Commit 955afd57 authored Nov 06, 2024 by Zijian Zhang Committed by Martin KaFai Lau Nov 06, 2024

bpf, sockmap: Fix sk_msg_reset_curr



Found in the test_txmsg_pull in test_sockmap,
```
txmsg_cork = 512; // corking is importrant here
opt->iov_length = 3;
opt->iov_count = 1;
opt->rate = 512; // sendmsg will be invoked 512 times
```
The first sendmsg will send an sk_msg with size 3, and bpf_msg_pull_data
will be invoked the first time. sk_msg_reset_curr will reset the copybreak
from 3 to 0. In the second sendmsg, since we are in the stage of corking,
psock->cork will be reused in func sk_msg_alloc. msg->sg.copybreak is 0
now, the second msg will overwrite the first msg. As a result, we could
not pass the data integrity test.

The same problem happens in push and pop test. Thus, fix sk_msg_reset_curr
to restore the correct copybreak.

Fixes: bb9aefde ("bpf: sockmap, updating the sg structure should also update curr")
Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com>
Link: https://lore.kernel.org/r/20241106222520.527076-9-zijianzhang@bytedance.com


Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>

parent 5d609ba2

net/core/filter.c

+9 −11

Original line number	Diff line number	Diff line
		@@ -2604,18 +2604,16 @@ BPF_CALL_2(bpf_msg_cork_bytes, struct sk_msg *, msg, u32, bytes)

		static void sk_msg_reset_curr(struct sk_msg *msg)
		{
		u32 i = msg->sg.start;
		u32 len = 0;

		do {
		len += sk_msg_elem(msg, i)->length;
		sk_msg_iter_var_next(i);
		if (len >= msg->sg.size)
		break;
		} while (i != msg->sg.end);
		if (!msg->sg.size) {
		msg->sg.curr = msg->sg.start;
		msg->sg.copybreak = 0;
		} else {
		u32 i = msg->sg.end;

		sk_msg_iter_var_prev(i);
		msg->sg.curr = i;
		msg->sg.copybreak = 0;
		msg->sg.copybreak = msg->sg.data[i].length;
		}
		}

		static const struct bpf_func_proto bpf_msg_cork_bytes_proto = {