Hi! In the PR Joseph said that the C standard for va_arg talks about pointers to object type and as a function type is not object type, it is invalid.
The following patch diagnoses it in the FE, instead of ICEing later on when optimizations are turned on (and with -O0 doing something weird at runtime). Bootstrapped/regtested on powerpc64le-linux, ok for trunk? 2022-04-07 Jakub Jelinek <ja...@redhat.com> PR c/105149 * c-typeck.cc (c_build_va_arg): Reject function types. * gcc.dg/pr105149.c: New test. --- gcc/c/c-typeck.cc.jj 2022-03-23 10:29:20.172967605 +0100 +++ gcc/c/c-typeck.cc 2022-04-07 12:26:17.236469809 +0200 @@ -15896,6 +15896,12 @@ c_build_va_arg (location_t loc1, tree ex "type %qT", type); return error_mark_node; } + else if (TREE_CODE (type) == FUNCTION_TYPE) + { + error_at (loc2, "second argument to %<va_arg%> is a function type %qT", + type); + return error_mark_node; + } else if (warn_cxx_compat && TREE_CODE (type) == ENUMERAL_TYPE) warning_at (loc2, OPT_Wc___compat, "C++ requires promoted type, not enum type, in %<va_arg%>"); --- gcc/testsuite/gcc.dg/pr105149.c.jj 2022-04-07 12:39:08.202711511 +0200 +++ gcc/testsuite/gcc.dg/pr105149.c 2022-04-07 12:38:21.418364583 +0200 @@ -0,0 +1,16 @@ +/* PR c/105149 */ +/* { dg-do compile } */ +/* { dg-options "-O2" } */ + +#include <stdarg.h> + +void +foo (int s, ...) +{ + int e; + va_list ap; + + va_start (ap, s); + e = va_arg (ap, int (void)) (); /* { dg-error "second argument to 'va_arg' is a function type" } */ + va_end (ap); +} Jakub