From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id 596C13864C5C; Thu, 23 May 2024 10:39:25 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 596C13864C5C DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1716460765; bh=AqhM5TQN9r/a2TQObchrBEM8OedG3k6J5WY1zb4nVrE=; h=From:To:Subject:Date:In-Reply-To:References:From; b=NHV9KBGLmsbUmNvTZRBt8P8BCPTEl8s4fnkCOjp2uvT6S1FB2Wy3wdEEZ7P1BYuQv Gir3KerB2ud8bxD7v2Yt2IXiKc/7RfniuTP2H2rRM5Okx4obExr2sIhRKYU3Yqcojk o3qIOQeYKfMovCKuJ1oj5pWtdQ1ijl3BhWjF4wVY= From: "slyfox at gcc dot gnu.org" To: gcc-bugs@gcc.gnu.org Subject: [Bug target/115161] [15 Regression] highway-1.0.7 miscompilation of some SSE2 intrinsics Date: Thu, 23 May 2024 10:39:22 +0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: gcc X-Bugzilla-Component: target X-Bugzilla-Version: 15.0 X-Bugzilla-Keywords: wrong-code X-Bugzilla-Severity: normal X-Bugzilla-Who: slyfox at gcc dot gnu.org X-Bugzilla-Status: NEW X-Bugzilla-Resolution: X-Bugzilla-Priority: P3 X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org X-Bugzilla-Target-Milestone: 15.0 X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: Message-ID: In-Reply-To: References: Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/ Auto-Submitted: auto-generated MIME-Version: 1.0 List-Id: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D115161 --- Comment #21 from Sergei Trofimovich --- Shrunk the example down to a single simpler function while preserving the original masking intent: ```c cat bug.cc #include #include #include __attribute__((noipa)) static void assert_eq_p(void * l, void * r) { char lb[16]; char rb[16]; __builtin_memcpy(lb, l, 16); __builtin_memcpy(rb, r, 16); if (__builtin_memcmp(lb, rb, 16) !=3D 0) __builtin_trap(); } __attribute__((noipa)) static void assert_eq(__m128i l, __m128i r) { assert_eq_p(&l, &r); } int main() { const __m128i su =3D _mm_set1_epi32(0x4f800000); const __m128 sf =3D _mm_castsi128_ps(su); const __m128 overflow_mask_f32 =3D _mm_cmpge_ps(sf, _mm_set1_ps(2147483648.0f)); const __m128i overflow_mask =3D _mm_castps_si128(overflow_mask_f32); const __m128i conv =3D _mm_cvttps_epi32(sf); // overflows const __m128i yes =3D _mm_set1_epi32(INT32_MAX); const __m128i a =3D _mm_and_si128(overflow_mask, yes); const __m128i na =3D _mm_andnot_si128(overflow_mask, conv); const __m128i conv_masked =3D _mm_or_si128(a, na); const __m128i actual =3D _mm_cmpeq_epi32(conv_masked, _mm_set1_epi32(INT32_MAX)); const __m128i expected =3D _mm_set1_epi32(-1); assert_eq(expected, actual); } ``` The discrepancy: Ok: $ /tmp/gb/gcc/xg++ -Wall -B/tmp/gb/gcc bug.cc -o bug -O0 && ./bug Bad: $ /tmp/gb/gcc/xg++ -Wall -B/tmp/gb/gcc bug.cc -o bug -O2 && ./bug Illegal instruction (core dumped)=