Merge tag 'mlx5-updates-2020-03-29' of git://git.kernel.org/pub/scm/l…

…inux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2020-03-29

1) mlx5 core level updates for mkey APIs + migrate some code to mlx5_ib

2) Use a separate work queue for fib event handling

3) Move new eswitch chains files to a new directory, to prepare for
upcoming E-Switch and offloads features.

4) Support indr block setup (TC_SETUP_FT) in Flow Table mode.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/infiniband/hw/mlx5/main.c

@@ -6392,6 +6392,7 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
 	spin_lock_init(&dev->reset_flow_resource_lock);
 	xa_init(&dev->odp_mkeys);
 	xa_init(&dev->sig_mrs);
+	spin_lock_init(&dev->mkey_lock);
 
 	spin_lock_init(&dev->dm.lock);
 	dev->dm.dev = mdev;

drivers/infiniband/hw/mlx5/mlx5_ib.h

@@ -994,6 +994,11 @@ struct mlx5_ib_dev {
 	/* sync used page count stats
 	 */
 	struct mlx5_ib_resources	devr;
+
+	/* protect mkey key part */
+	spinlock_t			mkey_lock;
+	u8				mkey_key;
+
 	struct mlx5_mr_cache		cache;
 	struct timer_list		delay_timer;
 	/* Prevents soft lock on massive reg MRs */

drivers/infiniband/hw/mlx5/mr.c

@@ -47,6 +47,46 @@ enum {
 
 #define MLX5_UMR_ALIGN 2048
 
+static void
+create_mkey_callback(int status, struct mlx5_async_work *context);
+
+static void
+assign_mkey_variant(struct mlx5_ib_dev *dev, struct mlx5_core_mkey *mkey,
+		    u32 *in)
+{
+	void *mkc;
+	u8 key;
+
+	spin_lock_irq(&dev->mkey_lock);
+	key = dev->mkey_key++;
+	spin_unlock_irq(&dev->mkey_lock);
+
+	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+	MLX5_SET(mkc, mkc, mkey_7_0, key);
+	mkey->key = key;
+}
+
+static int
+mlx5_ib_create_mkey(struct mlx5_ib_dev *dev, struct mlx5_core_mkey *mkey,
+		    u32 *in, int inlen)
+{
+	assign_mkey_variant(dev, mkey, in);
+	return mlx5_core_create_mkey(dev->mdev, mkey, in, inlen);
+}
+
+static int
+mlx5_ib_create_mkey_cb(struct mlx5_ib_dev *dev,
+		       struct mlx5_core_mkey *mkey,
+		       struct mlx5_async_ctx *async_ctx,
+		       u32 *in, int inlen, u32 *out, int outlen,
+		       struct mlx5_async_work *context)
+{
+	MLX5_SET(create_mkey_in, in, opcode, MLX5_CMD_OP_CREATE_MKEY);
+	assign_mkey_variant(dev, mkey, in);
+	return mlx5_cmd_exec_cb(async_ctx, in, inlen, out, outlen,
+				create_mkey_callback, context);
+}
+
 static void clean_mr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr);
 static void dereg_mr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr);
 static int mr_cache_max_order(struct mlx5_ib_dev *dev);
@@ -79,15 +119,14 @@ static bool use_umr_mtt_update(struct mlx5_ib_mr *mr, u64 start, u64 length)
 		length + (start & (MLX5_ADAPTER_PAGE_SIZE - 1));
 }
 
-static void reg_mr_callback(int status, struct mlx5_async_work *context)
+static void create_mkey_callback(int status, struct mlx5_async_work *context)
 {
 	struct mlx5_ib_mr *mr =
 		container_of(context, struct mlx5_ib_mr, cb_work);
 	struct mlx5_ib_dev *dev = mr->dev;
 	struct mlx5_mr_cache *cache = &dev->cache;
 	int c = order2idx(dev, mr->order);
 	struct mlx5_cache_ent *ent = &cache->ent[c];
-	u8 key;
 	unsigned long flags;
 
 	spin_lock_irqsave(&ent->lock, flags);
@@ -102,10 +141,8 @@ static void reg_mr_callback(int status, struct mlx5_async_work *context)
 	}
 
 	mr->mmkey.type = MLX5_MKEY_MR;
-	spin_lock_irqsave(&dev->mdev->priv.mkey_lock, flags);
-	key = dev->mdev->priv.mkey_key++;
-	spin_unlock_irqrestore(&dev->mdev->priv.mkey_lock, flags);
-	mr->mmkey.key = mlx5_idx_to_mkey(MLX5_GET(create_mkey_out, mr->out, mkey_index)) | key;
+	mr->mmkey.key |= mlx5_idx_to_mkey(
+		MLX5_GET(create_mkey_out, mr->out, mkey_index));
 
 	cache->last_add = jiffies;
 
@@ -163,10 +200,10 @@ static int add_keys(struct mlx5_ib_dev *dev, int c, int num)
 		spin_lock_irq(&ent->lock);
 		ent->pending++;
 		spin_unlock_irq(&ent->lock);
-		err = mlx5_core_create_mkey_cb(dev->mdev, &mr->mmkey,
+		err = mlx5_ib_create_mkey_cb(dev, &mr->mmkey,
 					       &dev->async_ctx, in, inlen,
 					       mr->out, sizeof(mr->out),
-					       reg_mr_callback, &mr->cb_work);
+					       &mr->cb_work);
 		if (err) {
 			spin_lock_irq(&ent->lock);
 			ent->pending--;
@@ -685,7 +722,6 @@ struct ib_mr *mlx5_ib_get_dma_mr(struct ib_pd *pd, int acc)
 {
 	struct mlx5_ib_dev *dev = to_mdev(pd->device);
 	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
-	struct mlx5_core_dev *mdev = dev->mdev;
 	struct mlx5_ib_mr *mr;
 	void *mkc;
 	u32 *in;
@@ -707,7 +743,7 @@ struct ib_mr *mlx5_ib_get_dma_mr(struct ib_pd *pd, int acc)
 	MLX5_SET(mkc, mkc, length64, 1);
 	set_mkc_access_pd_addr_fields(mkc, acc, 0, pd);
 
-	err = mlx5_core_create_mkey(mdev, &mr->mmkey, in, inlen);
+	err = mlx5_ib_create_mkey(dev, &mr->mmkey, in, inlen);
 	if (err)
 		goto err_in;
 
@@ -1097,7 +1133,7 @@ static struct mlx5_ib_mr *reg_create(struct ib_mr *ibmr, struct ib_pd *pd,
 			 get_octo_len(virt_addr, length, page_shift));
 	}
 
-	err = mlx5_core_create_mkey(dev->mdev, &mr->mmkey, in, inlen);
+	err = mlx5_ib_create_mkey(dev, &mr->mmkey, in, inlen);
 	if (err) {
 		mlx5_ib_warn(dev, "create mkey failed\n");
 		goto err_2;
@@ -1137,7 +1173,6 @@ static struct ib_mr *mlx5_ib_get_dm_mr(struct ib_pd *pd, u64 start_addr,
 {
 	struct mlx5_ib_dev *dev = to_mdev(pd->device);
 	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
-	struct mlx5_core_dev *mdev = dev->mdev;
 	struct mlx5_ib_mr *mr;
 	void *mkc;
 	u32 *in;
@@ -1160,7 +1195,7 @@ static struct ib_mr *mlx5_ib_get_dm_mr(struct ib_pd *pd, u64 start_addr,
 	MLX5_SET64(mkc, mkc, len, length);
 	set_mkc_access_pd_addr_fields(mkc, acc, start_addr, pd);
 
-	err = mlx5_core_create_mkey(mdev, &mr->mmkey, in, inlen);
+	err = mlx5_ib_create_mkey(dev, &mr->mmkey, in, inlen);
 	if (err)
 		goto err_in;
 
@@ -1638,7 +1673,7 @@ static int _mlx5_alloc_mkey_descs(struct ib_pd *pd, struct mlx5_ib_mr *mr,
 
 	mlx5_set_umr_free_mkey(pd, in, ndescs, access_mode, page_shift);
 
-	err = mlx5_core_create_mkey(dev->mdev, &mr->mmkey, in, inlen);
+	err = mlx5_ib_create_mkey(dev, &mr->mmkey, in, inlen);
 	if (err)
 		goto err_free_descs;
 
@@ -1905,7 +1940,7 @@ struct ib_mw *mlx5_ib_alloc_mw(struct ib_pd *pd, enum ib_mw_type type,
 	MLX5_SET(mkc, mkc, en_rinval, !!((type == IB_MW_TYPE_2)));
 	MLX5_SET(mkc, mkc, qpn, 0xffffff);
 
-	err = mlx5_core_create_mkey(dev->mdev, &mw->mmkey, in, inlen);
+	err = mlx5_ib_create_mkey(dev, &mw->mmkey, in, inlen);
 	if (err)
 		goto free;
 

drivers/net/ethernet/mellanox/mlx5/core/Makefile

@@ -43,7 +43,7 @@ mlx5_core-$(CONFIG_MLX5_TC_CT)	     += en/tc_ct.o
 # Core extra
 #
 mlx5_core-$(CONFIG_MLX5_ESWITCH)   += eswitch.o eswitch_offloads.o eswitch_offloads_termtbl.o \
-				      ecpf.o rdma.o eswitch_offloads_chains.o
+				      ecpf.o rdma.o esw/chains.o
 mlx5_core-$(CONFIG_MLX5_MPFS)      += lib/mpfs.o
 mlx5_core-$(CONFIG_VXLAN)          += lib/vxlan.o
 mlx5_core-$(CONFIG_PTP_1588_CLOCK) += lib/clock.o

drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c

@@ -13,11 +13,11 @@
 #include <net/netfilter/nf_flow_table.h>
 #include <linux/workqueue.h>
 
+#include "esw/chains.h"
 #include "en/tc_ct.h"
 #include "en.h"
 #include "en_tc.h"
 #include "en_rep.h"
-#include "eswitch_offloads_chains.h"
 
 #define MLX5_CT_ZONE_BITS (mlx5e_tc_attr_to_reg_mappings[ZONE_TO_REG].mlen * 8)
 #define MLX5_CT_ZONE_MASK GENMASK(MLX5_CT_ZONE_BITS - 1, 0)

drivers/net/ethernet/mellanox/mlx5/core/en_rep.c

@@ -41,7 +41,7 @@
 #include <net/ipv6_stubs.h>
 
 #include "eswitch.h"
-#include "eswitch_offloads_chains.h"
+#include "esw/chains.h"
 #include "en.h"
 #include "en_rep.h"
 #include "en_tc.h"
@@ -694,9 +694,9 @@ static void mlx5e_rep_indr_clean_block_privs(struct mlx5e_rep_priv *rpriv)
 static int
 mlx5e_rep_indr_offload(struct net_device *netdev,
 		       struct flow_cls_offload *flower,
-		       struct mlx5e_rep_indr_block_priv *indr_priv)
+		       struct mlx5e_rep_indr_block_priv *indr_priv,
+		       unsigned long flags)
 {
-	unsigned long flags = MLX5_TC_FLAG(EGRESS) | MLX5_TC_FLAG(ESW_OFFLOAD);
 	struct mlx5e_priv *priv = netdev_priv(indr_priv->rpriv->netdev);
 	int err = 0;
 
@@ -717,20 +717,68 @@ mlx5e_rep_indr_offload(struct net_device *netdev,
 	return err;
 }
 
-static int mlx5e_rep_indr_setup_block_cb(enum tc_setup_type type,
-					 void *type_data, void *indr_priv)
+static int mlx5e_rep_indr_setup_tc_cb(enum tc_setup_type type,
+				      void *type_data, void *indr_priv)
 {
+	unsigned long flags = MLX5_TC_FLAG(EGRESS) | MLX5_TC_FLAG(ESW_OFFLOAD);
 	struct mlx5e_rep_indr_block_priv *priv = indr_priv;
 
 	switch (type) {
 	case TC_SETUP_CLSFLOWER:
-		return mlx5e_rep_indr_offload(priv->netdev, type_data, priv);
+		return mlx5e_rep_indr_offload(priv->netdev, type_data, priv,
+					      flags);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_rep_indr_setup_ft_cb(enum tc_setup_type type,
+				      void *type_data, void *indr_priv)
+{
+	struct mlx5e_rep_indr_block_priv *priv = indr_priv;
+	struct flow_cls_offload *f = type_data;
+	struct flow_cls_offload tmp;
+	struct mlx5e_priv *mpriv;
+	struct mlx5_eswitch *esw;
+	unsigned long flags;
+	int err;
+
+	mpriv = netdev_priv(priv->rpriv->netdev);
+	esw = mpriv->mdev->priv.eswitch;
+
+	flags = MLX5_TC_FLAG(EGRESS) |
+		MLX5_TC_FLAG(ESW_OFFLOAD) |
+		MLX5_TC_FLAG(FT_OFFLOAD);
+
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		memcpy(&tmp, f, sizeof(*f));
+
+		/* Re-use tc offload path by moving the ft flow to the
+		 * reserved ft chain.
+		 *
+		 * FT offload can use prio range [0, INT_MAX], so we normalize
+		 * it to range [1, mlx5_esw_chains_get_prio_range(esw)]
+		 * as with tc, where prio 0 isn't supported.
+		 *
+		 * We only support chain 0 of FT offload.
+		 */
+		if (!mlx5_esw_chains_prios_supported(esw) ||
+		    tmp.common.prio >= mlx5_esw_chains_get_prio_range(esw) ||
+		    tmp.common.chain_index)
+			return -EOPNOTSUPP;
+
+		tmp.common.chain_index = mlx5_esw_chains_get_ft_chain(esw);
+		tmp.common.prio++;
+		err = mlx5e_rep_indr_offload(priv->netdev, &tmp, priv, flags);
+		memcpy(&f->stats, &tmp.stats, sizeof(f->stats));
+		return err;
 	default:
 		return -EOPNOTSUPP;
 	}
 }
 
-static void mlx5e_rep_indr_tc_block_unbind(void *cb_priv)
+static void mlx5e_rep_indr_block_unbind(void *cb_priv)
 {
 	struct mlx5e_rep_indr_block_priv *indr_priv = cb_priv;
 
@@ -741,9 +789,10 @@ static void mlx5e_rep_indr_tc_block_unbind(void *cb_priv)
 static LIST_HEAD(mlx5e_block_cb_list);
 
 static int
-mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
-			      struct mlx5e_rep_priv *rpriv,
-			      struct flow_block_offload *f)
+mlx5e_rep_indr_setup_block(struct net_device *netdev,
+			   struct mlx5e_rep_priv *rpriv,
+			   struct flow_block_offload *f,
+			   flow_setup_cb_t *setup_cb)
 {
 	struct mlx5e_rep_indr_block_priv *indr_priv;
 	struct flow_block_cb *block_cb;
@@ -769,9 +818,8 @@ mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
 		list_add(&indr_priv->list,
 			 &rpriv->uplink_priv.tc_indr_block_priv_list);
 
-		block_cb = flow_block_cb_alloc(mlx5e_rep_indr_setup_block_cb,
-					       indr_priv, indr_priv,
-					       mlx5e_rep_indr_tc_block_unbind);
+		block_cb = flow_block_cb_alloc(setup_cb, indr_priv, indr_priv,
+					       mlx5e_rep_indr_block_unbind);
 		if (IS_ERR(block_cb)) {
 			list_del(&indr_priv->list);
 			kfree(indr_priv);
@@ -786,9 +834,7 @@ mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
 		if (!indr_priv)
 			return -ENOENT;
 
-		block_cb = flow_block_cb_lookup(f->block,
-						mlx5e_rep_indr_setup_block_cb,
-						indr_priv);
+		block_cb = flow_block_cb_lookup(f->block, setup_cb, indr_priv);
 		if (!block_cb)
 			return -ENOENT;
 
@@ -802,13 +848,16 @@ mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
 }
 
 static
-int mlx5e_rep_indr_setup_tc_cb(struct net_device *netdev, void *cb_priv,
-			       enum tc_setup_type type, void *type_data)
+int mlx5e_rep_indr_setup_cb(struct net_device *netdev, void *cb_priv,
+			    enum tc_setup_type type, void *type_data)
 {
 	switch (type) {
 	case TC_SETUP_BLOCK:
-		return mlx5e_rep_indr_setup_tc_block(netdev, cb_priv,
-						      type_data);
+		return mlx5e_rep_indr_setup_block(netdev, cb_priv, type_data,
+						  mlx5e_rep_indr_setup_tc_cb);
+	case TC_SETUP_FT:
+		return mlx5e_rep_indr_setup_block(netdev, cb_priv, type_data,
+						  mlx5e_rep_indr_setup_ft_cb);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -820,7 +869,7 @@ static int mlx5e_rep_indr_register_block(struct mlx5e_rep_priv *rpriv,
 	int err;
 
 	err = __flow_indr_block_cb_register(netdev, rpriv,
-					    mlx5e_rep_indr_setup_tc_cb,
+					    mlx5e_rep_indr_setup_cb,
 					    rpriv);
 	if (err) {
 		struct mlx5e_priv *priv = netdev_priv(rpriv->netdev);
@@ -834,7 +883,7 @@ static int mlx5e_rep_indr_register_block(struct mlx5e_rep_priv *rpriv,
 static void mlx5e_rep_indr_unregister_block(struct mlx5e_rep_priv *rpriv,
 					    struct net_device *netdev)
 {
-	__flow_indr_block_cb_unregister(netdev, mlx5e_rep_indr_setup_tc_cb,
+	__flow_indr_block_cb_unregister(netdev, mlx5e_rep_indr_setup_cb,
 					rpriv);
 }
 

drivers/net/ethernet/mellanox/mlx5/core/en_tc.c

@@ -51,7 +51,7 @@
 #include "en_rep.h"
 #include "en_tc.h"
 #include "eswitch.h"
-#include "eswitch_offloads_chains.h"
+#include "esw/chains.h"
 #include "fs_core.h"
 #include "en/port.h"
 #include "en/tc_tun.h"

drivers/net/ethernet/mellanox/mlx5/core/esw/Makefile

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+subdir-ccflags-y += -I$(src)/..

drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c → drivers/net/ethernet/mellanox/mlx5/core/esw/chains.c

@@ -5,7 +5,7 @@
 #include <linux/mlx5/mlx5_ifc.h>
 #include <linux/mlx5/fs.h>
 
-#include "eswitch_offloads_chains.h"
+#include "esw/chains.h"
 #include "en/mapping.h"
 #include "mlx5_core.h"
 #include "fs_core.h"

drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.h → drivers/net/ethernet/mellanox/mlx5/core/esw/chains.h

@@ -4,6 +4,8 @@
 #ifndef __ML5_ESW_CHAINS_H__
 #define __ML5_ESW_CHAINS_H__
 
+#include "eswitch.h"
+
 bool
 mlx5_esw_chains_prios_supported(struct mlx5_eswitch *esw);
 bool