net: connection: Unconditionally forward packets when handling packet sockets #93246
Conversation
| * with this packet regardless the result. | ||
| */ | ||
| raw_pkt_continue = true; | ||
| continue; /* L2 not processed yet */ |
There was a problem hiding this comment.
I think we still need to continue here (skip this "connection") if the condition is satisfied? We don't want to report raw packets to datagram packet sockets.
subsys/net/ip/connection.c
Outdated
| @@ -713,6 +705,11 @@ enum net_verdict net_conn_packet_input(struct net_pkt *pkt, uint16_t proto, enum | |||
|
|
|||
| k_mutex_unlock(&conn_lock); | |||
|
|
|||
| /* We shall continue with raw packets regardless the result. */ | |||
| if (type == SOCK_RAW) { | |||
| raw_pkt_continue = true; | |||
There was a problem hiding this comment.
To be honest, when working with packet sockets I found this raw_pkt_continue quite fishy and I'm tempted to suggest we should just get rid of it. It seems a bit odd that packet socket, which delivers packet clones anyway, makes a decission whether the net stack should proceed with packet processing or not, it seems quite error prone (as this example shows).
Especially that with generalized L3 handling, it seems that the same issue will arise with packets delivered at the DGRA level. IMHO we should just return NET_CONTINUE here unconditionally and let the net stack decide how to proceed with the packet.
There was a problem hiding this comment.
Fully agree.
Can we even make them return void instead of NET_CONTINUE?
There was a problem hiding this comment.
The API's internal so we could update the function signature too.
ClaCodes
force-pushed
the
fix/packet_sockets
branch
3 times, most recently
from
e69c9a8 to
c61f82f
Compare
Store the flag in the packet meta-data so that processing may be deferred if necessary. Signed-off-by: Cla Mattia Galliard <clamattia@gmail.com>
Use the l2_processed-flag to decide whether a network packet needs to be processed by an L2-handler. This could be used in the future to requeue packets for later processing by a different traffic class queue. Signed-off-by: Cla Mattia Galliard <clamattia@gmail.com>
Specify the socket type, when inputing a packet into a packet-socket. Signed-off-by: Cla Mattia Galliard <clamattia@gmail.com>
ClaCodes
force-pushed
the
fix/packet_sockets
branch
from
c61f82f to
64dc2b6
Compare
When handling packets for inputing into packet-sockets, unconditionally forward them, so that they may be handled by the rest of the network stack after. Signed-off-by: Cla Mattia Galliard <clamattia@gmail.com>
ClaCodes
force-pushed
the
fix/packet_sockets
branch
from
64dc2b6 to
c287e48
Compare
|
When handling packets for inputing into raw-packet-sockets, unconditionally
forward them to upper layer, so that it can be handled by L2-handlers.
Based on: #93050
Fixes: #93245