Linux Ethernet Bridging - Jan 2021 - [Bridge] [PATCH net-next 08/14] net: bridge: multicast: add EHT host delete function

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Hi,
This set adds explicit host tracking support for IGMPv3/MLDv2. The
already present per-port fast leave flag is used to enable it since that
is the primary goal of EHT, to track a group and its S,Gs usage per-host
and when left without any interested hosts delete them before the standard
timers. The EHT code is pretty self-contained and not enabled by default.
There is no new uAPI added, all of the functionality is currently hidden
behind the fast leave flag. In the future that will change (more below).
The host tracking uses two new sets per port group: one having an entry for
each host which contains that host's view of the group (source list and
filter mode), and one set which contains an entry for each source having
an internal set which contains an entry for each host that has reported
an interest for that source. RB trees are used for all sets so they're
compact when not used and fast when we need to do lookups.
To illustrate it:
 [ bridge port group ]
  ` [ host set (rb) ]
   ` [ host entry with a list of sources and filter mode ]
  ` [ source set (rb) ]
   ` [ source entry ]
    ` [ source host set (rb) ]
     ` [ source host entry with a timer ]

The number of tracked sources per host is limited to the maximum total
number of S,G entries per port group - PG_SRC_ENT_LIMIT (currently 32).
The number of hosts is unlimited, I think the argument that a local
attacker can exhaust the memory/cause high CPU usage can be applied to
fdb entries as well which are unlimited. In the future if needed we can
add an option to limit these, but I don't think it's necessary for a
start. All of the new sets are protected by the bridge's multicast lock.
I'm pretty sure we'll be changing the cases and improving the
convergence time in the future, but this seems like a good start.

I'll post self-tests as a separate patch-set.

Patch breakdown:
 patch 1 -  4: minor cleanups and preparations for EHT
 patch      5: adds the new structures which will be used in the
               following patches
 patch      6: adds support to create, destroy and lookup host entries
 patch      7: adds support to create, delete and lokup source set entries
 patch      8: adds a host "delete" function which is just a
host's
               source list flush since that would automatically delete
               the host
 patch 9 - 10: add support for handling all IGMPv3/MLDv2 report types
               more information can be found in the individual patches
 patch     11: optmizes a specific TO_INCLUDE use-case with host timeouts
 patch     12: handles per-host filter mode changing (include <-> exclude)
 patch     13: pulls out block group deletion since now it can be
               deleted in both filter modes
 patch     14: marks deletions done due to fast leave

Future plans:
 - export host information
 - add an option to reduce queries
 - add an option to limit the number of host entries
 - tune more fast leave cases for quicker convergence

By the way I think this is the first open-source EHT implementation, I
couldn't find any while researching it. :)

Thanks,
 Nik


Nikolay Aleksandrov (14):
  net: bridge: multicast: rename src_size to addr_size
  net: bridge: multicast: pass host src address to IGMPv3/MLDv2
    functions
  net: bridge: multicast: __grp_src_block_incl can modify pg
  net: bridge: multicast: calculate idx position without changing ptr
  net: bridge: multicast: add EHT structures and definitions
  net: bridge: multicast: add EHT host handling functions
  net: bridge: multicast: add EHT source set handling functions
  net: bridge: multicast: add EHT host delete function
  net: bridge: multicast: add EHT allow/block handling
  net: bridge: multicast: add EHT include and exclude handling
  net: bridge: multicast: optimize TO_INCLUDE EHT timeouts
  net: bridge: multicast: add EHT host filter_mode handling
  net: bridge: multicast: handle block pg delete for all cases
  net: bridge: multicast: mark IGMPv3/MLDv2 fast-leave deletes

 net/bridge/Makefile               |   2 +-
 net/bridge/br_multicast.c         | 254 +++++----
 net/bridge/br_multicast_eht.c     | 856 ++++++++++++++++++++++++++++++
 net/bridge/br_private.h           |   6 +
 net/bridge/br_private_mcast_eht.h |  65 +++
 5 files changed, 1093 insertions(+), 90 deletions(-)
 create mode 100644 net/bridge/br_multicast_eht.c
 create mode 100644 net/bridge/br_private_mcast_eht.h

-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Rename src_size argument to addr_size in preparation for passing host
address as an argument to IGMPv3/MLDv2 functions.
No functional change.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c | 78 +++++++++++++++++++--------------------
 1 file changed, 39 insertions(+), 39 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 257ac4e25f6d..3ae2cef6f7ec 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -1800,7 +1800,7 @@ static void __grp_send_query_and_rexmit(struct
net_bridge_port_group *pg)
  * EXCLUDE (X,Y)  ALLOW (A)     EXCLUDE (X+A,Y-A)        (A)=GMI
  */
 static bool br_multicast_isinc_allow(struct net_bridge_port_group *pg,
-				     void *srcs, u32 nsrcs, size_t src_size)
+				     void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
 	struct net_bridge_group_src *ent;
@@ -1812,7 +1812,7 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (!ent) {
 			ent = br_multicast_new_group_src(pg, &src_ip);
@@ -1822,7 +1822,7 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg,
 
 		if (ent)
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	return changed;
@@ -1834,7 +1834,7 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg,
  *                                                       Group Timer=GMI
  */
 static void __grp_src_isexc_incl(struct net_bridge_port_group *pg,
-				 void *srcs, u32 nsrcs, size_t src_size)
+				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
 	struct br_ip src_ip;
@@ -1846,7 +1846,7 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent)
 			ent->flags &= ~BR_SGRP_F_DELETE;
@@ -1854,7 +1854,7 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg,
 			ent = br_multicast_new_group_src(pg, &src_ip);
 		if (ent)
 			br_multicast_fwd_src_handle(ent);
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	__grp_src_delete_marked(pg);
@@ -1867,7 +1867,7 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg,
  *                                                       Group Timer=GMI
  */
 static bool __grp_src_isexc_excl(struct net_bridge_port_group *pg,
-				 void *srcs, u32 nsrcs, size_t src_size)
+				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
 	struct net_bridge_group_src *ent;
@@ -1882,7 +1882,7 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags &= ~BR_SGRP_F_DELETE;
@@ -1894,7 +1894,7 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg,
 				changed = true;
 			}
 		}
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	if (__grp_src_delete_marked(pg))
@@ -1904,19 +1904,19 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg,
 }
 
 static bool br_multicast_isexc(struct net_bridge_port_group *pg,
-			       void *srcs, u32 nsrcs, size_t src_size)
+			       void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_isexc_incl(pg, srcs, nsrcs, src_size);
+		__grp_src_isexc_incl(pg, srcs, nsrcs, addr_size);
 		br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
 		changed = true;
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_isexc_excl(pg, srcs, nsrcs, src_size);
+		changed = __grp_src_isexc_excl(pg, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -1931,7 +1931,7 @@ static bool br_multicast_isexc(struct
net_bridge_port_group *pg,
  *                                                       Send Q(G,A-B)
  */
 static bool __grp_src_toin_incl(struct net_bridge_port_group *pg,
-				void *srcs, u32 nsrcs, size_t src_size)
+				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
 	u32 src_idx, to_send = pg->src_ents;
@@ -1946,7 +1946,7 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags &= ~BR_SGRP_F_SEND;
@@ -1958,7 +1958,7 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg,
 		}
 		if (ent)
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -1973,7 +1973,7 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg,
  *                                                       Send Q(G)
  */
 static bool __grp_src_toin_excl(struct net_bridge_port_group *pg,
-				void *srcs, u32 nsrcs, size_t src_size)
+				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
 	u32 src_idx, to_send = pg->src_ents;
@@ -1989,7 +1989,7 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			if (timer_pending(&ent->timer)) {
@@ -2003,7 +2003,7 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg,
 		}
 		if (ent)
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -2015,16 +2015,16 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg,
 }
 
 static bool br_multicast_toin(struct net_bridge_port_group *pg,
-			      void *srcs, u32 nsrcs, size_t src_size)
+			      void *srcs, u32 nsrcs, size_t addr_size)
 {
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		changed = __grp_src_toin_incl(pg, srcs, nsrcs, src_size);
+		changed = __grp_src_toin_incl(pg, srcs, nsrcs, addr_size);
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_toin_excl(pg, srcs, nsrcs, src_size);
+		changed = __grp_src_toin_excl(pg, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -2038,7 +2038,7 @@ static bool br_multicast_toin(struct net_bridge_port_group
*pg,
  *                                                       Group Timer=GMI
  */
 static void __grp_src_toex_incl(struct net_bridge_port_group *pg,
-				void *srcs, u32 nsrcs, size_t src_size)
+				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2050,7 +2050,7 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags = (ent->flags & ~BR_SGRP_F_DELETE) |
@@ -2061,7 +2061,7 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg,
 		}
 		if (ent)
 			br_multicast_fwd_src_handle(ent);
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	__grp_src_delete_marked(pg);
@@ -2077,7 +2077,7 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg,
  *                                                       Group Timer=GMI
  */
 static bool __grp_src_toex_excl(struct net_bridge_port_group *pg,
-				void *srcs, u32 nsrcs, size_t src_size)
+				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2090,7 +2090,7 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags &= ~BR_SGRP_F_DELETE;
@@ -2105,7 +2105,7 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg,
 			ent->flags |= BR_SGRP_F_SEND;
 			to_send++;
 		}
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	if (__grp_src_delete_marked(pg))
@@ -2117,19 +2117,19 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg,
 }
 
 static bool br_multicast_toex(struct net_bridge_port_group *pg,
-			      void *srcs, u32 nsrcs, size_t src_size)
+			      void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_toex_incl(pg, srcs, nsrcs, src_size);
+		__grp_src_toex_incl(pg, srcs, nsrcs, addr_size);
 		br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
 		changed = true;
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_toex_excl(pg, srcs, nsrcs, src_size);
+		changed = __grp_src_toex_excl(pg, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -2143,7 +2143,7 @@ static bool br_multicast_toex(struct net_bridge_port_group
*pg,
  * INCLUDE (A)    BLOCK (B)     INCLUDE (A)              Send Q(G,A*B)
  */
 static void __grp_src_block_incl(struct net_bridge_port_group *pg,
-				 void *srcs, u32 nsrcs, size_t src_size)
+				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2155,13 +2155,13 @@ static void __grp_src_block_incl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags |= BR_SGRP_F_SEND;
 			to_send++;
 		}
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -2176,7 +2176,7 @@ static void __grp_src_block_incl(struct
net_bridge_port_group *pg,
  *                                                       Send Q(G,A-Y)
  */
 static bool __grp_src_block_excl(struct net_bridge_port_group *pg,
-				 void *srcs, u32 nsrcs, size_t src_size)
+				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2189,7 +2189,7 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, src_size);
+		memcpy(&src_ip.src, srcs, addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (!ent) {
 			ent = br_multicast_new_group_src(pg, &src_ip);
@@ -2202,7 +2202,7 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg,
 			ent->flags |= BR_SGRP_F_SEND;
 			to_send++;
 		}
-		srcs += src_size;
+		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -2212,16 +2212,16 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg,
 }
 
 static bool br_multicast_block(struct net_bridge_port_group *pg,
-			       void *srcs, u32 nsrcs, size_t src_size)
+			       void *srcs, u32 nsrcs, size_t addr_size)
 {
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_block_incl(pg, srcs, nsrcs, src_size);
+		__grp_src_block_incl(pg, srcs, nsrcs, addr_size);
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_block_excl(pg, srcs, nsrcs, src_size);
+		changed = __grp_src_block_excl(pg, srcs, nsrcs, addr_size);
 		break;
 	}
 
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

We need to pass the host address so later it can be used for explicit
host tracking. No functional change.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c | 90 +++++++++++++++++++++------------------
 1 file changed, 49 insertions(+), 41 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 3ae2cef6f7ec..861545094d67 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -1799,7 +1799,7 @@ static void __grp_send_query_and_rexmit(struct
net_bridge_port_group *pg)
  * INCLUDE (A)    ALLOW (B)     INCLUDE (A+B)            (B)=GMI
  * EXCLUDE (X,Y)  ALLOW (A)     EXCLUDE (X+A,Y-A)        (A)=GMI
  */
-static bool br_multicast_isinc_allow(struct net_bridge_port_group *pg,
+static bool br_multicast_isinc_allow(struct net_bridge_port_group *pg, void
*h_addr,
 				     void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
@@ -1833,7 +1833,7 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg,
  *                                                       Delete (A-B)
  *                                                       Group Timer=GMI
  */
-static void __grp_src_isexc_incl(struct net_bridge_port_group *pg,
+static void __grp_src_isexc_incl(struct net_bridge_port_group *pg, void
*h_addr,
 				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
@@ -1866,7 +1866,7 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg,
  *                                                       Delete (Y-A)
  *                                                       Group Timer=GMI
  */
-static bool __grp_src_isexc_excl(struct net_bridge_port_group *pg,
+static bool __grp_src_isexc_excl(struct net_bridge_port_group *pg, void
*h_addr,
 				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
@@ -1903,7 +1903,7 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg,
 	return changed;
 }
 
-static bool br_multicast_isexc(struct net_bridge_port_group *pg,
+static bool br_multicast_isexc(struct net_bridge_port_group *pg, void *h_addr,
 			       void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
@@ -1911,12 +1911,12 @@ static bool br_multicast_isexc(struct
net_bridge_port_group *pg,
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_isexc_incl(pg, srcs, nsrcs, addr_size);
+		__grp_src_isexc_incl(pg, h_addr, srcs, nsrcs, addr_size);
 		br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
 		changed = true;
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_isexc_excl(pg, srcs, nsrcs, addr_size);
+		changed = __grp_src_isexc_excl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -1930,7 +1930,7 @@ static bool br_multicast_isexc(struct
net_bridge_port_group *pg,
  * INCLUDE (A)    TO_IN (B)     INCLUDE (A+B)            (B)=GMI
  *                                                       Send Q(G,A-B)
  */
-static bool __grp_src_toin_incl(struct net_bridge_port_group *pg,
+static bool __grp_src_toin_incl(struct net_bridge_port_group *pg, void *h_addr,
 				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
@@ -1972,7 +1972,7 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg,
  *                                                       Send Q(G,X-A)
  *                                                       Send Q(G)
  */
-static bool __grp_src_toin_excl(struct net_bridge_port_group *pg,
+static bool __grp_src_toin_excl(struct net_bridge_port_group *pg, void *h_addr,
 				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
@@ -2014,17 +2014,17 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg,
 	return changed;
 }
 
-static bool br_multicast_toin(struct net_bridge_port_group *pg,
+static bool br_multicast_toin(struct net_bridge_port_group *pg, void *h_addr,
 			      void *srcs, u32 nsrcs, size_t addr_size)
 {
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		changed = __grp_src_toin_incl(pg, srcs, nsrcs, addr_size);
+		changed = __grp_src_toin_incl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_toin_excl(pg, srcs, nsrcs, addr_size);
+		changed = __grp_src_toin_excl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -2037,7 +2037,7 @@ static bool br_multicast_toin(struct net_bridge_port_group
*pg,
  *                                                       Send Q(G,A*B)
  *                                                       Group Timer=GMI
  */
-static void __grp_src_toex_incl(struct net_bridge_port_group *pg,
+static void __grp_src_toex_incl(struct net_bridge_port_group *pg, void *h_addr,
 				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
@@ -2076,7 +2076,7 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg,
  *                                                       Send Q(G,A-Y)
  *                                                       Group Timer=GMI
  */
-static bool __grp_src_toex_excl(struct net_bridge_port_group *pg,
+static bool __grp_src_toex_excl(struct net_bridge_port_group *pg, void *h_addr,
 				void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
@@ -2116,7 +2116,7 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg,
 	return changed;
 }
 
-static bool br_multicast_toex(struct net_bridge_port_group *pg,
+static bool br_multicast_toex(struct net_bridge_port_group *pg, void *h_addr,
 			      void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge *br = pg->key.port->br;
@@ -2124,12 +2124,12 @@ static bool br_multicast_toex(struct
net_bridge_port_group *pg,
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_toex_incl(pg, srcs, nsrcs, addr_size);
+		__grp_src_toex_incl(pg, h_addr, srcs, nsrcs, addr_size);
 		br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
 		changed = true;
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_toex_excl(pg, srcs, nsrcs, addr_size);
+		changed = __grp_src_toex_excl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -2142,7 +2142,7 @@ static bool br_multicast_toex(struct net_bridge_port_group
*pg,
 /* State          Msg type      New state                Actions
  * INCLUDE (A)    BLOCK (B)     INCLUDE (A)              Send Q(G,A*B)
  */
-static void __grp_src_block_incl(struct net_bridge_port_group *pg,
+static void __grp_src_block_incl(struct net_bridge_port_group *pg, void
*h_addr,
 				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
@@ -2175,7 +2175,7 @@ static void __grp_src_block_incl(struct
net_bridge_port_group *pg,
  * EXCLUDE (X,Y)  BLOCK (A)     EXCLUDE (X+(A-Y),Y)      (A-X-Y)=Group Timer
  *                                                       Send Q(G,A-Y)
  */
-static bool __grp_src_block_excl(struct net_bridge_port_group *pg,
+static bool __grp_src_block_excl(struct net_bridge_port_group *pg, void
*h_addr,
 				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
@@ -2211,17 +2211,17 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg,
 	return changed;
 }
 
-static bool br_multicast_block(struct net_bridge_port_group *pg,
+static bool br_multicast_block(struct net_bridge_port_group *pg, void *h_addr,
 			       void *srcs, u32 nsrcs, size_t addr_size)
 {
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_block_incl(pg, srcs, nsrcs, addr_size);
+		__grp_src_block_incl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_block_excl(pg, srcs, nsrcs, addr_size);
+		changed = __grp_src_block_excl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	}
 
@@ -2257,8 +2257,8 @@ static int br_ip4_multicast_igmp3_report(struct net_bridge
*br,
 	struct igmpv3_report *ih;
 	struct igmpv3_grec *grec;
 	int i, len, num, type;
+	__be32 group, *h_addr;
 	bool changed = false;
-	__be32 group;
 	int err = 0;
 	u16 nsrcs;
 
@@ -2318,32 +2318,33 @@ static int br_ip4_multicast_igmp3_report(struct
net_bridge *br,
 		pg = br_multicast_find_port(mdst, port, src);
 		if (!pg || (pg->flags & MDB_PG_FLAGS_PERMANENT))
 			goto unlock_continue;
-		/* reload grec */
+		/* reload grec and host addr */
 		grec = (void *)(skb->data + len - sizeof(*grec) - (nsrcs * 4));
+		h_addr = &ip_hdr(skb)->saddr;
 		switch (type) {
 		case IGMPV3_ALLOW_NEW_SOURCES:
-			changed = br_multicast_isinc_allow(pg, grec->grec_src,
+			changed = br_multicast_isinc_allow(pg, h_addr, grec->grec_src,
 							   nsrcs, sizeof(__be32));
 			break;
 		case IGMPV3_MODE_IS_INCLUDE:
-			changed = br_multicast_isinc_allow(pg, grec->grec_src, nsrcs,
-							   sizeof(__be32));
+			changed = br_multicast_isinc_allow(pg, h_addr, grec->grec_src,
+							   nsrcs, sizeof(__be32));
 			break;
 		case IGMPV3_MODE_IS_EXCLUDE:
-			changed = br_multicast_isexc(pg, grec->grec_src, nsrcs,
-						     sizeof(__be32));
+			changed = br_multicast_isexc(pg, h_addr, grec->grec_src,
+						     nsrcs, sizeof(__be32));
 			break;
 		case IGMPV3_CHANGE_TO_INCLUDE:
-			changed = br_multicast_toin(pg, grec->grec_src, nsrcs,
-						    sizeof(__be32));
+			changed = br_multicast_toin(pg, h_addr, grec->grec_src,
+						    nsrcs, sizeof(__be32));
 			break;
 		case IGMPV3_CHANGE_TO_EXCLUDE:
-			changed = br_multicast_toex(pg, grec->grec_src, nsrcs,
-						    sizeof(__be32));
+			changed = br_multicast_toex(pg, h_addr, grec->grec_src,
+						    nsrcs, sizeof(__be32));
 			break;
 		case IGMPV3_BLOCK_OLD_SOURCES:
-			changed = br_multicast_block(pg, grec->grec_src, nsrcs,
-						     sizeof(__be32));
+			changed = br_multicast_block(pg, h_addr, grec->grec_src,
+						     nsrcs, sizeof(__be32));
 			break;
 		}
 		if (changed)
@@ -2367,6 +2368,7 @@ static int br_ip6_multicast_mld2_report(struct net_bridge
*br,
 	unsigned int nsrcs_offset;
 	const unsigned char *src;
 	struct icmp6hdr *icmp6h;
+	struct in6_addr *h_addr;
 	struct mld2_grec *grec;
 	unsigned int grec_len;
 	bool changed = false;
@@ -2445,30 +2447,36 @@ static int br_ip6_multicast_mld2_report(struct
net_bridge *br,
 		pg = br_multicast_find_port(mdst, port, src);
 		if (!pg || (pg->flags & MDB_PG_FLAGS_PERMANENT))
 			goto unlock_continue;
+		h_addr = &ipv6_hdr(skb)->saddr;
 		switch (grec->grec_type) {
 		case MLD2_ALLOW_NEW_SOURCES:
-			changed = br_multicast_isinc_allow(pg, grec->grec_src,
-							   nsrcs,
+			changed = br_multicast_isinc_allow(pg, h_addr,
+							   grec->grec_src, nsrcs,
 							   sizeof(struct in6_addr));
 			break;
 		case MLD2_MODE_IS_INCLUDE:
-			changed = br_multicast_isinc_allow(pg, grec->grec_src, nsrcs,
+			changed = br_multicast_isinc_allow(pg, h_addr,
+							   grec->grec_src, nsrcs,
 							   sizeof(struct in6_addr));
 			break;
 		case MLD2_MODE_IS_EXCLUDE:
-			changed = br_multicast_isexc(pg, grec->grec_src, nsrcs,
+			changed = br_multicast_isexc(pg, h_addr,
+						     grec->grec_src, nsrcs,
 						     sizeof(struct in6_addr));
 			break;
 		case MLD2_CHANGE_TO_INCLUDE:
-			changed = br_multicast_toin(pg, grec->grec_src, nsrcs,
+			changed = br_multicast_toin(pg, h_addr,
+						    grec->grec_src, nsrcs,
 						    sizeof(struct in6_addr));
 			break;
 		case MLD2_CHANGE_TO_EXCLUDE:
-			changed = br_multicast_toex(pg, grec->grec_src, nsrcs,
+			changed = br_multicast_toex(pg, h_addr,
+						    grec->grec_src, nsrcs,
 						    sizeof(struct in6_addr));
 			break;
 		case MLD2_BLOCK_OLD_SOURCES:
-			changed = br_multicast_block(pg, grec->grec_src, nsrcs,
+			changed = br_multicast_block(pg, h_addr,
+						     grec->grec_src, nsrcs,
 						     sizeof(struct in6_addr));
 			break;
 		}
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Prepare __grp_src_block_incl() for being able to cause a notification
due to changes. Currently it cannot happen, but EHT would change that
since we'll be deleting sources immediately. Make sure that if the pg is
deleted we don't return true as that would cause the caller to access
freed pg. This patch shouldn't cause any functional change.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 861545094d67..79569a398669 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -2142,11 +2142,12 @@ static bool br_multicast_toex(struct
net_bridge_port_group *pg, void *h_addr,
 /* State          Msg type      New state                Actions
  * INCLUDE (A)    BLOCK (B)     INCLUDE (A)              Send Q(G,A*B)
  */
-static void __grp_src_block_incl(struct net_bridge_port_group *pg, void
*h_addr,
+static bool __grp_src_block_incl(struct net_bridge_port_group *pg, void
*h_addr,
 				 void *srcs, u32 nsrcs, size_t addr_size)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
+	bool changed = false;
 	struct br_ip src_ip;
 
 	hlist_for_each_entry(ent, &pg->src_list, node)
@@ -2167,8 +2168,15 @@ static void __grp_src_block_incl(struct
net_bridge_port_group *pg, void *h_addr,
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
 
-	if (pg->filter_mode == MCAST_INCLUDE &&
hlist_empty(&pg->src_list))
+	if (pg->filter_mode == MCAST_INCLUDE &&
hlist_empty(&pg->src_list)) {
 		br_multicast_find_del_pg(pg->key.port->br, pg);
+		/* a notification has already been sent and we shouldn't access
+		 * pg after the delete thus we have to return false
+		 */
+		changed = false;
+	}
+
+	return changed;
 }
 
 /* State          Msg type      New state                Actions
@@ -2218,7 +2226,7 @@ static bool br_multicast_block(struct
net_bridge_port_group *pg, void *h_addr,
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_block_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_block_incl(pg, h_addr, srcs, nsrcs, addr_size);
 		break;
 	case MCAST_EXCLUDE:
 		changed = __grp_src_block_excl(pg, h_addr, srcs, nsrcs, addr_size);
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

We need to preserve the srcs pointer since we'll be passing it for EHT
handling later.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c | 27 +++++++++------------------
 1 file changed, 9 insertions(+), 18 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 79569a398669..f8b685ae56d4 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -1812,7 +1812,7 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg, void *h_a
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (!ent) {
 			ent = br_multicast_new_group_src(pg, &src_ip);
@@ -1822,7 +1822,6 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg, void *h_a
 
 		if (ent)
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
-		srcs += addr_size;
 	}
 
 	return changed;
@@ -1846,7 +1845,7 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent)
 			ent->flags &= ~BR_SGRP_F_DELETE;
@@ -1854,7 +1853,6 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg, void *h_addr,
 			ent = br_multicast_new_group_src(pg, &src_ip);
 		if (ent)
 			br_multicast_fwd_src_handle(ent);
-		srcs += addr_size;
 	}
 
 	__grp_src_delete_marked(pg);
@@ -1882,7 +1880,7 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags &= ~BR_SGRP_F_DELETE;
@@ -1894,7 +1892,6 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg, void *h_addr,
 				changed = true;
 			}
 		}
-		srcs += addr_size;
 	}
 
 	if (__grp_src_delete_marked(pg))
@@ -1946,7 +1943,7 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags &= ~BR_SGRP_F_SEND;
@@ -1958,7 +1955,6 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 		if (ent)
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
-		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -1989,7 +1985,7 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			if (timer_pending(&ent->timer)) {
@@ -2003,7 +1999,6 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 		if (ent)
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
-		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -2050,7 +2045,7 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags = (ent->flags & ~BR_SGRP_F_DELETE) |
@@ -2061,7 +2056,6 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 		if (ent)
 			br_multicast_fwd_src_handle(ent);
-		srcs += addr_size;
 	}
 
 	__grp_src_delete_marked(pg);
@@ -2090,7 +2084,7 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags &= ~BR_SGRP_F_DELETE;
@@ -2105,7 +2099,6 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg, void *h_addr,
 			ent->flags |= BR_SGRP_F_SEND;
 			to_send++;
 		}
-		srcs += addr_size;
 	}
 
 	if (__grp_src_delete_marked(pg))
@@ -2156,13 +2149,12 @@ static bool __grp_src_block_incl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (ent) {
 			ent->flags |= BR_SGRP_F_SEND;
 			to_send++;
 		}
-		srcs += addr_size;
 	}
 
 	if (to_send)
@@ -2197,7 +2189,7 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg, void *h_addr,
 	memset(&src_ip, 0, sizeof(src_ip));
 	src_ip.proto = pg->key.addr.proto;
 	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
-		memcpy(&src_ip.src, srcs, addr_size);
+		memcpy(&src_ip.src, srcs + (src_idx * addr_size), addr_size);
 		ent = br_multicast_find_group_src(pg, &src_ip);
 		if (!ent) {
 			ent = br_multicast_new_group_src(pg, &src_ip);
@@ -2210,7 +2202,6 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg, void *h_addr,
 			ent->flags |= BR_SGRP_F_SEND;
 			to_send++;
 		}
-		srcs += addr_size;
 	}
 
 	if (to_send)
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Add EHT structures for tracking hosts and sources per group. We keep one
set for each host which has all of the host's S,G entries, and one set for
each multicast source which has all hosts that have joined that S,G. For
each host, source entry we record the filter_mode and we keep an expiry
timer. There is also one global expiry timer per source set, it is
updated with each set entry update, it will be later used to lower the
set's timer instead of lowering each entry's timer separately.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c         |  1 +
 net/bridge/br_private.h           |  2 ++
 net/bridge/br_private_mcast_eht.h | 50 +++++++++++++++++++++++++++++++
 3 files changed, 53 insertions(+)
 create mode 100644 net/bridge/br_private_mcast_eht.h

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index f8b685ae56d4..3aaa6adbff82 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -33,6 +33,7 @@
 #endif
 
 #include "br_private.h"
+#include "br_private_mcast_eht.h"
 
 static const struct rhashtable_params br_mdb_rht_params = {
 	.head_offset = offsetof(struct net_bridge_mdb_entry, rhnode),
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index d62c6e1af64a..0bf4c544a5da 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -252,6 +252,8 @@ struct net_bridge_port_group {
 	struct timer_list		timer;
 	struct timer_list		rexmit_timer;
 	struct hlist_node		mglist;
+	struct rb_root			eht_set_tree;
+	struct rb_root			eht_host_tree;
 
 	struct rhash_head		rhnode;
 	struct net_bridge_mcast_gc	mcast_gc;
diff --git a/net/bridge/br_private_mcast_eht.h
b/net/bridge/br_private_mcast_eht.h
new file mode 100644
index 000000000000..0c9c4267969d
--- /dev/null
+++ b/net/bridge/br_private_mcast_eht.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (c) 2020, Nikolay Aleksandrov <nikolay at nvidia.com>
+ */
+#ifndef _BR_PRIVATE_MCAST_EHT_H_
+#define _BR_PRIVATE_MCAST_EHT_H_
+
+union net_bridge_eht_addr {
+	__be32				ip4;
+#if IS_ENABLED(CONFIG_IPV6)
+	struct in6_addr			ip6;
+#endif
+};
+
+/* single host's list of set entries and filter_mode */
+struct net_bridge_group_eht_host {
+	struct rb_node			rb_node;
+
+	union net_bridge_eht_addr	h_addr;
+	struct hlist_head		set_entries;
+	unsigned int			num_entries;
+	unsigned char			filter_mode;
+	struct net_bridge_port_group	*pg;
+};
+
+/* (host, src entry) added to a per-src set and host's list */
+struct net_bridge_group_eht_set_entry {
+	struct rb_node			rb_node;
+	struct hlist_node		host_list;
+
+	union net_bridge_eht_addr	h_addr;
+	struct timer_list		timer;
+	struct net_bridge		*br;
+	struct net_bridge_group_eht_set	*eht_set;
+	struct net_bridge_group_eht_host *h_parent;
+	struct net_bridge_mcast_gc	mcast_gc;
+};
+
+/* per-src set */
+struct net_bridge_group_eht_set {
+	struct rb_node			rb_node;
+
+	union net_bridge_eht_addr	src_addr;
+	struct rb_root			entry_tree;
+	struct timer_list		timer;
+	struct net_bridge_port_group	*pg;
+	struct net_bridge		*br;
+	struct net_bridge_mcast_gc	mcast_gc;
+};
+
+#endif /* _BR_PRIVATE_MCAST_EHT_H_ */
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Add functions to create, destroy and lookup an EHT host. These are
per-host entries contained in the eht_host_tree in net_bridge_port_group
which are used to store a list of all sources (S,G) entries joined for that
group by each host, the host's current filter mode and total number of
joined entries.
No functional changes yet, these would be used in later patches.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/Makefile           |   2 +-
 net/bridge/br_multicast.c     |   1 +
 net/bridge/br_multicast_eht.c | 115 ++++++++++++++++++++++++++++++++++
 3 files changed, 117 insertions(+), 1 deletion(-)
 create mode 100644 net/bridge/br_multicast_eht.c

diff --git a/net/bridge/Makefile b/net/bridge/Makefile
index 4702702a74d3..7fb9a021873b 100644
--- a/net/bridge/Makefile
+++ b/net/bridge/Makefile
@@ -18,7 +18,7 @@ br_netfilter-y := br_netfilter_hooks.o
 br_netfilter-$(subst m,y,$(CONFIG_IPV6)) += br_netfilter_ipv6.o
 obj-$(CONFIG_BRIDGE_NETFILTER) += br_netfilter.o
 
-bridge-$(CONFIG_BRIDGE_IGMP_SNOOPING) += br_multicast.o br_mdb.o
+bridge-$(CONFIG_BRIDGE_IGMP_SNOOPING) += br_multicast.o br_mdb.o
br_multicast_eht.o
 
 bridge-$(CONFIG_BRIDGE_VLAN_FILTERING) += br_vlan.o br_vlan_tunnel.o
br_vlan_options.o
 
diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 3aaa6adbff82..dc6e879dc840 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -1173,6 +1173,7 @@ struct net_bridge_port_group *br_multicast_new_port_group(
 	p->flags = flags;
 	p->filter_mode = filter_mode;
 	p->rt_protocol = rt_protocol;
+	p->eht_host_tree = RB_ROOT;
 	p->mcast_gc.destroy = br_multicast_destroy_port_group;
 	INIT_HLIST_HEAD(&p->src_list);
 
diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
new file mode 100644
index 000000000000..5cebca45e72c
--- /dev/null
+++ b/net/bridge/br_multicast_eht.c
@@ -0,0 +1,115 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+// Copyright (c) 2020, Nikolay Aleksandrov <nikolay at nvidia.com>
+#include <linux/err.h>
+#include <linux/export.h>
+#include <linux/if_ether.h>
+#include <linux/igmp.h>
+#include <linux/in.h>
+#include <linux/jhash.h>
+#include <linux/kernel.h>
+#include <linux/log2.h>
+#include <linux/netdevice.h>
+#include <linux/netfilter_bridge.h>
+#include <linux/random.h>
+#include <linux/rculist.h>
+#include <linux/skbuff.h>
+#include <linux/slab.h>
+#include <linux/timer.h>
+#include <linux/inetdevice.h>
+#include <linux/mroute.h>
+#include <net/ip.h>
+#include <net/switchdev.h>
+#if IS_ENABLED(CONFIG_IPV6)
+#include <linux/icmpv6.h>
+#include <net/ipv6.h>
+#include <net/mld.h>
+#include <net/ip6_checksum.h>
+#include <net/addrconf.h>
+#endif
+
+#include "br_private.h"
+#include "br_private_mcast_eht.h"
+
+static struct net_bridge_group_eht_host *
+br_multicast_eht_host_lookup(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr)
+{
+	struct rb_node *node = pg->eht_host_tree.rb_node;
+
+	while (node) {
+		struct net_bridge_group_eht_host *this;
+		int result;
+
+		this = rb_entry(node, struct net_bridge_group_eht_host,
+				rb_node);
+		result = memcmp(h_addr, &this->h_addr, sizeof(*h_addr));
+		if (result < 0)
+			node = node->rb_left;
+		else if (result > 0)
+			node = node->rb_right;
+		else
+			return this;
+	}
+
+	return NULL;
+}
+
+static int br_multicast_eht_host_filter_mode(struct net_bridge_port_group *pg,
+					     union net_bridge_eht_addr *h_addr)
+{
+	struct net_bridge_group_eht_host *eht_host;
+
+	eht_host = br_multicast_eht_host_lookup(pg, h_addr);
+	if (!eht_host)
+		return MCAST_INCLUDE;
+
+	return eht_host->filter_mode;
+}
+
+static void __eht_destroy_host(struct net_bridge_group_eht_host *eht_host)
+{
+	WARN_ON(!hlist_empty(&eht_host->set_entries));
+
+	rb_erase(&eht_host->rb_node, &eht_host->pg->eht_host_tree);
+	RB_CLEAR_NODE(&eht_host->rb_node);
+	kfree(eht_host);
+}
+
+static struct net_bridge_group_eht_host *
+__eht_lookup_create_host(struct net_bridge_port_group *pg,
+			 union net_bridge_eht_addr *h_addr,
+			 unsigned char filter_mode)
+{
+	struct rb_node **link = &pg->eht_host_tree.rb_node, *parent = NULL;
+	struct net_bridge_group_eht_host *eht_host;
+
+	while (*link) {
+		struct net_bridge_group_eht_host *this;
+		int result;
+
+		this = rb_entry(*link, struct net_bridge_group_eht_host,
+				rb_node);
+		result = memcmp(h_addr, &this->h_addr, sizeof(*h_addr));
+		parent = *link;
+		if (result < 0)
+			link = &((*link)->rb_left);
+		else if (result > 0)
+			link = &((*link)->rb_right);
+		else
+			return this;
+	}
+
+	eht_host = kzalloc(sizeof(*eht_host), GFP_ATOMIC);
+	if (!eht_host)
+		return NULL;
+
+	memcpy(&eht_host->h_addr, h_addr, sizeof(*h_addr));
+	INIT_HLIST_HEAD(&eht_host->set_entries);
+	eht_host->pg = pg;
+	eht_host->filter_mode = filter_mode;
+
+	rb_link_node(&eht_host->rb_node, parent, link);
+	rb_insert_color(&eht_host->rb_node, &pg->eht_host_tree);
+
+	return eht_host;
+}
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Add EHT source set and set-entry create, delete and lookup functions.
These allow to manipulate source sets which contain their own host sets
with entries which joined that S,G. We're limiting the maximum number of
tracked S,G entries per host to PG_SRC_ENT_LIMIT (currently 32) which is
the current maximum of S,G entries for a group. There's a per-set timer
which will be used to destroy the whole set later.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c         |   2 +
 net/bridge/br_multicast_eht.c     | 321 ++++++++++++++++++++++++++++++
 net/bridge/br_private_mcast_eht.h |   2 +
 3 files changed, 325 insertions(+)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index dc6e879dc840..ac363b104239 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -594,6 +594,7 @@ void br_multicast_del_pg(struct net_bridge_mdb_entry *mp,
 
 	rcu_assign_pointer(*pp, pg->next);
 	hlist_del_init(&pg->mglist);
+	br_multicast_eht_clean_sets(pg);
 	hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
 		br_multicast_del_group_src(ent);
 	br_mdb_notify(br->dev, mp, pg, RTM_DELMDB);
@@ -1174,6 +1175,7 @@ struct net_bridge_port_group *br_multicast_new_port_group(
 	p->filter_mode = filter_mode;
 	p->rt_protocol = rt_protocol;
 	p->eht_host_tree = RB_ROOT;
+	p->eht_set_tree = RB_ROOT;
 	p->mcast_gc.destroy = br_multicast_destroy_port_group;
 	INIT_HLIST_HEAD(&p->src_list);
 
diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index 5cebca45e72c..f4bbf2dc9fc8 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -30,6 +30,15 @@
 #include "br_private.h"
 #include "br_private_mcast_eht.h"
 
+static bool br_multicast_del_eht_set_entry(struct net_bridge_port_group *pg,
+					   union net_bridge_eht_addr *src_addr,
+					   union net_bridge_eht_addr *h_addr);
+static void br_multicast_create_eht_set_entry(struct net_bridge_port_group *pg,
+					      union net_bridge_eht_addr *src_addr,
+					      union net_bridge_eht_addr *h_addr,
+					      int filter_mode,
+					      bool allow_zero_src);
+
 static struct net_bridge_group_eht_host *
 br_multicast_eht_host_lookup(struct net_bridge_port_group *pg,
 			     union net_bridge_eht_addr *h_addr)
@@ -66,6 +75,54 @@ static int br_multicast_eht_host_filter_mode(struct
net_bridge_port_group *pg,
 	return eht_host->filter_mode;
 }
 
+static struct net_bridge_group_eht_set_entry *
+br_multicast_eht_set_entry_lookup(struct net_bridge_group_eht_set *eht_set,
+				  union net_bridge_eht_addr *h_addr)
+{
+	struct rb_node *node = eht_set->entry_tree.rb_node;
+
+	while (node) {
+		struct net_bridge_group_eht_set_entry *this;
+		int result;
+
+	this = rb_entry(node, struct net_bridge_group_eht_set_entry,
+			rb_node);
+	result = memcmp(h_addr, &this->h_addr, sizeof(*h_addr));
+	if (result < 0)
+		node = node->rb_left;
+	else if (result > 0)
+		node = node->rb_right;
+	else
+		return this;
+	}
+
+	return NULL;
+}
+
+static struct net_bridge_group_eht_set *
+br_multicast_eht_set_lookup(struct net_bridge_port_group *pg,
+			    union net_bridge_eht_addr *src_addr)
+{
+	struct rb_node *node = pg->eht_set_tree.rb_node;
+
+	while (node) {
+		struct net_bridge_group_eht_set *this;
+		int result;
+
+		this = rb_entry(node, struct net_bridge_group_eht_set,
+				rb_node);
+		result = memcmp(src_addr, &this->src_addr, sizeof(*src_addr));
+		if (result < 0)
+			node = node->rb_left;
+		else if (result > 0)
+			node = node->rb_right;
+		else
+			return this;
+	}
+
+	return NULL;
+}
+
 static void __eht_destroy_host(struct net_bridge_group_eht_host *eht_host)
 {
 	WARN_ON(!hlist_empty(&eht_host->set_entries));
@@ -75,6 +132,107 @@ static void __eht_destroy_host(struct
net_bridge_group_eht_host *eht_host)
 	kfree(eht_host);
 }
 
+static void br_multicast_destroy_eht_set_entry(struct net_bridge_mcast_gc *gc)
+{
+	struct net_bridge_group_eht_set_entry *set_h;
+
+	set_h = container_of(gc, struct net_bridge_group_eht_set_entry, mcast_gc);
+	WARN_ON(!RB_EMPTY_NODE(&set_h->rb_node));
+
+	del_timer_sync(&set_h->timer);
+	kfree(set_h);
+}
+
+static void br_multicast_destroy_eht_set(struct net_bridge_mcast_gc *gc)
+{
+	struct net_bridge_group_eht_set *eht_set;
+
+	eht_set = container_of(gc, struct net_bridge_group_eht_set, mcast_gc);
+	WARN_ON(!RB_EMPTY_NODE(&eht_set->rb_node));
+	WARN_ON(!RB_EMPTY_ROOT(&eht_set->entry_tree));
+
+	del_timer_sync(&eht_set->timer);
+	kfree(eht_set);
+}
+
+static void __eht_del_set_entry(struct net_bridge_group_eht_set_entry *set_h)
+{
+	struct net_bridge_group_eht_host *eht_host = set_h->h_parent;
+	union net_bridge_eht_addr zero_addr;
+
+	rb_erase(&set_h->rb_node, &set_h->eht_set->entry_tree);
+	RB_CLEAR_NODE(&set_h->rb_node);
+	hlist_del_init(&set_h->host_list);
+	memset(&zero_addr, 0, sizeof(zero_addr));
+	if (memcmp(&set_h->h_addr, &zero_addr, sizeof(zero_addr)))
+		eht_host->num_entries--;
+	hlist_add_head(&set_h->mcast_gc.gc_node,
&set_h->br->mcast_gc_list);
+	queue_work(system_long_wq, &set_h->br->mcast_gc_work);
+
+	if (hlist_empty(&eht_host->set_entries))
+		__eht_destroy_host(eht_host);
+}
+
+static void br_multicast_del_eht_set(struct net_bridge_group_eht_set *eht_set)
+{
+	struct net_bridge_group_eht_set_entry *set_h;
+	struct rb_node *node;
+
+	while ((node = rb_first(&eht_set->entry_tree))) {
+		set_h = rb_entry(node, struct net_bridge_group_eht_set_entry,
+				 rb_node);
+		__eht_del_set_entry(set_h);
+	}
+
+	rb_erase(&eht_set->rb_node, &eht_set->pg->eht_set_tree);
+	RB_CLEAR_NODE(&eht_set->rb_node);
+	hlist_add_head(&eht_set->mcast_gc.gc_node,
&eht_set->br->mcast_gc_list);
+	queue_work(system_long_wq, &eht_set->br->mcast_gc_work);
+}
+
+void br_multicast_eht_clean_sets(struct net_bridge_port_group *pg)
+{
+	struct net_bridge_group_eht_set *eht_set;
+	struct rb_node *node;
+
+	while ((node = rb_first(&pg->eht_set_tree))) {
+		eht_set = rb_entry(node, struct net_bridge_group_eht_set,
+				   rb_node);
+		br_multicast_del_eht_set(eht_set);
+	}
+}
+
+static void br_multicast_eht_set_entry_expired(struct timer_list *t)
+{
+	struct net_bridge_group_eht_set_entry *set_h = from_timer(set_h, t, timer);
+	struct net_bridge *br = set_h->br;
+
+	spin_lock(&br->multicast_lock);
+	if (RB_EMPTY_NODE(&set_h->rb_node) ||
timer_pending(&set_h->timer))
+		goto out;
+
+	br_multicast_del_eht_set_entry(set_h->eht_set->pg,
+				       &set_h->eht_set->src_addr,
+				       &set_h->h_addr);
+out:
+	spin_unlock(&br->multicast_lock);
+}
+
+static void br_multicast_eht_set_expired(struct timer_list *t)
+{
+	struct net_bridge_group_eht_set *eht_set = from_timer(eht_set, t,
+							      timer);
+	struct net_bridge *br = eht_set->br;
+
+	spin_lock(&br->multicast_lock);
+	if (RB_EMPTY_NODE(&eht_set->rb_node) ||
timer_pending(&eht_set->timer))
+		goto out;
+
+	br_multicast_del_eht_set(eht_set);
+out:
+	spin_unlock(&br->multicast_lock);
+}
+
 static struct net_bridge_group_eht_host *
 __eht_lookup_create_host(struct net_bridge_port_group *pg,
 			 union net_bridge_eht_addr *h_addr,
@@ -113,3 +271,166 @@ __eht_lookup_create_host(struct net_bridge_port_group *pg,
 
 	return eht_host;
 }
+
+static struct net_bridge_group_eht_set_entry *
+__eht_lookup_create_set_entry(struct net_bridge *br,
+			      struct net_bridge_group_eht_set *eht_set,
+			      struct net_bridge_group_eht_host *eht_host,
+			      bool allow_zero_src)
+{
+	struct rb_node **link = &eht_set->entry_tree.rb_node, *parent = NULL;
+	struct net_bridge_group_eht_set_entry *set_h;
+
+	while (*link) {
+		struct net_bridge_group_eht_set_entry *this;
+		int result;
+
+		this = rb_entry(*link, struct net_bridge_group_eht_set_entry,
+				rb_node);
+		result = memcmp(&eht_host->h_addr, &this->h_addr,
+				sizeof(union net_bridge_eht_addr));
+		parent = *link;
+		if (result < 0)
+			link = &((*link)->rb_left);
+		else if (result > 0)
+			link = &((*link)->rb_right);
+		else
+			return this;
+	}
+
+	/* always allow auto-created zero entry */
+	if (!allow_zero_src && eht_host->num_entries >=
PG_SRC_ENT_LIMIT)
+		return NULL;
+
+	set_h = kzalloc(sizeof(*set_h), GFP_ATOMIC);
+	if (!set_h)
+		return NULL;
+
+	memcpy(&set_h->h_addr, &eht_host->h_addr,
+	       sizeof(union net_bridge_eht_addr));
+	set_h->mcast_gc.destroy = br_multicast_destroy_eht_set_entry;
+	set_h->eht_set = eht_set;
+	set_h->h_parent = eht_host;
+	set_h->br = br;
+	timer_setup(&set_h->timer, br_multicast_eht_set_entry_expired, 0);
+
+	hlist_add_head(&set_h->host_list, &eht_host->set_entries);
+	rb_link_node(&set_h->rb_node, parent, link);
+	rb_insert_color(&set_h->rb_node, &eht_set->entry_tree);
+	/* we must not count the auto-created zero entry otherwise we won't be
+	 * able to track the full list of PG_SRC_ENT_LIMIT entries
+	 */
+	if (!allow_zero_src)
+		eht_host->num_entries++;
+
+	return set_h;
+}
+
+static struct net_bridge_group_eht_set *
+__eht_lookup_create_set(struct net_bridge_port_group *pg,
+			union net_bridge_eht_addr *src_addr)
+{
+	struct rb_node **link = &pg->eht_set_tree.rb_node, *parent = NULL;
+	struct net_bridge_group_eht_set *eht_set;
+
+	while (*link) {
+		struct net_bridge_group_eht_set *this;
+		int result;
+
+		this = rb_entry(*link, struct net_bridge_group_eht_set,
+				rb_node);
+		result = memcmp(src_addr, &this->src_addr, sizeof(*src_addr));
+		parent = *link;
+		if (result < 0)
+			link = &((*link)->rb_left);
+		else if (result > 0)
+			link = &((*link)->rb_right);
+		else
+			return this;
+	}
+
+	eht_set = kzalloc(sizeof(*eht_set), GFP_ATOMIC);
+	if (!eht_set)
+		return NULL;
+
+	memcpy(&eht_set->src_addr, src_addr, sizeof(*src_addr));
+	eht_set->mcast_gc.destroy = br_multicast_destroy_eht_set;
+	eht_set->pg = pg;
+	eht_set->br = pg->key.port->br;
+	eht_set->entry_tree = RB_ROOT;
+	timer_setup(&eht_set->timer, br_multicast_eht_set_expired, 0);
+
+	rb_link_node(&eht_set->rb_node, parent, link);
+	rb_insert_color(&eht_set->rb_node, &pg->eht_set_tree);
+
+	return eht_set;
+}
+
+static void br_multicast_create_eht_set_entry(struct net_bridge_port_group *pg,
+					      union net_bridge_eht_addr *src_addr,
+					      union net_bridge_eht_addr *h_addr,
+					      int filter_mode,
+					      bool allow_zero_src)
+{
+	struct net_bridge_group_eht_set_entry *set_h;
+	struct net_bridge_group_eht_host *eht_host;
+	struct net_bridge *br = pg->key.port->br;
+	struct net_bridge_group_eht_set *eht_set;
+	union net_bridge_eht_addr zero_addr;
+
+	memset(&zero_addr, 0, sizeof(zero_addr));
+	if (!allow_zero_src && !memcmp(src_addr, &zero_addr,
sizeof(zero_addr)))
+		return;
+
+	eht_set = __eht_lookup_create_set(pg, src_addr);
+	if (!eht_set)
+		return;
+
+	eht_host = __eht_lookup_create_host(pg, h_addr, filter_mode);
+	if (!eht_host)
+		goto fail_host;
+
+	set_h = __eht_lookup_create_set_entry(br, eht_set, eht_host,
+					      allow_zero_src);
+	if (!set_h)
+		goto fail_set_entry;
+
+	mod_timer(&set_h->timer, jiffies + br_multicast_gmi(br));
+	mod_timer(&eht_set->timer, jiffies + br_multicast_gmi(br));
+
+	return;
+
+fail_set_entry:
+	if (hlist_empty(&eht_host->set_entries))
+		__eht_destroy_host(eht_host);
+fail_host:
+	if (RB_EMPTY_ROOT(&eht_set->entry_tree))
+		br_multicast_del_eht_set(eht_set);
+}
+
+static bool br_multicast_del_eht_set_entry(struct net_bridge_port_group *pg,
+					   union net_bridge_eht_addr *src_addr,
+					   union net_bridge_eht_addr *h_addr)
+{
+	struct net_bridge_group_eht_set_entry *set_h;
+	struct net_bridge_group_eht_set *eht_set;
+	bool set_deleted = false;
+
+	eht_set = br_multicast_eht_set_lookup(pg, src_addr);
+	if (!eht_set)
+		goto out;
+
+	set_h = br_multicast_eht_set_entry_lookup(eht_set, h_addr);
+	if (!set_h)
+		goto out;
+
+	__eht_del_set_entry(set_h);
+
+	if (RB_EMPTY_ROOT(&eht_set->entry_tree)) {
+		br_multicast_del_eht_set(eht_set);
+		set_deleted = true;
+	}
+
+out:
+	return set_deleted;
+}
diff --git a/net/bridge/br_private_mcast_eht.h
b/net/bridge/br_private_mcast_eht.h
index 0c9c4267969d..bba507c9acb0 100644
--- a/net/bridge/br_private_mcast_eht.h
+++ b/net/bridge/br_private_mcast_eht.h
@@ -47,4 +47,6 @@ struct net_bridge_group_eht_set {
 	struct net_bridge_mcast_gc	mcast_gc;
 };
 
+void br_multicast_eht_clean_sets(struct net_bridge_port_group *pg);
+
 #endif /* _BR_PRIVATE_MCAST_EHT_H_ */
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Now that we can delete set entries, we can use that to remove EHT hosts.
Since the group's host set entries exist only when there are related
source set entries we just have to flush all source set entries
joined by the host set entry and it will be automatically removed.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast_eht.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index f4bbf2dc9fc8..409fced7eae2 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -434,3 +434,20 @@ static bool br_multicast_del_eht_set_entry(struct
net_bridge_port_group *pg,
 out:
 	return set_deleted;
 }
+
+static void br_multicast_del_eht_host(struct net_bridge_port_group *pg,
+				      union net_bridge_eht_addr *h_addr)
+{
+	struct net_bridge_group_eht_set_entry *set_h;
+	struct net_bridge_group_eht_host *eht_host;
+	struct hlist_node *tmp;
+
+	eht_host = br_multicast_eht_host_lookup(pg, h_addr);
+	if (!eht_host)
+		return;
+
+	hlist_for_each_entry_safe(set_h, tmp, &eht_host->set_entries,
host_list)
+		br_multicast_del_eht_set_entry(set_h->eht_set->pg,
+					       &set_h->eht_set->src_addr,
+					       &set_h->h_addr);
+}
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Add support for IGMPv3/MLDv2 allow/block EHT handling. Similar to how
the reports are processed we have 2 cases when the group is in include
or exclude mode, these are processed as follows:
 - group include
  - allow: create missing entries
  - block: remove existing matching entries and remove the corresponding
    S,G entries if there are no more set host entries, then possibly
    delete the whole group if there are no more S,G entries

 - group exclude
  - allow
    - host include: create missing entries
    - host exclude: remove existing matching entries and remove the
      corresponding S,G entries if there are no more set host entries
  - block
    - host include: remove existing matching entries and remove the
      corresponding S,G entries if there are no more set host entries,
      then possibly delete the whole group if there are no more S,G entries
    - host exclude: create missing entries

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c         |  43 +++--
 net/bridge/br_multicast_eht.c     | 252 ++++++++++++++++++++++++++++++
 net/bridge/br_private.h           |   3 +
 net/bridge/br_private_mcast_eht.h |   6 +
 4 files changed, 290 insertions(+), 14 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index ac363b104239..3b8c5d1d0c55 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -560,7 +560,7 @@ static void br_multicast_destroy_group_src(struct
net_bridge_mcast_gc *gc)
 	kfree_rcu(src, rcu);
 }
 
-static void br_multicast_del_group_src(struct net_bridge_group_src *src)
+void br_multicast_del_group_src(struct net_bridge_group_src *src)
 {
 	struct net_bridge *br = src->pg->key.port->br;
 
@@ -1092,7 +1092,7 @@ static void br_multicast_group_src_expired(struct
timer_list *t)
 	spin_unlock(&br->multicast_lock);
 }
 
-static struct net_bridge_group_src *
+struct net_bridge_group_src *
 br_multicast_find_group_src(struct net_bridge_port_group *pg, struct br_ip *ip)
 {
 	struct net_bridge_group_src *ent;
@@ -1804,7 +1804,8 @@ static void __grp_send_query_and_rexmit(struct
net_bridge_port_group *pg)
  * EXCLUDE (X,Y)  ALLOW (A)     EXCLUDE (X+A,Y-A)        (A)=GMI
  */
 static bool br_multicast_isinc_allow(struct net_bridge_port_group *pg, void
*h_addr,
-				     void *srcs, u32 nsrcs, size_t addr_size)
+				     void *srcs, u32 nsrcs, size_t addr_size,
+				     int grec_type)
 {
 	struct net_bridge *br = pg->key.port->br;
 	struct net_bridge_group_src *ent;
@@ -1828,6 +1829,9 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg, void *h_a
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	return changed;
 }
 
@@ -2140,7 +2144,7 @@ static bool br_multicast_toex(struct net_bridge_port_group
*pg, void *h_addr,
  * INCLUDE (A)    BLOCK (B)     INCLUDE (A)              Send Q(G,A*B)
  */
 static bool __grp_src_block_incl(struct net_bridge_port_group *pg, void
*h_addr,
-				 void *srcs, u32 nsrcs, size_t addr_size)
+				 void *srcs, u32 nsrcs, size_t addr_size, int grec_type)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2161,6 +2165,9 @@ static bool __grp_src_block_incl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
 
@@ -2180,7 +2187,7 @@ static bool __grp_src_block_incl(struct
net_bridge_port_group *pg, void *h_addr,
  *                                                       Send Q(G,A-Y)
  */
 static bool __grp_src_block_excl(struct net_bridge_port_group *pg, void
*h_addr,
-				 void *srcs, u32 nsrcs, size_t addr_size)
+				 void *srcs, u32 nsrcs, size_t addr_size, int grec_type)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2208,6 +2215,9 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
 
@@ -2215,16 +2225,18 @@ static bool __grp_src_block_excl(struct
net_bridge_port_group *pg, void *h_addr,
 }
 
 static bool br_multicast_block(struct net_bridge_port_group *pg, void *h_addr,
-			       void *srcs, u32 nsrcs, size_t addr_size)
+			       void *srcs, u32 nsrcs, size_t addr_size, int grec_type)
 {
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		changed = __grp_src_block_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_block_incl(pg, h_addr, srcs, nsrcs, addr_size,
+					       grec_type);
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_block_excl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_block_excl(pg, h_addr, srcs, nsrcs, addr_size,
+					       grec_type);
 		break;
 	}
 
@@ -2327,11 +2339,11 @@ static int br_ip4_multicast_igmp3_report(struct
net_bridge *br,
 		switch (type) {
 		case IGMPV3_ALLOW_NEW_SOURCES:
 			changed = br_multicast_isinc_allow(pg, h_addr, grec->grec_src,
-							   nsrcs, sizeof(__be32));
+							   nsrcs, sizeof(__be32), type);
 			break;
 		case IGMPV3_MODE_IS_INCLUDE:
 			changed = br_multicast_isinc_allow(pg, h_addr, grec->grec_src,
-							   nsrcs, sizeof(__be32));
+							   nsrcs, sizeof(__be32), type);
 			break;
 		case IGMPV3_MODE_IS_EXCLUDE:
 			changed = br_multicast_isexc(pg, h_addr, grec->grec_src,
@@ -2347,7 +2359,7 @@ static int br_ip4_multicast_igmp3_report(struct net_bridge
*br,
 			break;
 		case IGMPV3_BLOCK_OLD_SOURCES:
 			changed = br_multicast_block(pg, h_addr, grec->grec_src,
-						     nsrcs, sizeof(__be32));
+						     nsrcs, sizeof(__be32), type);
 			break;
 		}
 		if (changed)
@@ -2455,12 +2467,14 @@ static int br_ip6_multicast_mld2_report(struct
net_bridge *br,
 		case MLD2_ALLOW_NEW_SOURCES:
 			changed = br_multicast_isinc_allow(pg, h_addr,
 							   grec->grec_src, nsrcs,
-							   sizeof(struct in6_addr));
+							   sizeof(struct in6_addr),
+							   grec->grec_type);
 			break;
 		case MLD2_MODE_IS_INCLUDE:
 			changed = br_multicast_isinc_allow(pg, h_addr,
 							   grec->grec_src, nsrcs,
-							   sizeof(struct in6_addr));
+							   sizeof(struct in6_addr),
+							   grec->grec_type);
 			break;
 		case MLD2_MODE_IS_EXCLUDE:
 			changed = br_multicast_isexc(pg, h_addr,
@@ -2480,7 +2494,8 @@ static int br_ip6_multicast_mld2_report(struct net_bridge
*br,
 		case MLD2_BLOCK_OLD_SOURCES:
 			changed = br_multicast_block(pg, h_addr,
 						     grec->grec_src, nsrcs,
-						     sizeof(struct in6_addr));
+						     sizeof(struct in6_addr),
+						     grec->grec_type);
 			break;
 		}
 		if (changed)
diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index 409fced7eae2..43f60388df16 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -366,6 +366,21 @@ __eht_lookup_create_set(struct net_bridge_port_group *pg,
 	return eht_set;
 }
 
+static void br_multicast_ip_src_to_eht_addr(const struct br_ip *src,
+					    union net_bridge_eht_addr *dest)
+{
+	switch (src->proto) {
+	case htons(ETH_P_IP):
+		dest->ip4 = src->src.ip4;
+		break;
+#if IS_ENABLED(CONFIG_IPV6)
+	case htons(ETH_P_IPV6):
+		memcpy(&dest->ip6, &src->src.ip6, sizeof(struct in6_addr));
+		break;
+#endif
+	}
+}
+
 static void br_multicast_create_eht_set_entry(struct net_bridge_port_group *pg,
 					      union net_bridge_eht_addr *src_addr,
 					      union net_bridge_eht_addr *h_addr,
@@ -451,3 +466,240 @@ static void br_multicast_del_eht_host(struct
net_bridge_port_group *pg,
 					       &set_h->eht_set->src_addr,
 					       &set_h->h_addr);
 }
+
+static void __eht_allow_incl(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     size_t addr_size)
+{
+	union net_bridge_eht_addr eht_src_addr;
+	u32 src_idx;
+
+	memset(&eht_src_addr, 0, sizeof(eht_src_addr));
+	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
+		memcpy(&eht_src_addr, srcs + (src_idx * addr_size), addr_size);
+		br_multicast_create_eht_set_entry(pg, &eht_src_addr, h_addr,
+						  MCAST_INCLUDE,
+						  false);
+	}
+}
+
+static bool __eht_allow_excl(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     size_t addr_size)
+{
+	bool changed = false, host_excl = false;
+	union net_bridge_eht_addr eht_src_addr;
+	struct net_bridge_group_src *src_ent;
+	struct br_ip src_ip;
+	u32 src_idx;
+
+	host_excl = !!(br_multicast_eht_host_filter_mode(pg, h_addr) ==
MCAST_EXCLUDE);
+	memset(&eht_src_addr, 0, sizeof(eht_src_addr));
+	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
+		memcpy(&eht_src_addr, srcs + (src_idx * addr_size), addr_size);
+		if (!host_excl) {
+			br_multicast_create_eht_set_entry(pg, &eht_src_addr, h_addr,
+							  MCAST_INCLUDE,
+							  false);
+		} else {
+			if (!br_multicast_del_eht_set_entry(pg, &eht_src_addr,
+							    h_addr))
+				continue;
+			memcpy(&src_ip, srcs + (src_idx * addr_size), addr_size);
+			src_ent = br_multicast_find_group_src(pg, &src_ip);
+			if (!src_ent)
+				continue;
+			br_multicast_del_group_src(src_ent);
+			changed = true;
+		}
+	}
+
+	return changed;
+}
+
+static bool br_multicast_eht_allow(struct net_bridge_port_group *pg,
+				   union net_bridge_eht_addr *h_addr,
+				   void *srcs,
+				   u32 nsrcs,
+				   size_t addr_size)
+{
+	bool changed = false;
+
+	switch (br_multicast_eht_host_filter_mode(pg, h_addr)) {
+	case MCAST_INCLUDE:
+		__eht_allow_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		break;
+	case MCAST_EXCLUDE:
+		changed = __eht_allow_excl(pg, h_addr, srcs, nsrcs, addr_size);
+		break;
+	}
+
+	return changed;
+}
+
+static bool __eht_block_incl(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     size_t addr_size)
+{
+	union net_bridge_eht_addr eht_src_addr;
+	struct net_bridge_group_src *src_ent;
+	bool changed = false;
+	struct br_ip src_ip;
+	u32 src_idx;
+
+	memset(&eht_src_addr, 0, sizeof(eht_src_addr));
+	memset(&src_ip, 0, sizeof(src_ip));
+	src_ip.proto = pg->key.addr.proto;
+	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
+		memcpy(&eht_src_addr, srcs + (src_idx * addr_size), addr_size);
+		if (!br_multicast_del_eht_set_entry(pg, &eht_src_addr, h_addr))
+			continue;
+		memcpy(&src_ip, srcs + (src_idx * addr_size), addr_size);
+		src_ent = br_multicast_find_group_src(pg, &src_ip);
+		if (!src_ent)
+			continue;
+		br_multicast_del_group_src(src_ent);
+		changed = true;
+	}
+
+	return changed;
+}
+
+static bool __eht_block_excl(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     size_t addr_size)
+{
+	bool changed = false, host_excl = false;
+	union net_bridge_eht_addr eht_src_addr;
+	struct net_bridge_group_src *src_ent;
+	struct br_ip src_ip;
+	u32 src_idx;
+
+	host_excl = !!(br_multicast_eht_host_filter_mode(pg, h_addr) ==
MCAST_EXCLUDE);
+	memset(&eht_src_addr, 0, sizeof(eht_src_addr));
+	memset(&src_ip, 0, sizeof(src_ip));
+	src_ip.proto = pg->key.addr.proto;
+	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
+		memcpy(&eht_src_addr, srcs + (src_idx * addr_size), addr_size);
+		if (host_excl) {
+			br_multicast_create_eht_set_entry(pg, &eht_src_addr, h_addr,
+							  MCAST_EXCLUDE,
+							  false);
+		} else {
+			if (!br_multicast_del_eht_set_entry(pg, &eht_src_addr,
+							    h_addr))
+				continue;
+			memcpy(&src_ip, srcs + (src_idx * addr_size), addr_size);
+			src_ent = br_multicast_find_group_src(pg, &src_ip);
+			if (!src_ent)
+				continue;
+			br_multicast_del_group_src(src_ent);
+			changed = true;
+		}
+	}
+
+	return changed;
+}
+
+static bool br_multicast_eht_block(struct net_bridge_port_group *pg,
+				   union net_bridge_eht_addr *h_addr,
+				   void *srcs,
+				   u32 nsrcs,
+				   size_t addr_size)
+{
+	bool changed = false;
+
+	switch (br_multicast_eht_host_filter_mode(pg, h_addr)) {
+	case MCAST_INCLUDE:
+		changed = __eht_block_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		break;
+	case MCAST_EXCLUDE:
+		changed = __eht_block_excl(pg, h_addr, srcs, nsrcs, addr_size);
+		break;
+	}
+
+	return changed;
+}
+
+static bool __eht_ip4_handle(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     int grec_type)
+{
+	bool changed = false;
+
+	switch (grec_type) {
+	case IGMPV3_ALLOW_NEW_SOURCES:
+		br_multicast_eht_allow(pg, h_addr, srcs, nsrcs, sizeof(__be32));
+		break;
+	case IGMPV3_BLOCK_OLD_SOURCES:
+		changed = br_multicast_eht_block(pg, h_addr, srcs, nsrcs,
+						 sizeof(__be32));
+		break;
+	}
+
+	return changed;
+}
+
+#if IS_ENABLED(CONFIG_IPV6)
+static bool __eht_ip6_handle(struct net_bridge_port_group *pg,
+			     union net_bridge_eht_addr *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     int grec_type)
+{
+	bool changed = false;
+
+	switch (grec_type) {
+	case MLD2_ALLOW_NEW_SOURCES:
+		br_multicast_eht_allow(pg, h_addr, srcs, nsrcs,
+				       sizeof(struct in6_addr));
+		break;
+	case MLD2_BLOCK_OLD_SOURCES:
+		changed = br_multicast_eht_block(pg, h_addr, srcs, nsrcs,
+						 sizeof(struct in6_addr));
+		break;
+	}
+
+	return changed;
+}
+#endif
+
+/* true means an entry was deleted */
+bool br_multicast_eht_handle(struct net_bridge_port_group *pg,
+			     void *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     size_t addr_size,
+			     int grec_type)
+{
+	bool eht_enabled = !!(pg->key.port->flags &
BR_MULTICAST_FAST_LEAVE);
+	union net_bridge_eht_addr eht_host_addr;
+	bool changed = false;
+
+	if (!eht_enabled)
+		goto out;
+
+	memset(&eht_host_addr, 0, sizeof(eht_host_addr));
+	memcpy(&eht_host_addr, h_addr, addr_size);
+	if (addr_size == sizeof(__be32))
+		changed = __eht_ip4_handle(pg, &eht_host_addr, srcs, nsrcs,
+					   grec_type);
+#if IS_ENABLED(CONFIG_IPV6)
+	else
+		changed = __eht_ip6_handle(pg, &eht_host_addr, srcs, nsrcs,
+					   grec_type);
+#endif
+
+out:
+	return changed;
+}
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 0bf4c544a5da..cad967690e9f 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -848,6 +848,9 @@ void br_multicast_star_g_handle_mode(struct
net_bridge_port_group *pg,
 				     u8 filter_mode);
 void br_multicast_sg_add_exclude_ports(struct net_bridge_mdb_entry *star_mp,
 				       struct net_bridge_port_group *sg);
+struct net_bridge_group_src *
+br_multicast_find_group_src(struct net_bridge_port_group *pg, struct br_ip
*ip);
+void br_multicast_del_group_src(struct net_bridge_group_src *src);
 
 static inline bool br_group_is_l2(const struct br_ip *group)
 {
diff --git a/net/bridge/br_private_mcast_eht.h
b/net/bridge/br_private_mcast_eht.h
index bba507c9acb0..92933822301d 100644
--- a/net/bridge/br_private_mcast_eht.h
+++ b/net/bridge/br_private_mcast_eht.h
@@ -48,5 +48,11 @@ struct net_bridge_group_eht_set {
 };
 
 void br_multicast_eht_clean_sets(struct net_bridge_port_group *pg);
+bool br_multicast_eht_handle(struct net_bridge_port_group *pg,
+			     void *h_addr,
+			     void *srcs,
+			     u32 nsrcs,
+			     size_t addr_size,
+			     int grec_type);
 
 #endif /* _BR_PRIVATE_MCAST_EHT_H_ */
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Add support for IGMPv3/MLDv2 include and exclude EHT handling. Similar to
how the reports are processed we have 2 cases when the group is in include
or exclude mode, these are processed as follows:
 - group include
  - is_include: create missing entries
  - to_include: flush existing entries and create a new set from the
    report, obviously if the src set is empty then we delete the group

 - group exclude
  - is_exclude: create missing entries
  - to_exclude: flush existing entries and create a new set from the
    report, any empty source set entries are removed

If the group is in a different mode then we just flush all entries reported
by the host and we create a new set with the new mode entries created from
the report. If the report is include type, the source list is empty and
the group has empty sources' set then we remove it. Any source set entries
which are empty are removed as well. If the group is in exclude mode it
can exist without any S,G entries (allowing for all traffic to pass).

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c         |  84 ++++++++++++++++++-------
 net/bridge/br_multicast_eht.c     | 100 +++++++++++++++++++++++++++++-
 net/bridge/br_private_mcast_eht.h |   7 +++
 3 files changed, 168 insertions(+), 23 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 3b8c5d1d0c55..9cfc004312ab 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -1841,7 +1841,8 @@ static bool br_multicast_isinc_allow(struct
net_bridge_port_group *pg, void *h_a
  *                                                       Group Timer=GMI
  */
 static void __grp_src_isexc_incl(struct net_bridge_port_group *pg, void
*h_addr,
-				 void *srcs, u32 nsrcs, size_t addr_size)
+				 void *srcs, u32 nsrcs, size_t addr_size,
+				 int grec_type)
 {
 	struct net_bridge_group_src *ent;
 	struct br_ip src_ip;
@@ -1863,6 +1864,8 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg, void *h_addr,
 			br_multicast_fwd_src_handle(ent);
 	}
 
+	br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type);
+
 	__grp_src_delete_marked(pg);
 }
 
@@ -1873,7 +1876,8 @@ static void __grp_src_isexc_incl(struct
net_bridge_port_group *pg, void *h_addr,
  *                                                       Group Timer=GMI
  */
 static bool __grp_src_isexc_excl(struct net_bridge_port_group *pg, void
*h_addr,
-				 void *srcs, u32 nsrcs, size_t addr_size)
+				 void *srcs, u32 nsrcs, size_t addr_size,
+				 int grec_type)
 {
 	struct net_bridge *br = pg->key.port->br;
 	struct net_bridge_group_src *ent;
@@ -1902,6 +1906,9 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	if (__grp_src_delete_marked(pg))
 		changed = true;
 
@@ -1909,19 +1916,22 @@ static bool __grp_src_isexc_excl(struct
net_bridge_port_group *pg, void *h_addr,
 }
 
 static bool br_multicast_isexc(struct net_bridge_port_group *pg, void *h_addr,
-			       void *srcs, u32 nsrcs, size_t addr_size)
+			       void *srcs, u32 nsrcs, size_t addr_size,
+			       int grec_type)
 {
 	struct net_bridge *br = pg->key.port->br;
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_isexc_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		__grp_src_isexc_incl(pg, h_addr, srcs, nsrcs, addr_size,
+				     grec_type);
 		br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
 		changed = true;
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_isexc_excl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_isexc_excl(pg, h_addr, srcs, nsrcs, addr_size,
+					       grec_type);
 		break;
 	}
 
@@ -1936,7 +1946,8 @@ static bool br_multicast_isexc(struct
net_bridge_port_group *pg, void *h_addr,
  *                                                       Send Q(G,A-B)
  */
 static bool __grp_src_toin_incl(struct net_bridge_port_group *pg, void *h_addr,
-				void *srcs, u32 nsrcs, size_t addr_size)
+				void *srcs, u32 nsrcs, size_t addr_size,
+				int grec_type)
 {
 	struct net_bridge *br = pg->key.port->br;
 	u32 src_idx, to_send = pg->src_ents;
@@ -1965,6 +1976,9 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg, void *h_addr,
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
 
@@ -1977,7 +1991,8 @@ static bool __grp_src_toin_incl(struct
net_bridge_port_group *pg, void *h_addr,
  *                                                       Send Q(G)
  */
 static bool __grp_src_toin_excl(struct net_bridge_port_group *pg, void *h_addr,
-				void *srcs, u32 nsrcs, size_t addr_size)
+				void *srcs, u32 nsrcs, size_t addr_size,
+				int grec_type)
 {
 	struct net_bridge *br = pg->key.port->br;
 	u32 src_idx, to_send = pg->src_ents;
@@ -2009,6 +2024,9 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg, void *h_addr,
 			__grp_src_mod_timer(ent, now + br_multicast_gmi(br));
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
 
@@ -2018,19 +2036,30 @@ static bool __grp_src_toin_excl(struct
net_bridge_port_group *pg, void *h_addr,
 }
 
 static bool br_multicast_toin(struct net_bridge_port_group *pg, void *h_addr,
-			      void *srcs, u32 nsrcs, size_t addr_size)
+			      void *srcs, u32 nsrcs, size_t addr_size,
+			      int grec_type)
 {
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		changed = __grp_src_toin_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_toin_incl(pg, h_addr, srcs, nsrcs, addr_size,
+					      grec_type);
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_toin_excl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_toin_excl(pg, h_addr, srcs, nsrcs, addr_size,
+					      grec_type);
 		break;
 	}
 
+	if (br_multicast_eht_should_del_pg(pg)) {
+		br_multicast_find_del_pg(pg->key.port->br, pg);
+		/* a notification has already been sent and we shouldn't
+		 * access pg after the delete so we have to return false
+		 */
+		changed = false;
+	}
+
 	return changed;
 }
 
@@ -2041,7 +2070,8 @@ static bool br_multicast_toin(struct net_bridge_port_group
*pg, void *h_addr,
  *                                                       Group Timer=GMI
  */
 static void __grp_src_toex_incl(struct net_bridge_port_group *pg, void *h_addr,
-				void *srcs, u32 nsrcs, size_t addr_size)
+				void *srcs, u32 nsrcs, size_t addr_size,
+				int grec_type)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2066,6 +2096,8 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg, void *h_addr,
 			br_multicast_fwd_src_handle(ent);
 	}
 
+	br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type);
+
 	__grp_src_delete_marked(pg);
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
@@ -2079,7 +2111,8 @@ static void __grp_src_toex_incl(struct
net_bridge_port_group *pg, void *h_addr,
  *                                                       Group Timer=GMI
  */
 static bool __grp_src_toex_excl(struct net_bridge_port_group *pg, void *h_addr,
-				void *srcs, u32 nsrcs, size_t addr_size)
+				void *srcs, u32 nsrcs, size_t addr_size,
+				int grec_type)
 {
 	struct net_bridge_group_src *ent;
 	u32 src_idx, to_send = 0;
@@ -2109,6 +2142,9 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg, void *h_addr,
 		}
 	}
 
+	if (br_multicast_eht_handle(pg, h_addr, srcs, nsrcs, addr_size, grec_type))
+		changed = true;
+
 	if (__grp_src_delete_marked(pg))
 		changed = true;
 	if (to_send)
@@ -2118,19 +2154,22 @@ static bool __grp_src_toex_excl(struct
net_bridge_port_group *pg, void *h_addr,
 }
 
 static bool br_multicast_toex(struct net_bridge_port_group *pg, void *h_addr,
-			      void *srcs, u32 nsrcs, size_t addr_size)
+			      void *srcs, u32 nsrcs, size_t addr_size,
+			      int grec_type)
 {
 	struct net_bridge *br = pg->key.port->br;
 	bool changed = false;
 
 	switch (pg->filter_mode) {
 	case MCAST_INCLUDE:
-		__grp_src_toex_incl(pg, h_addr, srcs, nsrcs, addr_size);
+		__grp_src_toex_incl(pg, h_addr, srcs, nsrcs, addr_size,
+				    grec_type);
 		br_multicast_star_g_handle_mode(pg, MCAST_EXCLUDE);
 		changed = true;
 		break;
 	case MCAST_EXCLUDE:
-		changed = __grp_src_toex_excl(pg, h_addr, srcs, nsrcs, addr_size);
+		changed = __grp_src_toex_excl(pg, h_addr, srcs, nsrcs, addr_size,
+					      grec_type);
 		break;
 	}
 
@@ -2347,15 +2386,15 @@ static int br_ip4_multicast_igmp3_report(struct
net_bridge *br,
 			break;
 		case IGMPV3_MODE_IS_EXCLUDE:
 			changed = br_multicast_isexc(pg, h_addr, grec->grec_src,
-						     nsrcs, sizeof(__be32));
+						     nsrcs, sizeof(__be32), type);
 			break;
 		case IGMPV3_CHANGE_TO_INCLUDE:
 			changed = br_multicast_toin(pg, h_addr, grec->grec_src,
-						    nsrcs, sizeof(__be32));
+						    nsrcs, sizeof(__be32), type);
 			break;
 		case IGMPV3_CHANGE_TO_EXCLUDE:
 			changed = br_multicast_toex(pg, h_addr, grec->grec_src,
-						    nsrcs, sizeof(__be32));
+						    nsrcs, sizeof(__be32), type);
 			break;
 		case IGMPV3_BLOCK_OLD_SOURCES:
 			changed = br_multicast_block(pg, h_addr, grec->grec_src,
@@ -2479,17 +2518,20 @@ static int br_ip6_multicast_mld2_report(struct
net_bridge *br,
 		case MLD2_MODE_IS_EXCLUDE:
 			changed = br_multicast_isexc(pg, h_addr,
 						     grec->grec_src, nsrcs,
-						     sizeof(struct in6_addr));
+						     sizeof(struct in6_addr),
+						     grec->grec_type);
 			break;
 		case MLD2_CHANGE_TO_INCLUDE:
 			changed = br_multicast_toin(pg, h_addr,
 						    grec->grec_src, nsrcs,
-						    sizeof(struct in6_addr));
+						    sizeof(struct in6_addr),
+						    grec->grec_type);
 			break;
 		case MLD2_CHANGE_TO_EXCLUDE:
 			changed = br_multicast_toex(pg, h_addr,
 						    grec->grec_src, nsrcs,
-						    sizeof(struct in6_addr));
+						    sizeof(struct in6_addr),
+						    grec->grec_type);
 			break;
 		case MLD2_BLOCK_OLD_SOURCES:
 			changed = br_multicast_block(pg, h_addr,
diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index 43f60388df16..861ae63f4a1c 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -629,13 +629,79 @@ static bool br_multicast_eht_block(struct
net_bridge_port_group *pg,
 	return changed;
 }
 
+/* flush_entries is true when changing mode */
+static bool __eht_inc_exc(struct net_bridge_port_group *pg,
+			  union net_bridge_eht_addr *h_addr,
+			  void *srcs,
+			  u32 nsrcs,
+			  size_t addr_size,
+			  unsigned char filter_mode,
+			  bool to_report)
+{
+	bool changed = false, flush_entries = to_report;
+	union net_bridge_eht_addr eht_src_addr;
+	u32 src_idx;
+
+	if (br_multicast_eht_host_filter_mode(pg, h_addr) != filter_mode)
+		flush_entries = true;
+
+	memset(&eht_src_addr, 0, sizeof(eht_src_addr));
+	/* if we're changing mode del host and its entries */
+	if (flush_entries)
+		br_multicast_del_eht_host(pg, h_addr);
+	for (src_idx = 0; src_idx < nsrcs; src_idx++) {
+		memcpy(&eht_src_addr, srcs + (src_idx * addr_size), addr_size);
+		br_multicast_create_eht_set_entry(pg, &eht_src_addr, h_addr,
+						  filter_mode, false);
+	}
+	/* we can be missing sets only if we've deleted some entries */
+	if (flush_entries) {
+		struct net_bridge_group_src *src_ent;
+		struct hlist_node *tmp;
+
+		hlist_for_each_entry_safe(src_ent, tmp, &pg->src_list, node) {
+			br_multicast_ip_src_to_eht_addr(&src_ent->addr,
+							&eht_src_addr);
+			if (!br_multicast_eht_set_lookup(pg, &eht_src_addr)) {
+				br_multicast_del_group_src(src_ent);
+				changed = true;
+				continue;
+			}
+		}
+	}
+
+	return changed;
+}
+
+static bool br_multicast_eht_inc(struct net_bridge_port_group *pg,
+				 union net_bridge_eht_addr *h_addr,
+				 void *srcs,
+				 u32 nsrcs,
+				 size_t addr_size,
+				 bool to_report)
+{
+	return __eht_inc_exc(pg, h_addr, srcs, nsrcs, addr_size, MCAST_INCLUDE,
+			     to_report);
+}
+
+static bool br_multicast_eht_exc(struct net_bridge_port_group *pg,
+				 union net_bridge_eht_addr *h_addr,
+				 void *srcs,
+				 u32 nsrcs,
+				 size_t addr_size,
+				 bool to_report)
+{
+	return __eht_inc_exc(pg, h_addr, srcs, nsrcs, addr_size, MCAST_EXCLUDE,
+			     to_report);
+}
+
 static bool __eht_ip4_handle(struct net_bridge_port_group *pg,
 			     union net_bridge_eht_addr *h_addr,
 			     void *srcs,
 			     u32 nsrcs,
 			     int grec_type)
 {
-	bool changed = false;
+	bool changed = false, to_report = false;
 
 	switch (grec_type) {
 	case IGMPV3_ALLOW_NEW_SOURCES:
@@ -645,6 +711,20 @@ static bool __eht_ip4_handle(struct net_bridge_port_group
*pg,
 		changed = br_multicast_eht_block(pg, h_addr, srcs, nsrcs,
 						 sizeof(__be32));
 		break;
+	case IGMPV3_CHANGE_TO_INCLUDE:
+		to_report = true;
+		fallthrough;
+	case IGMPV3_MODE_IS_INCLUDE:
+		changed = br_multicast_eht_inc(pg, h_addr, srcs, nsrcs,
+					       sizeof(__be32), to_report);
+		break;
+	case IGMPV3_CHANGE_TO_EXCLUDE:
+		to_report = true;
+		fallthrough;
+	case IGMPV3_MODE_IS_EXCLUDE:
+		changed = br_multicast_eht_exc(pg, h_addr, srcs, nsrcs,
+					       sizeof(__be32), to_report);
+		break;
 	}
 
 	return changed;
@@ -657,7 +737,7 @@ static bool __eht_ip6_handle(struct net_bridge_port_group
*pg,
 			     u32 nsrcs,
 			     int grec_type)
 {
-	bool changed = false;
+	bool changed = false, to_report = false;
 
 	switch (grec_type) {
 	case MLD2_ALLOW_NEW_SOURCES:
@@ -668,6 +748,22 @@ static bool __eht_ip6_handle(struct net_bridge_port_group
*pg,
 		changed = br_multicast_eht_block(pg, h_addr, srcs, nsrcs,
 						 sizeof(struct in6_addr));
 		break;
+	case MLD2_CHANGE_TO_INCLUDE:
+		to_report = true;
+		fallthrough;
+	case MLD2_MODE_IS_INCLUDE:
+		changed = br_multicast_eht_inc(pg, h_addr, srcs, nsrcs,
+					       sizeof(struct in6_addr),
+					       to_report);
+		break;
+	case MLD2_CHANGE_TO_EXCLUDE:
+		to_report = true;
+		fallthrough;
+	case MLD2_MODE_IS_EXCLUDE:
+		changed = br_multicast_eht_exc(pg, h_addr, srcs, nsrcs,
+					       sizeof(struct in6_addr),
+					       to_report);
+		break;
 	}
 
 	return changed;
diff --git a/net/bridge/br_private_mcast_eht.h
b/net/bridge/br_private_mcast_eht.h
index 92933822301d..9daffa3ad8d5 100644
--- a/net/bridge/br_private_mcast_eht.h
+++ b/net/bridge/br_private_mcast_eht.h
@@ -55,4 +55,11 @@ bool br_multicast_eht_handle(struct net_bridge_port_group
*pg,
 			     size_t addr_size,
 			     int grec_type);
 
+static inline bool
+br_multicast_eht_should_del_pg(const struct net_bridge_port_group *pg)
+{
+	return !!((pg->key.port->flags & BR_MULTICAST_FAST_LEAVE) &&
+		  RB_EMPTY_ROOT(&pg->eht_host_tree));
+}
+
 #endif /* _BR_PRIVATE_MCAST_EHT_H_ */
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

This is an optimization specifically for TO_INCLUDE which sends queries
for the older entries and thus lowers the S,G timers to LMQT. If we have
the following situation for a group in either include or exclude mode:
 - host A was interested in srcs X and Y, but is timing out
 - host B sends TO_INCLUDE src Z, the bridge lowers X and Y's timeouts
   to LMQT
 - host B sends BLOCK src Z after LMQT time has passed
 => since host B is the last host we can delete the group, but if we
    still have host A's EHT entries for X and Y (i.e. if they weren't
    lowered to LMQT previously) then we'll have to wait another LMQT
    time before deleting the group, with this optimization we can
    directly remove it regardless of the group mode as there are no more
    interested hosts

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast_eht.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index 861ae63f4a1c..fee3060d0495 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -656,6 +656,8 @@ static bool __eht_inc_exc(struct net_bridge_port_group *pg,
 	}
 	/* we can be missing sets only if we've deleted some entries */
 	if (flush_entries) {
+		struct net_bridge *br = pg->key.port->br;
+		struct net_bridge_group_eht_set *eht_set;
 		struct net_bridge_group_src *src_ent;
 		struct hlist_node *tmp;
 
@@ -667,6 +669,25 @@ static bool __eht_inc_exc(struct net_bridge_port_group *pg,
 				changed = true;
 				continue;
 			}
+			/* this is an optimization for TO_INCLUDE where we lower
+			 * the set's timeout to LMQT to catch timeout hosts:
+			 * - host A (timing out): set entries X, Y
+			 * - host B: set entry Z (new from current TO_INCLUDE)
+			 *           sends BLOCK Z after LMQT but host A's EHT
+			 *           entries still exist (unless lowered to LMQT
+			 *           so they can timeout with the S,Gs)
+			 * => we wait another LMQT, when we can just delete the
+			 *    group immediately
+			 */
+			if (!(src_ent->flags & BR_SGRP_F_SEND) ||
+			    filter_mode != MCAST_INCLUDE ||
+			    !to_report)
+				continue;
+			eht_set = br_multicast_eht_set_lookup(pg,
+							      &eht_src_addr);
+			if (!eht_set)
+				continue;
+			mod_timer(&eht_set->timer, jiffies + br_multicast_lmqt(br));
 		}
 	}
 
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

We should be able to handle host filter mode changing. For exclude mode
we must create a zero-src entry so the group will be kept even without
any S,G entries (non-zero source sets). That entry doesn't count to the
entry limit and can always be created, its timer is refreshed on new
exclude reports and if we change the host filter mode to include then it
gets removed and we rely only on the non-zero source sets.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast_eht.c | 42 +++++++++++++++++++++++++++++++----
 1 file changed, 38 insertions(+), 4 deletions(-)

diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index fee3060d0495..64ccbd4ae9d9 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -381,6 +381,30 @@ static void br_multicast_ip_src_to_eht_addr(const struct
br_ip *src,
 	}
 }
 
+static void br_eht_convert_host_filter_mode(struct net_bridge_port_group *pg,
+					    union net_bridge_eht_addr *h_addr,
+					    int filter_mode)
+{
+	struct net_bridge_group_eht_host *eht_host;
+	union net_bridge_eht_addr zero_addr;
+
+	eht_host = br_multicast_eht_host_lookup(pg, h_addr);
+	if (eht_host)
+		eht_host->filter_mode = filter_mode;
+
+	memset(&zero_addr, 0, sizeof(zero_addr));
+	switch (filter_mode) {
+	case MCAST_INCLUDE:
+		br_multicast_del_eht_set_entry(pg, &zero_addr, h_addr);
+		break;
+	case MCAST_EXCLUDE:
+		br_multicast_create_eht_set_entry(pg, &zero_addr, h_addr,
+						  MCAST_EXCLUDE,
+						  true);
+		break;
+	}
+}
+
 static void br_multicast_create_eht_set_entry(struct net_bridge_port_group *pg,
 					      union net_bridge_eht_addr *src_addr,
 					      union net_bridge_eht_addr *h_addr,
@@ -701,8 +725,13 @@ static bool br_multicast_eht_inc(struct
net_bridge_port_group *pg,
 				 size_t addr_size,
 				 bool to_report)
 {
-	return __eht_inc_exc(pg, h_addr, srcs, nsrcs, addr_size, MCAST_INCLUDE,
-			     to_report);
+	bool changed;
+
+	changed = __eht_inc_exc(pg, h_addr, srcs, nsrcs, addr_size,
+				MCAST_INCLUDE, to_report);
+	br_eht_convert_host_filter_mode(pg, h_addr, MCAST_INCLUDE);
+
+	return changed;
 }
 
 static bool br_multicast_eht_exc(struct net_bridge_port_group *pg,
@@ -712,8 +741,13 @@ static bool br_multicast_eht_exc(struct
net_bridge_port_group *pg,
 				 size_t addr_size,
 				 bool to_report)
 {
-	return __eht_inc_exc(pg, h_addr, srcs, nsrcs, addr_size, MCAST_EXCLUDE,
-			     to_report);
+	bool changed;
+
+	changed = __eht_inc_exc(pg, h_addr, srcs, nsrcs, addr_size,
+				MCAST_EXCLUDE, to_report);
+	br_eht_convert_host_filter_mode(pg, h_addr, MCAST_EXCLUDE);
+
+	return changed;
 }
 
 static bool __eht_ip4_handle(struct net_bridge_port_group *pg,
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

A block report can result in empty source and host sets for both include
and exclude groups so if there are no hosts left we can safely remove
the group. Pull the block group handling so it can cover both cases and
add a check if EHT requires the delete.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 9cfc004312ab..47afb1e11daf 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -2210,14 +2210,6 @@ static bool __grp_src_block_incl(struct
net_bridge_port_group *pg, void *h_addr,
 	if (to_send)
 		__grp_src_query_marked_and_rexmit(pg);
 
-	if (pg->filter_mode == MCAST_INCLUDE &&
hlist_empty(&pg->src_list)) {
-		br_multicast_find_del_pg(pg->key.port->br, pg);
-		/* a notification has already been sent and we shouldn't access
-		 * pg after the delete thus we have to return false
-		 */
-		changed = false;
-	}
-
 	return changed;
 }
 
@@ -2279,6 +2271,15 @@ static bool br_multicast_block(struct
net_bridge_port_group *pg, void *h_addr,
 		break;
 	}
 
+	if ((pg->filter_mode == MCAST_INCLUDE &&
hlist_empty(&pg->src_list)) ||
+	    br_multicast_eht_should_del_pg(pg)) {
+		br_multicast_find_del_pg(pg->key.port->br, pg);
+		/* a notification has already been sent and we shouldn't
+		 * access pg after the delete so we have to return false
+		 */
+		changed = false;
+	}
+
 	return changed;
 }
 
-- 
2.29.2

From: Nikolay Aleksandrov <nikolay at nvidia.com>

Mark groups which were deleted due to fast leave/EHT.

Signed-off-by: Nikolay Aleksandrov <nikolay at nvidia.com>
---
 net/bridge/br_multicast.c     | 21 ++++++++++++++-------
 net/bridge/br_multicast_eht.c |  8 ++++----
 net/bridge/br_private.h       |  3 ++-
 3 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 47afb1e11daf..df5db6a58e95 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -442,7 +442,8 @@ static void br_multicast_fwd_src_add(struct
net_bridge_group_src *src)
 	br_multicast_sg_add_exclude_ports(star_mp, sg);
 }
 
-static void br_multicast_fwd_src_remove(struct net_bridge_group_src *src)
+static void br_multicast_fwd_src_remove(struct net_bridge_group_src *src,
+					bool fastleave)
 {
 	struct net_bridge_port_group *p, *pg = src->pg;
 	struct net_bridge_port_group __rcu **pp;
@@ -467,6 +468,8 @@ static void br_multicast_fwd_src_remove(struct
net_bridge_group_src *src)
 		    (p->flags & MDB_PG_FLAGS_PERMANENT))
 			break;
 
+		if (fastleave)
+			p->flags |= MDB_PG_FLAGS_FAST_LEAVE;
 		br_multicast_del_pg(mp, p, pp);
 		break;
 	}
@@ -560,11 +563,12 @@ static void br_multicast_destroy_group_src(struct
net_bridge_mcast_gc *gc)
 	kfree_rcu(src, rcu);
 }
 
-void br_multicast_del_group_src(struct net_bridge_group_src *src)
+void br_multicast_del_group_src(struct net_bridge_group_src *src,
+				bool fastleave)
 {
 	struct net_bridge *br = src->pg->key.port->br;
 
-	br_multicast_fwd_src_remove(src);
+	br_multicast_fwd_src_remove(src, fastleave);
 	hlist_del_init_rcu(&src->node);
 	src->pg->src_ents--;
 	hlist_add_head(&src->mcast_gc.gc_node, &br->mcast_gc_list);
@@ -596,7 +600,7 @@ void br_multicast_del_pg(struct net_bridge_mdb_entry *mp,
 	hlist_del_init(&pg->mglist);
 	br_multicast_eht_clean_sets(pg);
 	hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
-		br_multicast_del_group_src(ent);
+		br_multicast_del_group_src(ent, false);
 	br_mdb_notify(br->dev, mp, pg, RTM_DELMDB);
 	if (!br_multicast_is_star_g(&mp->addr)) {
 		rhashtable_remove_fast(&br->sg_port_tbl, &pg->rhnode,
@@ -653,7 +657,7 @@ static void br_multicast_port_group_expired(struct
timer_list *t)
 	pg->filter_mode = MCAST_INCLUDE;
 	hlist_for_each_entry_safe(src_ent, tmp, &pg->src_list, node) {
 		if (!timer_pending(&src_ent->timer)) {
-			br_multicast_del_group_src(src_ent);
+			br_multicast_del_group_src(src_ent, false);
 			changed = true;
 		}
 	}
@@ -1080,7 +1084,7 @@ static void br_multicast_group_src_expired(struct
timer_list *t)
 
 	pg = src->pg;
 	if (pg->filter_mode == MCAST_INCLUDE) {
-		br_multicast_del_group_src(src);
+		br_multicast_del_group_src(src, false);
 		if (!hlist_empty(&pg->src_list))
 			goto out;
 		br_multicast_find_del_pg(br, pg);
@@ -1704,7 +1708,7 @@ static int __grp_src_delete_marked(struct
net_bridge_port_group *pg)
 
 	hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
 		if (ent->flags & BR_SGRP_F_DELETE) {
-			br_multicast_del_group_src(ent);
+			br_multicast_del_group_src(ent, false);
 			deleted++;
 		}
 
@@ -2053,6 +2057,7 @@ static bool br_multicast_toin(struct net_bridge_port_group
*pg, void *h_addr,
 	}
 
 	if (br_multicast_eht_should_del_pg(pg)) {
+		pg->flags |= MDB_PG_FLAGS_FAST_LEAVE;
 		br_multicast_find_del_pg(pg->key.port->br, pg);
 		/* a notification has already been sent and we shouldn't
 		 * access pg after the delete so we have to return false
@@ -2273,6 +2278,8 @@ static bool br_multicast_block(struct
net_bridge_port_group *pg, void *h_addr,
 
 	if ((pg->filter_mode == MCAST_INCLUDE &&
hlist_empty(&pg->src_list)) ||
 	    br_multicast_eht_should_del_pg(pg)) {
+		if (br_multicast_eht_should_del_pg(pg))
+			pg->flags |= MDB_PG_FLAGS_FAST_LEAVE;
 		br_multicast_find_del_pg(pg->key.port->br, pg);
 		/* a notification has already been sent and we shouldn't
 		 * access pg after the delete so we have to return false
diff --git a/net/bridge/br_multicast_eht.c b/net/bridge/br_multicast_eht.c
index 64ccbd4ae9d9..a4fa1760bc8a 100644
--- a/net/bridge/br_multicast_eht.c
+++ b/net/bridge/br_multicast_eht.c
@@ -537,7 +537,7 @@ static bool __eht_allow_excl(struct net_bridge_port_group
*pg,
 			src_ent = br_multicast_find_group_src(pg, &src_ip);
 			if (!src_ent)
 				continue;
-			br_multicast_del_group_src(src_ent);
+			br_multicast_del_group_src(src_ent, true);
 			changed = true;
 		}
 	}
@@ -588,7 +588,7 @@ static bool __eht_block_incl(struct net_bridge_port_group
*pg,
 		src_ent = br_multicast_find_group_src(pg, &src_ip);
 		if (!src_ent)
 			continue;
-		br_multicast_del_group_src(src_ent);
+		br_multicast_del_group_src(src_ent, true);
 		changed = true;
 	}
 
@@ -625,7 +625,7 @@ static bool __eht_block_excl(struct net_bridge_port_group
*pg,
 			src_ent = br_multicast_find_group_src(pg, &src_ip);
 			if (!src_ent)
 				continue;
-			br_multicast_del_group_src(src_ent);
+			br_multicast_del_group_src(src_ent, true);
 			changed = true;
 		}
 	}
@@ -689,7 +689,7 @@ static bool __eht_inc_exc(struct net_bridge_port_group *pg,
 			br_multicast_ip_src_to_eht_addr(&src_ent->addr,
 							&eht_src_addr);
 			if (!br_multicast_eht_set_lookup(pg, &eht_src_addr)) {
-				br_multicast_del_group_src(src_ent);
+				br_multicast_del_group_src(src_ent, true);
 				changed = true;
 				continue;
 			}
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index cad967690e9f..0e26ba623006 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -850,7 +850,8 @@ void br_multicast_sg_add_exclude_ports(struct
net_bridge_mdb_entry *star_mp,
 				       struct net_bridge_port_group *sg);
 struct net_bridge_group_src *
 br_multicast_find_group_src(struct net_bridge_port_group *pg, struct br_ip
*ip);
-void br_multicast_del_group_src(struct net_bridge_group_src *src);
+void br_multicast_del_group_src(struct net_bridge_group_src *src,
+				bool fastleave);
 
 static inline bool br_group_is_l2(const struct br_ip *group)
 {
-- 
2.29.2

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

On Wed, 20 Jan 2021 16:51:49 +0200 you wrote:
> From: Nikolay Aleksandrov <nikolay at nvidia.com>
> 
> Hi,
> This set adds explicit host tracking support for IGMPv3/MLDv2. The
> already present per-port fast leave flag is used to enable it since that
> is the primary goal of EHT, to track a group and its S,Gs usage per-host
> and when left without any interested hosts delete them before the standard
> timers. The EHT code is pretty self-contained and not enabled by default.
> There is no new uAPI added, all of the functionality is currently hidden
> behind the fast leave flag. In the future that will change (more below).
> The host tracking uses two new sets per port group: one having an entry for
> each host which contains that host's view of the group (source list and
> filter mode), and one set which contains an entry for each source having
> an internal set which contains an entry for each host that has reported
> an interest for that source. RB trees are used for all sets so they're
> compact when not used and fast when we need to do lookups.
> To illustrate it:
>  [ bridge port group ]
>   ` [ host set (rb) ]
>    ` [ host entry with a list of sources and filter mode ]
>   ` [ source set (rb) ]
>    ` [ source entry ]
>     ` [ source host set (rb) ]
>      ` [ source host entry with a timer ]
> 
> [...]
Here is the summary with links:
  - [net-next,01/14] net: bridge: multicast: rename src_size to addr_size
    https://git.kernel.org/netdev/net-next/c/9e10b9e65699
  - [net-next,02/14] net: bridge: multicast: pass host src address to
IGMPv3/MLDv2 functions
    https://git.kernel.org/netdev/net-next/c/54bea7219641
  - [net-next,03/14] net: bridge: multicast: __grp_src_block_incl can modify pg
    https://git.kernel.org/netdev/net-next/c/0ad57c99e857
  - [net-next,04/14] net: bridge: multicast: calculate idx position without
changing ptr
    https://git.kernel.org/netdev/net-next/c/e7cfcf2c18c5
  - [net-next,05/14] net: bridge: multicast: add EHT structures and definitions
    https://git.kernel.org/netdev/net-next/c/8f07b831197e
  - [net-next,06/14] net: bridge: multicast: add EHT host handling functions
    https://git.kernel.org/netdev/net-next/c/5b1632887930
  - [net-next,07/14] net: bridge: multicast: add EHT source set handling
functions
    https://git.kernel.org/netdev/net-next/c/baa74d39ca39
  - [net-next,08/14] net: bridge: multicast: add EHT host delete function
    https://git.kernel.org/netdev/net-next/c/dba6b0a5ca21
  - [net-next,09/14] net: bridge: multicast: add EHT allow/block handling
    https://git.kernel.org/netdev/net-next/c/474ddb37fa3a
  - [net-next,10/14] net: bridge: multicast: add EHT include and exclude
handling
    https://git.kernel.org/netdev/net-next/c/ddc255d993d8
  - [net-next,11/14] net: bridge: multicast: optimize TO_INCLUDE EHT timeouts
    https://git.kernel.org/netdev/net-next/c/b66bf55bbc1c
  - [net-next,12/14] net: bridge: multicast: add EHT host filter_mode handling
    https://git.kernel.org/netdev/net-next/c/c9739016a032
  - [net-next,13/14] net: bridge: multicast: handle block pg delete for all
cases
    https://git.kernel.org/netdev/net-next/c/e87e4b5caa5d
  - [net-next,14/14] net: bridge: multicast: mark IGMPv3/MLDv2 fast-leave
deletes
    https://git.kernel.org/netdev/net-next/c/d5a1022283c3

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