtime.
>
> On unsupported devices (LAN9500/9512/9514) this patch claims
> to support the feature but if enabled it will always return
> failure to the autosuspend call (and fill up the kernel log
> with a message every 2s).
>
> Suggestions on how best to indicate this capability at runtime
> instead of compile-time would be appreciated, so we don't have
> to repeatedly fail if accidentally enabled. Or maybe this is
> actually the best way?
The ID register can be read inside bind(), so you may set
smsc95xx_info.manage_power as smsc95xx_manage_power
only for LAN9500A devices.
One disadvantage of above idea is that the link down can't trigger
runtime suspend via mange_power way(USB auto-suspend), but
we still can introduce explicit link change based runtime suspend for
non-LAN9500A devices.
>
> We should also be able to identify devices supporting
> autosuspend from the USB VID/PID if this would help.
>
> UPDATE: reposting this to a wider audience due to lack of
> feedback last time round
>
> Signed-off-by: Steve Glendinning <steve.glendinning-nksJyM/082jR7s880joybQ@public.gmane.org>
> ---
> drivers/net/usb/smsc95xx.c | 136 +++++++++++++++++++++++++++++++++++++++++++-
> 1 file changed, 135 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/usb/smsc95xx.c b/drivers/net/usb/smsc95xx.c
> index 9b73670..d9c6674 100644
> --- a/drivers/net/usb/smsc95xx.c
> +++ b/drivers/net/usb/smsc95xx.c
> @@ -55,6 +55,13 @@
> #define FEATURE_PHY_NLP_CROSSOVER (0x02)
> #define FEATURE_AUTOSUSPEND (0x04)
>
> +#define SUSPEND_SUSPEND0 (0x01)
> +#define SUSPEND_SUSPEND1 (0x02)
> +#define SUSPEND_SUSPEND2 (0x04)
> +#define SUSPEND_SUSPEND3 (0x08)
> +#define SUSPEND_ALLMODES (SUSPEND_SUSPEND0 | SUSPEND_SUSPEND1 | \
> + SUSPEND_SUSPEND2 | SUSPEND_SUSPEND3)
> +
> struct smsc95xx_priv {
> u32 mac_cr;
> u32 hash_hi;
> @@ -62,6 +69,7 @@ struct smsc95xx_priv {
> u32 wolopts;
> spinlock_t mac_cr_lock;
> u8 features;
> + u8 suspend_flags;
> };
>
> static bool turbo_mode = true;
> @@ -1341,6 +1349,8 @@ static int smsc95xx_enter_suspend0(struct usbnet *dev)
> if (ret < 0)
> netdev_warn(dev->net, "Error reading PM_CTRL\n");
>
> + pdata->suspend_flags |= SUSPEND_SUSPEND0;
> +
> return ret;
> }
>
> @@ -1393,11 +1403,14 @@ static int smsc95xx_enter_suspend1(struct usbnet *dev)
> if (ret < 0)
> netdev_warn(dev->net, "Error writing PM_CTRL\n");
>
> + pdata->suspend_flags |= SUSPEND_SUSPEND1;
> +
> return ret;
> }
>
> static int smsc95xx_enter_suspend2(struct usbnet *dev)
> {
> + struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
> u32 val;
> int ret;
>
> @@ -1414,9 +1427,96 @@ static int smsc95xx_enter_suspend2(struct usbnet *dev)
> if (ret < 0)
> netdev_warn(dev->net, "Error writing PM_CTRL\n");
>
> + pdata->suspend_flags |= SUSPEND_SUSPEND2;
> +
> return ret;
> }
>
> +static int smsc95xx_enter_suspend3(struct usbnet *dev)
> +{
> + struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
> + u32 val;
> + int ret;
> +
> + ret = smsc95xx_read_reg_nopm(dev, RX_FIFO_INF, &val);
> + if (ret < 0) {
> + netdev_warn(dev->net, "Error reading RX_FIFO_INF");
> + return ret;
> + }
> +
> + if (val & 0xFFFF) {
> + netdev_info(dev->net, "rx fifo not empty in autosuspend");
> + return -EBUSY;
> + }
> +
> + ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
> + if (ret < 0) {
> + netdev_warn(dev->net, "Error reading PM_CTRL");
> + return ret;
> + }
> +
> + val &= ~(PM_CTL_SUS_MODE_ | PM_CTL_WUPS_ | PM_CTL_PHY_RST_);
> + val |= PM_CTL_SUS_MODE_3 | PM_CTL_RES_CLR_WKP_STS;
> +
> + ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
> + if (ret < 0) {
> + netdev_warn(dev->net, "Error writing PM_CTRL");
> + return ret;
> + }
> +
> + /* clear wol status */
> + val &= ~PM_CTL_WUPS_;
> + val |= PM_CTL_WUPS_WOL_;
> +
> + ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
> + if (ret < 0) {
> + netdev_warn(dev->net, "Error writing PM_CTRL");
> + return ret;
> + }
> +
> + pdata->suspend_flags |= SUSPEND_SUSPEND3;
> +
> + return 0;
> +}
> +
> +static int smsc95xx_autosuspend(struct usbnet *dev, u32 link_up)
> +{
> + int ret;
> +
> + if (!netif_running(dev->net)) {
> + /* interface is ifconfig down so fully power down hw */
> + netdev_dbg(dev->net, "autosuspend entering SUSPEND2");
> + return smsc95xx_enter_suspend2(dev);
> + }
> +
> + if (!link_up) {
> + /* link is down so enter EDPD mode */
> + netdev_dbg(dev->net, "autosuspend entering SUSPEND1");
> +
> + /* enable PHY wakeup events for if cable is attached */
> + ret = smsc95xx_enable_phy_wakeup_interrupts(dev,
> + PHY_INT_MASK_ANEG_COMP_);
> + if (ret < 0) {
> + netdev_warn(dev->net, "error enabling PHY wakeup ints");
> + return ret;
> + }
> +
> + netdev_info(dev->net, "entering SUSPEND1 mode");
> + return smsc95xx_enter_suspend1(dev);
> + }
> +
> + /* enable PHY wakeup events so we remote wakeup if cable is pulled */
> + ret = smsc95xx_enable_phy_wakeup_interrupts(dev,
> + PHY_INT_MASK_LINK_DOWN_);
> + if (ret < 0) {
> + netdev_warn(dev->net, "error enabling PHY wakeup ints");
> + return ret;
> + }
> +
> + netdev_dbg(dev->net, "autosuspend entering SUSPEND3");
> + return smsc95xx_enter_suspend3(dev);
> +}
> +
> static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
> {
> struct usbnet *dev = usb_get_intfdata(intf);
> @@ -1424,15 +1524,35 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
> u32 val, link_up;
> int ret;
>
> + /* TODO: don't indicate this feature to usb framework if
> + * our current hardware doesn't have the capability
> + */
> + if ((message.event == PM_EVENT_AUTO_SUSPEND) &&
> + (!(pdata->features & FEATURE_AUTOSUSPEND))) {
> + netdev_warn(dev->net, "autosuspend not supported");
> + return -EBUSY;
> + }
> +
> ret = usbnet_suspend(intf, message);
> if (ret < 0) {
> netdev_warn(dev->net, "usbnet_suspend error\n");
> return ret;
> }
>
> + if (pdata->suspend_flags) {
> + netdev_warn(dev->net, "error during last resume");
> + pdata->suspend_flags = 0;
> + }
> +
> /* determine if link is up using only _nopm functions */
> link_up = smsc95xx_link_ok_nopm(dev);
>
> + if (message.event == PM_EVENT_AUTO_SUSPEND) {
> + ret = smsc95xx_autosuspend(dev, link_up);
> + goto done;
> + }
> +
> + /* if we get this far we're not autosuspending */
> /* if no wol options set, or if link is down and we're not waking on
> * PHY activity, enter lowest power SUSPEND2 mode
> */
> @@ -1694,12 +1814,18 @@ static int smsc95xx_resume(struct usb_interface *intf)
> {
> struct usbnet *dev = usb_get_intfdata(intf);
> struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
> + u8 suspend_flags = pdata->suspend_flags;
> int ret;
> u32 val;
>
> BUG_ON(!dev);
>
> - if (pdata->wolopts) {
> + netdev_dbg(dev->net, "resume suspend_flags=0x%02x", suspend_flags);
> +
> + /* do this first to ensure it's cleared even in error case */
> + pdata->suspend_flags = 0;
> +
> + if (suspend_flags & SUSPEND_ALLMODES) {
> /* clear wake-up sources */
> ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
> if (ret < 0) {
> @@ -1891,6 +2017,12 @@ static struct sk_buff *smsc95xx_tx_fixup(struct usbnet *dev,
> return skb;
> }
>
> +static int smsc95xx_manage_power(struct usbnet *dev, int on)
> +{
> + dev->intf->needs_remote_wakeup = on;
> + return 0;
> +}
> +
> static const struct driver_info smsc95xx_info = {
> .description = "smsc95xx USB 2.0 Ethernet",
> .bind = smsc95xx_bind,
> @@ -1900,6 +2032,7 @@ static const struct driver_info smsc95xx_info = {
> .rx_fixup = smsc95xx_rx_fixup,
> .tx_fixup = smsc95xx_tx_fixup,
> .status = smsc95xx_status,
> + .manage_power = smsc95xx_manage_power,
> .flags = FLAG_ETHER | FLAG_SEND_ZLP | FLAG_LINK_INTR,
> };
>
> @@ -2007,6 +2140,7 @@ static struct usb_driver smsc95xx_driver = {
> .reset_resume = smsc95xx_resume,
> .disconnect = usbnet_disconnect,
> .disable_hub_initiated_lpm = 1,
> + .supports_autosuspend = 1,
> };
>
> module_usb_driver(smsc95xx_driver);
> --
> 1.7.10.4
>
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [net-next:master 195/198] net/bridge/br_mdb.c:79:35: sparse: incompatible types in comparison expression (different address spaces)
From: Cong Wang @ 2012-12-10 12:44 UTC (permalink / raw)
To: kbuild test robot; +Cc: netdev
In-Reply-To: <1355143004.11752.19.camel@cr0>
On Mon, 2012-12-10 at 20:36 +0800, Cong Wang wrote:
> On Sun, 2012-12-09 at 10:54 +0800, kbuild test robot wrote:
> > tree: git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git master
> > head: 9ecb9aabaf634677c77af467f4e3028b09d7bcda
> > commit: ee07c6e7a6f8a25c18f0a6b18152fbd7499245f6 [195/198] bridge: export multicast database via netlink
> >
> >
> > sparse warnings:
> >
> > + net/bridge/br_mdb.c:79:35: sparse: incompatible types in comparison expression (different address spaces)
>
> Hi, Fengguang,
>
> I am not sure if I understand this warning correctly. Does the following
> patch fix it?
>
Hmm, no, probably this one:
diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index 2528328..cd6735c 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -63,7 +63,7 @@ static int br_mdb_fill_info(struct sk_buff *skb,
struct netlink_callback *cb,
for (i = 0; i < mdb->max; i++) {
struct hlist_node *h;
struct net_bridge_mdb_entry *mp;
- struct net_bridge_port_group *p, **pp;
+ struct net_bridge_port_group __rcu *p, **pp;
struct net_bridge_port *port;
hlist_for_each_entry_rcu(mp, h, &mdb->mhash[i],
hlist[mdb->ver]) {
^ permalink raw reply related
* Re: [net-next:master 195/198] net/bridge/br_mdb.c:79:35: sparse: incompatible types in comparison expression (different address spaces)
From: Cong Wang @ 2012-12-10 12:36 UTC (permalink / raw)
To: kbuild test robot; +Cc: netdev
In-Reply-To: <50c3fd5e.qB65a7/5e+IZT2ix%fengguang.wu@intel.com>
On Sun, 2012-12-09 at 10:54 +0800, kbuild test robot wrote:
> tree: git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git master
> head: 9ecb9aabaf634677c77af467f4e3028b09d7bcda
> commit: ee07c6e7a6f8a25c18f0a6b18152fbd7499245f6 [195/198] bridge: export multicast database via netlink
>
>
> sparse warnings:
>
> + net/bridge/br_mdb.c:79:35: sparse: incompatible types in comparison expression (different address spaces)
Hi, Fengguang,
I am not sure if I understand this warning correctly. Does the following
patch fix it?
Thanks!
---------------->
diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index 2528328..0bc0e13 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -77,7 +77,7 @@ static int br_mdb_fill_info(struct sk_buff *skb,
struct netlink_callback *cb,
}
for (pp = &mp->ports;
- (p = rcu_dereference(*pp)) != NULL;
+ (p = rcu_dereference_protected(*pp, 1)) !=
NULL;
pp = &p->next) {
port = p->port;
if (port) {
^ permalink raw reply related
* [Patch net-next] virtio_net: fix a typo in virtnet_alloc_queues()
From: Cong Wang @ 2012-12-10 12:24 UTC (permalink / raw)
To: netdev; +Cc: Jason Wang, David S. Miller, Cong Wang
Obviously it should check !vi->rq.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
---
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index a644eeb..68d64f0 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1347,7 +1347,7 @@ static int virtnet_alloc_queues(struct virtnet_info *vi)
if (!vi->sq)
goto err_sq;
vi->rq = kzalloc(sizeof(*vi->rq) * vi->max_queue_pairs, GFP_KERNEL);
- if (!vi->sq)
+ if (!vi->rq)
goto err_rq;
INIT_DELAYED_WORK(&vi->refill, refill_work);
^ permalink raw reply related
* [Patch net-next] bridge: fix seq check in br_mdb_dump()
From: Cong Wang @ 2012-12-10 12:15 UTC (permalink / raw)
To: netdev
Cc: Herbert Xu, Stephen Hemminger, David S. Miller, Thomas Graf,
Jesper Dangaard Brouer, Cong Wang
From: Cong Wang <amwang@redhat.com>
In case of rehashing, introduce a global variable 'br_mdb_rehash_seq'
which gets increased every time when rehashing, and assign
net->dev_base_seq + br_mdb_rehash_seq to cb->seq.
In theory cb->seq could be wrapped to zero, but this is not
easy to fix, as net->dev_base_seq is not visible inside
br_mdb_rehash(). In practice, this is rare.
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Stephen Hemminger <shemminger@vyatta.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Graf <tgraf@suug.ch>
Cc: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Cong Wang <amwang@redhat.com>
---
net/bridge/br_mdb.c | 6 ++----
net/bridge/br_multicast.c | 2 ++
net/bridge/br_private.h | 1 +
3 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index edc0d73..ccc43a9 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -117,10 +117,8 @@ static int br_mdb_dump(struct sk_buff *skb, struct netlink_callback *cb)
rcu_read_lock();
- /* TODO: in case of rehashing, we need to check
- * consistency for dumping.
- */
- cb->seq = net->dev_base_seq;
+ /* In theory this could be wrapped to 0... */
+ cb->seq = net->dev_base_seq + br_mdb_rehash_seq;
for_each_netdev_rcu(net, dev) {
if (dev->priv_flags & IFF_EBRIDGE) {
diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 68e375a..847b98a1 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -37,6 +37,7 @@
rcu_dereference_protected(X, lockdep_is_held(&br->multicast_lock))
static void br_multicast_start_querier(struct net_bridge *br);
+unsigned int br_mdb_rehash_seq;
#if IS_ENABLED(CONFIG_IPV6)
static inline int ipv6_is_transient_multicast(const struct in6_addr *addr)
@@ -338,6 +339,7 @@ static int br_mdb_rehash(struct net_bridge_mdb_htable __rcu **mdbp, int max,
return err;
}
+ br_mdb_rehash_seq++;
call_rcu_bh(&mdb->rcu, br_mdb_free);
out:
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index ae0a6ec..f95b766 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -411,6 +411,7 @@ extern int br_ioctl_deviceless_stub(struct net *net, unsigned int cmd, void __us
/* br_multicast.c */
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
+extern unsigned int br_mdb_rehash_seq;
extern int br_multicast_rcv(struct net_bridge *br,
struct net_bridge_port *port,
struct sk_buff *skb);
^ permalink raw reply related
* Re: [PATCH][RFC] smsc95xx: enable dynamic autosuspend (RFC)
From: Oliver Neukum @ 2012-12-10 12:09 UTC (permalink / raw)
To: Steve Glendinning
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, Ming Lei,
linux-usb-u79uwXL29TY76Z2rM5mHXA,
gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r
In-Reply-To: <1355140301-10518-1-git-send-email-steve.glendinning-nksJyM/082jR7s880joybQ@public.gmane.org>
On Monday 10 December 2012 11:51:41 Steve Glendinning wrote:
> This is a work in-progress patch. It's not yet complete but
> I thought I'd share it for comments, feedback and testing.
>
> This patch enables dynamic autosuspend for all devices
> supported by the driver, but it will only actually work on
> LAN9500A (as this has a new SUSPEND3 mode for this purpose).
So this is a problem with remote wakeup on older hardware?
> Unfortunately we don't know if the connected device supports
> this feature until we query its ID register at runtime.
>
> On unsupported devices (LAN9500/9512/9514) this patch claims
> to support the feature but if enabled it will always return
> failure to the autosuspend call (and fill up the kernel log
> with a message every 2s).
>
> Suggestions on how best to indicate this capability at runtime
> instead of compile-time would be appreciated, so we don't have
> to repeatedly fail if accidentally enabled. Or maybe this is
> actually the best way?
If this is a problem with remote wakeup, you should up the
pm counter (usb_autopm_get_noresume()) in .manage_power
That was the reason I implemented this is a callback and not as
a helper in usbnet.
Regards
Oliver
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH][RFC] smsc95xx: enable dynamic autosuspend (RFC)
From: Steve Glendinning @ 2012-12-10 11:51 UTC (permalink / raw)
To: netdev; +Cc: Ming Lei, Oliver Neukum, linux-usb, gregkh, Steve Glendinning
In-Reply-To: <1353607526-19307-6-git-send-email-steve.glendinning@shawell.net>
This is a work in-progress patch. It's not yet complete but
I thought I'd share it for comments, feedback and testing.
This patch enables dynamic autosuspend for all devices
supported by the driver, but it will only actually work on
LAN9500A (as this has a new SUSPEND3 mode for this purpose).
Unfortunately we don't know if the connected device supports
this feature until we query its ID register at runtime.
On unsupported devices (LAN9500/9512/9514) this patch claims
to support the feature but if enabled it will always return
failure to the autosuspend call (and fill up the kernel log
with a message every 2s).
Suggestions on how best to indicate this capability at runtime
instead of compile-time would be appreciated, so we don't have
to repeatedly fail if accidentally enabled. Or maybe this is
actually the best way?
We should also be able to identify devices supporting
autosuspend from the USB VID/PID if this would help.
UPDATE: reposting this to a wider audience due to lack of
feedback last time round
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
---
drivers/net/usb/smsc95xx.c | 136 +++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 135 insertions(+), 1 deletion(-)
diff --git a/drivers/net/usb/smsc95xx.c b/drivers/net/usb/smsc95xx.c
index 9b73670..d9c6674 100644
--- a/drivers/net/usb/smsc95xx.c
+++ b/drivers/net/usb/smsc95xx.c
@@ -55,6 +55,13 @@
#define FEATURE_PHY_NLP_CROSSOVER (0x02)
#define FEATURE_AUTOSUSPEND (0x04)
+#define SUSPEND_SUSPEND0 (0x01)
+#define SUSPEND_SUSPEND1 (0x02)
+#define SUSPEND_SUSPEND2 (0x04)
+#define SUSPEND_SUSPEND3 (0x08)
+#define SUSPEND_ALLMODES (SUSPEND_SUSPEND0 | SUSPEND_SUSPEND1 | \
+ SUSPEND_SUSPEND2 | SUSPEND_SUSPEND3)
+
struct smsc95xx_priv {
u32 mac_cr;
u32 hash_hi;
@@ -62,6 +69,7 @@ struct smsc95xx_priv {
u32 wolopts;
spinlock_t mac_cr_lock;
u8 features;
+ u8 suspend_flags;
};
static bool turbo_mode = true;
@@ -1341,6 +1349,8 @@ static int smsc95xx_enter_suspend0(struct usbnet *dev)
if (ret < 0)
netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ pdata->suspend_flags |= SUSPEND_SUSPEND0;
+
return ret;
}
@@ -1393,11 +1403,14 @@ static int smsc95xx_enter_suspend1(struct usbnet *dev)
if (ret < 0)
netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ pdata->suspend_flags |= SUSPEND_SUSPEND1;
+
return ret;
}
static int smsc95xx_enter_suspend2(struct usbnet *dev)
{
+ struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
u32 val;
int ret;
@@ -1414,9 +1427,96 @@ static int smsc95xx_enter_suspend2(struct usbnet *dev)
if (ret < 0)
netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ pdata->suspend_flags |= SUSPEND_SUSPEND2;
+
return ret;
}
+static int smsc95xx_enter_suspend3(struct usbnet *dev)
+{
+ struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
+ u32 val;
+ int ret;
+
+ ret = smsc95xx_read_reg_nopm(dev, RX_FIFO_INF, &val);
+ if (ret < 0) {
+ netdev_warn(dev->net, "Error reading RX_FIFO_INF");
+ return ret;
+ }
+
+ if (val & 0xFFFF) {
+ netdev_info(dev->net, "rx fifo not empty in autosuspend");
+ return -EBUSY;
+ }
+
+ ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
+ if (ret < 0) {
+ netdev_warn(dev->net, "Error reading PM_CTRL");
+ return ret;
+ }
+
+ val &= ~(PM_CTL_SUS_MODE_ | PM_CTL_WUPS_ | PM_CTL_PHY_RST_);
+ val |= PM_CTL_SUS_MODE_3 | PM_CTL_RES_CLR_WKP_STS;
+
+ ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
+ if (ret < 0) {
+ netdev_warn(dev->net, "Error writing PM_CTRL");
+ return ret;
+ }
+
+ /* clear wol status */
+ val &= ~PM_CTL_WUPS_;
+ val |= PM_CTL_WUPS_WOL_;
+
+ ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
+ if (ret < 0) {
+ netdev_warn(dev->net, "Error writing PM_CTRL");
+ return ret;
+ }
+
+ pdata->suspend_flags |= SUSPEND_SUSPEND3;
+
+ return 0;
+}
+
+static int smsc95xx_autosuspend(struct usbnet *dev, u32 link_up)
+{
+ int ret;
+
+ if (!netif_running(dev->net)) {
+ /* interface is ifconfig down so fully power down hw */
+ netdev_dbg(dev->net, "autosuspend entering SUSPEND2");
+ return smsc95xx_enter_suspend2(dev);
+ }
+
+ if (!link_up) {
+ /* link is down so enter EDPD mode */
+ netdev_dbg(dev->net, "autosuspend entering SUSPEND1");
+
+ /* enable PHY wakeup events for if cable is attached */
+ ret = smsc95xx_enable_phy_wakeup_interrupts(dev,
+ PHY_INT_MASK_ANEG_COMP_);
+ if (ret < 0) {
+ netdev_warn(dev->net, "error enabling PHY wakeup ints");
+ return ret;
+ }
+
+ netdev_info(dev->net, "entering SUSPEND1 mode");
+ return smsc95xx_enter_suspend1(dev);
+ }
+
+ /* enable PHY wakeup events so we remote wakeup if cable is pulled */
+ ret = smsc95xx_enable_phy_wakeup_interrupts(dev,
+ PHY_INT_MASK_LINK_DOWN_);
+ if (ret < 0) {
+ netdev_warn(dev->net, "error enabling PHY wakeup ints");
+ return ret;
+ }
+
+ netdev_dbg(dev->net, "autosuspend entering SUSPEND3");
+ return smsc95xx_enter_suspend3(dev);
+}
+
static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
{
struct usbnet *dev = usb_get_intfdata(intf);
@@ -1424,15 +1524,35 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
u32 val, link_up;
int ret;
+ /* TODO: don't indicate this feature to usb framework if
+ * our current hardware doesn't have the capability
+ */
+ if ((message.event == PM_EVENT_AUTO_SUSPEND) &&
+ (!(pdata->features & FEATURE_AUTOSUSPEND))) {
+ netdev_warn(dev->net, "autosuspend not supported");
+ return -EBUSY;
+ }
+
ret = usbnet_suspend(intf, message);
if (ret < 0) {
netdev_warn(dev->net, "usbnet_suspend error\n");
return ret;
}
+ if (pdata->suspend_flags) {
+ netdev_warn(dev->net, "error during last resume");
+ pdata->suspend_flags = 0;
+ }
+
/* determine if link is up using only _nopm functions */
link_up = smsc95xx_link_ok_nopm(dev);
+ if (message.event == PM_EVENT_AUTO_SUSPEND) {
+ ret = smsc95xx_autosuspend(dev, link_up);
+ goto done;
+ }
+
+ /* if we get this far we're not autosuspending */
/* if no wol options set, or if link is down and we're not waking on
* PHY activity, enter lowest power SUSPEND2 mode
*/
@@ -1694,12 +1814,18 @@ static int smsc95xx_resume(struct usb_interface *intf)
{
struct usbnet *dev = usb_get_intfdata(intf);
struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
+ u8 suspend_flags = pdata->suspend_flags;
int ret;
u32 val;
BUG_ON(!dev);
- if (pdata->wolopts) {
+ netdev_dbg(dev->net, "resume suspend_flags=0x%02x", suspend_flags);
+
+ /* do this first to ensure it's cleared even in error case */
+ pdata->suspend_flags = 0;
+
+ if (suspend_flags & SUSPEND_ALLMODES) {
/* clear wake-up sources */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
if (ret < 0) {
@@ -1891,6 +2017,12 @@ static struct sk_buff *smsc95xx_tx_fixup(struct usbnet *dev,
return skb;
}
+static int smsc95xx_manage_power(struct usbnet *dev, int on)
+{
+ dev->intf->needs_remote_wakeup = on;
+ return 0;
+}
+
static const struct driver_info smsc95xx_info = {
.description = "smsc95xx USB 2.0 Ethernet",
.bind = smsc95xx_bind,
@@ -1900,6 +2032,7 @@ static const struct driver_info smsc95xx_info = {
.rx_fixup = smsc95xx_rx_fixup,
.tx_fixup = smsc95xx_tx_fixup,
.status = smsc95xx_status,
+ .manage_power = smsc95xx_manage_power,
.flags = FLAG_ETHER | FLAG_SEND_ZLP | FLAG_LINK_INTR,
};
@@ -2007,6 +2140,7 @@ static struct usb_driver smsc95xx_driver = {
.reset_resume = smsc95xx_resume,
.disconnect = usbnet_disconnect,
.disable_hub_initiated_lpm = 1,
+ .supports_autosuspend = 1,
};
module_usb_driver(smsc95xx_driver);
--
1.7.10.4
^ permalink raw reply related
* RE: ipgre rss is broken since gro
From: Dmitry Kravkov @ 2012-12-10 11:32 UTC (permalink / raw)
To: Eric Dumazet; +Cc: netdev@vger.kernel.org
In-Reply-To: <CANn89iKtWuN=CRn-JBzmmo8jCCJQhZjAT8jA+9dgfmup99O03g@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 2635 bytes --]
> -----Original Message-----
> From: Eric Dumazet [mailto:edumazet@google.com]
> Sent: Monday, December 10, 2012 1:27 AM
> To: Dmitry Kravkov
> Cc: netdev@vger.kernel.org
> Subject: Re: ipgre rss is broken since gro
>
> On Sun, Dec 9, 2012 at 12:49 PM, Dmitry Kravkov <dmitry@broadcom.com>
> wrote:
>
> > for this item: drop_watch does not show any drops (i've disable all
> > other interfaces for clear env)
> > I will explain a little bit more the setup:
> > bnx2x device (under testing) is configured for RSS for IPGRE packets.
> > Sending multiple (3) TCP_STREAM causes ip_gre interface to disappear
> > packets (even ICMP).
> > This is not happening with single TCP_STREAM, or before gro_cell
> > introduction.
> >
> I dont know, I tried a bnx2x setup, and 100 tcp flows, no special problem.
Current bnx2x do not apply RSS for GRE, non GRE RSS is working w/o problem.
>
> If you receive a lot of packets on a single RX queue, they might be
> dropped because cpu cant cope with the load
> (This has nothing to do with GRE or GRO )
>
CPU is not loaded at all
> cat /proc/net/softnet_stat
Please find attached.
For gre interface RX and DROP statistics are advancing simultaneously (by one each ICMP request):
[root@ ~]# ifconfig gre
gre Link encap:UNSPEC HWaddr C0-A8-0A-40-73-72-83-D2-00-00-00-00-00-00-00-00
inet addr:8.0.0.1 P-t-P:8.0.0.1 Mask:255.255.255.0
inet6 addr: fe80::5efe:c0a8:a40/64 Scope:Link
UP POINTOPOINT RUNNING NOARP MTU:1476 Metric:1
RX packets:1646824 errors:0 dropped:51610 overruns:0 frame:0
TX packets:140519 errors:1 dropped:0 overruns:0 carrier:1
collisions:0 txqueuelen:0
RX bytes:2357650904 (2.1 GiB) TX bytes:7309072 (6.9 MiB)
[root@ ~]# ifconfig gre
gre Link encap:UNSPEC HWaddr C0-A8-0A-40-73-72-83-82-00-00-00-00-00-00-00-00
inet addr:8.0.0.1 P-t-P:8.0.0.1 Mask:255.255.255.0
inet6 addr: fe80::5efe:c0a8:a40/64 Scope:Link
UP POINTOPOINT RUNNING NOARP MTU:1476 Metric:1
RX packets:1646826 errors:0 dropped:51612 overruns:0 frame:0
TX packets:140519 errors:1 dropped:0 overruns:0 carrier:1
collisions:0 txqueuelen:0
RX bytes:2357651072 (2.1 GiB) TX bytes:7309072 (6.9 MiB)
[root@ ~]# tcpdump -i gre
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on gre, link-type LINUX_SLL (Linux cooked), capture size 65535 bytes
^C
0 packets captured
0 packets received by filter
0 packets dropped by kernel
2 packets dropped by interface
[-- Attachment #2: stat1 --]
[-- Type: application/octet-stream, Size: 3600 bytes --]
001b96a3 00000000 00001cb5 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000036f 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000181 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000001e6 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000012e 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000007b 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000004 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000a 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000002 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000029e 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000043 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000015 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000c 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000010 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000008 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000007 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00008192 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00005682 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000066bd 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00002ab5 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00001d11 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000165cf 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000005 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000203 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000009b 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000c914 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000003c 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000b 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000788 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00001dab 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000972f 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000015 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000c 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[-- Attachment #3: stat2 --]
[-- Type: application/octet-stream, Size: 3600 bytes --]
001b96a3 00000000 00001cb5 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000036f 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000181 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000001e6 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000012e 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000007b 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000004 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000a 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000002 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000029e 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000043 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000015 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000c 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000010 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000008 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000007 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00008192 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00005682 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000066bd 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00002ab5 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00001d11 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
000165cf 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000005 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000203 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000009b 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000c919 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000003c 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000b 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000789 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00001dab 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000972f 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000015 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0000000c 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
^ permalink raw reply
* Re: [PATCH 1/1] net: ethernet: davinci_cpdma: Add boundary for rx and tx descriptors
From: Mugunthan V N @ 2012-12-10 11:28 UTC (permalink / raw)
To: Christian Riesch; +Cc: netdev, davem, linux-arm-kernel, linux-omap, s.hauer
In-Reply-To: <CABkLObp1cZrNR65KJmOJXCB+W_1ZriCJQdL1TEmkJCptXHFDXw@mail.gmail.com>
On 12/10/2012 1:41 PM, Christian Riesch wrote:
> Hi,
>
> On Mon, Dec 10, 2012 at 8:37 AM, Mugunthan V N <mugunthanvnm@ti.com> wrote:
>> When there is heavy transmission traffic in the CPDMA, then Rx descriptors
>> memory is also utilized as tx desc memory this leads to reduced rx desc memory
>> which leads to poor performance.
>>
>> This patch adds boundary for tx and rx descriptors in bd ram dividing the
>> descriptor memory to ensure that during heavy transmission tx doesn't use
>> rx descriptors.
>>
>> This patch is already applied to davinci_emac driver, since CPSW and
>> davici_dmac uses the same CPDMA, moving the boundry seperation from
>> Davinci EMAC driver to CPDMA driver which was done in the following
>> commit
>>
>> commit 86d8c07ff2448eb4e860e50f34ef6ee78e45c40c
>> Author: Sascha Hauer <s.hauer@pengutronix.de>
>> Date: Tue Jan 3 05:27:47 2012 +0000
>>
>> net/davinci: do not use all descriptors for tx packets
>>
>> The driver uses a shared pool for both rx and tx descriptors.
>> During open it queues fixed number of 128 descriptors for receive
>> packets. For each received packet it tries to queue another
>> descriptor. If this fails the descriptor is lost for rx.
>> The driver has no limitation on tx descriptors to use, so it
>> can happen during a nmap / ping -f attack that the driver
>> allocates all descriptors for tx and looses all rx descriptors.
>> The driver stops working then.
>> To fix this limit the number of tx descriptors used to half of
>> the descriptors available, the rx path uses the other half.
>>
>> Tested on a custom board using nmap / ping -f to the board from
>> two different hosts.
>>
>> Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
>> ---
>> drivers/net/ethernet/ti/davinci_cpdma.c | 20 ++++++++++++++------
>> drivers/net/ethernet/ti/davinci_emac.c | 8 --------
>> 2 files changed, 14 insertions(+), 14 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/ti/davinci_cpdma.c b/drivers/net/ethernet/ti/davinci_cpdma.c
>> index 4995673..d37f546 100644
>> --- a/drivers/net/ethernet/ti/davinci_cpdma.c
>> +++ b/drivers/net/ethernet/ti/davinci_cpdma.c
>> @@ -105,13 +105,13 @@ struct cpdma_ctlr {
>> };
>>
>> struct cpdma_chan {
>> + struct cpdma_desc __iomem *head, *tail;
>> + void __iomem *hdp, *cp, *rxfree;
>> enum cpdma_state state;
>> struct cpdma_ctlr *ctlr;
>> int chan_num;
>> spinlock_t lock;
>> - struct cpdma_desc __iomem *head, *tail;
>> int count;
>> - void __iomem *hdp, *cp, *rxfree;
> Why?
Its just a code clean-up to have iomem variables at one place.
>
>> u32 mask;
>> cpdma_handler_fn handler;
>> enum dma_data_direction dir;
>> @@ -217,7 +217,7 @@ desc_from_phys(struct cpdma_desc_pool *pool, dma_addr_t dma)
>> }
>>
>> static struct cpdma_desc __iomem *
>> -cpdma_desc_alloc(struct cpdma_desc_pool *pool, int num_desc)
>> +cpdma_desc_alloc(struct cpdma_desc_pool *pool, int num_desc, bool is_rx)
>> {
>> unsigned long flags;
>> int index;
>> @@ -225,8 +225,14 @@ cpdma_desc_alloc(struct cpdma_desc_pool *pool, int num_desc)
>>
>> spin_lock_irqsave(&pool->lock, flags);
>>
>> - index = bitmap_find_next_zero_area(pool->bitmap, pool->num_desc, 0,
>> - num_desc, 0);
>> + if (is_rx) {
>> + index = bitmap_find_next_zero_area(pool->bitmap,
>> + pool->num_desc/2, 0, num_desc, 0);
>> + } else {
>> + index = bitmap_find_next_zero_area(pool->bitmap,
>> + pool->num_desc, pool->num_desc/2, num_desc, 0);
>> + }
> Would it make sense to use two separate pools for rx and tx instead,
> struct cpdma_desc_pool *rxpool, *txpool? It would result in using
> separate spinlocks for rx and tx, could this be an advantage? (I am a
> newbie in this field...)
I don't think separating pool will give an advantage, the same is
achieved by separating
the pool into first half and last half.
Regards
Mugunthan V N
^ permalink raw reply
* Re: [PATCH 1/1] net: ethernet: davinci_cpdma: Add boundary for rx and tx descriptors
From: Mugunthan V N @ 2012-12-10 11:29 UTC (permalink / raw)
To: Christian Riesch; +Cc: netdev, davem, linux-arm-kernel, linux-omap, s.hauer
In-Reply-To: <CABkLObqOEMKMD3df5pEDJWsGPA=oAj34Yz867D7p2hGdTr0WCA@mail.gmail.com>
On 12/10/2012 1:54 PM, Christian Riesch wrote:
> Hi again,
>
> On Mon, Dec 10, 2012 at 8:37 AM, Mugunthan V N <mugunthanvnm@ti.com> wrote:
>> When there is heavy transmission traffic in the CPDMA, then Rx descriptors
>> memory is also utilized as tx desc memory this leads to reduced rx desc memory
>> which leads to poor performance.
>>
> "poor performance" is an understatement, see Sascha's description of
> his patch. At initialization of the driver, half of the descriptors in
> the pool are allocated for rx. When a packet arrives, one of the rx
> descriptors is released and a new one is allocated. If tx allocates
> this descriptor in the meantime, it is lost for rx forever! If tx
> consumes all rx descriptors this way, the rx channel is dead!
>
> Regards, Christian
>
>> This patch adds boundary for tx and rx descriptors in bd ram dividing the
>> descriptor memory to ensure that during heavy transmission tx doesn't use
>> rx descriptors.
>>
>> This patch is already applied to davinci_emac driver, since CPSW and
>> davici_dmac uses the same CPDMA, moving the boundry seperation from
>> Davinci EMAC driver to CPDMA driver which was done in the following
>> commit
>>
>> commit 86d8c07ff2448eb4e860e50f34ef6ee78e45c40c
>> Author: Sascha Hauer <s.hauer@pengutronix.de>
>> Date: Tue Jan 3 05:27:47 2012 +0000
>>
>> net/davinci: do not use all descriptors for tx packets
>>
>> The driver uses a shared pool for both rx and tx descriptors.
>> During open it queues fixed number of 128 descriptors for receive
>> packets. For each received packet it tries to queue another
>> descriptor. If this fails the descriptor is lost for rx.
>> The driver has no limitation on tx descriptors to use, so it
>> can happen during a nmap / ping -f attack that the driver
>> allocates all descriptors for tx and looses all rx descriptors.
>> The driver stops working then.
>> To fix this limit the number of tx descriptors used to half of
>> the descriptors available, the rx path uses the other half.
>>
>> Tested on a custom board using nmap / ping -f to the board from
>> two different hosts.
>>
>> Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Will change the commit description and resubmit the patch.
Regards
Mugunthan V N
^ permalink raw reply
* [net-next:master 204/207] drivers/net/virtio_net.c:1312 virtnet_alloc_queues() error: potential null dereference 'vi->rq'. (kzalloc returns null)
From: kbuild test robot @ 2012-12-10 11:14 UTC (permalink / raw)
To: Jason Wang; +Cc: netdev, Krishna Kumar
tree: git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git master
head: 65d2897c0f1b240420d657f41e561239fa10ba94
commit: 986a4f4d452dec004697f667439d27c3fda9c928 [204/207] virtio_net: multiqueue support
smatch warnings:
+ drivers/net/virtio_net.c:1312 virtnet_alloc_queues() error: potential null dereference 'vi->rq'. (kzalloc returns null)
vim +1312 drivers/net/virtio_net.c
986a4f4d Jason Wang 2012-12-07 1296 return ret;
986a4f4d Jason Wang 2012-12-07 1297 }
986a4f4d Jason Wang 2012-12-07 1298
986a4f4d Jason Wang 2012-12-07 1299 static int virtnet_alloc_queues(struct virtnet_info *vi)
986a4f4d Jason Wang 2012-12-07 1300 {
986a4f4d Jason Wang 2012-12-07 1301 int i;
986a4f4d Jason Wang 2012-12-07 1302
986a4f4d Jason Wang 2012-12-07 1303 vi->sq = kzalloc(sizeof(*vi->sq) * vi->max_queue_pairs, GFP_KERNEL);
986a4f4d Jason Wang 2012-12-07 1304 if (!vi->sq)
986a4f4d Jason Wang 2012-12-07 1305 goto err_sq;
986a4f4d Jason Wang 2012-12-07 1306 vi->rq = kzalloc(sizeof(*vi->rq) * vi->max_queue_pairs, GFP_KERNEL);
986a4f4d Jason Wang 2012-12-07 1307 if (!vi->sq)
986a4f4d Jason Wang 2012-12-07 1308 goto err_rq;
986a4f4d Jason Wang 2012-12-07 1309
986a4f4d Jason Wang 2012-12-07 1310 INIT_DELAYED_WORK(&vi->refill, refill_work);
986a4f4d Jason Wang 2012-12-07 1311 for (i = 0; i < vi->max_queue_pairs; i++) {
986a4f4d Jason Wang 2012-12-07 @1312 vi->rq[i].pages = NULL;
986a4f4d Jason Wang 2012-12-07 1313 netif_napi_add(vi->dev, &vi->rq[i].napi, virtnet_poll,
986a4f4d Jason Wang 2012-12-07 1314 napi_weight);
986a4f4d Jason Wang 2012-12-07 1315
986a4f4d Jason Wang 2012-12-07 1316 sg_init_table(vi->rq[i].sg, ARRAY_SIZE(vi->rq[i].sg));
986a4f4d Jason Wang 2012-12-07 1317 sg_init_table(vi->sq[i].sg, ARRAY_SIZE(vi->sq[i].sg));
986a4f4d Jason Wang 2012-12-07 1318 }
986a4f4d Jason Wang 2012-12-07 1319
986a4f4d Jason Wang 2012-12-07 1320 return 0;
---
0-DAY kernel build testing backend Open Source Technology Center
Fengguang Wu, Yuanhan Liu Intel Corporation
^ permalink raw reply
* Re: [PATCH] ipv4: ip_check_defrag must not modify skb before unsharing
From: Eric Leblond @ 2012-12-10 11:02 UTC (permalink / raw)
To: Johannes Berg
Cc: David Miller, netdev, linux-wireless, linville, Eric Dumazet
In-Reply-To: <1355132466.9857.6.camel@jlt4.sipsolutions.net>
Hello,
On Mon, 2012-12-10 at 10:41 +0100, Johannes Berg wrote:
> From: Johannes Berg <johannes.berg@intel.com>
>
> ip_check_defrag() might be called from af_packet within the
> RX path where shared SKBs are used, so it must not modify
> the input SKB before it has unshared it for defragmentation.
> Use skb_copy_bits() to get the IP header and only pull in
> everything later.
>
> The same is true for the other caller in macvlan as it is
> called from dev->rx_handler which can also get a shared SKB.
I've applied the patch and built a new kernel. I did not manage to get
it crashed when using the two techniques (suspend to ram and down/up
interface) that were working well to crash kernel without the patch.
BR,
> Reported-by: Eric Leblond <eric@regit.org>
> Cc: stable@vger.kernel.org
> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
> ---
> For some versions of the kernel, this code goes into af_packet.c
>
> net/ipv4/ip_fragment.c | 19 +++++++++----------
> 1 file changed, 9 insertions(+), 10 deletions(-)
>
> diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
> index 448e685..8d5cc75 100644
> --- a/net/ipv4/ip_fragment.c
> +++ b/net/ipv4/ip_fragment.c
> @@ -707,28 +707,27 @@ EXPORT_SYMBOL(ip_defrag);
>
> struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
> {
> - const struct iphdr *iph;
> + struct iphdr iph;
> u32 len;
>
> if (skb->protocol != htons(ETH_P_IP))
> return skb;
>
> - if (!pskb_may_pull(skb, sizeof(struct iphdr)))
> + if (!skb_copy_bits(skb, 0, &iph, sizeof(iph)))
> return skb;
>
> - iph = ip_hdr(skb);
> - if (iph->ihl < 5 || iph->version != 4)
> + if (iph.ihl < 5 || iph.version != 4)
> return skb;
> - if (!pskb_may_pull(skb, iph->ihl*4))
> - return skb;
> - iph = ip_hdr(skb);
> - len = ntohs(iph->tot_len);
> - if (skb->len < len || len < (iph->ihl * 4))
> +
> + len = ntohs(iph.tot_len);
> + if (skb->len < len || len < (iph.ihl * 4))
> return skb;
>
> - if (ip_is_fragment(ip_hdr(skb))) {
> + if (ip_is_fragment(&iph)) {
> skb = skb_share_check(skb, GFP_ATOMIC);
> if (skb) {
> + if (!pskb_may_pull(skb, iph.ihl*4))
> + return skb;
> if (pskb_trim_rcsum(skb, len))
> return skb;
> memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
--
Eric Leblond <eric@regit.org>
Blog: https://home.regit.org/
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox