linux/net
Martin KaFai Lau 7449197d60 bpf: Keep the (rcv) timestamp behavior for the existing tc-bpf@ingress
The current tc-bpf@ingress reads and writes the __sk_buff->tstamp
as a (rcv) timestamp which currently could either be 0 (not available)
or ktime_get_real().  This patch is to backward compatible with the
(rcv) timestamp expectation at ingress.  If the skb->tstamp has
the delivery_time, the bpf insn rewrite will read 0 for tc-bpf
running at ingress as it is not available.  When writing at ingress,
it will also clear the skb->mono_delivery_time bit.

/* BPF_READ: a = __sk_buff->tstamp */
if (!skb->tc_at_ingress || !skb->mono_delivery_time)
	a = skb->tstamp;
else
	a = 0

/* BPF_WRITE: __sk_buff->tstamp = a */
if (skb->tc_at_ingress)
	skb->mono_delivery_time = 0;
skb->tstamp = a;

[ A note on the BPF_CGROUP_INET_INGRESS which can also access
  skb->tstamp.  At that point, the skb is delivered locally
  and skb_clear_delivery_time() has already been done,
  so the skb->tstamp will only have the (rcv) timestamp. ]

If the tc-bpf@egress writes 0 to skb->tstamp, the skb->mono_delivery_time
has to be cleared also.  It could be done together during
convert_ctx_access().  However, the latter patch will also expose
the skb->mono_delivery_time bit as __sk_buff->delivery_time_type.
Changing the delivery_time_type in the background may surprise
the user, e.g. the 2nd read on __sk_buff->delivery_time_type
may need a READ_ONCE() to avoid compiler optimization.  Thus,
in expecting the needs in the latter patch, this patch does a
check on !skb->tstamp after running the tc-bpf and clears the
skb->mono_delivery_time bit if needed.  The earlier discussion
on v4 [0].

The bpf insn rewrite requires the skb's mono_delivery_time bit and
tc_at_ingress bit.  They are moved up in sk_buff so that bpf rewrite
can be done at a fixed offset.  tc_skip_classify is moved together with
tc_at_ingress.  To get one bit for mono_delivery_time, csum_not_inet is
moved down and this bit is currently used by sctp.

[0]: https://lore.kernel.org/bpf/20220217015043.khqwqklx45c4m4se@kafai-mbp.dhcp.thefacebook.com/

Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-03 14:38:48 +00:00
..
6lowpan net: don't include ndisc.h from ipv6.h 2022-02-04 14:15:11 -08:00
9p virtio,vdpa,qemu_fw_cfg: features, cleanups, fixes 2022-01-18 10:05:48 +02:00
802 net: 802: Use memset_startat() to clear struct fields 2021-11-19 11:23:23 +00:00
8021q Revert "vlan: move dev_put into vlan_dev_uninit" 2022-02-23 15:21:13 -08:00
appletalk
atm proc: remove PDE_DATA() completely 2022-01-22 08:33:37 +02:00
ax25 Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-10 17:29:56 -08:00
batman-adv This cleanup patchset includes the following patches: 2022-03-02 21:58:03 -08:00
bluetooth bluetooth-next pull request for net-next: 2022-01-28 13:39:07 -08:00
bpf bpf: Make remote_port field in struct bpf_sk_lookup 16-bit wide 2022-02-09 11:40:45 -08:00
bpfilter
bridge net: Add skb_clear_tstamp() to keep the mono delivery_time 2022-03-03 14:38:48 +00:00
caif caif: cleanup double word in comment 2022-02-07 12:06:54 +00:00
can Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-24 17:54:25 -08:00
ceph libceph: optionally use bounce buffer on recv path in crc mode 2022-02-02 18:50:36 +01:00
core bpf: Keep the (rcv) timestamp behavior for the existing tc-bpf@ingress 2022-03-03 14:38:48 +00:00
dcb
dccp dccp: remove max48() 2022-01-27 13:53:27 +00:00
decnet net: decnet: use time_is_before_jiffies() instead of open coding it 2022-02-28 13:21:32 +00:00
dns_resolver
dsa net: dsa: felix: migrate host FDB and MDB entries when changing tag proto 2022-03-03 14:15:31 +00:00
ethernet gro: remove rcu_read_lock/rcu_read_unlock from gro_complete handlers 2021-11-24 17:21:42 -08:00
ethtool ethtool: add support to set/get completion queue event size 2022-02-23 20:33:05 -08:00
hsr flow_dissector: Add support for HSR 2022-03-02 22:44:49 -08:00
ieee802154 net: ipv6: Handle delivery_time in ipv6 defrag 2022-03-03 14:38:48 +00:00
ife
ipv4 net: Postpone skb_clear_delivery_time() until knowing the skb is delivered locally 2022-03-03 14:38:48 +00:00
ipv6 net: Postpone skb_clear_delivery_time() until knowing the skb is delivered locally 2022-03-03 14:38:48 +00:00
iucv s390/iucv: sort out physical vs virtual pointers usage 2022-02-22 16:09:13 -08:00
kcm net: Don't include filter.h from net/sock.h 2021-12-29 08:48:14 -08:00
key
l2tp l2tp: add netns refcount tracker to l2tp_dfs_seq_data 2021-12-10 06:38:27 -08:00
l3mdev
lapb
llc sock: Use sock_owned_by_user_nocheck() instead of sk_lock.owned. 2021-12-10 19:43:00 -08:00
mac80211 Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-17 11:44:20 -08:00
mac802154
mctp mctp: Avoid warning if unregister notifies twice 2022-02-25 22:23:23 -08:00
mpls net: mpls: Fix GCC 12 warning 2022-02-10 15:29:39 +00:00
mptcp Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-24 17:54:25 -08:00
ncsi all: replace find_next{,_zero}_bit with find_first{,_zero}_bit where appropriate 2022-01-15 08:47:31 -08:00
netfilter net: Get rcv tstamp if needed in nfnetlink_{log, queue}.c 2022-03-03 14:38:48 +00:00
netlabel lsm: security_task_getsecid_subj() -> security_current_getsecid_subj() 2021-11-22 17:52:47 -05:00
netlink net: Don't include filter.h from net/sock.h 2021-12-29 08:48:14 -08:00
netrom netrom: fix api breakage in nr_setsockopt() 2022-01-07 14:11:05 +00:00
nfc nfc: llcp: Revert "NFC: Keep socket alive until the DISC PDU is actually sent" 2022-03-03 10:43:37 +00:00
nsh
openvswitch net: Add skb_clear_tstamp() to keep the mono delivery_time 2022-03-03 14:38:48 +00:00
packet net: Handle delivery_time in skb->tstamp during network tapping with af_packet 2022-03-03 14:38:48 +00:00
phonet phonet/pep: refuse to enable an unbound pipe 2021-12-20 11:49:51 +00:00
psample
qrtr bus: mhi: core: Add an API for auto queueing buffers for DL channel 2021-12-17 17:17:14 +01:00
rds Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2021-12-16 16:13:19 -08:00
rfkill rfkill: allow to get the software rfkill state 2021-12-20 11:02:38 +01:00
rose net: Don't include filter.h from net/sock.h 2021-12-29 08:48:14 -08:00
rxrpc rxrpc: Adjust retransmission backoff 2022-01-22 02:03:24 +00:00
sched bpf: Keep the (rcv) timestamp behavior for the existing tc-bpf@ingress 2022-03-03 14:38:48 +00:00
sctp Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-01-05 14:36:10 -08:00
smc net: smc: fix different types in min() 2022-03-01 16:43:27 -08:00
strparser bpf: sockmap, strparser, and tls are reusing qdisc_skb_cb and colliding 2021-11-09 01:05:28 +01:00
sunrpc Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-10 17:29:56 -08:00
switchdev net: switchdev: remove lag_mod_cb from switchdev_handle_fdb_event_to_device 2022-02-24 21:31:43 -08:00
tipc Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-24 17:54:25 -08:00
tls tls: cap the output scatter list to something reasonable 2022-02-04 10:14:07 +00:00
unix Merge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next 2022-01-24 15:42:29 -08:00
vmw_vsock vsock: remove vsock from connected table when connect is interrupted by a signal 2022-02-17 08:56:02 -08:00
wireless Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-02-17 11:44:20 -08:00
x25 net: x25: drop harmless check of !more 2021-12-09 18:35:11 -08:00
xdp i40e: xsk: Move tmp desc array from driver to pool 2022-01-27 17:25:32 +01:00
xfrm net: Add skb_clear_tstamp() to keep the mono delivery_time 2022-03-03 14:38:48 +00:00
compat.c
devres.c
Kconfig page_pool: Add allocation stats 2022-03-03 09:55:28 +00:00
Kconfig.debug net: add networking namespace refcount tracker 2021-12-10 06:38:26 -08:00
Makefile
socket.c net: fix documentation for kernel_getsockname 2022-02-14 14:01:19 +00:00
sysctl_net.c sections: move and rename core_kernel_data() to is_kernel_core_data() 2021-11-09 10:02:50 -08:00