Merge branch 'wireguard-patches-for-5-20-rc1'

Jason A. Donenfeld says:

====================
wireguard patches for 5.20-rc1

I had planned to send these out eventually as net.git patches, but as
you emailed earlier, I figure there's no harm in just doing this now for
net-next.git. Please apply the following small fixes:

1) Rather than using msleep() in order to approximate ktime_get_coarse_
   boottime_ns(), instead use an hrtimer, rounded heuristically.

2) An update in selftest config fragments, from Lukas.

3) Linus noticed that a debugging WARN_ON() to detect (impossible) stack
   corruption would still allow the corruption to happen, making it harder
   to get the report about the corruption subsequently.

4) Support for User Mode Linux in the test suite. This depends on some
   UML patches that are slated for 5.20. Richard hasn't sent his pull
   in, but they're in his tree, so I assume it'll happen.
====================

Link: https://lore.kernel.org/r/20220802125613.340848-1-Jason@zx2c4.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 3 insertions, 3 deletions

diff --git a/drivers/net/wireguard/selftest/allowedips.c b/drivers/net/wireguard/selftest/allowedips.c
index e173204..41db10f 100644
--- a/drivers/net/wireguard/selftest/allowedips.c
+++ b/drivers/net/wireguard/selftest/allowedips.c
@@ -593,10 +593,10 @@ bool __init wg_allowedips_selftest(void)
 	wg_allowedips_remove_by_peer(&t, a, &mutex);
 	test_negative(4, a, 192, 168, 0, 1);
 
-	/* These will hit the WARN_ON(len >= 128) in free_node if something
-	 * goes wrong.
+	/* These will hit the WARN_ON(len >= MAX_ALLOWEDIPS_BITS) in free_node
+	 * if something goes wrong.
 	 */
-	for (i = 0; i < 128; ++i) {
+	for (i = 0; i < MAX_ALLOWEDIPS_BITS; ++i) {
 		part = cpu_to_be64(~(1LLU << (i % 64)));
 		memset(&ip, 0xff, 16);
 		memcpy((u8 *)&ip + (i < 64) * 8, &part, 8);