usock: fix off-by-one in nanosecond normalization in poll_restart()

When tv_nsec equals exactly 1000000000 (1 second), the > comparison
misses this case, leaving an invalid timespec value that could cause
incorrect timeout calculations. Use >= to properly normalize, matching
the correct pattern already used in usock_inet_timeout().

Link: https://github.com/openwrt/libubox/pull/42
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit e7c13bf8cbca063e0975c5e57abc2d704e23548c)

1 files changed, 1 insertions, 1 deletions

diff --git a/usock.c b/usock.c
index 2ac84de..18ba0fe 100644
--- a/usock.c
+++ b/usock.c
@@ -112,7 +112,7 @@ static int poll_restart(struct pollfd *fds, int nfds, int timeout)
 	clock_gettime(CLOCK_MONOTONIC, &ts);
 
 	ts.tv_nsec += msec * 1000000;
-	if (ts.tv_nsec > 1000000000) {
+	if (ts.tv_nsec >= 1000000000) {
 		ts.tv_sec++;
 		ts.tv_nsec -= 1000000000;
 	}