Merge branch 'net-reduce-sk_filter-and-friends-bloat'

Eric Dumazet says:

====================
net: reduce sk_filter() (and friends) bloat

Some functions return an error by value, and a drop_reason
by an output parameter. This extra parameter can force stack canaries.

A drop_reason is enough and more efficient.

This series reduces bloat by 678 bytes on x86_64:

$ scripts/bloat-o-meter -t vmlinux.old vmlinux.final
add/remove: 0/0 grow/shrink: 3/18 up/down: 79/-757 (-678)
Function                                     old     new   delta
vsock_queue_rcv_skb                           50      79     +29
ipmr_cache_report                           1290    1315     +25
ip6mr_cache_report                          1322    1347     +25
tcp_v6_rcv                                  3169    3167      -2
packet_rcv_spkt                              329     327      -2
unix_dgram_sendmsg                          1731    1726      -5
netlink_unicast                              957     945     -12
netlink_dump                                1372    1359     -13
sk_filter_trim_cap                           889     858     -31
netlink_broadcast_filtered                  1633    1595     -38
tcp_v4_rcv                                  3152    3111     -41
raw_rcv_skb                                  122      80     -42
ping_queue_rcv_skb                           109      61     -48
ping_rcv                                     215     162     -53
rawv6_rcv_skb                                278     224     -54
__sk_receive_skb                             690     632     -58
raw_rcv                                      591     527     -64
udpv6_queue_rcv_one_skb                      935     869     -66
udp_queue_rcv_one_skb                        919     853     -66
tun_net_xmit                                1146    1074     -72
sock_queue_rcv_skb_reason                    166      76     -90
Total: Before=29722890, After=29722212, chg -0.00%

Future conversions from sock_queue_rcv_skb() to sock_queue_rcv_skb_reason()
can be done later.
====================

Link: https://patch.msgid.link/20260409145625.2306224-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

diff --cc include/net/tcp.h
index 23ce0329b2304e6bf1866bc495a9e057ae040b8d,49f45bcff917942e993c627dd3d0017369186f67..0f09429ff4cba1484ede374e969f13eb561b12ab
--- 1/include/net/tcp.h
--- 2/include/net/tcp.h
+++ b/include/net/tcp.h
@@@ -1680,10 -1680,11 +1680,10 @@@ static inline bool tcp_checksum_complet
                __skb_checksum_complete(skb);
  }
  
 -bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb,
 -                   enum skb_drop_reason *reason);
 +enum skb_drop_reason tcp_add_backlog(struct sock *sk, struct sk_buff *skb);
  
- static inline int tcp_filter(struct sock *sk, struct sk_buff *skb,
-                            enum skb_drop_reason *reason)
+ static inline enum skb_drop_reason
+ tcp_filter(struct sock *sk, struct sk_buff *skb)
  {
        const struct tcphdr *th = (const struct tcphdr *)skb->data;
  

diff --cc net/ipv4/tcp_ipv4.c
Simple merge

diff --cc net/ipv6/tcp_ipv6.c
Simple merge