* [PATCH net-next 11/17] mlxsw: spectrum_acl: Add A-TCAM initialization
From: Ido Schimmel @ 2018-07-25 6:24 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
Initialize the A-TCAM as part of the driver's initialization routine.
Specifically, initialize the eRP tables so that A-TCAM regions will be
able to perform allocations of eRP tables upon rule insertion in
subsequent patches.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
.../net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c | 7 +++++++
.../net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c | 12 ++++++++++++
.../net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h | 4 ++++
3 files changed, 23 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c
index 33787b154e74..2442decd0652 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c
@@ -39,6 +39,7 @@
#include "core_acl_flex_actions.h"
struct mlxsw_sp2_acl_tcam {
+ struct mlxsw_sp_acl_atcam atcam;
u32 kvdl_index;
unsigned int kvdl_count;
};
@@ -100,9 +101,14 @@ static int mlxsw_sp2_acl_tcam_init(struct mlxsw_sp *mlxsw_sp, void *priv,
if (err)
goto err_pgcr_write;
+ err = mlxsw_sp_acl_atcam_init(mlxsw_sp, &tcam->atcam);
+ if (err)
+ goto err_atcam_init;
+
mlxsw_afa_block_destroy(afa_block);
return 0;
+err_atcam_init:
err_pgcr_write:
err_pefa_write:
err_afa_block_continue:
@@ -117,6 +123,7 @@ static void mlxsw_sp2_acl_tcam_fini(struct mlxsw_sp *mlxsw_sp, void *priv)
{
struct mlxsw_sp2_acl_tcam *tcam = priv;
+ mlxsw_sp_acl_atcam_fini(mlxsw_sp, &tcam->atcam);
mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ACTSET,
tcam->kvdl_count, tcam->kvdl_index);
}
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c
index a27d3b0f9fcb..89c78c62e7e5 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c
@@ -93,3 +93,15 @@ int mlxsw_sp_acl_atcam_region_init(struct mlxsw_sp *mlxsw_sp,
return 0;
}
+
+int mlxsw_sp_acl_atcam_init(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam)
+{
+ return mlxsw_sp_acl_erps_init(mlxsw_sp, atcam);
+}
+
+void mlxsw_sp_acl_atcam_fini(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam)
+{
+ mlxsw_sp_acl_erps_fini(mlxsw_sp, atcam);
+}
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
index 70094936ca43..ac6bdffd99a7 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
@@ -173,6 +173,10 @@ int mlxsw_sp_acl_atcam_region_associate(struct mlxsw_sp *mlxsw_sp,
u16 region_id);
int mlxsw_sp_acl_atcam_region_init(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl_tcam_region *region);
+int mlxsw_sp_acl_atcam_init(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam);
+void mlxsw_sp_acl_atcam_fini(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam);
struct mlxsw_sp_acl_erp;
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 10/17] mlxsw: spectrum_acl: Allow encoding a partial key
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
When working with 12 key blocks in the A-TCAM, rules are split into two
records, which constitute two lookups. The two records are linked using
a "large entry key ID". The ID is assigned to key blocks 6 to 11 and
resolved during the first lookup. The second lookup is performed using
the ID and the remaining key blocks.
Allow encoding a partial key so that it can be later used to check if an
ID can be reused.
This is done by adding two arguments to the existing encode function
that specify the range of the block indexes we would like to encode. The
key and mask arguments become optional, as we will not need to encode
both of them all the time.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
.../net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c | 10 ++++++----
.../net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h | 2 +-
.../net/ethernet/mellanox/mlxsw/spectrum_acl_ctcam.c | 5 ++++-
3 files changed, 11 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c
index 5f8485c7640e..9649b4d9349a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c
@@ -457,7 +457,7 @@ mlxsw_sp_afk_encode_one(const struct mlxsw_afk_element_inst *elinst,
void mlxsw_afk_encode(struct mlxsw_afk *mlxsw_afk,
struct mlxsw_afk_key_info *key_info,
struct mlxsw_afk_element_values *values,
- char *key, char *mask)
+ char *key, char *mask, int block_start, int block_end)
{
char block_mask[MLXSW_SP_AFK_KEY_BLOCK_MAX_SIZE];
char block_key[MLXSW_SP_AFK_KEY_BLOCK_MAX_SIZE];
@@ -465,7 +465,7 @@ void mlxsw_afk_encode(struct mlxsw_afk *mlxsw_afk,
enum mlxsw_afk_element element;
int block_index, i;
- for (i = 0; i < key_info->blocks_count; i++) {
+ for (i = block_start; i <= block_end; i++) {
memset(block_key, 0, MLXSW_SP_AFK_KEY_BLOCK_MAX_SIZE);
memset(block_mask, 0, MLXSW_SP_AFK_KEY_BLOCK_MAX_SIZE);
@@ -482,8 +482,10 @@ void mlxsw_afk_encode(struct mlxsw_afk *mlxsw_afk,
values->storage.mask);
}
- mlxsw_afk->ops->encode_block(block_key, i, key);
- mlxsw_afk->ops->encode_block(block_mask, i, mask);
+ if (key)
+ mlxsw_afk->ops->encode_block(block_key, i, key);
+ if (mask)
+ mlxsw_afk->ops->encode_block(block_mask, i, mask);
}
}
EXPORT_SYMBOL(mlxsw_afk_encode);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h
index 2ffde915349b..18d9bfed6001 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h
@@ -259,6 +259,6 @@ void mlxsw_afk_values_add_buf(struct mlxsw_afk_element_values *values,
void mlxsw_afk_encode(struct mlxsw_afk *mlxsw_afk,
struct mlxsw_afk_key_info *key_info,
struct mlxsw_afk_element_values *values,
- char *key, char *mask);
+ char *key, char *mask, int block_start, int block_end);
#endif
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_ctcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_ctcam.c
index ef0d4c0a5a1f..11b19272ab13 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_ctcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_ctcam.c
@@ -76,6 +76,7 @@ mlxsw_sp_acl_ctcam_region_entry_insert(struct mlxsw_sp *mlxsw_sp,
{
struct mlxsw_afk *afk = mlxsw_sp_acl_afk(mlxsw_sp->acl);
char ptce2_pl[MLXSW_REG_PTCE2_LEN];
+ unsigned int blocks_count;
char *act_set;
u32 priority;
char *mask;
@@ -91,7 +92,9 @@ mlxsw_sp_acl_ctcam_region_entry_insert(struct mlxsw_sp *mlxsw_sp,
region->tcam_region_info, offset, priority);
key = mlxsw_reg_ptce2_flex_key_blocks_data(ptce2_pl);
mask = mlxsw_reg_ptce2_mask_data(ptce2_pl);
- mlxsw_afk_encode(afk, region->key_info, &rulei->values, key, mask);
+ blocks_count = mlxsw_afk_key_info_blocks_count_get(region->key_info);
+ mlxsw_afk_encode(afk, region->key_info, &rulei->values, key, mask, 0,
+ blocks_count - 1);
/* Only the first action set belongs here, the rest is in KVD */
act_set = mlxsw_afa_block_first_set(rulei->act_block);
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 09/17] mlxsw: spectrum_acl: Extend Spectrum-2 region struct
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
In a similar fashion to Spectrum-1's region struct, Spectrum-2's struct
needs to store a pointer to the common region struct.
The pointer will be used in follow-up patches that implement rules
insertion and deletion.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c
index d7f1fb35ea2a..33787b154e74 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum2_acl_tcam.c
@@ -45,6 +45,7 @@ struct mlxsw_sp2_acl_tcam {
struct mlxsw_sp2_acl_tcam_region {
struct mlxsw_sp_acl_ctcam_region cregion;
+ struct mlxsw_sp_acl_tcam_region *region;
};
struct mlxsw_sp2_acl_tcam_chunk {
@@ -127,6 +128,8 @@ mlxsw_sp2_acl_tcam_region_init(struct mlxsw_sp *mlxsw_sp, void *region_priv,
struct mlxsw_sp2_acl_tcam_region *region = region_priv;
int err;
+ region->region = _region;
+
err = mlxsw_sp_acl_atcam_region_init(mlxsw_sp, _region);
if (err)
return err;
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 06/17] mlxsw: spectrum_acl: Implement common eRP core
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
When rules are inserted into the A-TCAM they are associated with a mask,
which is part of the lookup key: { masked key, mask ID, region ID }.
These masks are called rule patterns (RP) and the aggregation of several
masks into one (to be introduced in follow-up patch sets) is called an
extended RP (eRP).
When a packet undergoes a lookup in an ACL region it is masked by the
current set of eRPs used by the region, looking for an exact match.
Eventually, the rule with the highest priority is picked.
These eRPs are stored in several global banks to allow for lookup to
occur using several eRPs simultaneously.
At first, an ACL region will only require a single mask - upon the
insertion of the first rule. In this case, the region can use the
"master RP" which is composed by OR-ing all the masks used by the
region. This mask is a property of the region and thus there is no need
to use the above mentioned banks.
At some point, a second mask will be needed. In this case, the region
will need to allocate an eRP table from the above mentioned banks and
insert its masks there.
>From now on, upon lookup, the eRP table used by the region will be
fetched from the eRP banks - using {eRP bank, Index within the bank} -
and the eRPs present in the table will be used to mask the packet. Note
that masks with consecutive indexes are inserted into consecutive banks.
When rules are deleted and a region only needs a single mask once again
it can free its eRP table and use the master RP.
The above logic is implemented in the eRP core and represented using the
following state machine:
+------------+ create mask - as master RP +---------------+
| +--------------------------------> |
| no masks | | single mask |
| <--------------------------------+ |
+------------+ delete mask +-----+--^------+
| |
| |
create mask - | | delete mask -
create mask transition to use eRP | | transition to
+--------+ table | | use master RP
| | | |
| | | |
+----v--------+----+ create mask +----v--+-----+
| <-------------------------------+ |
| multiple masks | | two masks |
| +-------------------------------> |
+------------------+ delete mask - if two +-------------+
remaining
The code that actually configures rules in the A-TCAM will interface
with the eRP core by getting or putting an eRP based on the required
mask used by the rule.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/Kconfig | 1 +
drivers/net/ethernet/mellanox/mlxsw/Makefile | 2 +-
.../mellanox/mlxsw/spectrum_acl_erp.c | 1019 +++++++++++++++++
.../mellanox/mlxsw/spectrum_acl_tcam.h | 40 +
4 files changed, 1061 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
diff --git a/drivers/net/ethernet/mellanox/mlxsw/Kconfig b/drivers/net/ethernet/mellanox/mlxsw/Kconfig
index 82827a8d3d67..8a291eb36c64 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/Kconfig
+++ b/drivers/net/ethernet/mellanox/mlxsw/Kconfig
@@ -78,6 +78,7 @@ config MLXSW_SPECTRUM
depends on IPV6 || IPV6=n
depends on NET_IPGRE || NET_IPGRE=n
depends on IPV6_GRE || IPV6_GRE=n
+ select GENERIC_ALLOCATOR
select PARMAN
select MLXFW
default m
diff --git a/drivers/net/ethernet/mellanox/mlxsw/Makefile b/drivers/net/ethernet/mellanox/mlxsw/Makefile
index 4a1069166119..68fa44a41485 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/Makefile
+++ b/drivers/net/ethernet/mellanox/mlxsw/Makefile
@@ -18,7 +18,7 @@ mlxsw_spectrum-objs := spectrum.o spectrum_buffers.o \
spectrum1_kvdl.o spectrum2_kvdl.o \
spectrum_kvdl.o \
spectrum_acl_tcam.o spectrum_acl_ctcam.o \
- spectrum_acl_atcam.o \
+ spectrum_acl_atcam.o spectrum_acl_erp.o \
spectrum1_acl_tcam.o spectrum2_acl_tcam.o \
spectrum_acl.o \
spectrum_flower.o spectrum_cnt.o \
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
new file mode 100644
index 000000000000..960f29140b43
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
@@ -0,0 +1,1019 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
+ * Copyright (c) 2018 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2018 Ido Schimmel <idosch@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. Neither the names of the copyright holders nor the names of its
+ * contributors may be used to endorse or promote products derived from
+ * this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/bitmap.h>
+#include <linux/errno.h>
+#include <linux/genalloc.h>
+#include <linux/gfp.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/rhashtable.h>
+#include <linux/rtnetlink.h>
+#include <linux/slab.h>
+
+#include "core.h"
+#include "reg.h"
+#include "spectrum.h"
+#include "spectrum_acl_tcam.h"
+
+/* gen_pool_alloc() returns 0 when allocation fails, so use an offset */
+#define MLXSW_SP_ACL_ERP_GENALLOC_OFFSET 0x100
+#define MLXSW_SP_ACL_ERP_MAX_PER_REGION 16
+
+struct mlxsw_sp_acl_erp_core {
+ unsigned int erpt_entries_size[MLXSW_SP_ACL_ATCAM_REGION_TYPE_MAX + 1];
+ struct gen_pool *erp_tables;
+ struct mlxsw_sp *mlxsw_sp;
+ unsigned int num_erp_banks;
+};
+
+struct mlxsw_sp_acl_erp_key {
+ char mask[MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN];
+};
+
+struct mlxsw_sp_acl_erp {
+ struct mlxsw_sp_acl_erp_key key;
+ u8 id;
+ u8 index;
+ refcount_t refcnt;
+ DECLARE_BITMAP(mask_bitmap, MLXSW_SP_ACL_TCAM_MASK_LEN);
+ struct list_head list;
+ struct rhash_head ht_node;
+ struct mlxsw_sp_acl_erp_table *erp_table;
+};
+
+struct mlxsw_sp_acl_erp_master_mask {
+ DECLARE_BITMAP(bitmap, MLXSW_SP_ACL_TCAM_MASK_LEN);
+ unsigned int count[MLXSW_SP_ACL_TCAM_MASK_LEN];
+};
+
+struct mlxsw_sp_acl_erp_table {
+ struct mlxsw_sp_acl_erp_master_mask master_mask;
+ DECLARE_BITMAP(erp_id_bitmap, MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ DECLARE_BITMAP(erp_index_bitmap, MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ struct list_head atcam_erps_list;
+ struct rhashtable erp_ht;
+ struct mlxsw_sp_acl_erp_core *erp_core;
+ struct mlxsw_sp_acl_atcam_region *aregion;
+ const struct mlxsw_sp_acl_erp_table_ops *ops;
+ unsigned long base_index;
+ unsigned int num_atcam_erps;
+ unsigned int num_max_atcam_erps;
+};
+
+static const struct rhashtable_params mlxsw_sp_acl_erp_ht_params = {
+ .key_len = sizeof(struct mlxsw_sp_acl_erp_key),
+ .key_offset = offsetof(struct mlxsw_sp_acl_erp, key),
+ .head_offset = offsetof(struct mlxsw_sp_acl_erp, ht_node),
+};
+
+struct mlxsw_sp_acl_erp_table_ops {
+ struct mlxsw_sp_acl_erp *
+ (*erp_create)(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key);
+ void (*erp_destroy)(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp);
+};
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key);
+static void
+mlxsw_sp_acl_erp_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp);
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_second_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key);
+static void
+mlxsw_sp_acl_erp_second_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp);
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_first_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key);
+static void
+mlxsw_sp_acl_erp_first_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp);
+static void
+mlxsw_sp_acl_erp_no_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp);
+
+static const struct mlxsw_sp_acl_erp_table_ops erp_multiple_masks_ops = {
+ .erp_create = mlxsw_sp_acl_erp_mask_create,
+ .erp_destroy = mlxsw_sp_acl_erp_mask_destroy,
+};
+
+static const struct mlxsw_sp_acl_erp_table_ops erp_two_masks_ops = {
+ .erp_create = mlxsw_sp_acl_erp_mask_create,
+ .erp_destroy = mlxsw_sp_acl_erp_second_mask_destroy,
+};
+
+static const struct mlxsw_sp_acl_erp_table_ops erp_single_mask_ops = {
+ .erp_create = mlxsw_sp_acl_erp_second_mask_create,
+ .erp_destroy = mlxsw_sp_acl_erp_first_mask_destroy,
+};
+
+static const struct mlxsw_sp_acl_erp_table_ops erp_no_mask_ops = {
+ .erp_create = mlxsw_sp_acl_erp_first_mask_create,
+ .erp_destroy = mlxsw_sp_acl_erp_no_mask_destroy,
+};
+
+u8 mlxsw_sp_acl_erp_id(const struct mlxsw_sp_acl_erp *erp)
+{
+ return erp->id;
+}
+
+static unsigned int
+mlxsw_sp_acl_erp_table_entry_size(const struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_atcam_region *aregion = erp_table->aregion;
+ struct mlxsw_sp_acl_erp_core *erp_core = erp_table->erp_core;
+
+ return erp_core->erpt_entries_size[aregion->type];
+}
+
+static int mlxsw_sp_acl_erp_id_get(struct mlxsw_sp_acl_erp_table *erp_table,
+ u8 *p_id)
+{
+ u8 id;
+
+ id = find_first_zero_bit(erp_table->erp_id_bitmap,
+ MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ if (id < MLXSW_SP_ACL_ERP_MAX_PER_REGION) {
+ __set_bit(id, erp_table->erp_id_bitmap);
+ *p_id = id;
+ return 0;
+ }
+
+ return -ENOBUFS;
+}
+
+static void mlxsw_sp_acl_erp_id_put(struct mlxsw_sp_acl_erp_table *erp_table,
+ u8 id)
+{
+ __clear_bit(id, erp_table->erp_id_bitmap);
+}
+
+static void
+mlxsw_sp_acl_erp_master_mask_bit_set(unsigned long bit,
+ struct mlxsw_sp_acl_erp_master_mask *mask)
+{
+ if (mask->count[bit]++ == 0)
+ __set_bit(bit, mask->bitmap);
+}
+
+static void
+mlxsw_sp_acl_erp_master_mask_bit_clear(unsigned long bit,
+ struct mlxsw_sp_acl_erp_master_mask *mask)
+{
+ if (--mask->count[bit] == 0)
+ __clear_bit(bit, mask->bitmap);
+}
+
+static int
+mlxsw_sp_acl_erp_master_mask_update(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
+ struct mlxsw_sp *mlxsw_sp = region->mlxsw_sp;
+ char percr_pl[MLXSW_REG_PERCR_LEN];
+ char *master_mask;
+
+ mlxsw_reg_percr_pack(percr_pl, region->id);
+ master_mask = mlxsw_reg_percr_master_mask_data(percr_pl);
+ bitmap_to_arr32((u32 *) master_mask, erp_table->master_mask.bitmap,
+ MLXSW_SP_ACL_TCAM_MASK_LEN);
+
+ return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(percr), percr_pl);
+}
+
+static int
+mlxsw_sp_acl_erp_master_mask_set(struct mlxsw_sp_acl_erp_table *erp_table,
+ const struct mlxsw_sp_acl_erp *erp)
+{
+ unsigned long bit;
+ int err;
+
+ for_each_set_bit(bit, erp->mask_bitmap, MLXSW_SP_ACL_TCAM_MASK_LEN)
+ mlxsw_sp_acl_erp_master_mask_bit_set(bit,
+ &erp_table->master_mask);
+
+ err = mlxsw_sp_acl_erp_master_mask_update(erp_table);
+ if (err)
+ goto err_master_mask_update;
+
+ return 0;
+
+err_master_mask_update:
+ for_each_set_bit(bit, erp->mask_bitmap, MLXSW_SP_ACL_TCAM_MASK_LEN)
+ mlxsw_sp_acl_erp_master_mask_bit_clear(bit,
+ &erp_table->master_mask);
+ return err;
+}
+
+static int
+mlxsw_sp_acl_erp_master_mask_clear(struct mlxsw_sp_acl_erp_table *erp_table,
+ const struct mlxsw_sp_acl_erp *erp)
+{
+ unsigned long bit;
+ int err;
+
+ for_each_set_bit(bit, erp->mask_bitmap, MLXSW_SP_ACL_TCAM_MASK_LEN)
+ mlxsw_sp_acl_erp_master_mask_bit_clear(bit,
+ &erp_table->master_mask);
+
+ err = mlxsw_sp_acl_erp_master_mask_update(erp_table);
+ if (err)
+ goto err_master_mask_update;
+
+ return 0;
+
+err_master_mask_update:
+ for_each_set_bit(bit, erp->mask_bitmap, MLXSW_SP_ACL_TCAM_MASK_LEN)
+ mlxsw_sp_acl_erp_master_mask_bit_set(bit,
+ &erp_table->master_mask);
+ return err;
+}
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_generic_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key)
+{
+ struct mlxsw_sp_acl_erp *erp;
+ int err;
+
+ erp = kzalloc(sizeof(*erp), GFP_KERNEL);
+ if (!erp)
+ return ERR_PTR(-ENOMEM);
+
+ err = mlxsw_sp_acl_erp_id_get(erp_table, &erp->id);
+ if (err)
+ goto err_erp_id_get;
+
+ memcpy(&erp->key, key, sizeof(*key));
+ bitmap_from_arr32(erp->mask_bitmap, (u32 *) key->mask,
+ MLXSW_SP_ACL_TCAM_MASK_LEN);
+ list_add(&erp->list, &erp_table->atcam_erps_list);
+ refcount_set(&erp->refcnt, 1);
+ erp_table->num_atcam_erps++;
+ erp->erp_table = erp_table;
+
+ err = mlxsw_sp_acl_erp_master_mask_set(erp_table, erp);
+ if (err)
+ goto err_master_mask_set;
+
+ err = rhashtable_insert_fast(&erp_table->erp_ht, &erp->ht_node,
+ mlxsw_sp_acl_erp_ht_params);
+ if (err)
+ goto err_rhashtable_insert;
+
+ return erp;
+
+err_rhashtable_insert:
+ mlxsw_sp_acl_erp_master_mask_clear(erp_table, erp);
+err_master_mask_set:
+ erp_table->num_atcam_erps--;
+ list_del(&erp->list);
+ mlxsw_sp_acl_erp_id_put(erp_table, erp->id);
+err_erp_id_get:
+ kfree(erp);
+ return ERR_PTR(err);
+}
+
+static void
+mlxsw_sp_acl_erp_generic_destroy(struct mlxsw_sp_acl_erp *erp)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table = erp->erp_table;
+
+ rhashtable_remove_fast(&erp_table->erp_ht, &erp->ht_node,
+ mlxsw_sp_acl_erp_ht_params);
+ mlxsw_sp_acl_erp_master_mask_clear(erp_table, erp);
+ erp_table->num_atcam_erps--;
+ list_del(&erp->list);
+ mlxsw_sp_acl_erp_id_put(erp_table, erp->id);
+ kfree(erp);
+}
+
+static int
+mlxsw_sp_acl_erp_table_alloc(struct mlxsw_sp_acl_erp_core *erp_core,
+ unsigned int num_erps,
+ enum mlxsw_sp_acl_atcam_region_type region_type,
+ unsigned long *p_index)
+{
+ unsigned int num_rows, entry_size;
+
+ /* We only allow allocations of entire rows */
+ if (num_erps % erp_core->num_erp_banks != 0)
+ return -EINVAL;
+
+ entry_size = erp_core->erpt_entries_size[region_type];
+ num_rows = num_erps / erp_core->num_erp_banks;
+
+ *p_index = gen_pool_alloc(erp_core->erp_tables, num_rows * entry_size);
+ if (*p_index == 0)
+ return -ENOBUFS;
+ *p_index -= MLXSW_SP_ACL_ERP_GENALLOC_OFFSET;
+
+ return 0;
+}
+
+static void
+mlxsw_sp_acl_erp_table_free(struct mlxsw_sp_acl_erp_core *erp_core,
+ unsigned int num_erps,
+ enum mlxsw_sp_acl_atcam_region_type region_type,
+ unsigned long index)
+{
+ unsigned long base_index;
+ unsigned int entry_size;
+ size_t size;
+
+ entry_size = erp_core->erpt_entries_size[region_type];
+ base_index = index + MLXSW_SP_ACL_ERP_GENALLOC_OFFSET;
+ size = num_erps / erp_core->num_erp_banks * entry_size;
+ gen_pool_free(erp_core->erp_tables, base_index, size);
+}
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_table_master_rp(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ if (!list_is_singular(&erp_table->atcam_erps_list))
+ return NULL;
+
+ return list_first_entry(&erp_table->atcam_erps_list,
+ struct mlxsw_sp_acl_erp, list);
+}
+
+static int mlxsw_sp_acl_erp_index_get(struct mlxsw_sp_acl_erp_table *erp_table,
+ u8 *p_index)
+{
+ u8 index;
+
+ index = find_first_zero_bit(erp_table->erp_index_bitmap,
+ erp_table->num_max_atcam_erps);
+ if (index < erp_table->num_max_atcam_erps) {
+ __set_bit(index, erp_table->erp_index_bitmap);
+ *p_index = index;
+ return 0;
+ }
+
+ return -ENOBUFS;
+}
+
+static void mlxsw_sp_acl_erp_index_put(struct mlxsw_sp_acl_erp_table *erp_table,
+ u8 index)
+{
+ __clear_bit(index, erp_table->erp_index_bitmap);
+}
+
+static void
+mlxsw_sp_acl_erp_table_locate(const struct mlxsw_sp_acl_erp_table *erp_table,
+ const struct mlxsw_sp_acl_erp *erp,
+ u8 *p_erpt_bank, u8 *p_erpt_index)
+{
+ unsigned int entry_size = mlxsw_sp_acl_erp_table_entry_size(erp_table);
+ struct mlxsw_sp_acl_erp_core *erp_core = erp_table->erp_core;
+ unsigned int row;
+
+ *p_erpt_bank = erp->index % erp_core->num_erp_banks;
+ row = erp->index / erp_core->num_erp_banks;
+ *p_erpt_index = erp_table->base_index + row * entry_size;
+}
+
+static int
+mlxsw_sp_acl_erp_table_erp_add(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ enum mlxsw_reg_perpt_key_size key_size;
+ char perpt_pl[MLXSW_REG_PERPT_LEN];
+ u8 erpt_bank, erpt_index;
+
+ mlxsw_sp_acl_erp_table_locate(erp_table, erp, &erpt_bank, &erpt_index);
+ key_size = (enum mlxsw_reg_perpt_key_size) erp_table->aregion->type;
+ mlxsw_reg_perpt_pack(perpt_pl, erpt_bank, erpt_index, key_size, erp->id,
+ 0, erp_table->base_index, erp->index,
+ erp->key.mask);
+ mlxsw_reg_perpt_erp_vector_pack(perpt_pl, erp_table->erp_index_bitmap,
+ MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ mlxsw_reg_perpt_erp_vector_set(perpt_pl, erp->index, true);
+ return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(perpt), perpt_pl);
+}
+
+static void mlxsw_sp_acl_erp_table_erp_del(struct mlxsw_sp_acl_erp *erp)
+{
+ char empty_mask[MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN] = { 0 };
+ struct mlxsw_sp_acl_erp_table *erp_table = erp->erp_table;
+ struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ enum mlxsw_reg_perpt_key_size key_size;
+ char perpt_pl[MLXSW_REG_PERPT_LEN];
+ u8 erpt_bank, erpt_index;
+
+ mlxsw_sp_acl_erp_table_locate(erp_table, erp, &erpt_bank, &erpt_index);
+ key_size = (enum mlxsw_reg_perpt_key_size) erp_table->aregion->type;
+ mlxsw_reg_perpt_pack(perpt_pl, erpt_bank, erpt_index, key_size, erp->id,
+ 0, erp_table->base_index, erp->index, empty_mask);
+ mlxsw_reg_perpt_erp_vector_pack(perpt_pl, erp_table->erp_index_bitmap,
+ MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ mlxsw_reg_perpt_erp_vector_set(perpt_pl, erp->index, false);
+ mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(perpt), perpt_pl);
+}
+
+static int
+mlxsw_sp_acl_erp_table_enable(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
+ struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ char pererp_pl[MLXSW_REG_PERERP_LEN];
+
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, false, true, 0,
+ erp_table->base_index, 0);
+ mlxsw_reg_pererp_erp_vector_pack(pererp_pl, erp_table->erp_index_bitmap,
+ MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+
+ return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
+}
+
+static void
+mlxsw_sp_acl_erp_table_disable(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
+ struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ char pererp_pl[MLXSW_REG_PERERP_LEN];
+ struct mlxsw_sp_acl_erp *master_rp;
+
+ master_rp = mlxsw_sp_acl_erp_table_master_rp(erp_table);
+ if (!master_rp)
+ return;
+
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, false, false, 0, 0,
+ master_rp->id);
+ mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
+}
+
+static int
+mlxsw_sp_acl_erp_table_relocate(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_erp *erp;
+ int err;
+
+ list_for_each_entry(erp, &erp_table->atcam_erps_list, list) {
+ err = mlxsw_sp_acl_erp_table_erp_add(erp_table, erp);
+ if (err)
+ goto err_table_erp_add;
+ }
+
+ return 0;
+
+err_table_erp_add:
+ list_for_each_entry_continue_reverse(erp, &erp_table->atcam_erps_list,
+ list)
+ mlxsw_sp_acl_erp_table_erp_del(erp);
+ return err;
+}
+
+static int
+mlxsw_sp_acl_erp_table_expand(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ unsigned int num_erps, old_num_erps = erp_table->num_max_atcam_erps;
+ struct mlxsw_sp_acl_erp_core *erp_core = erp_table->erp_core;
+ unsigned long old_base_index = erp_table->base_index;
+ int err;
+
+ if (erp_table->num_atcam_erps < erp_table->num_max_atcam_erps)
+ return 0;
+
+ if (erp_table->num_max_atcam_erps == MLXSW_SP_ACL_ERP_MAX_PER_REGION)
+ return -ENOBUFS;
+
+ num_erps = old_num_erps + erp_core->num_erp_banks;
+ err = mlxsw_sp_acl_erp_table_alloc(erp_core, num_erps,
+ erp_table->aregion->type,
+ &erp_table->base_index);
+ if (err)
+ return err;
+ erp_table->num_max_atcam_erps = num_erps;
+
+ err = mlxsw_sp_acl_erp_table_relocate(erp_table);
+ if (err)
+ goto err_table_relocate;
+
+ err = mlxsw_sp_acl_erp_table_enable(erp_table);
+ if (err)
+ goto err_table_enable;
+
+ mlxsw_sp_acl_erp_table_free(erp_core, old_num_erps,
+ erp_table->aregion->type, old_base_index);
+
+ return 0;
+
+err_table_enable:
+err_table_relocate:
+ erp_table->num_max_atcam_erps = old_num_erps;
+ mlxsw_sp_acl_erp_table_free(erp_core, num_erps,
+ erp_table->aregion->type,
+ erp_table->base_index);
+ erp_table->base_index = old_base_index;
+ return err;
+}
+
+static int
+mlxsw_sp_acl_erp_region_table_trans(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_erp_core *erp_core = erp_table->erp_core;
+ struct mlxsw_sp_acl_erp *master_rp;
+ int err;
+
+ /* Initially, allocate a single eRP row. Expand later as needed */
+ err = mlxsw_sp_acl_erp_table_alloc(erp_core, erp_core->num_erp_banks,
+ erp_table->aregion->type,
+ &erp_table->base_index);
+ if (err)
+ return err;
+ erp_table->num_max_atcam_erps = erp_core->num_erp_banks;
+
+ /* Transition the sole RP currently configured (the master RP)
+ * to the eRP table
+ */
+ master_rp = mlxsw_sp_acl_erp_table_master_rp(erp_table);
+ if (!master_rp) {
+ err = -EINVAL;
+ goto err_table_master_rp;
+ }
+
+ /* Maintain the same eRP bank for the master RP, so that we
+ * wouldn't need to update the bloom filter
+ */
+ master_rp->index = master_rp->index % erp_core->num_erp_banks;
+ __set_bit(master_rp->index, erp_table->erp_index_bitmap);
+
+ err = mlxsw_sp_acl_erp_table_erp_add(erp_table, master_rp);
+ if (err)
+ goto err_table_master_rp_add;
+
+ err = mlxsw_sp_acl_erp_table_enable(erp_table);
+ if (err)
+ goto err_table_enable;
+
+ return 0;
+
+err_table_enable:
+ mlxsw_sp_acl_erp_table_erp_del(master_rp);
+err_table_master_rp_add:
+ __clear_bit(master_rp->index, erp_table->erp_index_bitmap);
+err_table_master_rp:
+ mlxsw_sp_acl_erp_table_free(erp_core, erp_table->num_max_atcam_erps,
+ erp_table->aregion->type,
+ erp_table->base_index);
+ return err;
+}
+
+static void
+mlxsw_sp_acl_erp_region_master_mask_trans(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ struct mlxsw_sp_acl_erp_core *erp_core = erp_table->erp_core;
+ struct mlxsw_sp_acl_erp *master_rp;
+
+ mlxsw_sp_acl_erp_table_disable(erp_table);
+ master_rp = mlxsw_sp_acl_erp_table_master_rp(erp_table);
+ if (!master_rp)
+ return;
+ mlxsw_sp_acl_erp_table_erp_del(master_rp);
+ __clear_bit(master_rp->index, erp_table->erp_index_bitmap);
+ mlxsw_sp_acl_erp_table_free(erp_core, erp_table->num_max_atcam_erps,
+ erp_table->aregion->type,
+ erp_table->base_index);
+}
+
+static int
+mlxsw_sp_acl_erp_region_erp_add(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
+ struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ char pererp_pl[MLXSW_REG_PERERP_LEN];
+
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, false, true, 0,
+ erp_table->base_index, 0);
+ mlxsw_reg_pererp_erp_vector_pack(pererp_pl, erp_table->erp_index_bitmap,
+ MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ mlxsw_reg_pererp_erpt_vector_set(pererp_pl, erp->index, true);
+
+ return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
+}
+
+static void mlxsw_sp_acl_erp_region_erp_del(struct mlxsw_sp_acl_erp *erp)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table = erp->erp_table;
+ struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
+ struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ char pererp_pl[MLXSW_REG_PERERP_LEN];
+
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, false, true, 0,
+ erp_table->base_index, 0);
+ mlxsw_reg_pererp_erp_vector_pack(pererp_pl, erp_table->erp_index_bitmap,
+ MLXSW_SP_ACL_ERP_MAX_PER_REGION);
+ mlxsw_reg_pererp_erpt_vector_set(pererp_pl, erp->index, false);
+
+ mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
+}
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key)
+{
+ struct mlxsw_sp_acl_erp *erp;
+ int err;
+
+ /* Expand the eRP table for the new eRP, if needed */
+ err = mlxsw_sp_acl_erp_table_expand(erp_table);
+ if (err)
+ return ERR_PTR(err);
+
+ erp = mlxsw_sp_acl_erp_generic_create(erp_table, key);
+ if (IS_ERR(erp))
+ return erp;
+
+ err = mlxsw_sp_acl_erp_index_get(erp_table, &erp->index);
+ if (err)
+ goto err_erp_index_get;
+
+ err = mlxsw_sp_acl_erp_table_erp_add(erp_table, erp);
+ if (err)
+ goto err_table_erp_add;
+
+ err = mlxsw_sp_acl_erp_region_erp_add(erp_table, erp);
+ if (err)
+ goto err_region_erp_add;
+
+ erp_table->ops = &erp_multiple_masks_ops;
+
+ return erp;
+
+err_region_erp_add:
+ mlxsw_sp_acl_erp_table_erp_del(erp);
+err_table_erp_add:
+ mlxsw_sp_acl_erp_index_put(erp_table, erp->index);
+err_erp_index_get:
+ mlxsw_sp_acl_erp_generic_destroy(erp);
+ return ERR_PTR(err);
+}
+
+static void
+mlxsw_sp_acl_erp_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ mlxsw_sp_acl_erp_region_erp_del(erp);
+ mlxsw_sp_acl_erp_table_erp_del(erp);
+ mlxsw_sp_acl_erp_index_put(erp_table, erp->index);
+ mlxsw_sp_acl_erp_generic_destroy(erp);
+
+ if (erp_table->num_atcam_erps == 2)
+ erp_table->ops = &erp_two_masks_ops;
+}
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_second_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key)
+{
+ struct mlxsw_sp_acl_erp *erp;
+ int err;
+
+ /* Transition to use eRP table instead of master mask */
+ err = mlxsw_sp_acl_erp_region_table_trans(erp_table);
+ if (err)
+ return ERR_PTR(err);
+
+ erp = mlxsw_sp_acl_erp_generic_create(erp_table, key);
+ if (IS_ERR(erp)) {
+ err = PTR_ERR(erp);
+ goto err_erp_create;
+ }
+
+ err = mlxsw_sp_acl_erp_index_get(erp_table, &erp->index);
+ if (err)
+ goto err_erp_index_get;
+
+ err = mlxsw_sp_acl_erp_table_erp_add(erp_table, erp);
+ if (err)
+ goto err_table_erp_add;
+
+ err = mlxsw_sp_acl_erp_region_erp_add(erp_table, erp);
+ if (err)
+ goto err_region_erp_add;
+
+ erp_table->ops = &erp_two_masks_ops;
+
+ return erp;
+
+err_region_erp_add:
+ mlxsw_sp_acl_erp_table_erp_del(erp);
+err_table_erp_add:
+ mlxsw_sp_acl_erp_index_put(erp_table, erp->index);
+err_erp_index_get:
+ mlxsw_sp_acl_erp_generic_destroy(erp);
+err_erp_create:
+ mlxsw_sp_acl_erp_region_master_mask_trans(erp_table);
+ return ERR_PTR(err);
+}
+
+static void
+mlxsw_sp_acl_erp_second_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ mlxsw_sp_acl_erp_region_erp_del(erp);
+ mlxsw_sp_acl_erp_table_erp_del(erp);
+ mlxsw_sp_acl_erp_index_put(erp_table, erp->index);
+ mlxsw_sp_acl_erp_generic_destroy(erp);
+ /* Transition to use master mask instead of eRP table */
+ mlxsw_sp_acl_erp_region_master_mask_trans(erp_table);
+
+ erp_table->ops = &erp_single_mask_ops;
+}
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_first_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key)
+{
+ struct mlxsw_sp_acl_erp *erp;
+
+ erp = mlxsw_sp_acl_erp_generic_create(erp_table, key);
+ if (IS_ERR(erp))
+ return erp;
+
+ erp_table->ops = &erp_single_mask_ops;
+
+ return erp;
+}
+
+static void
+mlxsw_sp_acl_erp_first_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ mlxsw_sp_acl_erp_generic_destroy(erp);
+ erp_table->ops = &erp_no_mask_ops;
+}
+
+static void
+mlxsw_sp_acl_erp_no_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ WARN_ON(1);
+}
+
+struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_get(struct mlxsw_sp_acl_atcam_region *aregion,
+ const char *mask)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table = aregion->erp_table;
+ struct mlxsw_sp_acl_erp_key key;
+ struct mlxsw_sp_acl_erp *erp;
+
+ /* eRPs are allocated from a shared resource, but currently all
+ * allocations are done under RTNL.
+ */
+ ASSERT_RTNL();
+
+ memcpy(key.mask, mask, MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
+ erp = rhashtable_lookup_fast(&erp_table->erp_ht, &key,
+ mlxsw_sp_acl_erp_ht_params);
+ if (erp) {
+ refcount_inc(&erp->refcnt);
+ return erp;
+ }
+
+ return erp_table->ops->erp_create(erp_table, &key);
+}
+
+void mlxsw_sp_acl_erp_put(struct mlxsw_sp_acl_atcam_region *aregion,
+ struct mlxsw_sp_acl_erp *erp)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table = aregion->erp_table;
+
+ ASSERT_RTNL();
+
+ if (!refcount_dec_and_test(&erp->refcnt))
+ return;
+
+ erp_table->ops->erp_destroy(erp_table, erp);
+}
+
+static struct mlxsw_sp_acl_erp_table *
+mlxsw_sp_acl_erp_table_create(struct mlxsw_sp_acl_atcam_region *aregion)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table;
+ int err;
+
+ erp_table = kzalloc(sizeof(*erp_table), GFP_KERNEL);
+ if (!erp_table)
+ return ERR_PTR(-ENOMEM);
+
+ err = rhashtable_init(&erp_table->erp_ht, &mlxsw_sp_acl_erp_ht_params);
+ if (err)
+ goto err_rhashtable_init;
+
+ erp_table->erp_core = aregion->atcam->erp_core;
+ erp_table->ops = &erp_no_mask_ops;
+ INIT_LIST_HEAD(&erp_table->atcam_erps_list);
+ erp_table->aregion = aregion;
+
+ return erp_table;
+
+err_rhashtable_init:
+ kfree(erp_table);
+ return ERR_PTR(err);
+}
+
+static void
+mlxsw_sp_acl_erp_table_destroy(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ WARN_ON(!list_empty(&erp_table->atcam_erps_list));
+ rhashtable_destroy(&erp_table->erp_ht);
+ kfree(erp_table);
+}
+
+static int
+mlxsw_sp_acl_erp_master_mask_init(struct mlxsw_sp_acl_atcam_region *aregion)
+{
+ struct mlxsw_sp *mlxsw_sp = aregion->region->mlxsw_sp;
+ char percr_pl[MLXSW_REG_PERCR_LEN];
+ char *master_mask;
+
+ mlxsw_reg_percr_pack(percr_pl, aregion->region->id);
+ master_mask = mlxsw_reg_percr_master_mask_data(percr_pl);
+ memset(master_mask, 0, MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
+ return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(percr), percr_pl);
+}
+
+static int
+mlxsw_sp_acl_erp_region_param_init(struct mlxsw_sp_acl_atcam_region *aregion)
+{
+ struct mlxsw_sp *mlxsw_sp = aregion->region->mlxsw_sp;
+ char pererp_pl[MLXSW_REG_PERERP_LEN];
+
+ mlxsw_reg_pererp_pack(pererp_pl, aregion->region->id, false, false, 0,
+ 0, 0);
+ return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
+}
+
+int mlxsw_sp_acl_erp_region_init(struct mlxsw_sp_acl_atcam_region *aregion)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table;
+ int err;
+
+ erp_table = mlxsw_sp_acl_erp_table_create(aregion);
+ if (IS_ERR(erp_table))
+ return PTR_ERR(erp_table);
+ aregion->erp_table = erp_table;
+
+ /* Initialize the region's master mask to all zeroes */
+ err = mlxsw_sp_acl_erp_master_mask_init(aregion);
+ if (err)
+ goto err_erp_master_mask_init;
+
+ /* Initialize the region to not use the eRP table */
+ err = mlxsw_sp_acl_erp_region_param_init(aregion);
+ if (err)
+ goto err_erp_region_param_init;
+
+ return 0;
+
+err_erp_region_param_init:
+err_erp_master_mask_init:
+ mlxsw_sp_acl_erp_table_destroy(erp_table);
+ return err;
+}
+
+void mlxsw_sp_acl_erp_region_fini(struct mlxsw_sp_acl_atcam_region *aregion)
+{
+ mlxsw_sp_acl_erp_table_destroy(aregion->erp_table);
+}
+
+static int
+mlxsw_sp_acl_erp_tables_sizes_query(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_erp_core *erp_core)
+{
+ unsigned int size;
+
+ if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, ACL_ERPT_ENTRIES_2KB) ||
+ !MLXSW_CORE_RES_VALID(mlxsw_sp->core, ACL_ERPT_ENTRIES_4KB) ||
+ !MLXSW_CORE_RES_VALID(mlxsw_sp->core, ACL_ERPT_ENTRIES_8KB) ||
+ !MLXSW_CORE_RES_VALID(mlxsw_sp->core, ACL_ERPT_ENTRIES_12KB))
+ return -EIO;
+
+ size = MLXSW_CORE_RES_GET(mlxsw_sp->core, ACL_ERPT_ENTRIES_2KB);
+ erp_core->erpt_entries_size[MLXSW_SP_ACL_ATCAM_REGION_TYPE_2KB] = size;
+
+ size = MLXSW_CORE_RES_GET(mlxsw_sp->core, ACL_ERPT_ENTRIES_4KB);
+ erp_core->erpt_entries_size[MLXSW_SP_ACL_ATCAM_REGION_TYPE_4KB] = size;
+
+ size = MLXSW_CORE_RES_GET(mlxsw_sp->core, ACL_ERPT_ENTRIES_8KB);
+ erp_core->erpt_entries_size[MLXSW_SP_ACL_ATCAM_REGION_TYPE_8KB] = size;
+
+ size = MLXSW_CORE_RES_GET(mlxsw_sp->core, ACL_ERPT_ENTRIES_12KB);
+ erp_core->erpt_entries_size[MLXSW_SP_ACL_ATCAM_REGION_TYPE_12KB] = size;
+
+ return 0;
+}
+
+static int mlxsw_sp_acl_erp_tables_init(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_erp_core *erp_core)
+{
+ unsigned int erpt_bank_size;
+ int err;
+
+ if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, ACL_MAX_ERPT_BANK_SIZE) ||
+ !MLXSW_CORE_RES_VALID(mlxsw_sp->core, ACL_MAX_ERPT_BANKS))
+ return -EIO;
+ erpt_bank_size = MLXSW_CORE_RES_GET(mlxsw_sp->core,
+ ACL_MAX_ERPT_BANK_SIZE);
+ erp_core->num_erp_banks = MLXSW_CORE_RES_GET(mlxsw_sp->core,
+ ACL_MAX_ERPT_BANKS);
+
+ erp_core->erp_tables = gen_pool_create(0, -1);
+ if (!erp_core->erp_tables)
+ return -ENOMEM;
+ gen_pool_set_algo(erp_core->erp_tables, gen_pool_best_fit, NULL);
+
+ err = gen_pool_add(erp_core->erp_tables,
+ MLXSW_SP_ACL_ERP_GENALLOC_OFFSET, erpt_bank_size,
+ -1);
+ if (err)
+ goto err_gen_pool_add;
+
+ /* Different regions require masks of different sizes */
+ err = mlxsw_sp_acl_erp_tables_sizes_query(mlxsw_sp, erp_core);
+ if (err)
+ goto err_erp_tables_sizes_query;
+
+ return 0;
+
+err_erp_tables_sizes_query:
+err_gen_pool_add:
+ gen_pool_destroy(erp_core->erp_tables);
+ return err;
+}
+
+static void mlxsw_sp_acl_erp_tables_fini(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_erp_core *erp_core)
+{
+ gen_pool_destroy(erp_core->erp_tables);
+}
+
+int mlxsw_sp_acl_erps_init(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam)
+{
+ struct mlxsw_sp_acl_erp_core *erp_core;
+ int err;
+
+ erp_core = kzalloc(sizeof(*erp_core), GFP_KERNEL);
+ if (!erp_core)
+ return -ENOMEM;
+ erp_core->mlxsw_sp = mlxsw_sp;
+ atcam->erp_core = erp_core;
+
+ err = mlxsw_sp_acl_erp_tables_init(mlxsw_sp, erp_core);
+ if (err)
+ goto err_erp_tables_init;
+
+ return 0;
+
+err_erp_tables_init:
+ kfree(erp_core);
+ return err;
+}
+
+void mlxsw_sp_acl_erps_fini(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam)
+{
+ mlxsw_sp_acl_erp_tables_fini(mlxsw_sp, atcam->erp_core);
+ kfree(atcam->erp_core);
+}
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
index 6403ec4f5267..af21f7c6c6df 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
@@ -92,6 +92,9 @@ mlxsw_sp_acl_tcam_profile_ops(struct mlxsw_sp *mlxsw_sp,
#define MLXSW_SP_ACL_TCAM_CATCHALL_PRIO (~0U)
+#define MLXSW_SP_ACL_TCAM_MASK_LEN \
+ (MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN * BITS_PER_BYTE)
+
struct mlxsw_sp_acl_tcam_group;
struct mlxsw_sp_acl_tcam_region {
@@ -144,9 +147,46 @@ mlxsw_sp_acl_ctcam_entry_offset(struct mlxsw_sp_acl_ctcam_entry *centry)
return centry->parman_item.index;
}
+enum mlxsw_sp_acl_atcam_region_type {
+ MLXSW_SP_ACL_ATCAM_REGION_TYPE_2KB,
+ MLXSW_SP_ACL_ATCAM_REGION_TYPE_4KB,
+ MLXSW_SP_ACL_ATCAM_REGION_TYPE_8KB,
+ MLXSW_SP_ACL_ATCAM_REGION_TYPE_12KB,
+ __MLXSW_SP_ACL_ATCAM_REGION_TYPE_MAX,
+};
+
+#define MLXSW_SP_ACL_ATCAM_REGION_TYPE_MAX \
+ (__MLXSW_SP_ACL_ATCAM_REGION_TYPE_MAX - 1)
+
+struct mlxsw_sp_acl_atcam {
+ struct mlxsw_sp_acl_erp_core *erp_core;
+};
+
+struct mlxsw_sp_acl_atcam_region {
+ struct mlxsw_sp_acl_tcam_region *region;
+ struct mlxsw_sp_acl_atcam *atcam;
+ enum mlxsw_sp_acl_atcam_region_type type;
+ struct mlxsw_sp_acl_erp_table *erp_table;
+};
+
int mlxsw_sp_acl_atcam_region_associate(struct mlxsw_sp *mlxsw_sp,
u16 region_id);
int mlxsw_sp_acl_atcam_region_init(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl_tcam_region *region);
+struct mlxsw_sp_acl_erp;
+
+u8 mlxsw_sp_acl_erp_id(const struct mlxsw_sp_acl_erp *erp);
+struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_get(struct mlxsw_sp_acl_atcam_region *aregion,
+ const char *mask);
+void mlxsw_sp_acl_erp_put(struct mlxsw_sp_acl_atcam_region *aregion,
+ struct mlxsw_sp_acl_erp *erp);
+int mlxsw_sp_acl_erp_region_init(struct mlxsw_sp_acl_atcam_region *aregion);
+void mlxsw_sp_acl_erp_region_fini(struct mlxsw_sp_acl_atcam_region *aregion);
+int mlxsw_sp_acl_erps_init(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam);
+void mlxsw_sp_acl_erps_fini(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_atcam *atcam);
+
#endif
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 08/17] mlxsw: spectrum_acl: Add support for C-TCAM eRPs
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
The number of eRPs that can be used by a single A-TCAM region is limited
to 16. When more eRPs are needed, an ordinary circuit TCAM (C-TCAM) can
be used to hold the extra eRPs.
Unlike the A-TCAM, only a single (last) lookup is performed in the
C-TCAM and not a lookup per-eRP. However, modeling the C-TCAM as extra
eRPs will allow us to easily introduce support for pruning in a
follow-up patch set and is also logically correct.
The following diagram depicts the relation between both TCAMs:
C-TCAM
+-------------------+ +--------------------+ +-----------+
| | | | | |
| eRP #1 (A-TCAM) +----> ... +----+ eRP #16 (A-TCAM) +----+ eRP #17 |
| | | | | ... |
+-------------------+ +--------------------+ | eRP #N |
| |
+-----------+
Lookup order is from left to right.
Extend the eRP core APIs with a C-TCAM parameter which indicates whether
the requested eRP is to be used with the C-TCAM or not.
Since the C-TCAM is only meant to absorb rules that can't fit in the
A-TCAM due to exceeded number of eRPs or key collision, an error is
returned when a C-TCAM eRP needs to be created when the eRP state
machine is in its initial state (i.e., 'no masks'). This should only
happen in the face of very unlikely errors when trying to push rules
into the A-TCAM.
In order not to perform unnecessary lookups, the eRP core will only
enable a C-TCAM lookup for a given region if it knows there are C-TCAM
eRPs present.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
.../mellanox/mlxsw/spectrum_acl_erp.c | 202 ++++++++++++++++--
.../mellanox/mlxsw/spectrum_acl_tcam.h | 2 +-
2 files changed, 191 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
index 86600c780c95..e26efa451d4a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
@@ -60,6 +60,7 @@ struct mlxsw_sp_acl_erp_core {
struct mlxsw_sp_acl_erp_key {
char mask[MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN];
+ bool ctcam;
};
struct mlxsw_sp_acl_erp {
@@ -90,6 +91,7 @@ struct mlxsw_sp_acl_erp_table {
unsigned long base_index;
unsigned int num_atcam_erps;
unsigned int num_max_atcam_erps;
+ unsigned int num_ctcam_erps;
};
static const struct rhashtable_params mlxsw_sp_acl_erp_ht_params = {
@@ -448,13 +450,14 @@ static void mlxsw_sp_acl_erp_table_erp_del(struct mlxsw_sp_acl_erp *erp)
}
static int
-mlxsw_sp_acl_erp_table_enable(struct mlxsw_sp_acl_erp_table *erp_table)
+mlxsw_sp_acl_erp_table_enable(struct mlxsw_sp_acl_erp_table *erp_table,
+ bool ctcam_le)
{
struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
char pererp_pl[MLXSW_REG_PERERP_LEN];
- mlxsw_reg_pererp_pack(pererp_pl, region->id, false, true, 0,
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, ctcam_le, true, 0,
erp_table->base_index, 0);
mlxsw_reg_pererp_erp_vector_pack(pererp_pl, erp_table->erp_index_bitmap,
MLXSW_SP_ACL_ERP_MAX_PER_REGION);
@@ -471,11 +474,12 @@ mlxsw_sp_acl_erp_table_disable(struct mlxsw_sp_acl_erp_table *erp_table)
struct mlxsw_sp_acl_erp *master_rp;
master_rp = mlxsw_sp_acl_erp_table_master_rp(erp_table);
- if (!master_rp)
- return;
-
+ /* It is possible we do not have a master RP when we disable the
+ * table when there are no rules in the A-TCAM and the last C-TCAM
+ * rule is deleted
+ */
mlxsw_reg_pererp_pack(pererp_pl, region->id, false, false, 0, 0,
- master_rp->id);
+ master_rp ? master_rp->id : 0);
mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
}
@@ -506,6 +510,7 @@ mlxsw_sp_acl_erp_table_expand(struct mlxsw_sp_acl_erp_table *erp_table)
unsigned int num_erps, old_num_erps = erp_table->num_max_atcam_erps;
struct mlxsw_sp_acl_erp_core *erp_core = erp_table->erp_core;
unsigned long old_base_index = erp_table->base_index;
+ bool ctcam_le = erp_table->num_ctcam_erps > 0;
int err;
if (erp_table->num_atcam_erps < erp_table->num_max_atcam_erps)
@@ -526,7 +531,7 @@ mlxsw_sp_acl_erp_table_expand(struct mlxsw_sp_acl_erp_table *erp_table)
if (err)
goto err_table_relocate;
- err = mlxsw_sp_acl_erp_table_enable(erp_table);
+ err = mlxsw_sp_acl_erp_table_enable(erp_table, ctcam_le);
if (err)
goto err_table_enable;
@@ -579,7 +584,7 @@ mlxsw_sp_acl_erp_region_table_trans(struct mlxsw_sp_acl_erp_table *erp_table)
if (err)
goto err_table_master_rp_add;
- err = mlxsw_sp_acl_erp_table_enable(erp_table);
+ err = mlxsw_sp_acl_erp_table_enable(erp_table, false);
if (err)
goto err_table_enable;
@@ -619,9 +624,10 @@ mlxsw_sp_acl_erp_region_erp_add(struct mlxsw_sp_acl_erp_table *erp_table,
{
struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ bool ctcam_le = erp_table->num_ctcam_erps > 0;
char pererp_pl[MLXSW_REG_PERERP_LEN];
- mlxsw_reg_pererp_pack(pererp_pl, region->id, false, true, 0,
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, ctcam_le, true, 0,
erp_table->base_index, 0);
mlxsw_reg_pererp_erp_vector_pack(pererp_pl, erp_table->erp_index_bitmap,
MLXSW_SP_ACL_ERP_MAX_PER_REGION);
@@ -635,9 +641,10 @@ static void mlxsw_sp_acl_erp_region_erp_del(struct mlxsw_sp_acl_erp *erp)
struct mlxsw_sp_acl_erp_table *erp_table = erp->erp_table;
struct mlxsw_sp_acl_tcam_region *region = erp_table->aregion->region;
struct mlxsw_sp *mlxsw_sp = erp_table->erp_core->mlxsw_sp;
+ bool ctcam_le = erp_table->num_ctcam_erps > 0;
char pererp_pl[MLXSW_REG_PERERP_LEN];
- mlxsw_reg_pererp_pack(pererp_pl, region->id, false, true, 0,
+ mlxsw_reg_pererp_pack(pererp_pl, region->id, ctcam_le, true, 0,
erp_table->base_index, 0);
mlxsw_reg_pererp_erp_vector_pack(pererp_pl, erp_table->erp_index_bitmap,
MLXSW_SP_ACL_ERP_MAX_PER_REGION);
@@ -646,6 +653,161 @@ static void mlxsw_sp_acl_erp_region_erp_del(struct mlxsw_sp_acl_erp *erp)
mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
}
+static int
+mlxsw_sp_acl_erp_region_ctcam_enable(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ /* No need to re-enable lookup in the C-TCAM */
+ if (erp_table->num_ctcam_erps > 1)
+ return 0;
+
+ return mlxsw_sp_acl_erp_table_enable(erp_table, true);
+}
+
+static void
+mlxsw_sp_acl_erp_region_ctcam_disable(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ /* Only disable C-TCAM lookup when last C-TCAM eRP is deleted */
+ if (erp_table->num_ctcam_erps > 1)
+ return;
+
+ mlxsw_sp_acl_erp_table_enable(erp_table, false);
+}
+
+static void
+mlxsw_sp_acl_erp_ctcam_table_ops_set(struct mlxsw_sp_acl_erp_table *erp_table)
+{
+ switch (erp_table->num_atcam_erps) {
+ case 2:
+ /* Keep using the eRP table, but correctly set the
+ * operations pointer so that when an A-TCAM eRP is
+ * deleted we will transition to use the master mask
+ */
+ erp_table->ops = &erp_two_masks_ops;
+ break;
+ case 1:
+ /* We only kept the eRP table because we had C-TCAM
+ * eRPs in use. Now that the last C-TCAM eRP is gone we
+ * can stop using the table and transition to use the
+ * master mask
+ */
+ mlxsw_sp_acl_erp_region_master_mask_trans(erp_table);
+ erp_table->ops = &erp_single_mask_ops;
+ break;
+ case 0:
+ /* There are no more eRPs of any kind used by the region
+ * so free its eRP table and transition to initial state
+ */
+ mlxsw_sp_acl_erp_table_disable(erp_table);
+ mlxsw_sp_acl_erp_table_free(erp_table->erp_core,
+ erp_table->num_max_atcam_erps,
+ erp_table->aregion->type,
+ erp_table->base_index);
+ erp_table->ops = &erp_no_mask_ops;
+ break;
+ default:
+ break;
+ }
+}
+
+static struct mlxsw_sp_acl_erp *
+__mlxsw_sp_acl_erp_ctcam_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key)
+{
+ struct mlxsw_sp_acl_erp *erp;
+ int err;
+
+ erp = kzalloc(sizeof(*erp), GFP_KERNEL);
+ if (!erp)
+ return ERR_PTR(-ENOMEM);
+
+ memcpy(&erp->key, key, sizeof(*key));
+ bitmap_from_arr32(erp->mask_bitmap, (u32 *) key->mask,
+ MLXSW_SP_ACL_TCAM_MASK_LEN);
+ refcount_set(&erp->refcnt, 1);
+ erp_table->num_ctcam_erps++;
+ erp->erp_table = erp_table;
+
+ err = mlxsw_sp_acl_erp_master_mask_set(erp_table, erp);
+ if (err)
+ goto err_master_mask_set;
+
+ err = rhashtable_insert_fast(&erp_table->erp_ht, &erp->ht_node,
+ mlxsw_sp_acl_erp_ht_params);
+ if (err)
+ goto err_rhashtable_insert;
+
+ err = mlxsw_sp_acl_erp_region_ctcam_enable(erp_table);
+ if (err)
+ goto err_erp_region_ctcam_enable;
+
+ /* When C-TCAM is used, the eRP table must be used */
+ erp_table->ops = &erp_multiple_masks_ops;
+
+ return erp;
+
+err_erp_region_ctcam_enable:
+ rhashtable_remove_fast(&erp_table->erp_ht, &erp->ht_node,
+ mlxsw_sp_acl_erp_ht_params);
+err_rhashtable_insert:
+ mlxsw_sp_acl_erp_master_mask_clear(erp_table, erp);
+err_master_mask_set:
+ erp_table->num_ctcam_erps--;
+ kfree(erp);
+ return ERR_PTR(err);
+}
+
+static struct mlxsw_sp_acl_erp *
+mlxsw_sp_acl_erp_ctcam_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
+ struct mlxsw_sp_acl_erp_key *key)
+{
+ struct mlxsw_sp_acl_erp *erp;
+ int err;
+
+ /* There is a special situation where we need to spill rules
+ * into the C-TCAM, yet the region is still using a master
+ * mask and thus not performing a lookup in the C-TCAM. This
+ * can happen when two rules that only differ in priority - and
+ * thus sharing the same key - are programmed. In this case
+ * we transition the region to use an eRP table
+ */
+ err = mlxsw_sp_acl_erp_region_table_trans(erp_table);
+ if (err)
+ return ERR_PTR(err);
+
+ erp = __mlxsw_sp_acl_erp_ctcam_mask_create(erp_table, key);
+ if (IS_ERR(erp)) {
+ err = PTR_ERR(erp);
+ goto err_erp_create;
+ }
+
+ return erp;
+
+err_erp_create:
+ mlxsw_sp_acl_erp_region_master_mask_trans(erp_table);
+ return ERR_PTR(err);
+}
+
+static void
+mlxsw_sp_acl_erp_ctcam_mask_destroy(struct mlxsw_sp_acl_erp *erp)
+{
+ struct mlxsw_sp_acl_erp_table *erp_table = erp->erp_table;
+
+ mlxsw_sp_acl_erp_region_ctcam_disable(erp_table);
+ rhashtable_remove_fast(&erp_table->erp_ht, &erp->ht_node,
+ mlxsw_sp_acl_erp_ht_params);
+ mlxsw_sp_acl_erp_master_mask_clear(erp_table, erp);
+ erp_table->num_ctcam_erps--;
+ kfree(erp);
+
+ /* Once the last C-TCAM eRP was destroyed, the state we
+ * transition to depends on the number of A-TCAM eRPs currently
+ * in use
+ */
+ if (erp_table->num_ctcam_erps > 0)
+ return;
+ mlxsw_sp_acl_erp_ctcam_table_ops_set(erp_table);
+}
+
static struct mlxsw_sp_acl_erp *
mlxsw_sp_acl_erp_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
struct mlxsw_sp_acl_erp_key *key)
@@ -653,6 +815,9 @@ mlxsw_sp_acl_erp_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
struct mlxsw_sp_acl_erp *erp;
int err;
+ if (key->ctcam)
+ return __mlxsw_sp_acl_erp_ctcam_mask_create(erp_table, key);
+
/* Expand the eRP table for the new eRP, if needed */
err = mlxsw_sp_acl_erp_table_expand(erp_table);
if (err)
@@ -691,12 +856,15 @@ static void
mlxsw_sp_acl_erp_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
struct mlxsw_sp_acl_erp *erp)
{
+ if (erp->key.ctcam)
+ return mlxsw_sp_acl_erp_ctcam_mask_destroy(erp);
+
mlxsw_sp_acl_erp_region_erp_del(erp);
mlxsw_sp_acl_erp_table_erp_del(erp);
mlxsw_sp_acl_erp_index_put(erp_table, erp->index);
mlxsw_sp_acl_erp_generic_destroy(erp);
- if (erp_table->num_atcam_erps == 2)
+ if (erp_table->num_atcam_erps == 2 && erp_table->num_ctcam_erps == 0)
erp_table->ops = &erp_two_masks_ops;
}
@@ -707,6 +875,9 @@ mlxsw_sp_acl_erp_second_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
struct mlxsw_sp_acl_erp *erp;
int err;
+ if (key->ctcam)
+ return mlxsw_sp_acl_erp_ctcam_mask_create(erp_table, key);
+
/* Transition to use eRP table instead of master mask */
err = mlxsw_sp_acl_erp_region_table_trans(erp_table);
if (err)
@@ -749,6 +920,9 @@ static void
mlxsw_sp_acl_erp_second_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
struct mlxsw_sp_acl_erp *erp)
{
+ if (erp->key.ctcam)
+ return mlxsw_sp_acl_erp_ctcam_mask_destroy(erp);
+
mlxsw_sp_acl_erp_region_erp_del(erp);
mlxsw_sp_acl_erp_table_erp_del(erp);
mlxsw_sp_acl_erp_index_put(erp_table, erp->index);
@@ -765,6 +939,9 @@ mlxsw_sp_acl_erp_first_mask_create(struct mlxsw_sp_acl_erp_table *erp_table,
{
struct mlxsw_sp_acl_erp *erp;
+ if (key->ctcam)
+ return ERR_PTR(-EINVAL);
+
erp = mlxsw_sp_acl_erp_generic_create(erp_table, key);
if (IS_ERR(erp))
return erp;
@@ -791,7 +968,7 @@ mlxsw_sp_acl_erp_no_mask_destroy(struct mlxsw_sp_acl_erp_table *erp_table,
struct mlxsw_sp_acl_erp *
mlxsw_sp_acl_erp_get(struct mlxsw_sp_acl_atcam_region *aregion,
- const char *mask)
+ const char *mask, bool ctcam)
{
struct mlxsw_sp_acl_erp_table *erp_table = aregion->erp_table;
struct mlxsw_sp_acl_erp_key key;
@@ -803,6 +980,7 @@ mlxsw_sp_acl_erp_get(struct mlxsw_sp_acl_atcam_region *aregion,
ASSERT_RTNL();
memcpy(key.mask, mask, MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
+ key.ctcam = ctcam;
erp = rhashtable_lookup_fast(&erp_table->erp_ht, &key,
mlxsw_sp_acl_erp_ht_params);
if (erp) {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
index af21f7c6c6df..70094936ca43 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
@@ -179,7 +179,7 @@ struct mlxsw_sp_acl_erp;
u8 mlxsw_sp_acl_erp_id(const struct mlxsw_sp_acl_erp *erp);
struct mlxsw_sp_acl_erp *
mlxsw_sp_acl_erp_get(struct mlxsw_sp_acl_atcam_region *aregion,
- const char *mask);
+ const char *mask, bool ctcam);
void mlxsw_sp_acl_erp_put(struct mlxsw_sp_acl_atcam_region *aregion,
struct mlxsw_sp_acl_erp *erp);
int mlxsw_sp_acl_erp_region_init(struct mlxsw_sp_acl_atcam_region *aregion);
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 07/17] mlxsw: spectrum_acl: Enable C-TCAM only mode in eRP core
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
Currently, no calls are performed into the eRP core, but in order to
make review easier we would like to gradually add these calls.
Have the eRP core initialize a region's master mask to all ones and
allow it to use an empty eRP table. This directs the lookup to the
C-TCAM and allows the C-TCAM only mode to continue working.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
index 960f29140b43..86600c780c95 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
@@ -879,7 +879,7 @@ mlxsw_sp_acl_erp_region_param_init(struct mlxsw_sp_acl_atcam_region *aregion)
struct mlxsw_sp *mlxsw_sp = aregion->region->mlxsw_sp;
char pererp_pl[MLXSW_REG_PERERP_LEN];
- mlxsw_reg_pererp_pack(pererp_pl, aregion->region->id, false, false, 0,
+ mlxsw_reg_pererp_pack(pererp_pl, aregion->region->id, true, true, 0,
0, 0);
return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
}
@@ -894,12 +894,16 @@ int mlxsw_sp_acl_erp_region_init(struct mlxsw_sp_acl_atcam_region *aregion)
return PTR_ERR(erp_table);
aregion->erp_table = erp_table;
- /* Initialize the region's master mask to all zeroes */
+ /* Initialize the region's master mask to all ones for C-TCAM
+ * only mode
+ */
err = mlxsw_sp_acl_erp_master_mask_init(aregion);
if (err)
goto err_erp_master_mask_init;
- /* Initialize the region to not use the eRP table */
+ /* Initialize the region to use the eRP table and enable C-TCAM
+ * lookup
+ */
err = mlxsw_sp_acl_erp_region_param_init(aregion);
if (err)
goto err_erp_region_param_init;
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 05/17] mlxsw: resources: Add Spectrum-2 eRP resources
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
Add the following resources to be used by A-TCAM code:
* Maximum number of eRP banks
* Maximum size of eRP bank
* Number of eRP entries required for a 2/4/8/12 key blocks mask
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/resources.h | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/resources.h b/drivers/net/ethernet/mellanox/mlxsw/resources.h
index 4f84c7c8117b..bf650f2cd5af 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/resources.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/resources.h
@@ -65,7 +65,13 @@ enum mlxsw_res_id {
MLXSW_RES_ID_ACL_FLEX_KEYS,
MLXSW_RES_ID_ACL_MAX_ACTION_PER_RULE,
MLXSW_RES_ID_ACL_ACTIONS_PER_SET,
+ MLXSW_RES_ID_ACL_MAX_ERPT_BANKS,
+ MLXSW_RES_ID_ACL_MAX_ERPT_BANK_SIZE,
MLXSW_RES_ID_ACL_MAX_LARGE_KEY_ID,
+ MLXSW_RES_ID_ACL_ERPT_ENTRIES_2KB,
+ MLXSW_RES_ID_ACL_ERPT_ENTRIES_4KB,
+ MLXSW_RES_ID_ACL_ERPT_ENTRIES_8KB,
+ MLXSW_RES_ID_ACL_ERPT_ENTRIES_12KB,
MLXSW_RES_ID_MAX_CPU_POLICERS,
MLXSW_RES_ID_MAX_VRS,
MLXSW_RES_ID_MAX_RIFS,
@@ -109,7 +115,13 @@ static u16 mlxsw_res_ids[] = {
[MLXSW_RES_ID_ACL_FLEX_KEYS] = 0x2910,
[MLXSW_RES_ID_ACL_MAX_ACTION_PER_RULE] = 0x2911,
[MLXSW_RES_ID_ACL_ACTIONS_PER_SET] = 0x2912,
+ [MLXSW_RES_ID_ACL_MAX_ERPT_BANKS] = 0x2940,
+ [MLXSW_RES_ID_ACL_MAX_ERPT_BANK_SIZE] = 0x2941,
[MLXSW_RES_ID_ACL_MAX_LARGE_KEY_ID] = 0x2942,
+ [MLXSW_RES_ID_ACL_ERPT_ENTRIES_2KB] = 0x2950,
+ [MLXSW_RES_ID_ACL_ERPT_ENTRIES_4KB] = 0x2951,
+ [MLXSW_RES_ID_ACL_ERPT_ENTRIES_8KB] = 0x2952,
+ [MLXSW_RES_ID_ACL_ERPT_ENTRIES_12KB] = 0x2953,
[MLXSW_RES_ID_MAX_CPU_POLICERS] = 0x2A13,
[MLXSW_RES_ID_MAX_VRS] = 0x2C01,
[MLXSW_RES_ID_MAX_RIFS] = 0x2C02,
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 04/17] mlxsw: resources: Add Spectrum-2 maximum large key ID resource
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
Add a resource to make sure we do not exceed the maximum number of
supported large key IDs in a region.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/resources.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/resources.h b/drivers/net/ethernet/mellanox/mlxsw/resources.h
index f672a7b71de7..4f84c7c8117b 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/resources.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/resources.h
@@ -65,6 +65,7 @@ enum mlxsw_res_id {
MLXSW_RES_ID_ACL_FLEX_KEYS,
MLXSW_RES_ID_ACL_MAX_ACTION_PER_RULE,
MLXSW_RES_ID_ACL_ACTIONS_PER_SET,
+ MLXSW_RES_ID_ACL_MAX_LARGE_KEY_ID,
MLXSW_RES_ID_MAX_CPU_POLICERS,
MLXSW_RES_ID_MAX_VRS,
MLXSW_RES_ID_MAX_RIFS,
@@ -108,6 +109,7 @@ static u16 mlxsw_res_ids[] = {
[MLXSW_RES_ID_ACL_FLEX_KEYS] = 0x2910,
[MLXSW_RES_ID_ACL_MAX_ACTION_PER_RULE] = 0x2911,
[MLXSW_RES_ID_ACL_ACTIONS_PER_SET] = 0x2912,
+ [MLXSW_RES_ID_ACL_MAX_LARGE_KEY_ID] = 0x2942,
[MLXSW_RES_ID_MAX_CPU_POLICERS] = 0x2A13,
[MLXSW_RES_ID_MAX_VRS] = 0x2C01,
[MLXSW_RES_ID_MAX_RIFS] = 0x2C02,
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 03/17] mlxsw: reg: Add Policy-Engine eRP Table Register
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
The register is used to add and delete eRPs from the eRP table.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/reg.h | 113 ++++++++++++++++++++++
1 file changed, 113 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index cf07436c911b..5acef249e776 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -2504,6 +2504,118 @@ static inline void mlxsw_reg_ptce2_pack(char *payload, bool valid,
mlxsw_reg_ptce2_tcam_region_info_memcpy_to(payload, tcam_region_info);
}
+/* PERPT - Policy-Engine ERP Table Register
+ * ----------------------------------------
+ * This register adds and removes eRPs from the eRP table.
+ */
+#define MLXSW_REG_PERPT_ID 0x3021
+#define MLXSW_REG_PERPT_LEN 0x80
+
+MLXSW_REG_DEFINE(perpt, MLXSW_REG_PERPT_ID, MLXSW_REG_PERPT_LEN);
+
+/* reg_perpt_erpt_bank
+ * eRP table bank.
+ * Range 0 .. cap_max_erp_table_banks - 1
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, perpt, erpt_bank, 0x00, 16, 4);
+
+/* reg_perpt_erpt_index
+ * Index to eRP table within the eRP bank.
+ * Range is 0 .. cap_max_erp_table_bank_size - 1
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, perpt, erpt_index, 0x00, 0, 8);
+
+enum mlxsw_reg_perpt_key_size {
+ MLXSW_REG_PERPT_KEY_SIZE_2KB,
+ MLXSW_REG_PERPT_KEY_SIZE_4KB,
+ MLXSW_REG_PERPT_KEY_SIZE_8KB,
+ MLXSW_REG_PERPT_KEY_SIZE_12KB,
+};
+
+/* reg_perpt_key_size
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, perpt, key_size, 0x04, 0, 4);
+
+/* reg_perpt_bf_bypass
+ * 0 - The eRP is used only if bloom filter state is set for the given
+ * rule.
+ * 1 - The eRP is used regardless of bloom filter state.
+ * The bypass is an OR condition of region_id or eRP. See PERCR.bf_bypass
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, perpt, bf_bypass, 0x08, 8, 1);
+
+/* reg_perpt_erp_id
+ * eRP ID for use by the rules.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, perpt, erp_id, 0x08, 0, 4);
+
+/* reg_perpt_erpt_base_bank
+ * Base eRP table bank, points to head of erp_vector
+ * Range is 0 .. cap_max_erp_table_banks - 1
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, perpt, erpt_base_bank, 0x0C, 16, 4);
+
+/* reg_perpt_erpt_base_index
+ * Base index to eRP table within the eRP bank
+ * Range is 0 .. cap_max_erp_table_bank_size - 1
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, perpt, erpt_base_index, 0x0C, 0, 8);
+
+/* reg_perpt_erp_index_in_vector
+ * eRP index in the vector.
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, perpt, erp_index_in_vector, 0x10, 0, 4);
+
+/* reg_perpt_erp_vector
+ * eRP vector.
+ * Access: OP
+ */
+MLXSW_ITEM_BIT_ARRAY(reg, perpt, erp_vector, 0x14, 4, 1);
+
+/* reg_perpt_mask
+ * Mask
+ * 0 - A-TCAM will ignore the bit in key
+ * 1 - A-TCAM will compare the bit in key
+ * Access: RW
+ */
+MLXSW_ITEM_BUF(reg, perpt, mask, 0x20, MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
+
+static inline void mlxsw_reg_perpt_erp_vector_pack(char *payload,
+ unsigned long *erp_vector,
+ unsigned long size)
+{
+ unsigned long bit;
+
+ for_each_set_bit(bit, erp_vector, size)
+ mlxsw_reg_perpt_erp_vector_set(payload, bit, true);
+}
+
+static inline void
+mlxsw_reg_perpt_pack(char *payload, u8 erpt_bank, u8 erpt_index,
+ enum mlxsw_reg_perpt_key_size key_size, u8 erp_id,
+ u8 erpt_base_bank, u8 erpt_base_index, u8 erp_index,
+ char *mask)
+{
+ MLXSW_REG_ZERO(perpt, payload);
+ mlxsw_reg_perpt_erpt_bank_set(payload, erpt_bank);
+ mlxsw_reg_perpt_erpt_index_set(payload, erpt_index);
+ mlxsw_reg_perpt_key_size_set(payload, key_size);
+ mlxsw_reg_perpt_bf_bypass_set(payload, true);
+ mlxsw_reg_perpt_erp_id_set(payload, erp_id);
+ mlxsw_reg_perpt_erpt_base_bank_set(payload, erpt_base_bank);
+ mlxsw_reg_perpt_erpt_base_index_set(payload, erpt_base_index);
+ mlxsw_reg_perpt_erp_index_in_vector_set(payload, erp_index);
+ mlxsw_reg_perpt_mask_memcpy_to(payload, mask);
+}
+
/* PERAR - Policy-Engine Region Association Register
* -------------------------------------------------
* This register associates a hw region for region_id's. Changing on the fly
@@ -8422,6 +8534,7 @@ static const struct mlxsw_reg_info *mlxsw_reg_infos[] = {
MLXSW_REG(prcr),
MLXSW_REG(pefa),
MLXSW_REG(ptce2),
+ MLXSW_REG(perpt),
MLXSW_REG(perar),
MLXSW_REG(ptce3),
MLXSW_REG(percr),
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 02/17] mlxsw: reg: Add Policy-Engine TCAM Entry Register Version 3
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
The register is used to configure rules in the A-TCAM.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/reg.h | 165 +++++++++++++++++++++-
1 file changed, 162 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index f4e16ac20fea..cf07436c911b 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -2466,14 +2466,14 @@ MLXSW_ITEM32(reg, ptce2, priority, 0x04, 0, 24);
MLXSW_ITEM_BUF(reg, ptce2, tcam_region_info, 0x10,
MLXSW_REG_PXXX_TCAM_REGION_INFO_LEN);
-#define MLXSW_REG_PTCE2_FLEX_KEY_BLOCKS_LEN 96
+#define MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN 96
/* reg_ptce2_flex_key_blocks
* ACL Key.
* Access: RW
*/
MLXSW_ITEM_BUF(reg, ptce2, flex_key_blocks, 0x20,
- MLXSW_REG_PTCE2_FLEX_KEY_BLOCKS_LEN);
+ MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
/* reg_ptce2_mask
* mask- in the same size as key. A bit that is set directs the TCAM
@@ -2482,7 +2482,7 @@ MLXSW_ITEM_BUF(reg, ptce2, flex_key_blocks, 0x20,
* Access: RW
*/
MLXSW_ITEM_BUF(reg, ptce2, mask, 0x80,
- MLXSW_REG_PTCE2_FLEX_KEY_BLOCKS_LEN);
+ MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
/* reg_ptce2_flex_action_set
* ACL action set.
@@ -2545,6 +2545,164 @@ static inline void mlxsw_reg_perar_pack(char *payload, u16 region_id,
mlxsw_reg_perar_hw_region_set(payload, hw_region);
}
+/* PTCE-V3 - Policy-Engine TCAM Entry Register Version 3
+ * -----------------------------------------------------
+ * This register is a new version of PTCE-V2 in order to support the
+ * A-TCAM. This register is not supported by SwitchX/-2 and Spectrum.
+ */
+#define MLXSW_REG_PTCE3_ID 0x3027
+#define MLXSW_REG_PTCE3_LEN 0xF0
+
+MLXSW_REG_DEFINE(ptce3, MLXSW_REG_PTCE3_ID, MLXSW_REG_PTCE3_LEN);
+
+/* reg_ptce3_v
+ * Valid.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ptce3, v, 0x00, 31, 1);
+
+enum mlxsw_reg_ptce3_op {
+ /* Write operation. Used to write a new entry to the table.
+ * All R/W fields are relevant for new entry. Activity bit is set
+ * for new entries. Write with v = 0 will delete the entry. Must
+ * not be used if an entry exists.
+ */
+ MLXSW_REG_PTCE3_OP_WRITE_WRITE = 0,
+ /* Update operation */
+ MLXSW_REG_PTCE3_OP_WRITE_UPDATE = 1,
+ /* Read operation */
+ MLXSW_REG_PTCE3_OP_QUERY_READ = 0,
+};
+
+/* reg_ptce3_op
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, ptce3, op, 0x00, 20, 3);
+
+/* reg_ptce3_priority
+ * Priority of the rule. Higher values win.
+ * For Spectrum-2 range is 1..cap_kvd_size - 1
+ * Note: Priority does not have to be unique per rule.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ptce3, priority, 0x04, 0, 24);
+
+/* reg_ptce3_tcam_region_info
+ * Opaque object that represents the TCAM region.
+ * Access: Index
+ */
+MLXSW_ITEM_BUF(reg, ptce3, tcam_region_info, 0x10,
+ MLXSW_REG_PXXX_TCAM_REGION_INFO_LEN);
+
+/* reg_ptce3_flex2_key_blocks
+ * ACL key. The key must be masked according to eRP (if exists) or
+ * according to master mask.
+ * Access: Index
+ */
+MLXSW_ITEM_BUF(reg, ptce3, flex2_key_blocks, 0x20,
+ MLXSW_REG_PTCEX_FLEX_KEY_BLOCKS_LEN);
+
+/* reg_ptce3_erp_id
+ * eRP ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ptce3, erp_id, 0x80, 0, 4);
+
+/* reg_ptce3_delta_start
+ * Start point of delta_value and delta_mask, in bits. Must not exceed
+ * num_key_blocks * 36 - 8. Reserved when delta_mask = 0.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ptce3, delta_start, 0x84, 0, 10);
+
+/* reg_ptce3_delta_mask
+ * Delta mask.
+ * 0 - Ignore relevant bit in delta_value
+ * 1 - Compare relevant bit in delta_value
+ * Delta mask must not be set for reserved fields in the key blocks.
+ * Note: No delta when no eRPs. Thus, for regions with
+ * PERERP.erpt_pointer_valid = 0 the delta mask must be 0.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ptce3, delta_mask, 0x88, 16, 8);
+
+/* reg_ptce3_delta_value
+ * Delta value.
+ * Bits which are masked by delta_mask must be 0.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ptce3, delta_value, 0x88, 0, 8);
+
+/* reg_ptce3_prune_vector
+ * Pruning vector relative to the PERPT.erp_id.
+ * Used for reducing lookups.
+ * 0 - NEED: Do a lookup using the eRP.
+ * 1 - PRUNE: Do not perform a lookup using the eRP.
+ * Maybe be modified by PEAPBL and PEAPBM.
+ * Note: In Spectrum-2, a region of 8 key blocks must be set to either
+ * all 1's or all 0's.
+ * Access: RW
+ */
+MLXSW_ITEM_BIT_ARRAY(reg, ptce3, prune_vector, 0x90, 4, 1);
+
+/* reg_ptce3_prune_ctcam
+ * Pruning on C-TCAM. Used for reducing lookups.
+ * 0 - NEED: Do a lookup in the C-TCAM.
+ * 1 - PRUNE: Do not perform a lookup in the C-TCAM.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ptce3, prune_ctcam, 0x94, 31, 1);
+
+/* reg_ptce3_large_exists
+ * Large entry key ID exists.
+ * Within the region:
+ * 0 - SINGLE: The large_entry_key_id is not currently in use.
+ * For rule insert: The MSB of the key (blocks 6..11) will be added.
+ * For rule delete: The MSB of the key will be removed.
+ * 1 - NON_SINGLE: The large_entry_key_id is currently in use.
+ * For rule insert: The MSB of the key (blocks 6..11) will not be added.
+ * For rule delete: The MSB of the key will not be removed.
+ * Access: WO
+ */
+MLXSW_ITEM32(reg, ptce3, large_exists, 0x98, 31, 1);
+
+/* reg_ptce3_large_entry_key_id
+ * Large entry key ID.
+ * A key for 12 key blocks rules. Reserved when region has less than 12 key
+ * blocks. Must be different for different keys which have the same common
+ * 6 key blocks (MSB, blocks 6..11) key within a region.
+ * Range is 0..cap_max_pe_large_key_id - 1
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ptce3, large_entry_key_id, 0x98, 0, 24);
+
+/* reg_ptce3_action_pointer
+ * Pointer to action.
+ * Range is 0..cap_max_kvd_action_sets - 1
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ptce3, action_pointer, 0xA0, 0, 24);
+
+static inline void mlxsw_reg_ptce3_pack(char *payload, bool valid,
+ enum mlxsw_reg_ptce3_op op,
+ u32 priority,
+ const char *tcam_region_info,
+ const char *key, u8 erp_id,
+ bool large_exists, u32 lkey_id,
+ u32 action_pointer)
+{
+ MLXSW_REG_ZERO(ptce3, payload);
+ mlxsw_reg_ptce3_v_set(payload, valid);
+ mlxsw_reg_ptce3_op_set(payload, op);
+ mlxsw_reg_ptce3_priority_set(payload, priority);
+ mlxsw_reg_ptce3_tcam_region_info_memcpy_to(payload, tcam_region_info);
+ mlxsw_reg_ptce3_flex2_key_blocks_memcpy_to(payload, key);
+ mlxsw_reg_ptce3_erp_id_set(payload, erp_id);
+ mlxsw_reg_ptce3_large_exists_set(payload, large_exists);
+ mlxsw_reg_ptce3_large_entry_key_id_set(payload, lkey_id);
+ mlxsw_reg_ptce3_action_pointer_set(payload, action_pointer);
+}
+
/* PERCR - Policy-Engine Region Configuration Register
* ---------------------------------------------------
* This register configures the region parameters. The region_id must be
@@ -8265,6 +8423,7 @@ static const struct mlxsw_reg_info *mlxsw_reg_infos[] = {
MLXSW_REG(pefa),
MLXSW_REG(ptce2),
MLXSW_REG(perar),
+ MLXSW_REG(ptce3),
MLXSW_REG(percr),
MLXSW_REG(pererp),
MLXSW_REG(iedr),
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 01/17] mlxsw: reg: Prepare PERERP register for A-TCAM usage
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
In-Reply-To: <20180725062406.3342-1-idosch@mellanox.com>
Before introducing A-TCAM support we need to make sure all the necessary
fields are configurable and not hard coded to values that worked for the
C-TCAM only use case.
This includes - for example - the ability to configure the eRP table
used by the TCAM region.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/reg.h | 22 ++++++++++++++++---
.../mellanox/mlxsw/spectrum_acl_atcam.c | 2 +-
2 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index 596fddfb3850..f4e16ac20fea 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -2663,12 +2663,28 @@ MLXSW_ITEM_BIT_ARRAY(reg, pererp, erpt_vector, 0x14, 4, 1);
*/
MLXSW_ITEM32(reg, pererp, master_rp_id, 0x18, 0, 4);
-static inline void mlxsw_reg_pererp_pack(char *payload, u16 region_id)
+static inline void mlxsw_reg_pererp_erp_vector_pack(char *payload,
+ unsigned long *erp_vector,
+ unsigned long size)
+{
+ unsigned long bit;
+
+ for_each_set_bit(bit, erp_vector, size)
+ mlxsw_reg_pererp_erpt_vector_set(payload, bit, true);
+}
+
+static inline void mlxsw_reg_pererp_pack(char *payload, u16 region_id,
+ bool ctcam_le, bool erpt_pointer_valid,
+ u8 erpt_bank_pointer, u8 erpt_pointer,
+ u8 master_rp_id)
{
MLXSW_REG_ZERO(pererp, payload);
mlxsw_reg_pererp_region_id_set(payload, region_id);
- mlxsw_reg_pererp_ctcam_le_set(payload, true);
- mlxsw_reg_pererp_erpt_pointer_valid_set(payload, true);
+ mlxsw_reg_pererp_ctcam_le_set(payload, ctcam_le);
+ mlxsw_reg_pererp_erpt_pointer_valid_set(payload, erpt_pointer_valid);
+ mlxsw_reg_pererp_erpt_bank_pointer_set(payload, erpt_bank_pointer);
+ mlxsw_reg_pererp_erpt_pointer_set(payload, erpt_pointer);
+ mlxsw_reg_pererp_master_rp_id_set(payload, master_rp_id);
}
/* IEDR - Infrastructure Entry Delete Register
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c
index b1b3b0e0c00e..a27d3b0f9fcb 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_atcam.c
@@ -72,7 +72,7 @@ mlxsw_sp_acl_atcam_region_erp_init(struct mlxsw_sp *mlxsw_sp,
{
char pererp_pl[MLXSW_REG_PERERP_LEN];
- mlxsw_reg_pererp_pack(pererp_pl, region_id);
+ mlxsw_reg_pererp_pack(pererp_pl, region_id, true, true, 0, 0, 0);
return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pererp), pererp_pl);
}
--
2.17.1
^ permalink raw reply related
* [PATCH net-next 00/17] mlxsw: Introduce algorithmic TCAM support
From: Ido Schimmel @ 2018-07-25 6:23 UTC (permalink / raw)
To: netdev; +Cc: davem, jiri, mlxsw, Ido Schimmel
The Spectrum-2 ASIC uses an algorithmic TCAM (A-TCAM) where multiple
exact matches lookups are performed instead of a single lookup as with
standard circuit TCAM (C-TCAM) memory. This allows for higher scale and
reduced power consumption.
The lookups are performed by masking a packet using different masks
(e.g., {dst_ip/24, ethtype}) defined for the region and looking for an
exact match. Eventually, the rule with the highest priority will be
picked.
Since the number of masks per-region is limited, the ASIC includes a
C-TCAM that can be used as a spill area for rules that do not fit into
the A-TCAM.
The driver currently uses a C-TCAM only mode which is similar to
Spectrum-1. However, this mode severely limits both the number of
supported ACL rules and the performance of the ACL lookup.
This patch set introduces initial support for the A-TCAM mode where the
C-TCAM is only used for rule spillage.
The first five patches add the registers and ASIC resources needed in
order to make use of the A-TCAM.
Next three patches are the "meat" and add the eRP core which is used to
manage the masks used by each ACL region. The individual commit messages
are lengthy and aim to thoroughly explain the subject.
The next seven patches perform small adjustments in the code and the
related data structures and are meant to prepare the code base to the
introduction of the A-TCAM in the last two patches.
Various A-TCAM optimization will be the focus of follow-up patch sets:
* Pruning - Used to reduce the number of lookups. Each rule will include
a prune vector that indicates which masks should not be considered for
further lookups as they cannot result in a higher priority match
* Bloom filter - Used to reduce the number of lookups. Before performing
a lookup with a given mask the ASIC will consult a bloom filter
(managed by the driver) that indicates whether a match might exist using
the considered mask
* Masks aggregation - Used to increase scale and reduce lookups. Masks
that only differ by up to eight consecutive bits (delta bits) can be
aggregated into a single mask. The delta bits then become a part of the
rule's key. For example, dst_ip/16 and dst_ip/17 can be represented as
dst_ip/16 with a delta bit of one. Rules using the aggregated mask then
specify whether the 17-th bit should be masked or not and its value
Ido Schimmel (17):
mlxsw: reg: Prepare PERERP register for A-TCAM usage
mlxsw: reg: Add Policy-Engine TCAM Entry Register Version 3
mlxsw: reg: Add Policy-Engine eRP Table Register
mlxsw: resources: Add Spectrum-2 maximum large key ID resource
mlxsw: resources: Add Spectrum-2 eRP resources
mlxsw: spectrum_acl: Implement common eRP core
mlxsw: spectrum_acl: Enable C-TCAM only mode in eRP core
mlxsw: spectrum_acl: Add support for C-TCAM eRPs
mlxsw: spectrum_acl: Extend Spectrum-2 region struct
mlxsw: spectrum_acl: Allow encoding a partial key
mlxsw: spectrum_acl: Add A-TCAM initialization
mlxsw: spectrum_acl: Encapsulate C-TCAM region in A-TCAM region
mlxsw: spectrum_acl: Make global TCAM resources available to regions
mlxsw: spectrum_acl: Add A-TCAM region initialization
mlxsw: spectrum_acl: Pass C-TCAM region and entry to insert function
mlxsw: spectrum_acl: Add A-TCAM rule insertion and deletion
mlxsw: spectrum_acl: Start using A-TCAM
drivers/net/ethernet/mellanox/mlxsw/Kconfig | 1 +
drivers/net/ethernet/mellanox/mlxsw/Makefile | 2 +-
.../mellanox/mlxsw/core_acl_flex_keys.c | 10 +-
.../mellanox/mlxsw/core_acl_flex_keys.h | 2 +-
drivers/net/ethernet/mellanox/mlxsw/reg.h | 301 ++++-
.../net/ethernet/mellanox/mlxsw/resources.h | 14 +
.../net/ethernet/mellanox/mlxsw/spectrum.h | 1 +
.../mellanox/mlxsw/spectrum1_acl_tcam.c | 24 +-
.../mellanox/mlxsw/spectrum2_acl_tcam.c | 82 +-
.../mellanox/mlxsw/spectrum_acl_atcam.c | 509 ++++++-
.../mellanox/mlxsw/spectrum_acl_ctcam.c | 40 +-
.../mellanox/mlxsw/spectrum_acl_erp.c | 1199 +++++++++++++++++
.../mellanox/mlxsw/spectrum_acl_tcam.c | 2 +-
.../mellanox/mlxsw/spectrum_acl_tcam.h | 117 +-
14 files changed, 2235 insertions(+), 69 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_erp.c
--
2.17.1
^ permalink raw reply
* Re: [PATCH v6 13/18] wireless/lib80211: Convert from ahash to shash
From: Johannes Berg @ 2018-07-25 7:24 UTC (permalink / raw)
To: Kees Cook, Herbert Xu
Cc: Arnd Bergmann, Eric Biggers, Gustavo A. R. Silva, Alasdair Kergon,
Rabin Vincent, Tim Chen, Rafael J. Wysocki, Pavel Machek,
Thomas Gleixner, Ingo Molnar, H. Peter Anvin, x86,
Philipp Reisner, Lars Ellenberg, Jens Axboe, Giovanni Cabiddu,
Mike Snitzer, Paul Mackerras, Greg Kroah-Hartman, David Howells,
Tudor-Dan
In-Reply-To: <20180724164936.37477-14-keescook@chromium.org>
On Tue, 2018-07-24 at 09:49 -0700, Kees Cook wrote:
> In preparing to remove all stack VLA usage from the kernel[1], this
> removes the discouraged use of AHASH_REQUEST_ON_STACK in favor of
> the smaller SHASH_DESC_ON_STACK by converting from ahash-wrapped-shash
> to direct shash. By removing a layer of indirection this both improves
> performance and reduces stack usage. The stack allocation will be made
> a fixed size in a later patch to the crypto subsystem.
>
I think you sent this before - it should be in net-next now.
johannes
^ permalink raw reply
* [PATCH v3] ipvs: fix race between ip_vs_conn_new() and ip_vs_del_dest()
From: Tan Hu @ 2018-07-25 7:23 UTC (permalink / raw)
To: wensong, horms, ja, pablo, kadlec, fw, davem
Cc: netdev, lvs-devel, netfilter-devel, coreteam, linux-kernel,
zhong.weidong, jiang.biao2
We came across infinite loop in ipvs when using ipvs in docker
env.
When ipvs receives new packets and cannot find an ipvs connection,
it will create a new connection, then if the dest is unavailable
(i.e. IP_VS_DEST_F_AVAILABLE), the packet will be dropped sliently.
But if the dropped packet is the first packet of this connection,
the connection control timer never has a chance to start and the
ipvs connection cannot be released. This will lead to memory leak, or
infinite loop in cleanup_net() when net namespace is released like
this:
ip_vs_conn_net_cleanup at ffffffffa0a9f31a [ip_vs]
__ip_vs_cleanup at ffffffffa0a9f60a [ip_vs]
ops_exit_list at ffffffff81567a49
cleanup_net at ffffffff81568b40
process_one_work at ffffffff810a851b
worker_thread at ffffffff810a9356
kthread at ffffffff810b0b6f
ret_from_fork at ffffffff81697a18
race condition:
CPU1 CPU2
ip_vs_in()
ip_vs_conn_new()
ip_vs_del_dest()
__ip_vs_unlink_dest()
~IP_VS_DEST_F_AVAILABLE
cp->dest && !IP_VS_DEST_F_AVAILABLE
__ip_vs_conn_put
...
cleanup_net ---> infinite looping
Fix this by checking whether the timer already started.
Signed-off-by: Tan Hu <tan.hu@zte.com.cn>
Reviewed-by: Jiang Biao <jiang.biao2@zte.com.cn>
---
v2: fix use-after-free in CONN_ONE_PACKET case suggested by Julian Anastasov
v3: remove trailing whitespace for patch checking
net/netfilter/ipvs/ip_vs_core.c | 15 +++++++++++----
1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index 0679dd1..a17104f 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -1972,13 +1972,20 @@ static int ip_vs_in_icmp_v6(struct netns_ipvs *ipvs, struct sk_buff *skb,
if (cp->dest && !(cp->dest->flags & IP_VS_DEST_F_AVAILABLE)) {
/* the destination server is not available */
- if (sysctl_expire_nodest_conn(ipvs)) {
+ __u32 flags = cp->flags;
+
+ /* when timer already started, silently drop the packet.*/
+ if (timer_pending(&cp->timer))
+ __ip_vs_conn_put(cp);
+ else
+ ip_vs_conn_put(cp);
+
+ if (sysctl_expire_nodest_conn(ipvs) &&
+ !(flags & IP_VS_CONN_F_ONE_PACKET)) {
/* try to expire the connection immediately */
ip_vs_conn_expire_now(cp);
}
- /* don't restart its timer, and silently
- drop the packet. */
- __ip_vs_conn_put(cp);
+
return NF_DROP;
}
--
1.8.3.1
^ permalink raw reply related
* RE: [PATCH net-next] tls: Fix improper revert in zerocopy_from_iter
From: Vakul Garg @ 2018-07-25 6:10 UTC (permalink / raw)
To: Doron Roberts-Kedes
Cc: David S . Miller, Dave Watson, Matt Mullins,
netdev@vger.kernel.org
In-Reply-To: <20180724201941.GB97592@doronrk-mbp.dhcp.thefacebook.com>
> -----Original Message-----
> From: Doron Roberts-Kedes [mailto:doronrk@fb.com]
> Sent: Wednesday, July 25, 2018 1:50 AM
> To: Vakul Garg <vakul.garg@nxp.com>
> Cc: David S . Miller <davem@davemloft.net>; Dave Watson
> <davejwatson@fb.com>; Matt Mullins <mmullins@fb.com>;
> netdev@vger.kernel.org
> Subject: Re: [PATCH net-next] tls: Fix improper revert in zerocopy_from_iter
>
> On Tue, Jul 24, 2018 at 05:13:26AM +0000, Vakul Garg wrote:
> >
> > > -----Original Message-----
> > > From: Doron Roberts-Kedes [mailto:doronrk@fb.com]
> > > Sent: Tuesday, July 24, 2018 3:50 AM @@ -811,6 +809,7 @@ int
> > > tls_sw_recvmsg(struct sock *sk,
> > > likely(!(flags & MSG_PEEK))) {
> > > struct scatterlist sgin[MAX_SKB_FRAGS + 1];
> > > int pages = 0;
> > > + int orig_chunk = chunk;
> > >
> > > zc = true;
> > > sg_init_table(sgin, MAX_SKB_FRAGS + 1);
> @@ -820,9 +819,11 @@
> > > int tls_sw_recvmsg(struct sock *sk,
> > > err = zerocopy_from_iter(sk, &msg-
> > > >msg_iter,
> > > to_copy, &pages,
> > > &chunk, &sgin[1],
> > > - MAX_SKB_FRAGS,
> > > false, true);
> > > - if (err < 0)
> > > + MAX_SKB_FRAGS,
> > > false);
> > > + if (err < 0) {
> > > + iov_iter_revert(&msg->msg_iter,
> > > chunk - orig_chunk);
> > > goto fallback_to_reg_recv;
> > > + }
> >
> > This assumes that msg_iter gets advanced even if zerocopy_from_iter()
> fails.
>
> Not sure I see what you mean. If msg_iter is not advanced then chunk -
> orig_chunk is 0, and the revert is a no-op.
>
As I said below, my comment was to improve code readability.
It takes a while to grasp that calling iov_iter_revert would result in no-op
if zerocopy_from_iter() fails.
> > It is easier from code readability perspective if functions upon failure do
> not leave any side effects for the caller to clean-up.
> > I suggest that iov_iter_revert() should be called from zerocopy_from_iter()
> itself if it is going to fail.
>
> Agreed that zerocopy_from_iter() should call iov_iter_revert(). I didn't do
> that because at first glace, the tx path seems to depend on the
> iov_iter_revert() being called as a result of either failed
> zerocopy_from_iter() or tls_push_record(). However, I think the latter path
> cannot actually be taken because tls_push_record appears never to return a
> positive value. This means that the code between the continue and
> fallback_to_reg_send should probably just be replaced with simply, goto
> send_end.
>
> After that change, its clear that iov_iter_revert() can be safely moved inside
> zerocopy_from_iter() in the error case.
>
> Pending your input, I'll plan on putting up a 2 part patch addressing each of
> these.
Please submit.
^ permalink raw reply
* Re: Re: [PATCH v2] ipvs: fix race between ip_vs_conn_new() andip_vs_del_dest()
From: tan.hu @ 2018-07-25 7:15 UTC (permalink / raw)
To: ja
Cc: wensong, horms, pablo, kadlec, fw, davem, netdev, lvs-devel,
netfilter-devel, coreteam, linux-kernel, zhong.weidong,
jiang.biao2
In-Reply-To: <alpine.LFD.2.20.1807250849100.2794@ja.home.ssi.bg>
[-- Attachment #1.1: Type: text/plain, Size: 2349 bytes --]
Thanks, patch-v3 has been sent. please check it again.
> Hello,
>
> On Wed, 25 Jul 2018, Tan Hu wrote:
>
> > We came across infinite loop in ipvs when using ipvs in docker
> > env.
> >
> > When ipvs receives new packets and cannot find an ipvs connection,
> > it will create a new connection, then if the dest is unavailable
> > (i.e. IP_VS_DEST_F_AVAILABLE), the packet will be dropped sliently.
> >
> > But if the dropped packet is the first packet of this connection,
> > the connection control timer never has a chance to start and the
> > ipvs connection cannot be released. This will lead to memory leak, or
> > infinite loop in cleanup_net() when net namespace is released like
> > this:
> >
> > ip_vs_conn_net_cleanup at ffffffffa0a9f31a [ip_vs]
> > __ip_vs_cleanup at ffffffffa0a9f60a [ip_vs]
> > ops_exit_list at ffffffff81567a49
> > cleanup_net at ffffffff81568b40
> > process_one_work at ffffffff810a851b
> > worker_thread at ffffffff810a9356
> > kthread at ffffffff810b0b6f
> > ret_from_fork at ffffffff81697a18
> >
> > race condition:
> > CPU1 CPU2
> > ip_vs_in()
> > ip_vs_conn_new()
> > ip_vs_del_dest()
> > __ip_vs_unlink_dest()
> > ~IP_VS_DEST_F_AVAILABLE
> > cp->dest && !IP_VS_DEST_F_AVAILABLE
> > __ip_vs_conn_put
> > ...
> > cleanup_net ---> infinite looping
> >
> > Fix this by checking whether the timer already started.
> >
> > Signed-off-by: Tan Hu <tan.hu@zte.com.cn>
> > Reviewed-by: Jiang Biao <jiang.biao2@zte.com.cn>
> > ---
> > v2: fix use-after-free in CONN_ONE_PACKET case suggested by Julian Anastasov
> >
> > net/netfilter/ipvs/ip_vs_core.c | 15 +++++++++++----
> > 1 file changed, 11 insertions(+), 4 deletions(-)
> >
> > diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
> > index 0679dd1..a17104f 100644
> > --- a/net/netfilter/ipvs/ip_vs_core.c
> > +++ b/net/netfilter/ipvs/ip_vs_core.c
> > @@ -1972,13 +1972,20 @@ static int ip_vs_in_icmp_v6(struct netns_ipvs *ipvs, struct sk_buff *skb,
> > if (cp->dest && !(cp->dest->flags & IP_VS_DEST_F_AVAILABLE)) {
> > /* the destination server is not available */
> >
> > - if (sysctl_expire_nodest_conn(ipvs)) {
> > + __u32 flags = cp->flags;
>
> Ops, now scripts/checkpatch.pl --strict /tmp/file.patch
> is complaining about extra trailing space in above line.
> You can also remove the empty line above the new code...
>
> Regards
^ permalink raw reply
* [PATCH net-next] lan743x: Make symbol lan743x_pm_ops static
From: Wei Yongjun @ 2018-07-25 6:11 UTC (permalink / raw)
To: Bryan Whitehead, Microchip Linux Driver Support
Cc: Wei Yongjun, netdev, kernel-janitors
Fixes the following sparse warning:
drivers/net/ethernet/microchip/lan743x_main.c:2944:25: warning:
symbol 'lan743x_pm_ops' was not declared. Should it be static?
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
---
drivers/net/ethernet/microchip/lan743x_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/microchip/lan743x_main.c b/drivers/net/ethernet/microchip/lan743x_main.c
index cd41911..bb323f2 100644
--- a/drivers/net/ethernet/microchip/lan743x_main.c
+++ b/drivers/net/ethernet/microchip/lan743x_main.c
@@ -2941,7 +2941,7 @@ static int lan743x_pm_resume(struct device *dev)
return 0;
}
-const struct dev_pm_ops lan743x_pm_ops = {
+static const struct dev_pm_ops lan743x_pm_ops = {
SET_SYSTEM_SLEEP_PM_OPS(lan743x_pm_suspend, lan743x_pm_resume)
};
#endif /*CONFIG_PM */
^ permalink raw reply related
* [PATCH net-next] net: igmp: make function __ip_mc_inc_group() static
From: Wei Yongjun @ 2018-07-25 6:06 UTC (permalink / raw)
To: Alexey Kuznetsov, Hideaki YOSHIFUJI, Hangbin Liu
Cc: Wei Yongjun, netdev, kernel-janitors
Fixes the following sparse warnings:
net/ipv4/igmp.c:1391:6: warning:
symbol '__ip_mc_inc_group' was not declared. Should it be static?
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
---
net/ipv4/igmp.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index 598333b..c9f8d46 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -1388,7 +1388,8 @@ static void ip_mc_hash_remove(struct in_device *in_dev,
/*
* A socket has joined a multicast group on device dev.
*/
-void __ip_mc_inc_group(struct in_device *in_dev, __be32 addr, unsigned int mode)
+static void __ip_mc_inc_group(struct in_device *in_dev, __be32 addr,
+ unsigned int mode)
{
struct ip_mc_list *im;
#ifdef CONFIG_IP_MULTICAST
^ permalink raw reply related
* [PATCH net-next] tcp: make function tcp_retransmit_stamp() static
From: Wei Yongjun @ 2018-07-25 6:06 UTC (permalink / raw)
To: Eric Dumazet, Jon Maxwell, Alexey Kuznetsov, Hideaki YOSHIFUJI
Cc: Wei Yongjun, netdev, kernel-janitors
Fixes the following sparse warnings:
net/ipv4/tcp_timer.c:25:5: warning:
symbol 'tcp_retransmit_stamp' was not declared. Should it be static?
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
---
net/ipv4/tcp_timer.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index a242f88..7fdf222 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -22,7 +22,7 @@
#include <linux/gfp.h>
#include <net/tcp.h>
-u32 tcp_retransmit_stamp(const struct sock *sk)
+static u32 tcp_retransmit_stamp(const struct sock *sk)
{
u32 start_ts = tcp_sk(sk)->retrans_stamp;
^ permalink raw reply related
* RE: [net-next v6 1/2] net/tls: Use socket data_ready callback on record availability
From: Vakul Garg @ 2018-07-25 5:51 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, borisp@mellanox.com, aviadye@mellanox.com,
davejwatson@fb.com
In-Reply-To: <20180724.131235.814794866288330271.davem@davemloft.net>
> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net]
> Sent: Wednesday, July 25, 2018 1:43 AM
> To: Vakul Garg <vakul.garg@nxp.com>
> Cc: netdev@vger.kernel.org; borisp@mellanox.com;
> aviadye@mellanox.com; davejwatson@fb.com
> Subject: Re: [net-next v6 1/2] net/tls: Use socket data_ready callback on
> record availability
>
> From: Vakul Garg <vakul.garg@nxp.com>
> Date: Tue, 24 Jul 2018 15:44:02 +0530
>
> > On receipt of a complete tls record, use socket's saved data_ready
> > callback instead of state_change callback.
> >
> > Signed-off-by: Vakul Garg <vakul.garg@nxp.com>
>
> I don't think this is correct.
>
> Here, the stream parser has given us a complete TLS record.
>
> But we haven't decrypted this packet yet. It sits on the stream parser's
> queue to be processed by tls_sw_recvmsg(), not the saved socket's receive
> queue.
I understand that at this point in code, the TLS record is still queued in encrypted
state. But the decryption happens inline when tls_sw_recvmsg() gets invokved.
So it should be ok to notify the waiting context about the availability of data as soon
as we could collect a full TLS record.
For new data availability notification, sk_data_ready callback should be more
more appropriate. It points to sock_def_readable() which wakes up specifically for
EPOLLIN event.
This is in contrast to the socket callback sk_state_change which points
to sock_def_wakeup() which issues a wakeup unconditionally (without event mask).
^ permalink raw reply
* 答复: [PATCH][v2] netfilter: use kvzalloc to allocate memory for hashtable
From: Li,Rongqing @ 2018-07-25 5:49 UTC (permalink / raw)
To: Eric Dumazet, netdev@vger.kernel.org, pablo@netfilter.org,
kadlec@blackhole.kfki.hu, fw@strlen.de,
netfilter-devel@vger.kernel.org, coreteam@netfilter.org,
edumazet@google.com
In-Reply-To: <c86c7a1d-863d-229c-0383-c2e50aa6c5a2@gmail.com>
> -----邮件原件-----
> 发件人: Eric Dumazet [mailto:eric.dumazet@gmail.com]
> 发送时间: 2018年7月25日 13:45
> 收件人: Li,Rongqing <lirongqing@baidu.com>; netdev@vger.kernel.org;
> pablo@netfilter.org; kadlec@blackhole.kfki.hu; fw@strlen.de; netfilter-
> devel@vger.kernel.org; coreteam@netfilter.org; edumazet@google.com
> 主题: Re: [PATCH][v2] netfilter: use kvzalloc to allocate memory for
> hashtable
>
>
>
> On 07/24/2018 10:34 PM, Li RongQing wrote:
> > nf_ct_alloc_hashtable is used to allocate memory for conntrack, NAT
> > bysrc and expectation hashtable. Assuming 64k bucket size, which means
> > 7th order page allocation, __get_free_pages, called by
> > nf_ct_alloc_hashtable, will trigger the direct memory reclaim and
> > stall for a long time, when system has lots of memory stress
>
> ...
>
> > sz = nr_slots * sizeof(struct hlist_nulls_head);
> > - hash = (void *)__get_free_pages(GFP_KERNEL | __GFP_NOWARN |
> __GFP_ZERO,
> > - get_order(sz));
> > - if (!hash)
> > - hash = vzalloc(sz);
> > + hash = kvzalloc(sz, GFP_KERNEL);
>
>
> You could remove the @sz computation and call
>
> hash = kvcalloc(nr_slots, sizeof(struct hlist_nulls_head), GFP_KERNEL);
>
> Thanks to kvmalloc_array() check, you also could remove the :
>
> if (nr_slots > (UINT_MAX / sizeof(struct hlist_nulls_head)))
> return NULL;
>
> That would remove a lot of stuff now we have proper helpers.
Ok, I will send v3
Thanks
-RongQing
^ permalink raw reply
* Re: [PATCH][v2] netfilter: use kvzalloc to allocate memory for hashtable
From: Eric Dumazet @ 2018-07-25 5:44 UTC (permalink / raw)
To: Li RongQing, netdev, pablo, kadlec, fw, netfilter-devel, coreteam,
edumazet
In-Reply-To: <1532496894-17200-1-git-send-email-lirongqing@baidu.com>
On 07/24/2018 10:34 PM, Li RongQing wrote:
> nf_ct_alloc_hashtable is used to allocate memory for conntrack,
> NAT bysrc and expectation hashtable. Assuming 64k bucket size,
> which means 7th order page allocation, __get_free_pages, called
> by nf_ct_alloc_hashtable, will trigger the direct memory reclaim
> and stall for a long time, when system has lots of memory stress
...
> sz = nr_slots * sizeof(struct hlist_nulls_head);
> - hash = (void *)__get_free_pages(GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO,
> - get_order(sz));
> - if (!hash)
> - hash = vzalloc(sz);
> + hash = kvzalloc(sz, GFP_KERNEL);
You could remove the @sz computation and call
hash = kvcalloc(nr_slots, sizeof(struct hlist_nulls_head), GFP_KERNEL);
Thanks to kvmalloc_array() check, you also could remove the :
if (nr_slots > (UINT_MAX / sizeof(struct hlist_nulls_head)))
return NULL;
That would remove a lot of stuff now we have proper helpers.
^ permalink raw reply
* Re: [PATCH bpf-next] bpf: add End.DT6 action to bpf_lwt_seg6_action helper
From: Daniel Borkmann @ 2018-07-25 5:40 UTC (permalink / raw)
To: Martin KaFai Lau, Mathieu Xhonneux; +Cc: netdev, alexei.starovoitov
In-Reply-To: <20180724171423.xjm4ikr6nfhkisms@kafai-mbp>
On 07/24/2018 07:14 PM, Martin KaFai Lau wrote:
> On Tue, Jul 24, 2018 at 04:59:54PM +0000, Mathieu Xhonneux wrote:
>> The seg6local LWT provides the End.DT6 action, which allows to
>> decapsulate an outer IPv6 header containing a Segment Routing Header
>> (SRH), full specification is available here:
>>
>> https://urldefense.proofpoint.com/v2/url?u=https-3A__tools.ietf.org_html_draft-2Dfilsfils-2Dspring-2Dsrv6-2Dnetwork-2Dprogramming-2D05&d=DwIBAg&c=5VD0RTtNlTh3ycd41b3MUw&r=VQnoQ7LvghIj0gVEaiQSUw&m=c61PGnhPMmCUcL5lpyBsxOmsBU2mU5KFY0-Ioo-pBC4&s=mzShtRc5ofzfknAuqoehbGN1ifA17aKihiVLJVfkuZ8&e=
>>
>> This patch adds this action now to the seg6local BPF
>> interface. Since it is not mandatory that the inner IPv6 header also
>> contains a SRH, seg6_bpf_srh_state has been extended with a pointer to
>> a possible SRH of the outermost IPv6 header. This helps assessing if the
>> validation must be triggered or not, and avoids some calls to
>> ipv6_find_hdr.
>>
>> Signed-off-by: Mathieu Xhonneux <m.xhonneux@gmail.com>
[...]
>> +
>> static int input_action_end_bpf(struct sk_buff *skb,
>> struct seg6_local_lwt *slwt)
>> {
>> struct seg6_bpf_srh_state *srh_state =
>> this_cpu_ptr(&seg6_bpf_srh_states);
>> - struct seg6_bpf_srh_state local_srh_state;
>> struct ipv6_sr_hdr *srh;
>> - int srhoff = 0;
>> int ret;
>>
>> srh = get_and_validate_srh(skb);
>> @@ -478,6 +499,7 @@ static int input_action_end_bpf(struct sk_buff *skb,
>> * which is also accessed by the bpf_lwt_seg6_* helpers
>> */
>> preempt_disable();
>> + srh_state->srh = srh;
>> srh_state->hdrlen = srh->hdrlen << 3;
>> srh_state->valid = 1;
>>
>> @@ -486,9 +508,6 @@ static int input_action_end_bpf(struct sk_buff *skb,
>> ret = bpf_prog_run_save_cb(slwt->bpf.prog, skb);
>> rcu_read_unlock();
>>
>> - local_srh_state = *srh_state;
>> - preempt_enable();
>> -
>> switch (ret) {
>> case BPF_OK:
>> case BPF_REDIRECT:
>> @@ -500,24 +519,17 @@ static int input_action_end_bpf(struct sk_buff *skb,
>> goto drop;
>> }
>>
>> - if (unlikely((local_srh_state.hdrlen & 7) != 0))
>> - goto drop;
>> -
>> - if (ipv6_find_hdr(skb, &srhoff, IPPROTO_ROUTING, NULL, NULL) < 0)
>> - goto drop;
>> - srh = (struct ipv6_sr_hdr *)(skb->data + srhoff);
>> - srh->hdrlen = (u8)(local_srh_state.hdrlen >> 3);
>> -
>> - if (!local_srh_state.valid &&
>> - unlikely(!seg6_validate_srh(srh, (srh->hdrlen + 1) << 3)))
>> + if (srh_state->srh && !seg6_bpf_has_valid_srh(skb))
>> goto drop;
>>
>> + preempt_enable();
>> if (ret != BPF_REDIRECT)
>> seg6_lookup_nexthop(skb, NULL, 0);
>>
>> return dst_input(skb);
>>
>> drop:
>> + preempt_enable();
> For this drop case at the beginning of this function:
>
> srh = get_and_validate_srh(skb);
> if (!srh)
> goto drop;
>
> preempt_disable() was not called yet?
Agree, this is buggy.
^ permalink raw reply
* Re: [PATCH rdma-next v2 0/8] Support mlx5 flow steering with RAW data
From: Leon Romanovsky @ 2018-07-25 5:37 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: Doug Ledford, RDMA mailing list, Yishai Hadas, Saeed Mahameed,
linux-netdev
In-Reply-To: <20180724201300.GQ532@ziepe.ca>
[-- Attachment #1: Type: text/plain, Size: 11819 bytes --]
On Tue, Jul 24, 2018 at 02:13:00PM -0600, Jason Gunthorpe wrote:
> On Tue, Jul 24, 2018 at 08:56:09AM +0300, Leon Romanovsky wrote:
> > On Mon, Jul 23, 2018 at 08:42:36PM -0600, Jason Gunthorpe wrote:
> > > On Mon, Jul 23, 2018 at 03:25:04PM +0300, Leon Romanovsky wrote:
> > > > From: Leon Romanovsky <leonro@mellanox.com>
> > > >
> > > > Changelog:
> > > > v1->v2:
> > > > * Fix matcher to use the correct size.
> > > > * Rephrase commit log of the first patch.
> > > > v0->v1:
> > > > * Fixed ADD_UVERBS_ATTRIBUTES_SIMPLE macro to pass the real address.
> > > > ?* Replaced UA_ALLOC_AND_COPY to regular copy_from
> > > > * Added UVERBS_ATTR_NO_DATA new macro for cleaner code.
> > > > * Used ib_dev from uobj when it exists.
> > > > * ib_is_destroy_retryable was replaced by ib_destroy_usecnt
> > > >
> > > > >From Yishai:
> > > >
> > > > This series introduces vendor create and destroy flow methods on the
> > > > uverbs flow object by using the KABI infra-structure.
> > > >
> > > > It's done in a way that enables the driver to get its specific device
> > > > attributes in a raw data to match its underlay specification while still
> > > > using the generic ib_flow object for cleanup and code sharing.
> > > >
> > > > In addition, a specific mlx5 matcher object and its create/destroy
> > > > methods were introduced. This object matches the underlay flow steering
> > > > mask specification and is used as part of mlx5 create flow input data.
> > > >
> > > > This series supports IB_QP/TIR as its flow steering destination as
> > > > applicable today via the ib_create_flow API, however, it adds also an
> > > > option to work with DEVX object which its destination can be both TIR
> > > > and flow table.
> > > >
> > > > Few changes were done in the mlx5 core layer to support forward
> > > > compatible for the device specification raw data and to support flow
> > > > table when the DEVX destination is used.
> > > >
> > > > As part of this series the default IB destroy handler
> > > > (i.e. uverbs_destroy_def_handler()) was exposed from IB core to be
> > > > used by the drivers and existing code was refactored to use it.
> > > >
> > > > Thanks
> > > >
> > > > Yishai Hadas (8):
> > > > net/mlx5: Add forward compatible support for the FTE match data
> > > > net/mlx5: Add support for flow table destination number
> > > > IB/mlx5: Introduce flow steering matcher object
> > > > IB: Consider ib_flow creation by the KABI infrastructure
> > > > IB/mlx5: Introduce vendor create and destroy flow methods
> > > > IB/mlx5: Support adding flow steering rule by raw data
> > > > IB/mlx5: Add support for a flow table destination
> > > > IB/mlx5: Expose vendor flow trees
> > >
> > > This seems fine to me. Can you send the mlx5 shared branch for the
> > > first two patches?
> >
> > I applied two first patches with Acked-by from Saeed to mlx5-next
> >
> > 664000b6bb43 net/mlx5: Add support for flow table destination number
> > 2aada6c0c96e net/mlx5: Add forward compatible support for the FTE match data
>
> Okay, I merged the mlx5 branch and applied the series to for-next.
>
> There was a trivial build failure with !CONFIG_INFINIBAND_USER_ACCESS
> and a bunch of annoying check patch warnings regarding tabs and
> leading white space. While I was fixing those I fixed the long lines
> too, they had no reason to be long.
>
> Also, I would like to keep the specs consistently formatted according
> to clang-format with 'BinPackParameters: true', so I reflowed them as
> well.
I'm using default VIM clang-format.py without anything in .clang-format.
Do you have an extra definitions there, except BinPackParameters?
Thanks for fixing.
>
> Please check and let me know if I made an error, diff is below.
>
> Jason
>
> diff --git a/drivers/infiniband/hw/mlx5/flow.c b/drivers/infiniband/hw/mlx5/flow.c
> index c94ee1a43e2c3f..ee398a9b5f26b0 100644
> --- a/drivers/infiniband/hw/mlx5/flow.c
> +++ b/drivers/infiniband/hw/mlx5/flow.c
> @@ -50,8 +50,8 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
> int inlen;
> bool dest_devx, dest_qp;
> struct ib_qp *qp = NULL;
> - struct ib_uobject *uobj = uverbs_attr_get_uobject(attrs,
> - MLX5_IB_ATTR_CREATE_FLOW_HANDLE);
> + struct ib_uobject *uobj =
> + uverbs_attr_get_uobject(attrs, MLX5_IB_ATTR_CREATE_FLOW_HANDLE);
> struct mlx5_ib_dev *dev = to_mdev(uobj->context->device);
>
> if (!capable(CAP_NET_RAW))
> @@ -66,7 +66,8 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
> return -EINVAL;
>
> if (dest_devx) {
> - devx_obj = uverbs_attr_get_obj(attrs, MLX5_IB_ATTR_CREATE_FLOW_DEST_DEVX);
> + devx_obj = uverbs_attr_get_obj(
> + attrs, MLX5_IB_ATTR_CREATE_FLOW_DEST_DEVX);
> if (IS_ERR(devx_obj))
> return PTR_ERR(devx_obj);
>
> @@ -97,8 +98,8 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
> if (dev->rep)
> return -ENOTSUPP;
>
> - cmd_in = uverbs_attr_get_alloced_ptr(attrs,
> - MLX5_IB_ATTR_CREATE_FLOW_MATCH_VALUE);
> + cmd_in = uverbs_attr_get_alloced_ptr(
> + attrs, MLX5_IB_ATTR_CREATE_FLOW_MATCH_VALUE);
> inlen = uverbs_attr_get_len(attrs,
> MLX5_IB_ATTR_CREATE_FLOW_MATCH_VALUE);
> fs_matcher = uverbs_attr_get_obj(attrs,
> @@ -127,9 +128,9 @@ static int flow_matcher_cleanup(struct ib_uobject *uobject,
> return 0;
> }
>
> -static int UVERBS_HANDLER(MLX5_IB_METHOD_FLOW_MATCHER_CREATE)(struct ib_device *ib_dev,
> - struct ib_uverbs_file *file,
> - struct uverbs_attr_bundle *attrs)
> +static int UVERBS_HANDLER(MLX5_IB_METHOD_FLOW_MATCHER_CREATE)(
> + struct ib_device *ib_dev, struct ib_uverbs_file *file,
> + struct uverbs_attr_bundle *attrs)
> {
> struct ib_uobject *uobj = uverbs_attr_get_uobject(
> attrs, MLX5_IB_ATTR_FLOW_MATCHER_CREATE_HANDLE);
> @@ -141,16 +142,16 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_FLOW_MATCHER_CREATE)(struct ib_device *
> if (!obj)
> return -ENOMEM;
>
> - obj->mask_len = uverbs_attr_get_len(attrs,
> - MLX5_IB_ATTR_FLOW_MATCHER_MATCH_MASK);
> + obj->mask_len = uverbs_attr_get_len(
> + attrs, MLX5_IB_ATTR_FLOW_MATCHER_MATCH_MASK);
> err = uverbs_copy_from(&obj->matcher_mask,
> attrs,
> MLX5_IB_ATTR_FLOW_MATCHER_MATCH_MASK);
> if (err)
> goto end;
>
> - obj->flow_type = uverbs_attr_get_enum_id(attrs,
> - MLX5_IB_ATTR_FLOW_MATCHER_FLOW_TYPE);
> + obj->flow_type = uverbs_attr_get_enum_id(
> + attrs, MLX5_IB_ATTR_FLOW_MATCHER_FLOW_TYPE);
>
> if (obj->flow_type == MLX5_IB_FLOW_TYPE_NORMAL) {
> err = uverbs_copy_from(&obj->priority,
> @@ -182,21 +183,22 @@ DECLARE_UVERBS_NAMED_METHOD(
> UVERBS_OBJECT_FLOW,
> UVERBS_ACCESS_NEW,
> UA_MANDATORY),
> - UVERBS_ATTR_PTR_IN(MLX5_IB_ATTR_CREATE_FLOW_MATCH_VALUE,
> - UVERBS_ATTR_SIZE(1, sizeof(struct mlx5_ib_match_params)),
> - UA_MANDATORY,
> - UA_ALLOC_AND_COPY),
> + UVERBS_ATTR_PTR_IN(
> + MLX5_IB_ATTR_CREATE_FLOW_MATCH_VALUE,
> + UVERBS_ATTR_SIZE(1, sizeof(struct mlx5_ib_match_params)),
> + UA_MANDATORY,
> + UA_ALLOC_AND_COPY),
> UVERBS_ATTR_IDR(MLX5_IB_ATTR_CREATE_FLOW_MATCHER,
> MLX5_IB_OBJECT_FLOW_MATCHER,
> UVERBS_ACCESS_READ,
> UA_MANDATORY),
> - UVERBS_ATTR_IDR(MLX5_IB_ATTR_CREATE_FLOW_DEST_QP, UVERBS_OBJECT_QP,
> + UVERBS_ATTR_IDR(MLX5_IB_ATTR_CREATE_FLOW_DEST_QP,
> + UVERBS_OBJECT_QP,
> UVERBS_ACCESS_READ),
> UVERBS_ATTR_IDR(MLX5_IB_ATTR_CREATE_FLOW_DEST_DEVX,
> MLX5_IB_OBJECT_DEVX_OBJ,
> UVERBS_ACCESS_READ));
>
> -
> DECLARE_UVERBS_NAMED_METHOD_DESTROY(
> MLX5_IB_METHOD_DESTROY_FLOW,
> UVERBS_ATTR_IDR(MLX5_IB_ATTR_CREATE_FLOW_HANDLE,
> @@ -212,36 +214,34 @@ ADD_UVERBS_METHODS(mlx5_ib_fs,
> DECLARE_UVERBS_NAMED_METHOD(
> MLX5_IB_METHOD_FLOW_MATCHER_CREATE,
> UVERBS_ATTR_IDR(MLX5_IB_ATTR_FLOW_MATCHER_CREATE_HANDLE,
> - MLX5_IB_OBJECT_FLOW_MATCHER,
> - UVERBS_ACCESS_NEW,
> - UA_MANDATORY),
> + MLX5_IB_OBJECT_FLOW_MATCHER,
> + UVERBS_ACCESS_NEW,
> + UA_MANDATORY),
> UVERBS_ATTR_PTR_IN(
> MLX5_IB_ATTR_FLOW_MATCHER_MATCH_MASK,
> UVERBS_ATTR_SIZE(1, sizeof(struct mlx5_ib_match_params)),
> UA_MANDATORY),
> - UVERBS_ATTR_ENUM_IN(
> - MLX5_IB_ATTR_FLOW_MATCHER_FLOW_TYPE,
> - mlx5_ib_flow_type,
> - UA_MANDATORY),
> - UVERBS_ATTR_PTR_IN(
> - MLX5_IB_ATTR_FLOW_MATCHER_MATCH_CRITERIA,
> - UVERBS_ATTR_TYPE(u8),
> - UA_MANDATORY));
> + UVERBS_ATTR_ENUM_IN(MLX5_IB_ATTR_FLOW_MATCHER_FLOW_TYPE,
> + mlx5_ib_flow_type,
> + UA_MANDATORY),
> + UVERBS_ATTR_PTR_IN(MLX5_IB_ATTR_FLOW_MATCHER_MATCH_CRITERIA,
> + UVERBS_ATTR_TYPE(u8),
> + UA_MANDATORY));
>
> DECLARE_UVERBS_NAMED_METHOD_DESTROY(
> MLX5_IB_METHOD_FLOW_MATCHER_DESTROY,
> UVERBS_ATTR_IDR(MLX5_IB_ATTR_FLOW_MATCHER_DESTROY_HANDLE,
> - MLX5_IB_OBJECT_FLOW_MATCHER,
> - UVERBS_ACCESS_DESTROY,
> - UA_MANDATORY));
> + MLX5_IB_OBJECT_FLOW_MATCHER,
> + UVERBS_ACCESS_DESTROY,
> + UA_MANDATORY));
>
> DECLARE_UVERBS_NAMED_OBJECT(MLX5_IB_OBJECT_FLOW_MATCHER,
> - UVERBS_TYPE_ALLOC_IDR(flow_matcher_cleanup),
> - &UVERBS_METHOD(MLX5_IB_METHOD_FLOW_MATCHER_CREATE),
> - &UVERBS_METHOD(MLX5_IB_METHOD_FLOW_MATCHER_DESTROY));
> + UVERBS_TYPE_ALLOC_IDR(flow_matcher_cleanup),
> + &UVERBS_METHOD(MLX5_IB_METHOD_FLOW_MATCHER_CREATE),
> + &UVERBS_METHOD(MLX5_IB_METHOD_FLOW_MATCHER_DESTROY));
>
> DECLARE_UVERBS_OBJECT_TREE(flow_objects,
> - &UVERBS_OBJECT(MLX5_IB_OBJECT_FLOW_MATCHER));
> + &UVERBS_OBJECT(MLX5_IB_OBJECT_FLOW_MATCHER));
>
> int mlx5_ib_get_flow_trees(const struct uverbs_object_tree_def **root)
> {
> diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
> index 5f08b69f8a4f60..ec8410d3c4eb2a 100644
> --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
> +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
> @@ -1232,11 +1232,9 @@ int mlx5_ib_devx_create(struct mlx5_ib_dev *dev,
> void mlx5_ib_devx_destroy(struct mlx5_ib_dev *dev,
> struct mlx5_ib_ucontext *context);
> const struct uverbs_object_tree_def *mlx5_ib_get_devx_tree(void);
> -struct mlx5_ib_flow_handler *mlx5_ib_raw_fs_rule_add(struct mlx5_ib_dev *dev,
> - struct mlx5_ib_flow_matcher *fs_matcher,
> - void *cmd_in,
> - int inlen, int dest_id,
> - int dest_type);
> +struct mlx5_ib_flow_handler *mlx5_ib_raw_fs_rule_add(
> + struct mlx5_ib_dev *dev, struct mlx5_ib_flow_matcher *fs_matcher,
> + void *cmd_in, int inlen, int dest_id, int dest_type);
> bool mlx5_ib_devx_is_flow_dest(void *obj, int *dest_id, int *dest_type);
> int mlx5_ib_get_flow_trees(const struct uverbs_object_tree_def **root);
> #else
> @@ -1247,17 +1245,22 @@ static inline void mlx5_ib_devx_destroy(struct mlx5_ib_dev *dev,
> struct mlx5_ib_ucontext *context) {}
> static inline const struct uverbs_object_tree_def *
> mlx5_ib_get_devx_tree(void) { return NULL; }
> -static inline struct mlx5_ib_flow_handler *
> -mlx5_ib_raw_fs_rule_add(struct mlx5_ib_dev *dev,
> - struct mlx5_ib_flow_matcher *fs_matcher,
> - void *cmd_in,
> - int inlen, int dest_id,
> - int dest_type) { return -EOPNOTSUPP; };
> -static inline bool
> -mlx5_ib_devx_is_flow_dest(void *obj, int *dest_id,
> - int *dest_type) { return false; };
> +static inline struct mlx5_ib_flow_handler *mlx5_ib_raw_fs_rule_add(
> + struct mlx5_ib_dev *dev, struct mlx5_ib_flow_matcher *fs_matcher,
> + void *cmd_in, int inlen, int dest_id, int dest_type)
> +{
> + return ERR_PTR(-EOPNOTSUPP);
> +}
> +static inline bool mlx5_ib_devx_is_flow_dest(void *obj, int *dest_id,
> + int *dest_type)
> +{
> + return false;
> +}
> static inline int
> -mlx5_ib_get_flow_trees(const struct uverbs_object_tree_def **root) { return 0; };
> +mlx5_ib_get_flow_trees(const struct uverbs_object_tree_def **root)
> +{
> + return 0;
> +}
> #endif
> static inline void init_query_mad(struct ib_smp *mad)
> {
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 801 bytes --]
^ permalink raw reply
* [PATCH][v2] netfilter: use kvzalloc to allocate memory for hashtable
From: Li RongQing @ 2018-07-25 5:34 UTC (permalink / raw)
To: netdev, pablo, kadlec, fw, netfilter-devel, coreteam, edumazet
nf_ct_alloc_hashtable is used to allocate memory for conntrack,
NAT bysrc and expectation hashtable. Assuming 64k bucket size,
which means 7th order page allocation, __get_free_pages, called
by nf_ct_alloc_hashtable, will trigger the direct memory reclaim
and stall for a long time, when system has lots of memory stress
so replace combination of __get_free_pages and vzalloc with
kvzalloc, which provides a fallback if no high order memory
is available, and do not retry to reclaim memory, reduce stall
and remove nf_ct_free_hashtable, since it is just a kvfree
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Signed-off-by: Wang Li <wangli39@baidu.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
---
include/net/netfilter/nf_conntrack.h | 2 --
net/netfilter/nf_conntrack_core.c | 23 +++++------------------
net/netfilter/nf_conntrack_expect.c | 2 +-
net/netfilter/nf_conntrack_helper.c | 4 ++--
net/netfilter/nf_nat_core.c | 4 ++--
5 files changed, 10 insertions(+), 25 deletions(-)
diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h
index a2b0ed025908..7e012312cd61 100644
--- a/include/net/netfilter/nf_conntrack.h
+++ b/include/net/netfilter/nf_conntrack.h
@@ -176,8 +176,6 @@ void nf_ct_netns_put(struct net *net, u8 nfproto);
*/
void *nf_ct_alloc_hashtable(unsigned int *sizep, int nulls);
-void nf_ct_free_hashtable(void *hash, unsigned int size);
-
int nf_conntrack_hash_check_insert(struct nf_conn *ct);
bool nf_ct_delete(struct nf_conn *ct, u32 pid, int report);
diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
index 8a113ca1eea2..191140c469d5 100644
--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -2022,16 +2022,6 @@ static int kill_all(struct nf_conn *i, void *data)
return net_eq(nf_ct_net(i), data);
}
-void nf_ct_free_hashtable(void *hash, unsigned int size)
-{
- if (is_vmalloc_addr(hash))
- vfree(hash);
- else
- free_pages((unsigned long)hash,
- get_order(sizeof(struct hlist_head) * size));
-}
-EXPORT_SYMBOL_GPL(nf_ct_free_hashtable);
-
void nf_conntrack_cleanup_start(void)
{
conntrack_gc_work.exiting = true;
@@ -2042,7 +2032,7 @@ void nf_conntrack_cleanup_end(void)
{
RCU_INIT_POINTER(nf_ct_hook, NULL);
cancel_delayed_work_sync(&conntrack_gc_work.dwork);
- nf_ct_free_hashtable(nf_conntrack_hash, nf_conntrack_htable_size);
+ kvfree(nf_conntrack_hash);
nf_conntrack_proto_fini();
nf_conntrack_seqadj_fini();
@@ -2120,10 +2110,7 @@ void *nf_ct_alloc_hashtable(unsigned int *sizep, int nulls)
return NULL;
sz = nr_slots * sizeof(struct hlist_nulls_head);
- hash = (void *)__get_free_pages(GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO,
- get_order(sz));
- if (!hash)
- hash = vzalloc(sz);
+ hash = kvzalloc(sz, GFP_KERNEL);
if (hash && nulls)
for (i = 0; i < nr_slots; i++)
@@ -2150,7 +2137,7 @@ int nf_conntrack_hash_resize(unsigned int hashsize)
old_size = nf_conntrack_htable_size;
if (old_size == hashsize) {
- nf_ct_free_hashtable(hash, hashsize);
+ kvfree(hash);
return 0;
}
@@ -2186,7 +2173,7 @@ int nf_conntrack_hash_resize(unsigned int hashsize)
local_bh_enable();
synchronize_net();
- nf_ct_free_hashtable(old_hash, old_size);
+ kvfree(old_hash);
return 0;
}
@@ -2350,7 +2337,7 @@ int nf_conntrack_init_start(void)
err_expect:
kmem_cache_destroy(nf_conntrack_cachep);
err_cachep:
- nf_ct_free_hashtable(nf_conntrack_hash, nf_conntrack_htable_size);
+ kvfree(nf_conntrack_hash);
return ret;
}
diff --git a/net/netfilter/nf_conntrack_expect.c b/net/netfilter/nf_conntrack_expect.c
index 3f586ba23d92..27b84231db10 100644
--- a/net/netfilter/nf_conntrack_expect.c
+++ b/net/netfilter/nf_conntrack_expect.c
@@ -712,5 +712,5 @@ void nf_conntrack_expect_fini(void)
{
rcu_barrier(); /* Wait for call_rcu() before destroy */
kmem_cache_destroy(nf_ct_expect_cachep);
- nf_ct_free_hashtable(nf_ct_expect_hash, nf_ct_expect_hsize);
+ kvfree(nf_ct_expect_hash);
}
diff --git a/net/netfilter/nf_conntrack_helper.c b/net/netfilter/nf_conntrack_helper.c
index d557a425289d..e24b762ffa1d 100644
--- a/net/netfilter/nf_conntrack_helper.c
+++ b/net/netfilter/nf_conntrack_helper.c
@@ -562,12 +562,12 @@ int nf_conntrack_helper_init(void)
return 0;
out_extend:
- nf_ct_free_hashtable(nf_ct_helper_hash, nf_ct_helper_hsize);
+ kvfree(nf_ct_helper_hash);
return ret;
}
void nf_conntrack_helper_fini(void)
{
nf_ct_extend_unregister(&helper_extend);
- nf_ct_free_hashtable(nf_ct_helper_hash, nf_ct_helper_hsize);
+ kvfree(nf_ct_helper_hash);
}
diff --git a/net/netfilter/nf_nat_core.c b/net/netfilter/nf_nat_core.c
index 6366f0c0b8c1..e2b196054dfc 100644
--- a/net/netfilter/nf_nat_core.c
+++ b/net/netfilter/nf_nat_core.c
@@ -1056,7 +1056,7 @@ static int __init nf_nat_init(void)
ret = nf_ct_extend_register(&nat_extend);
if (ret < 0) {
- nf_ct_free_hashtable(nf_nat_bysource, nf_nat_htable_size);
+ kvfree(nf_nat_bysource);
pr_err("Unable to register extension\n");
return ret;
}
@@ -1094,7 +1094,7 @@ static void __exit nf_nat_cleanup(void)
for (i = 0; i < NFPROTO_NUMPROTO; i++)
kfree(nf_nat_l4protos[i]);
synchronize_net();
- nf_ct_free_hashtable(nf_nat_bysource, nf_nat_htable_size);
+ kvfree(nf_nat_bysource);
unregister_pernet_subsys(&nat_net_ops);
}
--
2.16.2
^ permalink raw reply related
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