The first one is a bogus report of an uninitialised field passed as
input to hash functions for TCP connections, already seen with gcc 11.
The second one is a bogus stringop-overread which already happened
with gcc 12.1, but this time a "cute" pragma won't make it go away.
Both harmless, but noisy.
Based on pending series posted by David, if it matters.
Stefano Brivio (2):
Makefile: Extend noinline workarounds for LTO and -O2 to gcc 12
udp: Replace pragma to ignore bogus stringop-overread warning with
workaround
Makefile | 6 +++---
udp.c | 26 ++++++++++++++++++--------
util.h | 23 -----------------------
3 files changed, 21 insertions(+), 34 deletions(-)
--
2.35.1