Hi! The x86-64 and aarch64 psABIs (and the unwritten ia64 psABI part) say that the padding bits of _BitInt are undefined, while the expansion internally typically assumes that non-mode precision integers are sign/zero extended and extends after operations. We handle that mismatch with EXTEND_BITINT done when reading from untrusted sources like function arguments, reading _BitInt from memory etc. but otherwise keep relying on stuff being extended internally (say in pseudos). The return value of a function is an ABI boundary though too and we need to extend that too.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? 2024-03-15 Jakub Jelinek <ja...@redhat.com> PR middle-end/114332 * expr.cc (expand_expr_real_1): EXTEND_BITINT also CALL_EXPR results. --- gcc/expr.cc.jj 2024-03-04 19:20:27.120200885 +0100 +++ gcc/expr.cc 2024-03-14 19:22:50.623550538 +0100 @@ -12350,7 +12350,8 @@ expand_expr_real_1 (tree exp, rtx target return expand_builtin (exp, target, subtarget, tmode, ignore); } } - return expand_call (exp, target, ignore); + temp = expand_call (exp, target, ignore); + return EXTEND_BITINT (temp); case VIEW_CONVERT_EXPR: op0 = NULL_RTX; --- gcc/testsuite/gcc.dg/torture/bitint-64.c.jj 2024-03-14 19:50:41.051311949 +0100 +++ gcc/testsuite/gcc.dg/torture/bitint-64.c 2024-03-14 19:50:05.480806808 +0100 @@ -0,0 +1,22 @@ +/* PR middle-end/114332 */ +/* { dg-do run { target bitint } } */ +/* { dg-options "-std=c23 -fwrapv" } */ +/* { dg-skip-if "" { ! run_expensive_tests } { "*" } { "-O0" "-O2" } } */ +/* { dg-skip-if "" { ! run_expensive_tests } { "-flto" } { "" } } */ + +enum E { E22 = 22 } e = E22; + +_BitInt (5) +foo (void) +{ + _Atomic _BitInt (5) b = 0; + b += e; + return b; +} + +int +main () +{ + if (foo () != -10) + __builtin_abort (); +} Jakub