linux/net/ipv4
Eric Dumazet 31e1da773a tcp: avoid premature drops in tcp_add_backlog()
[ Upstream commit ec00ed472b ]

While testing TCP performance with latest trees,
I saw suspect SOCKET_BACKLOG drops.

tcp_add_backlog() computes its limit with :

    limit = (u32)READ_ONCE(sk->sk_rcvbuf) +
            (u32)(READ_ONCE(sk->sk_sndbuf) >> 1);
    limit += 64 * 1024;

This does not take into account that sk->sk_backlog.len
is reset only at the very end of __release_sock().

Both sk->sk_backlog.len and sk->sk_rmem_alloc could reach
sk_rcvbuf in normal conditions.

We should double sk->sk_rcvbuf contribution in the formula
to absorb bubbles in the backlog, which happen more often
for very fast flows.

This change maintains decent protection against abuses.

Fixes: c377411f24 ("net: sk_add_backlog() take rmem_alloc into account")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20240423125620.3309458-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2024-06-16 13:28:35 +02:00
..
bpfilter
netfilter treewide: Remove uninitialized_var() usage 2023-06-09 10:29:01 +02:00
af_inet.c tcp: add annotations around sk->sk_shutdown accesses 2024-03-01 13:13:34 +01:00
ah4.c
arp.c ipv4: Invalidate neighbour for broadcast address upon address addition 2022-04-15 14:18:33 +02:00
cipso_ipv4.c cipso: Fix data-races around sysctl. 2022-07-21 20:59:20 +02:00
datagram.c inet: stop leaking jiffies on the wire 2019-11-01 14:57:52 -07:00
devinet.c ipv4: properly combine dev_base_seq and ipv4.dev_addr_genid 2024-03-01 13:13:37 +01:00
esp4_offload.c xfrm: Linearize the skb after offloading if needed. 2023-06-28 10:18:38 +02:00
esp4.c net: ipv4: fix return value check in esp_remove_trailer 2023-10-25 11:53:21 +02:00
fib_frontend.c ipv4: Fix incorrect table ID in IOCTL path 2023-03-22 13:28:06 +01:00
fib_lookup.h
fib_notifier.c
fib_rules.c ipv6: fix memory leak in fib6_rule_suppress 2021-12-08 09:01:13 +01:00
fib_semantics.c ipv4: fib: annotate races around nh->nh_saddr_genid and nh->nh_saddr 2023-10-25 11:53:21 +02:00
fib_trie.c ipv4: annotate data-races around fi->fib_dead 2023-09-23 10:59:58 +02:00
fou.c fou: Fix IPv6 netlink policy 2020-01-29 16:45:22 +01:00
gre_demux.c erspan: fix version 1 check in gre_parse_header() 2021-01-12 20:16:15 +01:00
gre_offload.c net: gre: recompute gre csum for sctp over gre tunnels 2020-08-11 15:33:40 +02:00
icmp.c icmp: guard against too small mtu 2023-04-20 12:07:33 +02:00
igmp.c ipv4: igmp: fix refcnt uaf issue when receiving igmp query packet 2023-12-08 08:44:27 +01:00
inet_connection_sock.c tcp: properly terminate timers for kernel sockets 2024-04-13 12:51:33 +02:00
inet_diag.c inet_diag: fix kernel-infoleak for UDP sockets 2021-12-22 09:29:36 +01:00
inet_fragment.c inet: frags: annotate races around fqdir->dead and fqdir->high_thresh 2022-01-27 09:19:54 +01:00
inet_hashtables.c Revert "tcp: avoid the lookup process failing to get sk in ehash table" 2023-07-27 08:37:44 +02:00
inet_timewait_sock.c tcp: Fix NEW_SYN_RECV handling in inet_twsk_purge() 2024-05-02 16:18:37 +02:00
inetpeer.c inetpeer: Fix data-races around sysctl. 2022-07-21 20:59:20 +02:00
ip_forward.c
ip_fragment.c inet: frags: annotate races around fqdir->dead and fqdir->high_thresh 2022-01-27 09:19:54 +01:00
ip_gre.c ip_gre: do not report erspan version on GRE interface 2024-04-13 12:51:40 +02:00
ip_input.c tcp/udp: Make early_demux back namespacified. 2022-11-10 17:57:55 +01:00
ip_options.c
ip_output.c net: ipv4: fix a memleak in ip_setup_cork 2024-02-23 08:25:05 +01:00
ip_sockglue.c ipv{4,6}/raw: fix output xfrm lookup wrt protocol 2023-06-05 08:17:33 +02:00
ip_tunnel_core.c
ip_tunnel.c net: ip_tunnel: make sure to pull inner header in ip_tunnel_rcv() 2024-03-26 18:22:17 -04:00
ip_vti.c ip_vti: fix potential slab-use-after-free in decode_session6 2023-08-30 16:27:16 +02:00
ipcomp.c
ipconfig.c net: ipconfig: Don't override command-line hostnames or domains 2021-06-18 09:58:59 +02:00
ipip.c net: ipip: fix wrong address family in init error path 2020-06-03 08:20:52 +02:00
ipmr_base.c
ipmr.c ipmr,ip6mr: acquire RTNL before calling ip[6]mr_free_table() on failure path 2022-02-16 12:52:51 +01:00
Kconfig tcp: configurable source port perturb table size 2022-12-08 11:23:00 +01:00
Makefile
metrics.c ipv4: prevent potential spectre v1 gadget in ip_metrics_convert() 2023-02-06 07:52:45 +01:00
netfilter.c netfilter: use actual socket sk rather than skb sk when routing harder 2020-11-18 19:20:17 +01:00
netlink.c
nexthop.c nh: fix scope used to find saddr when adding non gw nh 2022-11-03 23:56:56 +09:00
ping.c ping: fix address binding wrt vrf 2022-05-18 09:47:28 +02:00
proc.c
protocol.c
raw_diag.c inet_diag: return classid for all socket types 2020-03-18 07:17:38 +01:00
raw.c ipv{4,6}/raw: fix output xfrm lookup wrt protocol 2023-06-05 08:17:33 +02:00
route.c ipv4/route: avoid unused-but-set-variable warning 2024-05-02 16:18:28 +02:00
syncookies.c tcp: fix cookie_init_timestamp() overflows 2023-11-20 10:30:10 +01:00
sysctl_net_ipv4.c tcp/udp: Make early_demux back namespacified. 2022-11-10 17:57:55 +01:00
tcp_bbr.c tcp_bbr: fix u32 wrap bug in round logic if bbr_init() called after 2B packets 2021-08-18 08:57:01 +02:00
tcp_bic.c
tcp_bpf.c bpf, sockmap: Fix data loss caused by using apply_bytes on ingress redirect 2023-01-18 11:41:08 +01:00
tcp_cdg.c tcp: cdg: allow tcp_cdg_release() to be called multiple times 2022-11-25 17:42:21 +01:00
tcp_cong.c net: Only allow init netns to set default tcp cong to a restricted algo 2021-05-14 09:44:33 +02:00
tcp_cubic.c tcp_cubic: fix spurious Hystart ACK train detections for not-cwnd-limited flows 2021-12-01 09:23:33 +01:00
tcp_dctcp.c
tcp_dctcp.h
tcp_diag.c
tcp_fastopen.c tcp: annotate data-races around fastopenq.max_qlen 2023-07-27 08:37:45 +02:00
tcp_highspeed.c
tcp_htcp.c
tcp_hybla.c
tcp_illinois.c
tcp_input.c tcp: defer shutdown(SEND_SHUTDOWN) for TCP_SYN_RECV sockets 2024-05-17 11:43:53 +02:00
tcp_ipv4.c tcp: avoid premature drops in tcp_add_backlog() 2024-06-16 13:28:35 +02:00
tcp_lp.c
tcp_metrics.c tcp_metrics: do not create an entry from tcp_init_metrics() 2023-11-20 10:30:09 +01:00
tcp_minisocks.c tcp: tcp_check_req() can be called from process context 2023-03-11 16:44:12 +01:00
tcp_nv.c
tcp_offload.c net, gro: Set inner transport header offset in tcp/udp GRO hook 2021-08-12 13:20:56 +02:00
tcp_output.c tcp: defer shutdown(SEND_SHUTDOWN) for TCP_SYN_RECV sockets 2024-05-17 11:43:53 +02:00
tcp_rate.c tcp: ensure to use the most recently sent skb when filling the rate sample 2022-05-09 09:03:24 +02:00
tcp_recovery.c tcp: fix excessive TLP and RACK timeouts from HZ rounding 2023-10-25 11:53:21 +02:00
tcp_scalable.c
tcp_timer.c net: fix the RTO timer retransmitting skb every 1ms if linear option is enabled 2023-08-30 16:27:19 +02:00
tcp_ulp.c bpf: Sockmap/tls, push write_space updates through ulp updates 2020-01-23 08:22:45 +01:00
tcp_vegas.c
tcp_vegas.h
tcp_veno.c
tcp_westwood.c
tcp_yeah.c
tcp.c tcp: defer shutdown(SEND_SHUTDOWN) for TCP_SYN_RECV sockets 2024-05-17 11:43:53 +02:00
tunnel4.c
udp_diag.c inet_diag: return classid for all socket types 2020-03-18 07:17:38 +01:00
udp_impl.h
udp_offload.c net, gro: Set inner transport header offset in tcp/udp GRO hook 2021-08-12 13:20:56 +02:00
udp_tunnel.c net/tunnel: wait until all sk_user_data reader finish before releasing the sock 2023-01-18 11:41:14 +01:00
udp.c udp: preserve the connected status if only UDP cmsg 2024-05-02 16:18:37 +02:00
udplite.c udplite: Fix NULL pointer dereference in __sk_mem_raise_allocated(). 2023-05-30 12:44:09 +01:00
xfrm4_input.c xfrm: Preserve vlan tags for transport mode software GRO 2024-05-17 11:43:53 +02:00
xfrm4_output.c xfrm: Always set XFRM_TRANSFORMED in xfrm{4,6}_output_finish 2020-04-29 16:33:11 +02:00
xfrm4_policy.c net: add bool confirm_neigh parameter for dst_ops.update_pmtu 2020-01-04 19:18:58 +01:00
xfrm4_protocol.c net: xfrm: unexport __init-annotated xfrm4_protocol_init() 2022-06-14 18:11:58 +02:00
xfrm4_state.c
xfrm4_tunnel.c