On Mon, May 11, 2020 at 02:34:04PM -0700, Maciej Żenczykowski wrote: > From: Maciej Żenczykowski <maze@xxxxxxxxxx> > > Fixes: > In file included from libip4tc.c:113: > In function ‘iptcc_compile_chain’, > inlined from ‘iptcc_compile_table’ at libiptc.c:1246:13, > inlined from ‘iptc_commit’ at libiptc.c:2575:8, > inlined from ‘iptc_commit’ at libiptc.c:2513:1: > libiptc.c:1172:2: warning: writing 16 bytes into a region of size 0 [-Wstringop-overflow=] > 1172 | memcpy(&foot->e.counters, &c->counters, sizeof(STRUCT_COUNTERS)); > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > In file included from ../include/libiptc/libiptc.h:12, > from libip4tc.c:29: > libiptc.c: In function ‘iptc_commit’: > ../include/linux/netfilter_ipv4/ip_tables.h:202:19: note: at offset 0 to object ‘entries’ with size 0 declared here > 202 | struct ipt_entry entries[0]; > | ^~~~~~~ > > Which was found via compilation on Fedora 32. > > Test: builds without warnings > Signed-off-by: Maciej Żenczykowski <maze@xxxxxxxxxx> > --- > libiptc/libiptc.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/libiptc/libiptc.c b/libiptc/libiptc.c > index 58882015..1a92b267 100644 > --- a/libiptc/libiptc.c > +++ b/libiptc/libiptc.c > @@ -1169,7 +1169,10 @@ static int iptcc_compile_chain(struct xtc_handle *h, STRUCT_REPLACE *repl, struc > else > foot->target.verdict = RETURN; > /* set policy-counters */ > +#pragma GCC diagnostic push > +#pragma GCC diagnostic ignored "-Wstringop-overflow" > memcpy(&foot->e.counters, &c->counters, sizeof(STRUCT_COUNTERS)); > +#pragma GCC diagnostic pop Probably make a casting from foot->e.counters to the counters object to make gcc happy?
