net/mlx5: Introduce mapping infra for mapping unique ids to data

Add a new interface for mapping data to a given id range (max_id),
and back again. It uses xarray as the id allocator and for finding a
given id. For locking it uses xa_lock (spin_lock) for add()/del(),
and rcu_read_lock for find().

This mapping interface also supports delaying the mapping removal via
a workqueue. This is for cases where we need the mapping to have
some grace period in regards to finding it back again, for example
for packets arriving from hardware that were marked with by a rule
with an old mapping that no longer exists.

Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

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

@@ -34,7 +34,7 @@ mlx5_core-$(CONFIG_MLX5_EN_ARFS)     += en_arfs.o
 mlx5_core-$(CONFIG_MLX5_EN_RXNFC)    += en_fs_ethtool.o
 mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) += en_dcbnl.o en/port_buffer.o
 mlx5_core-$(CONFIG_MLX5_ESWITCH)     += en_rep.o en_tc.o en/tc_tun.o lib/port_tun.o lag_mp.o \
-					lib/geneve.o en/tc_tun_vxlan.o en/tc_tun_gre.o \
+					lib/geneve.o en/mapping.o en/tc_tun_vxlan.o en/tc_tun_gre.o \
 					en/tc_tun_geneve.o diag/en_tc_tracepoint.o
 mlx5_core-$(CONFIG_PCI_HYPERV_INTERFACE) += en/hv_vhca_stats.o
 

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

@@ -0,0 +1,218 @@
+// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
+/* Copyright (c) 2018 Mellanox Technologies */
+
+#include <linux/jhash.h>
+#include <linux/slab.h>
+#include <linux/xarray.h>
+#include <linux/hashtable.h>
+
+#include "mapping.h"
+
+#define MAPPING_GRACE_PERIOD 2000
+
+struct mapping_ctx {
+	struct xarray xarray;
+	DECLARE_HASHTABLE(ht, 8);
+	struct mutex lock; /* Guards hashtable and xarray */
+	unsigned long max_id;
+	size_t data_size;
+	bool delayed_removal;
+	struct delayed_work dwork;
+	struct list_head pending_list;
+	spinlock_t pending_list_lock; /* Guards pending list */
+};
+
+struct mapping_item {
+	struct rcu_head rcu;
+	struct list_head list;
+	unsigned long timeout;
+	struct hlist_node node;
+	int cnt;
+	u32 id;
+	char data[];
+};
+
+int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id)
+{
+	struct mapping_item *mi;
+	int err = -ENOMEM;
+	u32 hash_key;
+
+	mutex_lock(&ctx->lock);
+
+	hash_key = jhash(data, ctx->data_size, 0);
+	hash_for_each_possible(ctx->ht, mi, node, hash_key) {
+		if (!memcmp(data, mi->data, ctx->data_size))
+			goto attach;
+	}
+
+	mi = kzalloc(sizeof(*mi) + ctx->data_size, GFP_KERNEL);
+	if (!mi)
+		goto err_alloc;
+
+	memcpy(mi->data, data, ctx->data_size);
+	hash_add(ctx->ht, &mi->node, hash_key);
+
+	err = xa_alloc(&ctx->xarray, &mi->id, mi, XA_LIMIT(1, ctx->max_id),
+		       GFP_KERNEL);
+	if (err)
+		goto err_assign;
+attach:
+	++mi->cnt;
+	*id = mi->id;
+
+	mutex_unlock(&ctx->lock);
+
+	return 0;
+
+err_assign:
+	hash_del(&mi->node);
+	kfree(mi);
+err_alloc:
+	mutex_unlock(&ctx->lock);
+
+	return err;
+}
+
+static void mapping_remove_and_free(struct mapping_ctx *ctx,
+				    struct mapping_item *mi)
+{
+	xa_erase(&ctx->xarray, mi->id);
+	kfree_rcu(mi, rcu);
+}
+
+static void mapping_free_item(struct mapping_ctx *ctx,
+			      struct mapping_item *mi)
+{
+	if (!ctx->delayed_removal) {
+		mapping_remove_and_free(ctx, mi);
+		return;
+	}
+
+	mi->timeout = jiffies + msecs_to_jiffies(MAPPING_GRACE_PERIOD);
+
+	spin_lock(&ctx->pending_list_lock);
+	list_add_tail(&mi->list, &ctx->pending_list);
+	spin_unlock(&ctx->pending_list_lock);
+
+	schedule_delayed_work(&ctx->dwork, MAPPING_GRACE_PERIOD);
+}
+
+int mapping_remove(struct mapping_ctx *ctx, u32 id)
+{
+	unsigned long index = id;
+	struct mapping_item *mi;
+	int err = -ENOENT;
+
+	mutex_lock(&ctx->lock);
+	mi = xa_load(&ctx->xarray, index);
+	if (!mi)
+		goto out;
+	err = 0;
+
+	if (--mi->cnt > 0)
+		goto out;
+
+	hash_del(&mi->node);
+	mapping_free_item(ctx, mi);
+out:
+	mutex_unlock(&ctx->lock);
+
+	return err;
+}
+
+int mapping_find(struct mapping_ctx *ctx, u32 id, void *data)
+{
+	unsigned long index = id;
+	struct mapping_item *mi;
+	int err = -ENOENT;
+
+	rcu_read_lock();
+	mi = xa_load(&ctx->xarray, index);
+	if (!mi)
+		goto err_find;
+
+	memcpy(data, mi->data, ctx->data_size);
+	err = 0;
+
+err_find:
+	rcu_read_unlock();
+	return err;
+}
+
+static void
+mapping_remove_and_free_list(struct mapping_ctx *ctx, struct list_head *list)
+{
+	struct mapping_item *mi;
+
+	list_for_each_entry(mi, list, list)
+		mapping_remove_and_free(ctx, mi);
+}
+
+static void mapping_work_handler(struct work_struct *work)
+{
+	unsigned long min_timeout = 0, now = jiffies;
+	struct mapping_item *mi, *next;
+	LIST_HEAD(pending_items);
+	struct mapping_ctx *ctx;
+
+	ctx = container_of(work, struct mapping_ctx, dwork.work);
+
+	spin_lock(&ctx->pending_list_lock);
+	list_for_each_entry_safe(mi, next, &ctx->pending_list, list) {
+		if (time_after(now, mi->timeout))
+			list_move(&mi->list, &pending_items);
+		else if (!min_timeout ||
+			 time_before(mi->timeout, min_timeout))
+			min_timeout = mi->timeout;
+	}
+	spin_unlock(&ctx->pending_list_lock);
+
+	mapping_remove_and_free_list(ctx, &pending_items);
+
+	if (min_timeout)
+		schedule_delayed_work(&ctx->dwork, abs(min_timeout - now));
+}
+
+static void mapping_flush_work(struct mapping_ctx *ctx)
+{
+	if (!ctx->delayed_removal)
+		return;
+
+	cancel_delayed_work_sync(&ctx->dwork);
+	mapping_remove_and_free_list(ctx, &ctx->pending_list);
+}
+
+struct mapping_ctx *
+mapping_create(size_t data_size, u32 max_id, bool delayed_removal)
+{
+	struct mapping_ctx *ctx;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return ERR_PTR(-ENOMEM);
+
+	ctx->max_id = max_id ? max_id : UINT_MAX;
+	ctx->data_size = data_size;
+
+	if (delayed_removal) {
+		INIT_DELAYED_WORK(&ctx->dwork, mapping_work_handler);
+		INIT_LIST_HEAD(&ctx->pending_list);
+		spin_lock_init(&ctx->pending_list_lock);
+		ctx->delayed_removal = true;
+	}
+
+	mutex_init(&ctx->lock);
+	xa_init_flags(&ctx->xarray, XA_FLAGS_ALLOC1);
+
+	return ctx;
+}
+
+void mapping_destroy(struct mapping_ctx *ctx)
+{
+	mapping_flush_work(ctx);
+	xa_destroy(&ctx->xarray);
+	mutex_destroy(&ctx->lock);
+
+	kfree(ctx);
+}

drivers/net/ethernet/mellanox/mlx5/core/en/mapping.h

@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2019 Mellanox Technologies */
+
+#ifndef __MLX5_MAPPING_H__
+#define __MLX5_MAPPING_H__
+
+struct mapping_ctx;
+
+int mapping_add(struct mapping_ctx *ctx, void *data, u32 *id);
+int mapping_remove(struct mapping_ctx *ctx, u32 id);
+int mapping_find(struct mapping_ctx *ctx, u32 id, void *data);
+
+/* mapping uses an xarray to map data to ids in add(), and for find().
+ * For locking, it uses a internal xarray spin lock for add()/remove(),
+ * find() uses rcu_read_lock().
+ * Choosing delayed_removal postpones the removal of a previously mapped
+ * id by MAPPING_GRACE_PERIOD milliseconds.
+ * This is to avoid races against hardware, where we mark the packet in
+ * hardware with a previous id, and quick remove() and add() reusing the same
+ * previous id. Then find() will get the new mapping instead of the old
+ * which was used to mark the packet.
+ */
+struct mapping_ctx *mapping_create(size_t data_size, u32 max_id,
+				   bool delayed_removal);
+void mapping_destroy(struct mapping_ctx *ctx);
+
+#endif /* __MLX5_MAPPING_H__ */