[PATCH net-next 0/8] mlxsw: Add Spectrum-4 support

[PATCH net-next 0/8] mlxsw: Add Spectrum-4 support

[Ido Schimmel]

This patchset adds Spectrum-4 support in mlxsw. It builds on top of a
previous patchset merged in commit 10184da91666 ("Merge branch
'mlxsw-Spectrum-4-prep'") and makes two additional changes before adding
Spectrum-4 support.

Patchset overview:

Patches #1-#2 add a few Spectrum-4 specific variants of existing ACL
keys. The new variants are needed because the size of certain key
elements (e.g., local port) was increased in Spectrum-4.

Patches #3-#6 are preparations.

Patch #7 implements the Spectrum-4 variant of the Bloom filter hash
function. The Bloom filter is used to optimize ACL lookups by
potentially skipping certain lookups if they are guaranteed not to
match. See additional info in merge commit ae6750e0a5ef ("Merge branch
'mlxsw-spectrum_acl-Add-Bloom-filter-support'").

Patch #8 finally adds Spectrum-4 support.

Amit Cohen (8):
  mlxsw: Rename virtual router flex key element
  mlxsw: Introduce flex key elements for Spectrum-4
  mlxsw: spectrum_acl_bloom_filter: Reorder functions to make the code
    more aesthetic
  mlxsw: spectrum_acl_bloom_filter: Make mlxsw_sp_acl_bf_key_encode()
    more flexible
  mlxsw: spectrum_acl_bloom_filter: Rename Spectrum-2 specific objects
    for future use
  mlxsw: Add operations structure for bloom filter calculation
  mlxsw: spectrum_acl_bloom_filter: Add support for Spectrum-4
    calculation
  mlxsw: spectrum: Extend to support Spectrum-4 ASIC

 drivers/net/ethernet/mellanox/mlxsw/Kconfig   |   2 +-
 .../mellanox/mlxsw/core_acl_flex_keys.c       |   4 +-
 .../mellanox/mlxsw/core_acl_flex_keys.h       |   4 +-
 drivers/net/ethernet/mellanox/mlxsw/pci.h     |   1 +
 drivers/net/ethernet/mellanox/mlxsw/reg.h     |   1 +
 .../net/ethernet/mellanox/mlxsw/spectrum.c    |  97 +++++
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |   6 +
 .../mellanox/mlxsw/spectrum2_mr_tcam.c        |  12 +-
 .../mlxsw/spectrum_acl_bloom_filter.c         | 351 +++++++++++++++---
 .../mellanox/mlxsw/spectrum_acl_flex_keys.c   |  46 ++-
 .../mellanox/mlxsw/spectrum_acl_tcam.h        |   6 +
 11 files changed, 468 insertions(+), 62 deletions(-)

-- 
2.33.1

[PATCH net-next 1/8] mlxsw: Rename virtual router flex key element

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

In Spectrum-4, the size of the virtual router ACL key element increased
from 11 bits to 12 bits.

In order to reuse the existing virtual router ACL key element
enumerators for Spectrum-4, rename 'VIRT_ROUTER_8_10' and
'VIRT_ROUTER_0_7' to 'VIRT_ROUTER_MSB' and 'VIRT_ROUTER_LSB',
respectively.

No functional changes intended.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c |  4 ++--
 .../net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h |  4 ++--
 .../net/ethernet/mellanox/mlxsw/spectrum2_mr_tcam.c  | 12 ++++++------
 .../ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c |  4 ++--
 4 files changed, 12 insertions(+), 12 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 f1b09c2f9eda..bd1a51a0a540 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.c
@@ -32,8 +32,8 @@ static const struct mlxsw_afk_element_info mlxsw_afk_element_infos[] = {
 	MLXSW_AFK_ELEMENT_INFO_U32(IP_TTL_, 0x18, 0, 8),
 	MLXSW_AFK_ELEMENT_INFO_U32(IP_ECN, 0x18, 9, 2),
 	MLXSW_AFK_ELEMENT_INFO_U32(IP_DSCP, 0x18, 11, 6),
-	MLXSW_AFK_ELEMENT_INFO_U32(VIRT_ROUTER_8_10, 0x18, 17, 3),
-	MLXSW_AFK_ELEMENT_INFO_U32(VIRT_ROUTER_0_7, 0x18, 20, 8),
+	MLXSW_AFK_ELEMENT_INFO_U32(VIRT_ROUTER_MSB, 0x18, 17, 3),
+	MLXSW_AFK_ELEMENT_INFO_U32(VIRT_ROUTER_LSB, 0x18, 20, 8),
 	MLXSW_AFK_ELEMENT_INFO_BUF(SRC_IP_96_127, 0x20, 4),
 	MLXSW_AFK_ELEMENT_INFO_BUF(SRC_IP_64_95, 0x24, 4),
 	MLXSW_AFK_ELEMENT_INFO_BUF(SRC_IP_32_63, 0x28, 4),
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 a47a17c04c62..3a037fe47211 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_keys.h
@@ -33,8 +33,8 @@ enum mlxsw_afk_element {
 	MLXSW_AFK_ELEMENT_IP_TTL_,
 	MLXSW_AFK_ELEMENT_IP_ECN,
 	MLXSW_AFK_ELEMENT_IP_DSCP,
-	MLXSW_AFK_ELEMENT_VIRT_ROUTER_8_10,
-	MLXSW_AFK_ELEMENT_VIRT_ROUTER_0_7,
+	MLXSW_AFK_ELEMENT_VIRT_ROUTER_MSB,
+	MLXSW_AFK_ELEMENT_VIRT_ROUTER_LSB,
 	MLXSW_AFK_ELEMENT_MAX,
 };
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum2_mr_tcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum2_mr_tcam.c
index a11d911302f1..e4f4cded2b6f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum2_mr_tcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum2_mr_tcam.c
@@ -45,8 +45,8 @@ static int mlxsw_sp2_mr_tcam_bind_group(struct mlxsw_sp *mlxsw_sp,
 }
 
 static const enum mlxsw_afk_element mlxsw_sp2_mr_tcam_usage_ipv4[] = {
-		MLXSW_AFK_ELEMENT_VIRT_ROUTER_8_10,
-		MLXSW_AFK_ELEMENT_VIRT_ROUTER_0_7,
+		MLXSW_AFK_ELEMENT_VIRT_ROUTER_MSB,
+		MLXSW_AFK_ELEMENT_VIRT_ROUTER_LSB,
 		MLXSW_AFK_ELEMENT_SRC_IP_0_31,
 		MLXSW_AFK_ELEMENT_DST_IP_0_31,
 };
@@ -89,8 +89,8 @@ static void mlxsw_sp2_mr_tcam_ipv4_fini(struct mlxsw_sp2_mr_tcam *mr_tcam)
 }
 
 static const enum mlxsw_afk_element mlxsw_sp2_mr_tcam_usage_ipv6[] = {
-		MLXSW_AFK_ELEMENT_VIRT_ROUTER_8_10,
-		MLXSW_AFK_ELEMENT_VIRT_ROUTER_0_7,
+		MLXSW_AFK_ELEMENT_VIRT_ROUTER_MSB,
+		MLXSW_AFK_ELEMENT_VIRT_ROUTER_LSB,
 		MLXSW_AFK_ELEMENT_SRC_IP_96_127,
 		MLXSW_AFK_ELEMENT_SRC_IP_64_95,
 		MLXSW_AFK_ELEMENT_SRC_IP_32_63,
@@ -189,10 +189,10 @@ mlxsw_sp2_mr_tcam_rule_parse(struct mlxsw_sp_acl_rule *rule,
 
 	rulei = mlxsw_sp_acl_rule_rulei(rule);
 	rulei->priority = priority;
-	mlxsw_sp_acl_rulei_keymask_u32(rulei, MLXSW_AFK_ELEMENT_VIRT_ROUTER_0_7,
+	mlxsw_sp_acl_rulei_keymask_u32(rulei, MLXSW_AFK_ELEMENT_VIRT_ROUTER_LSB,
 				       key->vrid, GENMASK(7, 0));
 	mlxsw_sp_acl_rulei_keymask_u32(rulei,
-				       MLXSW_AFK_ELEMENT_VIRT_ROUTER_8_10,
+				       MLXSW_AFK_ELEMENT_VIRT_ROUTER_MSB,
 				       key->vrid >> 8, GENMASK(2, 0));
 	switch (key->proto) {
 	case MLXSW_SP_L3_PROTO_IPV4:
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c
index 279c241f76f0..fa66316234c4 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c
@@ -168,8 +168,8 @@ static struct mlxsw_afk_element_inst mlxsw_sp_afk_element_info_ipv4_2[] = {
 };
 
 static struct mlxsw_afk_element_inst mlxsw_sp_afk_element_info_ipv4_4[] = {
-	MLXSW_AFK_ELEMENT_INST_U32(VIRT_ROUTER_0_7, 0x04, 24, 8),
-	MLXSW_AFK_ELEMENT_INST_U32(VIRT_ROUTER_8_10, 0x00, 0, 3),
+	MLXSW_AFK_ELEMENT_INST_U32(VIRT_ROUTER_LSB, 0x04, 24, 8),
+	MLXSW_AFK_ELEMENT_INST_U32(VIRT_ROUTER_MSB, 0x00, 0, 3),
 };
 
 static struct mlxsw_afk_element_inst mlxsw_sp_afk_element_info_ipv6_0[] = {
-- 
2.33.1

[PATCH net-next 2/8] mlxsw: Introduce flex key elements for Spectrum-4

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Spectrum-4 ASIC will support more virtual routers and local ports
compared to the existing ASICs. Therefore, the virtual router and local
port ACL key elements need to be increased.

Introduce new key elements for Spectrum-4 to be aligned with the elements
used already for other Spectrum ASICs.

The key blocks layout is the same for Spectrum-4, so use the existing
code for encode_block() and clear_block(), just create separate blocks.

Note that size of `VIRT_ROUTER_MSB` is 4 bits in Spectrum-4,
therefore declare it using `MLXSW_AFK_ELEMENT_INST_U32()`, in order to
be able to set `.avoid_size_check` to true.
Otherwise, `mlxsw_afk_blocks_check()` will fail and warn.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |  1 +
 .../mellanox/mlxsw/spectrum_acl_flex_keys.c   | 42 +++++++++++++++++++
 2 files changed, 43 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 8445fc5c9ea3..a49316d0bd37 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -1106,6 +1106,7 @@ extern const struct mlxsw_afa_ops mlxsw_sp2_act_afa_ops;
 /* spectrum_acl_flex_keys.c */
 extern const struct mlxsw_afk_ops mlxsw_sp1_afk_ops;
 extern const struct mlxsw_afk_ops mlxsw_sp2_afk_ops;
+extern const struct mlxsw_afk_ops mlxsw_sp4_afk_ops;
 
 /* spectrum_matchall.c */
 struct mlxsw_sp_mall_ops {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c
index fa66316234c4..00c32320f891 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.c
@@ -311,3 +311,45 @@ const struct mlxsw_afk_ops mlxsw_sp2_afk_ops = {
 	.encode_block	= mlxsw_sp2_afk_encode_block,
 	.clear_block	= mlxsw_sp2_afk_clear_block,
 };
+
+static struct mlxsw_afk_element_inst mlxsw_sp_afk_element_info_mac_5b[] = {
+	MLXSW_AFK_ELEMENT_INST_U32(VID, 0x04, 18, 12),
+	MLXSW_AFK_ELEMENT_INST_EXT_U32(SRC_SYS_PORT, 0x04, 0, 9, -1, true), /* RX_ACL_SYSTEM_PORT */
+};
+
+static struct mlxsw_afk_element_inst mlxsw_sp_afk_element_info_ipv4_4b[] = {
+	MLXSW_AFK_ELEMENT_INST_U32(VIRT_ROUTER_LSB, 0x04, 13, 8),
+	MLXSW_AFK_ELEMENT_INST_EXT_U32(VIRT_ROUTER_MSB, 0x04, 21, 4, 0, true),
+};
+
+static struct mlxsw_afk_element_inst mlxsw_sp_afk_element_info_ipv6_2b[] = {
+	MLXSW_AFK_ELEMENT_INST_BUF(DST_IP_96_127, 0x04, 4),
+};
+
+static const struct mlxsw_afk_block mlxsw_sp4_afk_blocks[] = {
+	MLXSW_AFK_BLOCK(0x10, mlxsw_sp_afk_element_info_mac_0),
+	MLXSW_AFK_BLOCK(0x11, mlxsw_sp_afk_element_info_mac_1),
+	MLXSW_AFK_BLOCK(0x12, mlxsw_sp_afk_element_info_mac_2),
+	MLXSW_AFK_BLOCK(0x13, mlxsw_sp_afk_element_info_mac_3),
+	MLXSW_AFK_BLOCK(0x14, mlxsw_sp_afk_element_info_mac_4),
+	MLXSW_AFK_BLOCK(0x1A, mlxsw_sp_afk_element_info_mac_5b),
+	MLXSW_AFK_BLOCK(0x38, mlxsw_sp_afk_element_info_ipv4_0),
+	MLXSW_AFK_BLOCK(0x39, mlxsw_sp_afk_element_info_ipv4_1),
+	MLXSW_AFK_BLOCK(0x3A, mlxsw_sp_afk_element_info_ipv4_2),
+	MLXSW_AFK_BLOCK(0x35, mlxsw_sp_afk_element_info_ipv4_4b),
+	MLXSW_AFK_BLOCK(0x40, mlxsw_sp_afk_element_info_ipv6_0),
+	MLXSW_AFK_BLOCK(0x41, mlxsw_sp_afk_element_info_ipv6_1),
+	MLXSW_AFK_BLOCK(0x47, mlxsw_sp_afk_element_info_ipv6_2b),
+	MLXSW_AFK_BLOCK(0x43, mlxsw_sp_afk_element_info_ipv6_3),
+	MLXSW_AFK_BLOCK(0x44, mlxsw_sp_afk_element_info_ipv6_4),
+	MLXSW_AFK_BLOCK(0x45, mlxsw_sp_afk_element_info_ipv6_5),
+	MLXSW_AFK_BLOCK(0x90, mlxsw_sp_afk_element_info_l4_0),
+	MLXSW_AFK_BLOCK(0x92, mlxsw_sp_afk_element_info_l4_2),
+};
+
+const struct mlxsw_afk_ops mlxsw_sp4_afk_ops = {
+	.blocks		= mlxsw_sp4_afk_blocks,
+	.blocks_count	= ARRAY_SIZE(mlxsw_sp4_afk_blocks),
+	.encode_block	= mlxsw_sp2_afk_encode_block,
+	.clear_block	= mlxsw_sp2_afk_clear_block,
+};
-- 
2.33.1

[PATCH net-next 3/8] mlxsw: spectrum_acl_bloom_filter: Reorder functions to make the code more aesthetic

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Currently, mlxsw_sp_acl_bf_rule_count_index_get() is implemented before
mlxsw_sp_acl_bf_index_get() but is used after it.

Adding a new function for Spectrum-4 would make them further apart still.
Fix by moving them around.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../mellanox/mlxsw/spectrum_acl_bloom_filter.c   | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
index dbd3bebf11ec..732c261f83cf 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
@@ -142,14 +142,6 @@ mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
 	*len = chunk_count * MLXSW_BLOOM_KEY_CHUNK_BYTES;
 }
 
-static unsigned int
-mlxsw_sp_acl_bf_rule_count_index_get(struct mlxsw_sp_acl_bf *bf,
-				     unsigned int erp_bank,
-				     unsigned int bf_index)
-{
-	return erp_bank * bf->bank_size + bf_index;
-}
-
 static unsigned int
 mlxsw_sp_acl_bf_index_get(struct mlxsw_sp_acl_bf *bf,
 			  struct mlxsw_sp_acl_atcam_region *aregion,
@@ -162,6 +154,14 @@ mlxsw_sp_acl_bf_index_get(struct mlxsw_sp_acl_bf *bf,
 	return mlxsw_sp_acl_bf_crc(bf_key, bf_size);
 }
 
+static unsigned int
+mlxsw_sp_acl_bf_rule_count_index_get(struct mlxsw_sp_acl_bf *bf,
+				     unsigned int erp_bank,
+				     unsigned int bf_index)
+{
+	return erp_bank * bf->bank_size + bf_index;
+}
+
 int
 mlxsw_sp_acl_bf_entry_add(struct mlxsw_sp *mlxsw_sp,
 			  struct mlxsw_sp_acl_bf *bf,
-- 
2.33.1

[PATCH net-next 4/8] mlxsw: spectrum_acl_bloom_filter: Make mlxsw_sp_acl_bf_key_encode() more flexible

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Spectrum-4 will calculate hash function for bloom filter differently from
the existing ASICs.

One of the changes is related to the way that the chunks will be build -
without padding.

As preparation for support of Spectrum-4 bloom filter, make
mlxsw_sp_acl_bf_key_encode() more flexible, so it will be able to use it
for Spectrum-4 as well.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../mlxsw/spectrum_acl_bloom_filter.c         | 36 +++++++++++++------
 1 file changed, 25 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
index 732c261f83cf..f0cd14a791df 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
@@ -116,9 +116,10 @@ static u16 mlxsw_sp_acl_bf_crc(const u8 *buffer, size_t len)
 }
 
 static void
-mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
-			   struct mlxsw_sp_acl_atcam_entry *aentry,
-			   char *output, u8 *len)
+__mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
+			     struct mlxsw_sp_acl_atcam_entry *aentry,
+			     char *output, u8 *len, u8 max_chunks, u8 pad_bytes,
+			     u8 key_offset, u8 chunk_key_len, u8 chunk_len)
 {
 	struct mlxsw_afk_key_info *key_info = aregion->region->key_info;
 	u8 chunk_index, chunk_count, block_count;
@@ -129,17 +130,30 @@ mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
 	chunk_count = 1 + ((block_count - 1) >> 2);
 	erp_region_id = cpu_to_be16(aentry->ht_key.erp_id |
 				   (aregion->region->id << 4));
-	for (chunk_index = MLXSW_BLOOM_KEY_CHUNKS - chunk_count;
-	     chunk_index < MLXSW_BLOOM_KEY_CHUNKS; chunk_index++) {
-		memset(chunk, 0, MLXSW_BLOOM_CHUNK_PAD_BYTES);
-		memcpy(chunk + MLXSW_BLOOM_CHUNK_PAD_BYTES, &erp_region_id,
+	for (chunk_index = max_chunks - chunk_count; chunk_index < max_chunks;
+	     chunk_index++) {
+		memset(chunk, 0, pad_bytes);
+		memcpy(chunk + pad_bytes, &erp_region_id,
 		       sizeof(erp_region_id));
-		memcpy(chunk + MLXSW_BLOOM_CHUNK_KEY_OFFSET,
+		memcpy(chunk + key_offset,
 		       &aentry->enc_key[chunk_key_offsets[chunk_index]],
-		       MLXSW_BLOOM_CHUNK_KEY_BYTES);
-		chunk += MLXSW_BLOOM_KEY_CHUNK_BYTES;
+		       chunk_key_len);
+		chunk += chunk_len;
 	}
-	*len = chunk_count * MLXSW_BLOOM_KEY_CHUNK_BYTES;
+	*len = chunk_count * chunk_len;
+}
+
+static void
+mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
+			   struct mlxsw_sp_acl_atcam_entry *aentry,
+			   char *output, u8 *len)
+{
+	__mlxsw_sp_acl_bf_key_encode(aregion, aentry, output, len,
+				     MLXSW_BLOOM_KEY_CHUNKS,
+				     MLXSW_BLOOM_CHUNK_PAD_BYTES,
+				     MLXSW_BLOOM_CHUNK_KEY_OFFSET,
+				     MLXSW_BLOOM_CHUNK_KEY_BYTES,
+				     MLXSW_BLOOM_KEY_CHUNK_BYTES);
 }
 
 static unsigned int
-- 
2.33.1

[PATCH net-next 5/8] mlxsw: spectrum_acl_bloom_filter: Rename Spectrum-2 specific objects for future use

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Spectrum-4 will calculate hash function for bloom filter differently from
the existing ASICs.

There are two changes:
1. Instead of using one hash function to calculate 16 bits output (CRC-16),
   two functions will be used.
2. The chunks will be built differently, without padding.

As preparation for support of Spectrum-4 bloom filter, rename CRC table
to include "sp2" prefix and "crc16", as next patch will add two additional
tables. In addition, rename all the dedicated functions and defines for
Spectrum-{2,3} to include "sp2" prefix.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../mlxsw/spectrum_acl_bloom_filter.c         | 50 +++++++++----------
 1 file changed, 25 insertions(+), 25 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
index f0cd14a791df..3a3c7683b725 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
@@ -27,7 +27,7 @@ struct mlxsw_sp_acl_bf {
  * +-------------------------+------------------------+------------------------+
  */
 #define MLXSW_BLOOM_KEY_CHUNKS 3
-#define MLXSW_BLOOM_KEY_LEN 69
+#define MLXSW_SP2_BLOOM_KEY_LEN 69
 
 /* Each chunk size is 23 bytes. 18 bytes of it contain 4 key blocks, each is
  * 36 bits, 2 bytes which hold eRP ID and region ID, and 3 bytes of zero
@@ -42,14 +42,14 @@ struct mlxsw_sp_acl_bf {
  * |    0    | region ID |  eRP ID  |      4 Key blocks (18 Bytes)      |
  * +---------+-----------+----------+-----------------------------------+
  */
-#define MLXSW_BLOOM_CHUNK_PAD_BYTES 3
-#define MLXSW_BLOOM_CHUNK_KEY_BYTES 18
-#define MLXSW_BLOOM_KEY_CHUNK_BYTES 23
+#define MLXSW_SP2_BLOOM_CHUNK_PAD_BYTES 3
+#define MLXSW_SP2_BLOOM_CHUNK_KEY_BYTES 18
+#define MLXSW_SP2_BLOOM_KEY_CHUNK_BYTES 23
 
 /* The offset of the key block within a chunk is 5 bytes as it comes after
  * 3 bytes of zero padding and 16 bits of region ID and eRP ID.
  */
-#define MLXSW_BLOOM_CHUNK_KEY_OFFSET 5
+#define MLXSW_SP2_BLOOM_CHUNK_KEY_OFFSET 5
 
 /* Each chunk contains 4 key blocks. Chunk 2 uses key blocks 11-8,
  * and we need to populate it with 4 key blocks copied from the entry encoded
@@ -66,7 +66,7 @@ static const u8 chunk_key_offsets[MLXSW_BLOOM_KEY_CHUNKS] = {2, 20, 38};
  * which is 0x8529 (1 + x^3 + x^5 + x^8 + x^10 + x^15 and
  * the implicit x^16).
  */
-static const u16 mlxsw_sp_acl_bf_crc_tab[256] = {
+static const u16 mlxsw_sp2_acl_bf_crc16_tab[256] = {
 0x0000, 0x8529, 0x8f7b, 0x0a52, 0x9bdf, 0x1ef6, 0x14a4, 0x918d,
 0xb297, 0x37be, 0x3dec, 0xb8c5, 0x2948, 0xac61, 0xa633, 0x231a,
 0xe007, 0x652e, 0x6f7c, 0xea55, 0x7bd8, 0xfef1, 0xf4a3, 0x718a,
@@ -101,17 +101,17 @@ static const u16 mlxsw_sp_acl_bf_crc_tab[256] = {
 0x0c4c, 0x8965, 0x8337, 0x061e, 0x9793, 0x12ba, 0x18e8, 0x9dc1,
 };
 
-static u16 mlxsw_sp_acl_bf_crc_byte(u16 crc, u8 c)
+static u16 mlxsw_sp2_acl_bf_crc16_byte(u16 crc, u8 c)
 {
-	return (crc << 8) ^ mlxsw_sp_acl_bf_crc_tab[(crc >> 8) ^ c];
+	return (crc << 8) ^ mlxsw_sp2_acl_bf_crc16_tab[(crc >> 8) ^ c];
 }
 
-static u16 mlxsw_sp_acl_bf_crc(const u8 *buffer, size_t len)
+static u16 mlxsw_sp2_acl_bf_crc(const u8 *buffer, size_t len)
 {
 	u16 crc = 0;
 
 	while (len--)
-		crc = mlxsw_sp_acl_bf_crc_byte(crc, *buffer++);
+		crc = mlxsw_sp2_acl_bf_crc16_byte(crc, *buffer++);
 	return crc;
 }
 
@@ -144,28 +144,28 @@ __mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
 }
 
 static void
-mlxsw_sp_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
-			   struct mlxsw_sp_acl_atcam_entry *aentry,
-			   char *output, u8 *len)
+mlxsw_sp2_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
+			    struct mlxsw_sp_acl_atcam_entry *aentry,
+			    char *output, u8 *len)
 {
 	__mlxsw_sp_acl_bf_key_encode(aregion, aentry, output, len,
 				     MLXSW_BLOOM_KEY_CHUNKS,
-				     MLXSW_BLOOM_CHUNK_PAD_BYTES,
-				     MLXSW_BLOOM_CHUNK_KEY_OFFSET,
-				     MLXSW_BLOOM_CHUNK_KEY_BYTES,
-				     MLXSW_BLOOM_KEY_CHUNK_BYTES);
+				     MLXSW_SP2_BLOOM_CHUNK_PAD_BYTES,
+				     MLXSW_SP2_BLOOM_CHUNK_KEY_OFFSET,
+				     MLXSW_SP2_BLOOM_CHUNK_KEY_BYTES,
+				     MLXSW_SP2_BLOOM_KEY_CHUNK_BYTES);
 }
 
 static unsigned int
-mlxsw_sp_acl_bf_index_get(struct mlxsw_sp_acl_bf *bf,
-			  struct mlxsw_sp_acl_atcam_region *aregion,
-			  struct mlxsw_sp_acl_atcam_entry *aentry)
+mlxsw_sp2_acl_bf_index_get(struct mlxsw_sp_acl_bf *bf,
+			   struct mlxsw_sp_acl_atcam_region *aregion,
+			   struct mlxsw_sp_acl_atcam_entry *aentry)
 {
-	char bf_key[MLXSW_BLOOM_KEY_LEN];
+	char bf_key[MLXSW_SP2_BLOOM_KEY_LEN];
 	u8 bf_size;
 
-	mlxsw_sp_acl_bf_key_encode(aregion, aentry, bf_key, &bf_size);
-	return mlxsw_sp_acl_bf_crc(bf_key, bf_size);
+	mlxsw_sp2_acl_bf_key_encode(aregion, aentry, bf_key, &bf_size);
+	return mlxsw_sp2_acl_bf_crc(bf_key, bf_size);
 }
 
 static unsigned int
@@ -190,7 +190,7 @@ mlxsw_sp_acl_bf_entry_add(struct mlxsw_sp *mlxsw_sp,
 
 	mutex_lock(&bf->lock);
 
-	bf_index = mlxsw_sp_acl_bf_index_get(bf, aregion, aentry);
+	bf_index = mlxsw_sp2_acl_bf_index_get(bf, aregion, aentry);
 	rule_index = mlxsw_sp_acl_bf_rule_count_index_get(bf, erp_bank,
 							  bf_index);
 
@@ -233,7 +233,7 @@ mlxsw_sp_acl_bf_entry_del(struct mlxsw_sp *mlxsw_sp,
 
 	mutex_lock(&bf->lock);
 
-	bf_index = mlxsw_sp_acl_bf_index_get(bf, aregion, aentry);
+	bf_index = mlxsw_sp2_acl_bf_index_get(bf, aregion, aentry);
 	rule_index = mlxsw_sp_acl_bf_rule_count_index_get(bf, erp_bank,
 							  bf_index);
 
-- 
2.33.1

[PATCH net-next 6/8] mlxsw: Add operations structure for bloom filter calculation

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Spectrum-4 will calculate hash function for bloom filter differently from
the existing ASICs.

There are two changes:
1. Instead of using one hash function to calculate 16 bits output (CRC-16),
   two functions will be used.
2. The chunks will be built differently, without padding.

As preparation for support of Spectrum-4 bloom filter, add 'ops'
structure to allow handling different calculation for different ASICs.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum.c            | 2 ++
 drivers/net/ethernet/mellanox/mlxsw/spectrum.h            | 4 ++++
 .../ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c   | 8 ++++++--
 drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h   | 6 ++++++
 4 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 5251f33af0fb..6e4265c86eb8 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -3155,6 +3155,7 @@ static int mlxsw_sp2_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->mr_tcam_ops = &mlxsw_sp2_mr_tcam_ops;
 	mlxsw_sp->acl_rulei_ops = &mlxsw_sp2_acl_rulei_ops;
 	mlxsw_sp->acl_tcam_ops = &mlxsw_sp2_acl_tcam_ops;
+	mlxsw_sp->acl_bf_ops = &mlxsw_sp2_acl_bf_ops;
 	mlxsw_sp->nve_ops_arr = mlxsw_sp2_nve_ops_arr;
 	mlxsw_sp->mac_mask = mlxsw_sp2_mac_mask;
 	mlxsw_sp->sb_vals = &mlxsw_sp2_sb_vals;
@@ -3184,6 +3185,7 @@ static int mlxsw_sp3_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->mr_tcam_ops = &mlxsw_sp2_mr_tcam_ops;
 	mlxsw_sp->acl_rulei_ops = &mlxsw_sp2_acl_rulei_ops;
 	mlxsw_sp->acl_tcam_ops = &mlxsw_sp2_acl_tcam_ops;
+	mlxsw_sp->acl_bf_ops = &mlxsw_sp2_acl_bf_ops;
 	mlxsw_sp->nve_ops_arr = mlxsw_sp2_nve_ops_arr;
 	mlxsw_sp->mac_mask = mlxsw_sp2_mac_mask;
 	mlxsw_sp->sb_vals = &mlxsw_sp2_sb_vals;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index a49316d0bd37..e7da6c83c442 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -190,6 +190,7 @@ struct mlxsw_sp {
 	const struct mlxsw_sp_mr_tcam_ops *mr_tcam_ops;
 	const struct mlxsw_sp_acl_rulei_ops *acl_rulei_ops;
 	const struct mlxsw_sp_acl_tcam_ops *acl_tcam_ops;
+	const struct mlxsw_sp_acl_bf_ops *acl_bf_ops;
 	const struct mlxsw_sp_nve_ops **nve_ops_arr;
 	const struct mlxsw_sp_sb_vals *sb_vals;
 	const struct mlxsw_sp_sb_ops *sb_ops;
@@ -1108,6 +1109,9 @@ extern const struct mlxsw_afk_ops mlxsw_sp1_afk_ops;
 extern const struct mlxsw_afk_ops mlxsw_sp2_afk_ops;
 extern const struct mlxsw_afk_ops mlxsw_sp4_afk_ops;
 
+/* spectrum_acl_bloom_filter.c */
+extern const struct mlxsw_sp_acl_bf_ops mlxsw_sp2_acl_bf_ops;
+
 /* spectrum_matchall.c */
 struct mlxsw_sp_mall_ops {
 	int (*sample_add)(struct mlxsw_sp *mlxsw_sp,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
index 3a3c7683b725..c6dab9615a0a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
@@ -190,7 +190,7 @@ mlxsw_sp_acl_bf_entry_add(struct mlxsw_sp *mlxsw_sp,
 
 	mutex_lock(&bf->lock);
 
-	bf_index = mlxsw_sp2_acl_bf_index_get(bf, aregion, aentry);
+	bf_index = mlxsw_sp->acl_bf_ops->index_get(bf, aregion, aentry);
 	rule_index = mlxsw_sp_acl_bf_rule_count_index_get(bf, erp_bank,
 							  bf_index);
 
@@ -233,7 +233,7 @@ mlxsw_sp_acl_bf_entry_del(struct mlxsw_sp *mlxsw_sp,
 
 	mutex_lock(&bf->lock);
 
-	bf_index = mlxsw_sp2_acl_bf_index_get(bf, aregion, aentry);
+	bf_index = mlxsw_sp->acl_bf_ops->index_get(bf, aregion, aentry);
 	rule_index = mlxsw_sp_acl_bf_rule_count_index_get(bf, erp_bank,
 							  bf_index);
 
@@ -281,3 +281,7 @@ void mlxsw_sp_acl_bf_fini(struct mlxsw_sp_acl_bf *bf)
 	mutex_destroy(&bf->lock);
 	kfree(bf);
 }
+
+const struct mlxsw_sp_acl_bf_ops mlxsw_sp2_acl_bf_ops = {
+	.index_get = mlxsw_sp2_acl_bf_index_get,
+};
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
index a41df10ade9b..edbbc89e7a71 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.h
@@ -287,6 +287,12 @@ void mlxsw_sp_acl_erps_fini(struct mlxsw_sp *mlxsw_sp,
 
 struct mlxsw_sp_acl_bf;
 
+struct mlxsw_sp_acl_bf_ops {
+	unsigned int (*index_get)(struct mlxsw_sp_acl_bf *bf,
+				  struct mlxsw_sp_acl_atcam_region *aregion,
+				  struct mlxsw_sp_acl_atcam_entry *aentry);
+};
+
 int
 mlxsw_sp_acl_bf_entry_add(struct mlxsw_sp *mlxsw_sp,
 			  struct mlxsw_sp_acl_bf *bf,
-- 
2.33.1

[PATCH net-next 7/8] mlxsw: spectrum_acl_bloom_filter: Add support for Spectrum-4 calculation

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Spectrum-4 will calculate hash function for bloom filter differently
from the existing ASICs.

First, two hash functions will be used to calculate 16 bits result.
The final result will be combination of the two results - 6 bits which
are result of CRC-6 will be used as MSB and 10 bits which are result of
CRC-10 will be used as LSB.

Second, while in Spectrum{2,3}, there is a padding in each chunk, so the
chunks use a sequence of whole bytes, in Spectrum-4 there is no padding,
so each chunk use 20 bytes minus 2 bits, so it is necessary to align the
chunks to be without holes.

Add dedicated 'mlxsw_sp_acl_bf_ops' for Spectrum-4 and add the required
tables for CRC calculations.

All the details are documented as part of the code for future use.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |   1 +
 .../mlxsw/spectrum_acl_bloom_filter.c         | 267 ++++++++++++++++--
 2 files changed, 252 insertions(+), 16 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index e7da6c83c442..bb2442e1f705 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -1111,6 +1111,7 @@ extern const struct mlxsw_afk_ops mlxsw_sp4_afk_ops;
 
 /* spectrum_acl_bloom_filter.c */
 extern const struct mlxsw_sp_acl_bf_ops mlxsw_sp2_acl_bf_ops;
+extern const struct mlxsw_sp_acl_bf_ops mlxsw_sp4_acl_bf_ops;
 
 /* spectrum_matchall.c */
 struct mlxsw_sp_mall_ops {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
index c6dab9615a0a..e2aced7ab454 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_bloom_filter.c
@@ -17,9 +17,9 @@ struct mlxsw_sp_acl_bf {
 };
 
 /* Bloom filter uses a crc-16 hash over chunks of data which contain 4 key
- * blocks, eRP ID and region ID. In Spectrum-2, region key is combined of up to
- * 12 key blocks, so there can be up to 3 chunks in the Bloom filter key,
- * depending on the actual number of key blocks used in the region.
+ * blocks, eRP ID and region ID. In Spectrum-2 and above, region key is combined
+ * of up to 12 key blocks, so there can be up to 3 chunks in the Bloom filter
+ * key, depending on the actual number of key blocks used in the region.
  * The layout of the Bloom filter key is as follows:
  *
  * +-------------------------+------------------------+------------------------+
@@ -27,6 +27,8 @@ struct mlxsw_sp_acl_bf {
  * +-------------------------+------------------------+------------------------+
  */
 #define MLXSW_BLOOM_KEY_CHUNKS 3
+
+/* Spectrum-2 and Spectrum-3 chunks */
 #define MLXSW_SP2_BLOOM_KEY_LEN 69
 
 /* Each chunk size is 23 bytes. 18 bytes of it contain 4 key blocks, each is
@@ -51,19 +53,9 @@ struct mlxsw_sp_acl_bf {
  */
 #define MLXSW_SP2_BLOOM_CHUNK_KEY_OFFSET 5
 
-/* Each chunk contains 4 key blocks. Chunk 2 uses key blocks 11-8,
- * and we need to populate it with 4 key blocks copied from the entry encoded
- * key. Since the encoded key contains a padding, key block 11 starts at offset
- * 2. block 7 that is used in chunk 1 starts at offset 20 as 4 key blocks take
- * 18 bytes.
- * This array defines key offsets for easy access when copying key blocks from
- * entry key to Bloom filter chunk.
- */
-static const u8 chunk_key_offsets[MLXSW_BLOOM_KEY_CHUNKS] = {2, 20, 38};
-
-/* This table is just the CRC of each possible byte. It is
- * computed, Msbit first, for the Bloom filter polynomial
- * which is 0x8529 (1 + x^3 + x^5 + x^8 + x^10 + x^15 and
+/* This table is just the CRC of each possible byte which is used for
+ * Spectrum-{2-3}. It is computed, Msbit first, for the Bloom filter
+ * polynomial which is 0x8529 (1 + x^3 + x^5 + x^8 + x^10 + x^15 and
  * the implicit x^16).
  */
 static const u16 mlxsw_sp2_acl_bf_crc16_tab[256] = {
@@ -101,6 +93,127 @@ static const u16 mlxsw_sp2_acl_bf_crc16_tab[256] = {
 0x0c4c, 0x8965, 0x8337, 0x061e, 0x9793, 0x12ba, 0x18e8, 0x9dc1,
 };
 
+/* Spectrum-4 chunks */
+#define MLXSW_SP4_BLOOM_KEY_LEN 60
+
+/* In Spectrum-4, there is no padding. Each chunk size is 20 bytes.
+ * 18 bytes of it contain 4 key blocks, each is 36 bits, and 2 bytes which hold
+ * eRP ID and region ID.
+ * The layout of each chunk is as follows:
+ *
+ * +----------------------+-----------------------------------+
+ * |        2 bytes       |              18 bytes             |
+ * +-----------+----------+-----------------------------------+
+ * |  157:148  | 147:144  |               143:0               |
+ * +---------+-----------+----------+-------------------------+
+ * | region ID |  eRP ID  |      4 Key blocks (18 Bytes)      |
+ * +-----------+----------+-----------------------------------+
+ */
+
+#define MLXSW_SP4_BLOOM_CHUNK_PAD_BYTES 0
+#define MLXSW_SP4_BLOOM_CHUNK_KEY_BYTES 18
+#define MLXSW_SP4_BLOOM_KEY_CHUNK_BYTES 20
+
+/* The offset of the key block within a chunk is 2 bytes as it comes after
+ * 16 bits of region ID and eRP ID.
+ */
+#define MLXSW_SP4_BLOOM_CHUNK_KEY_OFFSET 2
+
+/* For Spectrum-4, two hash functions are used, CRC-10 and CRC-6 based.
+ * The result is combination of the two calculations -
+ * 6 bit column are MSB (result of CRC-6),
+ * 10 bit row are LSB (result of CRC-10).
+ */
+
+/* This table is just the CRC of each possible byte which is used for
+ * Spectrum-4. It is computed, Msbit first, for the Bloom filter
+ * polynomial which is 0x1b (1 + x^1 + x^3 + x^4 and the implicit x^10).
+ */
+static const u16 mlxsw_sp4_acl_bf_crc10_tab[256] = {
+0x0000, 0x001b, 0x0036, 0x002d, 0x006c, 0x0077, 0x005a, 0x0041,
+0x00d8, 0x00c3, 0x00ee, 0x00f5, 0x00b4, 0x00af, 0x0082, 0x0099,
+0x01b0, 0x01ab, 0x0186, 0x019d, 0x01dc, 0x01c7, 0x01ea, 0x01f1,
+0x0168, 0x0173, 0x015e, 0x0145, 0x0104, 0x011f, 0x0132, 0x0129,
+0x0360, 0x037b, 0x0356, 0x034d, 0x030c, 0x0317, 0x033a, 0x0321,
+0x03b8, 0x03a3, 0x038e, 0x0395, 0x03d4, 0x03cf, 0x03e2, 0x03f9,
+0x02d0, 0x02cb, 0x02e6, 0x02fd, 0x02bc, 0x02a7, 0x028a, 0x0291,
+0x0208, 0x0213, 0x023e, 0x0225, 0x0264, 0x027f, 0x0252, 0x0249,
+0x02db, 0x02c0, 0x02ed, 0x02f6, 0x02b7, 0x02ac, 0x0281, 0x029a,
+0x0203, 0x0218, 0x0235, 0x022e, 0x026f, 0x0274, 0x0259, 0x0242,
+0x036b, 0x0370, 0x035d, 0x0346, 0x0307, 0x031c, 0x0331, 0x032a,
+0x03b3, 0x03a8, 0x0385, 0x039e, 0x03df, 0x03c4, 0x03e9, 0x03f2,
+0x01bb, 0x01a0, 0x018d, 0x0196, 0x01d7, 0x01cc, 0x01e1, 0x01fa,
+0x0163, 0x0178, 0x0155, 0x014e, 0x010f, 0x0114, 0x0139, 0x0122,
+0x000b, 0x0010, 0x003d, 0x0026, 0x0067, 0x007c, 0x0051, 0x004a,
+0x00d3, 0x00c8, 0x00e5, 0x00fe, 0x00bf, 0x00a4, 0x0089, 0x0092,
+0x01ad, 0x01b6, 0x019b, 0x0180, 0x01c1, 0x01da, 0x01f7, 0x01ec,
+0x0175, 0x016e, 0x0143, 0x0158, 0x0119, 0x0102, 0x012f, 0x0134,
+0x001d, 0x0006, 0x002b, 0x0030, 0x0071, 0x006a, 0x0047, 0x005c,
+0x00c5, 0x00de, 0x00f3, 0x00e8, 0x00a9, 0x00b2, 0x009f, 0x0084,
+0x02cd, 0x02d6, 0x02fb, 0x02e0, 0x02a1, 0x02ba, 0x0297, 0x028c,
+0x0215, 0x020e, 0x0223, 0x0238, 0x0279, 0x0262, 0x024f, 0x0254,
+0x037d, 0x0366, 0x034b, 0x0350, 0x0311, 0x030a, 0x0327, 0x033c,
+0x03a5, 0x03be, 0x0393, 0x0388, 0x03c9, 0x03d2, 0x03ff, 0x03e4,
+0x0376, 0x036d, 0x0340, 0x035b, 0x031a, 0x0301, 0x032c, 0x0337,
+0x03ae, 0x03b5, 0x0398, 0x0383, 0x03c2, 0x03d9, 0x03f4, 0x03ef,
+0x02c6, 0x02dd, 0x02f0, 0x02eb, 0x02aa, 0x02b1, 0x029c, 0x0287,
+0x021e, 0x0205, 0x0228, 0x0233, 0x0272, 0x0269, 0x0244, 0x025f,
+0x0016, 0x000d, 0x0020, 0x003b, 0x007a, 0x0061, 0x004c, 0x0057,
+0x00ce, 0x00d5, 0x00f8, 0x00e3, 0x00a2, 0x00b9, 0x0094, 0x008f,
+0x01a6, 0x01bd, 0x0190, 0x018b, 0x01ca, 0x01d1, 0x01fc, 0x01e7,
+0x017e, 0x0165, 0x0148, 0x0153, 0x0112, 0x0109, 0x0124, 0x013f,
+};
+
+/* This table is just the CRC of each possible byte which is used for
+ * Spectrum-4. It is computed, Msbit first, for the Bloom filter
+ * polynomial which is 0x2d (1 + x^2+ x^3 + x^5 and the implicit x^6).
+ */
+static const u8 mlxsw_sp4_acl_bf_crc6_tab[256] = {
+0x00, 0x2d, 0x37, 0x1a, 0x03, 0x2e, 0x34, 0x19,
+0x06, 0x2b, 0x31, 0x1c, 0x05, 0x28, 0x32, 0x1f,
+0x0c, 0x21, 0x3b, 0x16, 0x0f, 0x22, 0x38, 0x15,
+0x0a, 0x27, 0x3d, 0x10, 0x09, 0x24, 0x3e, 0x13,
+0x18, 0x35, 0x2f, 0x02, 0x1b, 0x36, 0x2c, 0x01,
+0x1e, 0x33, 0x29, 0x04, 0x1d, 0x30, 0x2a, 0x07,
+0x14, 0x39, 0x23, 0x0e, 0x17, 0x3a, 0x20, 0x0d,
+0x12, 0x3f, 0x25, 0x08, 0x11, 0x3c, 0x26, 0x0b,
+0x30, 0x1d, 0x07, 0x2a, 0x33, 0x1e, 0x04, 0x29,
+0x36, 0x1b, 0x01, 0x2c, 0x35, 0x18, 0x02, 0x2f,
+0x3c, 0x11, 0x0b, 0x26, 0x3f, 0x12, 0x08, 0x25,
+0x3a, 0x17, 0x0d, 0x20, 0x39, 0x14, 0x0e, 0x23,
+0x28, 0x05, 0x1f, 0x32, 0x2b, 0x06, 0x1c, 0x31,
+0x2e, 0x03, 0x19, 0x34, 0x2d, 0x00, 0x1a, 0x37,
+0x24, 0x09, 0x13, 0x3e, 0x27, 0x0a, 0x10, 0x3d,
+0x22, 0x0f, 0x15, 0x38, 0x21, 0x0c, 0x16, 0x3b,
+0x0d, 0x20, 0x3a, 0x17, 0x0e, 0x23, 0x39, 0x14,
+0x0b, 0x26, 0x3c, 0x11, 0x08, 0x25, 0x3f, 0x12,
+0x01, 0x2c, 0x36, 0x1b, 0x02, 0x2f, 0x35, 0x18,
+0x07, 0x2a, 0x30, 0x1d, 0x04, 0x29, 0x33, 0x1e,
+0x15, 0x38, 0x22, 0x0f, 0x16, 0x3b, 0x21, 0x0c,
+0x13, 0x3e, 0x24, 0x09, 0x10, 0x3d, 0x27, 0x0a,
+0x19, 0x34, 0x2e, 0x03, 0x1a, 0x37, 0x2d, 0x00,
+0x1f, 0x32, 0x28, 0x05, 0x1c, 0x31, 0x2b, 0x06,
+0x3d, 0x10, 0x0a, 0x27, 0x3e, 0x13, 0x09, 0x24,
+0x3b, 0x16, 0x0c, 0x21, 0x38, 0x15, 0x0f, 0x22,
+0x31, 0x1c, 0x06, 0x2b, 0x32, 0x1f, 0x05, 0x28,
+0x37, 0x1a, 0x00, 0x2d, 0x34, 0x19, 0x03, 0x2e,
+0x25, 0x08, 0x12, 0x3f, 0x26, 0x0b, 0x11, 0x3c,
+0x23, 0x0e, 0x14, 0x39, 0x20, 0x0d, 0x17, 0x3a,
+0x29, 0x04, 0x1e, 0x33, 0x2a, 0x07, 0x1d, 0x30,
+0x2f, 0x02, 0x18, 0x35, 0x2c, 0x01, 0x1b, 0x36,
+};
+
+/* Each chunk contains 4 key blocks. Chunk 2 uses key blocks 11-8,
+ * and we need to populate it with 4 key blocks copied from the entry encoded
+ * key. The original keys layout is same for Spectrum-{2,3,4}.
+ * Since the encoded key contains a 2 bytes padding, key block 11 starts at
+ * offset 2. block 7 that is used in chunk 1 starts at offset 20 as 4 key blocks
+ * take 18 bytes. See 'MLXSW_SP2_AFK_BLOCK_LAYOUT' for more details.
+ * This array defines key offsets for easy access when copying key blocks from
+ * entry key to Bloom filter chunk.
+ */
+static const u8 chunk_key_offsets[MLXSW_BLOOM_KEY_CHUNKS] = {2, 20, 38};
+
 static u16 mlxsw_sp2_acl_bf_crc16_byte(u16 crc, u8 c)
 {
 	return (crc << 8) ^ mlxsw_sp2_acl_bf_crc16_tab[(crc >> 8) ^ c];
@@ -168,6 +281,124 @@ mlxsw_sp2_acl_bf_index_get(struct mlxsw_sp_acl_bf *bf,
 	return mlxsw_sp2_acl_bf_crc(bf_key, bf_size);
 }
 
+static u16 mlxsw_sp4_acl_bf_crc10_byte(u16 crc, u8 c)
+{
+	u8 index = ((crc >> 2) ^ c) & 0xff;
+
+	return ((crc << 8) ^ mlxsw_sp4_acl_bf_crc10_tab[index]) & 0x3ff;
+}
+
+static u16 mlxsw_sp4_acl_bf_crc6_byte(u16 crc, u8 c)
+{
+	u8 index = (crc ^ c) & 0xff;
+
+	return ((crc << 6) ^ (mlxsw_sp4_acl_bf_crc6_tab[index] << 2)) & 0xfc;
+}
+
+static u16 mlxsw_sp4_acl_bf_crc(const u8 *buffer, size_t len)
+{
+	u16 crc_row = 0, crc_col = 0;
+
+	while (len--) {
+		crc_row = mlxsw_sp4_acl_bf_crc10_byte(crc_row, *buffer);
+		crc_col = mlxsw_sp4_acl_bf_crc6_byte(crc_col, *buffer);
+		buffer++;
+	}
+
+	crc_col >>= 2;
+
+	/* 6 bit column are MSB, 10 bit row are LSB */
+	return (crc_col << 10) | crc_row;
+}
+
+static void right_shift_array(char *arr, u8 len, u8 shift_bits)
+{
+	u8 byte_mask = 0xff >> shift_bits;
+	int i;
+
+	if (WARN_ON(!shift_bits || shift_bits >= 8))
+		return;
+
+	for (i = len - 1; i >= 0; i--) {
+		/* The first iteration looks like out-of-bounds access,
+		 * but actually references a buffer that the array is shifted
+		 * into. This move is legal as we never send the last chunk to
+		 * this function.
+		 */
+		arr[i + 1] &= byte_mask;
+		arr[i + 1] |= arr[i] << (8 - shift_bits);
+		arr[i] = arr[i] >> shift_bits;
+	}
+}
+
+static void mlxsw_sp4_bf_key_shift_chunks(u8 chunk_count, char *output)
+{
+	/* The chunks are suppoosed to be continuous, with no padding.
+	 * Since region ID and eRP ID use 14 bits, and not fully 2 bytes,
+	 * and in Spectrum-4 there is no padding, it is necessary to shift some
+	 * chunks 2 bits right.
+	 */
+	switch (chunk_count) {
+	case 2:
+		/* The chunks are copied as follow:
+		 * +-------------+-----------------+
+		 * | Chunk 0     |   Chunk 1       |
+		 * | IDs  | keys |(**) IDs  | keys |
+		 * +-------------+-----------------+
+		 * In (**), there are two unused bits, therefore, chunk 0 needs
+		 * to be shifted two bits right.
+		 */
+		right_shift_array(output, MLXSW_SP4_BLOOM_KEY_CHUNK_BYTES, 2);
+		break;
+	case 3:
+		/* The chunks are copied as follow:
+		 * +-------------+-----------------+-----------------+
+		 * | Chunk 0     |   Chunk 1       |   Chunk 2       |
+		 * | IDs  | keys |(**) IDs  | keys |(**) IDs  | keys |
+		 * +-------------+-----------------+-----------------+
+		 * In (**), there are two unused bits, therefore, chunk 1 needs
+		 * to be shifted two bits right and chunk 0 needs to be shifted
+		 * four bits right.
+		 */
+		right_shift_array(output + MLXSW_SP4_BLOOM_KEY_CHUNK_BYTES,
+				  MLXSW_SP4_BLOOM_KEY_CHUNK_BYTES, 2);
+		right_shift_array(output, MLXSW_SP4_BLOOM_KEY_CHUNK_BYTES, 4);
+		break;
+	default:
+		WARN_ON(chunk_count > MLXSW_BLOOM_KEY_CHUNKS);
+	}
+}
+
+static void
+mlxsw_sp4_acl_bf_key_encode(struct mlxsw_sp_acl_atcam_region *aregion,
+			    struct mlxsw_sp_acl_atcam_entry *aentry,
+			    char *output, u8 *len)
+{
+	struct mlxsw_afk_key_info *key_info = aregion->region->key_info;
+	u8 block_count = mlxsw_afk_key_info_blocks_count_get(key_info);
+	u8 chunk_count = 1 + ((block_count - 1) >> 2);
+
+	__mlxsw_sp_acl_bf_key_encode(aregion, aentry, output, len,
+				     MLXSW_BLOOM_KEY_CHUNKS,
+				     MLXSW_SP4_BLOOM_CHUNK_PAD_BYTES,
+				     MLXSW_SP4_BLOOM_CHUNK_KEY_OFFSET,
+				     MLXSW_SP4_BLOOM_CHUNK_KEY_BYTES,
+				     MLXSW_SP4_BLOOM_KEY_CHUNK_BYTES);
+	mlxsw_sp4_bf_key_shift_chunks(chunk_count, output);
+}
+
+static unsigned int
+mlxsw_sp4_acl_bf_index_get(struct mlxsw_sp_acl_bf *bf,
+			   struct mlxsw_sp_acl_atcam_region *aregion,
+			   struct mlxsw_sp_acl_atcam_entry *aentry)
+{
+	char bf_key[MLXSW_SP4_BLOOM_KEY_LEN] = {};
+	u8 bf_size;
+
+	mlxsw_sp4_acl_bf_key_encode(aregion, aentry, bf_key, &bf_size);
+	return mlxsw_sp4_acl_bf_crc(bf_key, bf_size);
+}
+
 static unsigned int
 mlxsw_sp_acl_bf_rule_count_index_get(struct mlxsw_sp_acl_bf *bf,
 				     unsigned int erp_bank,
@@ -285,3 +516,7 @@ void mlxsw_sp_acl_bf_fini(struct mlxsw_sp_acl_bf *bf)
 const struct mlxsw_sp_acl_bf_ops mlxsw_sp2_acl_bf_ops = {
 	.index_get = mlxsw_sp2_acl_bf_index_get,
 };
+
+const struct mlxsw_sp_acl_bf_ops mlxsw_sp4_acl_bf_ops = {
+	.index_get = mlxsw_sp4_acl_bf_index_get,
+};
-- 
2.33.1

[PATCH net-next 8/8] mlxsw: spectrum: Extend to support Spectrum-4 ASIC

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Extend existing driver for Spectrum, Spectrum-2 and Spectrum-3 ASICs
to support Spectrum-4 ASIC as well.

Currently there is no released firmware version for Spectrum-4, so the
driver is not enforcing a minimum version.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/Kconfig   |  2 +-
 drivers/net/ethernet/mellanox/mlxsw/pci.h     |  1 +
 drivers/net/ethernet/mellanox/mlxsw/reg.h     |  1 +
 .../net/ethernet/mellanox/mlxsw/spectrum.c    | 95 +++++++++++++++++++
 4 files changed, 98 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/Kconfig b/drivers/net/ethernet/mellanox/mlxsw/Kconfig
index d1ae248e125c..4683312861ac 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/Kconfig
+++ b/drivers/net/ethernet/mellanox/mlxsw/Kconfig
@@ -66,7 +66,7 @@ config MLXSW_SPECTRUM
 	default m
 	help
 	  This driver supports Mellanox Technologies
-	  Spectrum/Spectrum-2/Spectrum-3 Ethernet Switch ASICs.
+	  Spectrum/Spectrum-2/Spectrum-3/Spectrum-4 Ethernet Switch ASICs.
 
 	  To compile this driver as a module, choose M here: the
 	  module will be called mlxsw_spectrum.
diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci.h b/drivers/net/ethernet/mellanox/mlxsw/pci.h
index 9899c1a2ea8f..cacc2f9fa1d4 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/pci.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/pci.h
@@ -9,6 +9,7 @@
 #define PCI_DEVICE_ID_MELLANOX_SPECTRUM		0xcb84
 #define PCI_DEVICE_ID_MELLANOX_SPECTRUM2	0xcf6c
 #define PCI_DEVICE_ID_MELLANOX_SPECTRUM3	0xcf70
+#define PCI_DEVICE_ID_MELLANOX_SPECTRUM4	0xcf80
 
 #if IS_ENABLED(CONFIG_MLXSW_PCI)
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index c97d2c744725..24cc65018b41 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -3914,6 +3914,7 @@ MLXSW_ITEM32(reg, qeec, max_shaper_bs, 0x1C, 0, 6);
 #define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP1	5
 #define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP2	11
 #define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP3	11
+#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP4	11
 
 static inline void mlxsw_reg_qeec_pack(char *payload, u16 local_port,
 				       enum mlxsw_reg_qeec_hr hr, u8 index,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 6e4265c86eb8..aa411dec62f0 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -95,6 +95,7 @@ static const struct mlxsw_fw_rev mlxsw_sp3_fw_rev = {
 static const char mlxsw_sp1_driver_name[] = "mlxsw_spectrum";
 static const char mlxsw_sp2_driver_name[] = "mlxsw_spectrum2";
 static const char mlxsw_sp3_driver_name[] = "mlxsw_spectrum3";
+static const char mlxsw_sp4_driver_name[] = "mlxsw_spectrum4";
 
 static const unsigned char mlxsw_sp1_mac_mask[ETH_ALEN] = {
 	0xff, 0xff, 0xff, 0xff, 0xfc, 0x00
@@ -3202,6 +3203,36 @@ static int mlxsw_sp3_init(struct mlxsw_core *mlxsw_core,
 	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
 }
 
+static int mlxsw_sp4_init(struct mlxsw_core *mlxsw_core,
+			  const struct mlxsw_bus_info *mlxsw_bus_info,
+			  struct netlink_ext_ack *extack)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_core_driver_priv(mlxsw_core);
+
+	mlxsw_sp->switchdev_ops = &mlxsw_sp2_switchdev_ops;
+	mlxsw_sp->kvdl_ops = &mlxsw_sp2_kvdl_ops;
+	mlxsw_sp->afa_ops = &mlxsw_sp2_act_afa_ops;
+	mlxsw_sp->afk_ops = &mlxsw_sp4_afk_ops;
+	mlxsw_sp->mr_tcam_ops = &mlxsw_sp2_mr_tcam_ops;
+	mlxsw_sp->acl_rulei_ops = &mlxsw_sp2_acl_rulei_ops;
+	mlxsw_sp->acl_tcam_ops = &mlxsw_sp2_acl_tcam_ops;
+	mlxsw_sp->acl_bf_ops = &mlxsw_sp4_acl_bf_ops;
+	mlxsw_sp->nve_ops_arr = mlxsw_sp2_nve_ops_arr;
+	mlxsw_sp->mac_mask = mlxsw_sp2_mac_mask;
+	mlxsw_sp->sb_vals = &mlxsw_sp2_sb_vals;
+	mlxsw_sp->sb_ops = &mlxsw_sp3_sb_ops;
+	mlxsw_sp->port_type_speed_ops = &mlxsw_sp2_port_type_speed_ops;
+	mlxsw_sp->ptp_ops = &mlxsw_sp2_ptp_ops;
+	mlxsw_sp->span_ops = &mlxsw_sp3_span_ops;
+	mlxsw_sp->policer_core_ops = &mlxsw_sp2_policer_core_ops;
+	mlxsw_sp->trap_ops = &mlxsw_sp2_trap_ops;
+	mlxsw_sp->mall_ops = &mlxsw_sp2_mall_ops;
+	mlxsw_sp->router_ops = &mlxsw_sp2_router_ops;
+	mlxsw_sp->lowest_shaper_bs = MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP4;
+
+	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
+}
+
 static void mlxsw_sp_fini(struct mlxsw_core *mlxsw_core)
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_core_driver_priv(mlxsw_core);
@@ -3761,6 +3792,45 @@ static struct mlxsw_driver mlxsw_sp3_driver = {
 	.temp_warn_enabled		= true,
 };
 
+static struct mlxsw_driver mlxsw_sp4_driver = {
+	.kind				= mlxsw_sp4_driver_name,
+	.priv_size			= sizeof(struct mlxsw_sp),
+	.init				= mlxsw_sp4_init,
+	.fini				= mlxsw_sp_fini,
+	.basic_trap_groups_set		= mlxsw_sp_basic_trap_groups_set,
+	.port_split			= mlxsw_sp_port_split,
+	.port_unsplit			= mlxsw_sp_port_unsplit,
+	.sb_pool_get			= mlxsw_sp_sb_pool_get,
+	.sb_pool_set			= mlxsw_sp_sb_pool_set,
+	.sb_port_pool_get		= mlxsw_sp_sb_port_pool_get,
+	.sb_port_pool_set		= mlxsw_sp_sb_port_pool_set,
+	.sb_tc_pool_bind_get		= mlxsw_sp_sb_tc_pool_bind_get,
+	.sb_tc_pool_bind_set		= mlxsw_sp_sb_tc_pool_bind_set,
+	.sb_occ_snapshot		= mlxsw_sp_sb_occ_snapshot,
+	.sb_occ_max_clear		= mlxsw_sp_sb_occ_max_clear,
+	.sb_occ_port_pool_get		= mlxsw_sp_sb_occ_port_pool_get,
+	.sb_occ_tc_port_bind_get	= mlxsw_sp_sb_occ_tc_port_bind_get,
+	.trap_init			= mlxsw_sp_trap_init,
+	.trap_fini			= mlxsw_sp_trap_fini,
+	.trap_action_set		= mlxsw_sp_trap_action_set,
+	.trap_group_init		= mlxsw_sp_trap_group_init,
+	.trap_group_set			= mlxsw_sp_trap_group_set,
+	.trap_policer_init		= mlxsw_sp_trap_policer_init,
+	.trap_policer_fini		= mlxsw_sp_trap_policer_fini,
+	.trap_policer_set		= mlxsw_sp_trap_policer_set,
+	.trap_policer_counter_get	= mlxsw_sp_trap_policer_counter_get,
+	.txhdr_construct		= mlxsw_sp_txhdr_construct,
+	.resources_register		= mlxsw_sp2_resources_register,
+	.params_register		= mlxsw_sp2_params_register,
+	.params_unregister		= mlxsw_sp2_params_unregister,
+	.ptp_transmitted		= mlxsw_sp_ptp_transmitted,
+	.txhdr_len			= MLXSW_TXHDR_LEN,
+	.profile			= &mlxsw_sp2_config_profile,
+	.res_query_enabled		= true,
+	.fw_fatal_enabled		= true,
+	.temp_warn_enabled		= true,
+};
+
 bool mlxsw_sp_port_dev_check(const struct net_device *dev)
 {
 	return dev->netdev_ops == &mlxsw_sp_port_netdev_ops;
@@ -4928,6 +4998,16 @@ static struct pci_driver mlxsw_sp3_pci_driver = {
 	.id_table = mlxsw_sp3_pci_id_table,
 };
 
+static const struct pci_device_id mlxsw_sp4_pci_id_table[] = {
+	{PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_SPECTRUM4), 0},
+	{0, },
+};
+
+static struct pci_driver mlxsw_sp4_pci_driver = {
+	.name = mlxsw_sp4_driver_name,
+	.id_table = mlxsw_sp4_pci_id_table,
+};
+
 static int __init mlxsw_sp_module_init(void)
 {
 	int err;
@@ -4947,6 +5027,10 @@ static int __init mlxsw_sp_module_init(void)
 	if (err)
 		goto err_sp3_core_driver_register;
 
+	err = mlxsw_core_driver_register(&mlxsw_sp4_driver);
+	if (err)
+		goto err_sp4_core_driver_register;
+
 	err = mlxsw_pci_driver_register(&mlxsw_sp1_pci_driver);
 	if (err)
 		goto err_sp1_pci_driver_register;
@@ -4959,13 +5043,21 @@ static int __init mlxsw_sp_module_init(void)
 	if (err)
 		goto err_sp3_pci_driver_register;
 
+	err = mlxsw_pci_driver_register(&mlxsw_sp4_pci_driver);
+	if (err)
+		goto err_sp4_pci_driver_register;
+
 	return 0;
 
+err_sp4_pci_driver_register:
+	mlxsw_pci_driver_unregister(&mlxsw_sp3_pci_driver);
 err_sp3_pci_driver_register:
 	mlxsw_pci_driver_unregister(&mlxsw_sp2_pci_driver);
 err_sp2_pci_driver_register:
 	mlxsw_pci_driver_unregister(&mlxsw_sp1_pci_driver);
 err_sp1_pci_driver_register:
+	mlxsw_core_driver_unregister(&mlxsw_sp4_driver);
+err_sp4_core_driver_register:
 	mlxsw_core_driver_unregister(&mlxsw_sp3_driver);
 err_sp3_core_driver_register:
 	mlxsw_core_driver_unregister(&mlxsw_sp2_driver);
@@ -4979,9 +5071,11 @@ static int __init mlxsw_sp_module_init(void)
 
 static void __exit mlxsw_sp_module_exit(void)
 {
+	mlxsw_pci_driver_unregister(&mlxsw_sp4_pci_driver);
 	mlxsw_pci_driver_unregister(&mlxsw_sp3_pci_driver);
 	mlxsw_pci_driver_unregister(&mlxsw_sp2_pci_driver);
 	mlxsw_pci_driver_unregister(&mlxsw_sp1_pci_driver);
+	mlxsw_core_driver_unregister(&mlxsw_sp4_driver);
 	mlxsw_core_driver_unregister(&mlxsw_sp3_driver);
 	mlxsw_core_driver_unregister(&mlxsw_sp2_driver);
 	mlxsw_core_driver_unregister(&mlxsw_sp1_driver);
@@ -4998,6 +5092,7 @@ MODULE_DESCRIPTION("Mellanox Spectrum driver");
 MODULE_DEVICE_TABLE(pci, mlxsw_sp1_pci_id_table);
 MODULE_DEVICE_TABLE(pci, mlxsw_sp2_pci_id_table);
 MODULE_DEVICE_TABLE(pci, mlxsw_sp3_pci_id_table);
+MODULE_DEVICE_TABLE(pci, mlxsw_sp4_pci_id_table);
 MODULE_FIRMWARE(MLXSW_SP1_FW_FILENAME);
 MODULE_FIRMWARE(MLXSW_SP2_FW_FILENAME);
 MODULE_FIRMWARE(MLXSW_SP3_FW_FILENAME);
-- 
2.33.1

Re: [PATCH net-next 0/8] mlxsw: Add Spectrum-4 support

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (master)
by Jakub Kicinski <kuba@kernel.org>:

On Thu,  6 Jan 2022 18:06:44 +0200 you wrote:
> This patchset adds Spectrum-4 support in mlxsw. It builds on top of a
> previous patchset merged in commit 10184da91666 ("Merge branch
> 'mlxsw-Spectrum-4-prep'") and makes two additional changes before adding
> Spectrum-4 support.
> 
> Patchset overview:
> 
> [...]

Here is the summary with links:
  - [net-next,1/8] mlxsw: Rename virtual router flex key element
    https://git.kernel.org/netdev/net-next/c/6d5d8ebb881c
  - [net-next,2/8] mlxsw: Introduce flex key elements for Spectrum-4
    https://git.kernel.org/netdev/net-next/c/07ff135958dd
  - [net-next,3/8] mlxsw: spectrum_acl_bloom_filter: Reorder functions to make the code more aesthetic
    https://git.kernel.org/netdev/net-next/c/4711671297ec
  - [net-next,4/8] mlxsw: spectrum_acl_bloom_filter: Make mlxsw_sp_acl_bf_key_encode() more flexible
    https://git.kernel.org/netdev/net-next/c/5d5c3ba9e412
  - [net-next,5/8] mlxsw: spectrum_acl_bloom_filter: Rename Spectrum-2 specific objects for future use
    https://git.kernel.org/netdev/net-next/c/29409f363e2d
  - [net-next,6/8] mlxsw: Add operations structure for bloom filter calculation
    https://git.kernel.org/netdev/net-next/c/58723d2f7771
  - [net-next,7/8] mlxsw: spectrum_acl_bloom_filter: Add support for Spectrum-4 calculation
    https://git.kernel.org/netdev/net-next/c/852ee4191dd2
  - [net-next,8/8] mlxsw: spectrum: Extend to support Spectrum-4 ASIC
    https://git.kernel.org/netdev/net-next/c/4735402173e6

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html