From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id EBE2F3858C55; Thu, 13 Oct 2022 17:41:55 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org EBE2F3858C55 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1665682915; bh=LLcTkkGc+KgQ59RQ4X5+wpkyvoJVWa+D0EUjbf6nIS4=; h=From:To:Subject:Date:In-Reply-To:References:From; b=ia6H/JPaZRcniIGl5o7AfKJ+ctkP00MRZeJwggC8eNGCbsn6XF2o5L5J0qBiFlJeJ ChTB3/Lx9p/45Y/bXdDOycOcdqtKwABZ2TPuyBSJHDqu+XYt1mRPRzMKkkFwh5n7gv /+VwFLVP3yiSphUSQxCvAp3pXPV79h8VqR4unP60= From: "adhemerval.zanella at linaro dot org" To: glibc-bugs@sourceware.org Subject: [Bug libc/29651] glibc can't handle IRELATIVE in .rel.plt only on ARM32 and PPC Date: Thu, 13 Oct 2022 17:41:55 +0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: glibc X-Bugzilla-Component: libc X-Bugzilla-Version: unspecified X-Bugzilla-Keywords: X-Bugzilla-Severity: normal X-Bugzilla-Who: adhemerval.zanella at linaro dot org X-Bugzilla-Status: UNCONFIRMED X-Bugzilla-Resolution: X-Bugzilla-Priority: P2 X-Bugzilla-Assigned-To: unassigned at sourceware dot org X-Bugzilla-Target-Milestone: --- X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: cc Message-ID: In-Reply-To: References: Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Bugzilla-URL: http://sourceware.org/bugzilla/ Auto-Submitted: auto-generated MIME-Version: 1.0 List-Id: https://sourceware.org/bugzilla/show_bug.cgi?id=3D29651 Adhemerval Zanella changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |adhemerval.zanella at lina= ro dot o | |rg --- Comment #1 from Adhemerval Zanella --- It was already reported on binutils [1] and on lld [2]. The current behavior for ld on arm is to put IRELATIVE on .rel.dyn but I am not sure why exactly (Peter hinted it is a "legacy of ARM having a single .got section", and I am not sure if ld still have this limitation yet). So 'fixing' it won't gain us anything with current supported static linkers= .=20 It might improve support on arm ifunc itself (for instance to be able to ca= ll R_ARM_JUMP_SLOT functions, since ifunc will be reordered last), but what if= unc is able to do is not properly documented and has a lot of pitafalls that calling external functions (from current TU) is always 'tricky'. And even if we change it, a static-linker that only emits IRELATIVE on .rel= .plt will need to prevent to run on older glibc (as we did for DT_RELR with a ex= tra symbol). So the question is whether you need this changed. Do you need it for your = mold linker? [1] https://sourceware.org/bugzilla/show_bug.cgi?id=3D19891 [2] https://reviews.llvm.org/D26029 [3] https://sourceware.org/glibc/wiki/GNU_IFUNC --=20 You are receiving this mail because: You are on the CC list for the bug.=