Re: [PATCH net-next v5 6/6] selftests: drv-net: psp: add rekey tests

From: netdev-bot+sashiko

Date: Sat Nov 21 2026 - 18:03:48 EST


Thank you for your contribution! Sashiko AI review found 1 potential
issue(s) to consider.

Critical: 0 · High: 0 · Medium: 1 · Low: 0

If you believe none of the issues are legitimate, please reply to each
issue of Medium severity or higher and explain why the AI is wrong.
If any of them are legitimate and you plan to respin, please let patchwork
know by including "pw-bot: cr" as a separate line at the end of your reply
(one such reply per series is enough).

- [Medium] selftests: drv-net: psp: stale length n in run_session() after
tx assoc

--

> diff --git a/tools/testing/selftests/drivers/net/psp_responder.c b/tools/testing/selftests/drivers/net/psp_responder.c
> --- a/tools/testing/selftests/drivers/net/psp_responder.c
> +++ b/tools/testing/selftests/drivers/net/psp_responder.c
> @@ -247,6 +354,16 @@ run_session(struct ynl_sock *ys, struct opts *opts,
> fprintf(stderr, "WARN: echo but no data sock\n");
> send_ack(comm_sock);
> }
> + if (cmd("rx assoc"))
> + handle_rx_assoc(ys, data_sock,
> + comm_sock);
> + if (cmd_w_msg("tx assoc", struct assoc_msg)) {
> + handle_tx_assoc(ys, buf, data_sock,
> + comm_sock);
> + __consume(sizeof(struct assoc_msg));
> + }
> + if (cmd("key rotate"))
> + handle_key_rotate(ys, opts, comm_sock);
> if (cmd("data close")) {
> if (data_sock >= 0) {
> close(data_sock);

[Severity: Medium]
In run_session(), n is set from off once, before the do/while loop,
and is never updated. cmd() still tests n >= sz and __consume() tests
n == sz. For a single tx assoc message n is 49.

cmd("tx assoc") consumes the 9 byte string and moves the struct
assoc_msg to buf[0]. __consume(sizeof(struct assoc_msg)) then takes
the else branch. That leaves off == 0, but the 40 byte assoc_msg is
still in buf[0..39].

The checks that follow in the same iteration (key rotate, data close,
conn psp, conn clr, exit) still see n == 49. They memcmp() buf against
the SPI, version and key bytes of a message that was already handled.
Suppose those bytes match a command string. For example, an SPI of
0x65786974 with version 0 reads as "exit\0". The matching __consume()
then does off -= sz starting from zero, and memmove() gets a negative
length converted to a huge size_t, which crashes the responder.

The conn psp handler already has the same stale-n pattern, but it
only leaves the two version bytes behind. tx assoc is the first
command that leaves an arbitrary binary payload at the front of buf.

Would it make sense to have cmd() and __consume() check off rather
than n, perhaps as a small fix ahead of this patch? Then a command
could only match bytes that have not been consumed yet.

--
Sashiko AI review · https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20261118-psp-v5-0-b9ec840e0296%40gmail.com