From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id DF1773940CDE; Sun, 2 Jun 2024 20:48:06 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org DF1773940CDE DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1717361286; bh=SwRjYQjFHsQ2Opq3f0livk8pgJ/F0U0+F5NDil8Yh0U=; h=From:To:Subject:Date:In-Reply-To:References:From; b=v+0HsCOD5aiW9CItss6+29bOdqtEZdJina4EiinLOEUoO1YwdQl7OcRzYg8rG6jfS n2zHYOMz4zkA7Set2VjMR9N2jDCAQUlHEswcBHTM3d/DtCmwkfMGdSKS8AbfbER+wL yG4QaJBSLZdRQWhNHFLAiQUamj7a4iealo501KwQ= From: "gcc at sicherha dot de" To: gcc-bugs@gcc.gnu.org Subject: [Bug middle-end/110027] [11/12 regression] Stack objects with extended alignments (vectors etc) misaligned on detect_stack_use_after_return Date: Sun, 02 Jun 2024 20:48:04 +0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: gcc X-Bugzilla-Component: middle-end X-Bugzilla-Version: 13.1.1 X-Bugzilla-Keywords: wrong-code X-Bugzilla-Severity: normal X-Bugzilla-Who: gcc at sicherha dot de X-Bugzilla-Status: ASSIGNED X-Bugzilla-Resolution: X-Bugzilla-Priority: P2 X-Bugzilla-Assigned-To: jakub at gcc dot gnu.org X-Bugzilla-Target-Milestone: 11.5 X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: attachments.isobsolete attachments.created 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=3D110027 Christoph Erhardt changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #56169|0 |1 is obsolete| | --- Comment #26 from Christoph Erhardt --- Created attachment 58325 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=3D58325&action=3Dedit Reproducer program, new version Unfortunately, the fix appears to be incomplete. On x86_64 Fedora Linux 40 = with the latest GCC 14.1.1, my reproducer program still segfaults with a misalig= ned `vmovdqa64`. I have managed to further strip the program down to a now-manageable size -= see attachment. $ gcc -Wall -Wextra -g -Og -fsanitize=3Daddress -fno-stack-protector -mavx5= 12f a-repro.i $ ./a.out AddressSanitizer:DEADLYSIGNAL =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D =3D=3D107156=3D=3DERROR: AddressSanitizer: SEGV on unknown address (pc 0x00= 0000401463 bp 0x7ffd76144620 sp 0x7ffd76144200 T0) =3D=3D107156=3D=3DThe signal is caused by a READ memory access. =3D=3D107156=3D=3DHint: this fault was caused by a dereference of a high va= lue address (see register values below). Disassemble the provided pc to learn which register was used. #0 0x401463 in blake3_compress_subtree_wide /home/christoph/Projects/gcc-asan-stack-misalign/repro.c:52 #1 0x40158c in main /home/christoph/Projects/gcc-asan-stack-misalign/repro.c:57 #2 0x7f60dde3d087 in __libc_start_call_main (/lib64/libc.so.6+0x2a087) (BuildId: 4a92fcedbba6d6d2629ce066a2970017faa9995e) #3 0x7f60dde3d14a in __libc_start_main_alias_2 (/lib64/libc.so.6+0x2a14= a) (BuildId: 4a92fcedbba6d6d2629ce066a2970017faa9995e) #4 0x4010b4 in _start (/home/christoph/Projects/gcc-asan-stack-misalign/a.out+0x4010b4) (BuildId: 9f1d4d25413300b7347d0776d7087844a8d56649) AddressSanitizer can not provide additional info. SUMMARY: AddressSanitizer: SEGV /home/christoph/Projects/gcc-asan-stack-misalign/repro.c:52 in blake3_compress_subtree_wide =3D=3D107156=3D=3DABORTING=