From patchwork Tue Mar 3 12:16:07 2026 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mark Wielaard X-Patchwork-Id: 131004 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from vm01.sourceware.org (localhost [127.0.0.1]) by sourceware.org (Postfix) with ESMTP id 369014BA2E1B for ; Tue, 3 Mar 2026 12:16:50 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 369014BA2E1B X-Original-To: elfutils-devel@sourceware.org Delivered-To: elfutils-devel@sourceware.org Received: from gnu.wildebeest.org (gnu.wildebeest.org [45.83.234.184]) by sourceware.org (Postfix) with ESMTPS id 809094BA2E1B for ; Tue, 3 Mar 2026 12:16:35 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 809094BA2E1B Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=klomp.org Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=klomp.org ARC-Filter: OpenARC Filter v1.0.0 sourceware.org 809094BA2E1B Authentication-Results: server2.sourceware.org; arc=none smtp.remote-ip=45.83.234.184 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1772540195; cv=none; b=LJ/IzDwxbRLuqzKPqQreoP4BmOCMDubRw8tN8tIttGvEfT+PvTQNJfR9jz3tNSLBRr0Bzzm11z6IfY47B3dDF7bNSfwCCL+X9D6FpgY9o/4PWBfbic6aRKwjP1NzB1CWVK3+RldaU1Tg/uwkhJjVdlFlSQMx3J0MdWiPRZ7QugI= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1772540195; c=relaxed/simple; bh=Vr8QdX5vENUe1WatH2mBquvCd/VGC5H6fjMAeCETkWU=; h=From:To:Subject:Date:Message-ID:MIME-Version; b=Ulpd7RO550MP6x/VDwwJcwQ46BiqzGLUI7YPswxKxudHW3LasYvXxZfVdNaRiP6oLX4gbg0gXoPHrSFEkTig7ko9J3z2rzwQdk9NBaRYhIBhG8KhTbtdvsHPO+k8rWi26AX1T5yo47Nhpu8ubPf/q/a/1A73o5b/GzDKDT03vQA= ARC-Authentication-Results: i=1; server2.sourceware.org DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 809094BA2E1B Received: from 9950x.localdomain (82-217-174-174.cable.dynamic.v4.ziggo.nl [82.217.174.174]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by gnu.wildebeest.org (Postfix) with ESMTPSA id 033ED3032F85; Tue, 3 Mar 2026 13:16:33 +0100 (CET) Received: by 9950x.localdomain (Postfix, from userid 1000) id 525BDAA0035; Tue, 03 Mar 2026 13:16:32 +0100 (CET) From: Mark Wielaard To: elfutils-devel@sourceware.org Cc: Mark Wielaard Subject: [PATCH] libdwfl: Work around ET_REL files with sh_addr fields set to non-zero Date: Tue, 3 Mar 2026 13:16:07 +0100 Message-ID: <20260303121607.639207-1-mark@klomp.org> X-Mailer: git-send-email 2.53.0 MIME-Version: 1.0 X-Spam-Status: No, score=-7.0 required=5.0 tests=BAYES_00, GIT_PATCH_0, JMQ_SPF_NEUTRAL, KAM_DMARC_STATUS, RCVD_IN_BARRACUDACENTRAL, RCVD_IN_DNSWL_BLOCKED, RCVD_IN_VALIDITY_RPBL_BLOCKED, RCVD_IN_VALIDITY_SAFE_BLOCKED, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on sourceware.org X-BeenThere: elfutils-devel@sourceware.org X-Mailman-Version: 2.1.30 Precedence: list List-Id: Elfutils-devel mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: elfutils-devel-bounces~patchwork=sourceware.org@sourceware.org libdwfl tries to model the loading of modules (executables, shared libraries, the linux kernel and/or kernel modules) in memory. Depending on the (offline) load address it then also applies (simple) relocations for ET_REL (object files or kernel modules). Load addresses are normally represented through phdr segments for ET_EXEC or ET_DYN, but for ET_REL files (which don't have phdrs) the Elf section sh_addr fields are used. The sh_addr fields of the ET_REL Elf images are updated in two places __libdwfl_elf_address_range (through __libdwfl_report_elf) and __libdwfl_relocate (through dwfl_module_getelf and dwfl_module_getdwarf). Both rely on sh_addr being zero if no load address has been set yet, so the address layout for each (SHF_ALLOC) section is done only once. Recent linux kernels use a linker script that does set the sh_addr fields to (random, linker assigned) non-zero addresses. See commit 1ba9f8979426 ("vmlinux.lds: Unify TEXT_MAIN, DATA_MAIN, and related macros"). The sh_addr values seems unnecessary, but because they aren't zero anymore our layout/relocation code doesn't know it still has to figure out a "real" load value for these sections. Introduce __libdwfl_reset_sh_addr which resets all sh_addr fields to zero for SHF_ALLOC sections in ET_REL files. We don't call __libdwfl_reset_sh_addr on aux_sym files (from .gnu_debugdata) and when constructing an Elf from a core file. In all other cases we know (or assume) that the Elf file is being opened through libdw_open_elf (called indirectly through __libdw_open_elf, __libdwfl_report_offline, dwfl_module_getelf and dwfl_module_getdwarf) This technically changes the Elf that goes through libdwfl, but we would already update the sh_addr fields for ET_REL Elf sections in memory anyway to represent the load address as libdwfl would see them. So this isn't really a change in behavior (it just might update the sh_addr field twice). * libdwfl/libdwflP.h (__libdwfl_reset_sh_addr): Define new internal function. * libdwfl/relocate.c (__libdwfl_reset_sh_addr): New internal function. * libdwfl/open.c (libdw_open_elf): Call __libdwfl_reset_sh_addr. * libdwfl/dwfl_module_getdwarf.c (open_elf_file): Add comment. (find_aux_sym): Likewise. (find_dw): Likewise. * libdwfl/dwfl_segment_report_module.c (dwfl_segment_report_module): Likewise. * libdwfl/dwfl_report_elf.c (__libdwfl_report_elf): Likewise. Signed-off-by: Mark Wielaard --- libdwfl/dwfl_module_getdwarf.c | 3 +++ libdwfl/dwfl_report_elf.c | 3 +++ libdwfl/dwfl_segment_report_module.c | 2 ++ libdwfl/libdwflP.h | 4 ++++ libdwfl/open.c | 2 ++ libdwfl/relocate.c | 32 ++++++++++++++++++++++++++++ 6 files changed, 46 insertions(+) diff --git a/libdwfl/dwfl_module_getdwarf.c b/libdwfl/dwfl_module_getdwarf.c index 135132d69178..430c32367d01 100644 --- a/libdwfl/dwfl_module_getdwarf.c +++ b/libdwfl/dwfl_module_getdwarf.c @@ -57,6 +57,7 @@ open_elf_file (Elf **elf, int *fd, char **name) if (*fd < 0) return CBFAIL; + /* This will call __libdwfl_reset_sh_addr. */ return __libdw_open_file (fd, elf, true, false); } else if (unlikely (elf_kind (*elf) != ELF_K_ELF)) @@ -1004,6 +1005,7 @@ find_aux_sym (Dwfl_Module *mod __attribute__ ((unused)), free (buffer); else { + /* We don't call __libdwfl_reset_sh_addr here, should we? */ mod->aux_sym.elf = elf_memory (buffer, size); if (mod->aux_sym.elf == NULL) free (buffer); @@ -1423,6 +1425,7 @@ find_dw (Dwfl_Module *mod) switch (mod->dwerr) { case DWFL_E_NOERROR: + /* main.elf already should have had __libdwfl_reset_sh_addr called. */ mod->debug.elf = mod->main.elf; mod->debug.address_sync = mod->main.address_sync; diff --git a/libdwfl/dwfl_report_elf.c b/libdwfl/dwfl_report_elf.c index a76d36816c6f..7173375127a8 100644 --- a/libdwfl/dwfl_report_elf.c +++ b/libdwfl/dwfl_report_elf.c @@ -268,6 +268,9 @@ __libdwfl_report_elf (Dwfl *dwfl, const char *name, const char *file_name, /* Preinstall the open ELF handle for the module. */ if (m->main.elf == NULL) { + /* We assume all calls to __libdwfl_report_elf got their Elf + through __libdw_open_file which already called + __libdwfl_reset_sh_addr. */ m->main.elf = elf; m->main.vaddr = vaddr; m->main.address_sync = address_sync; diff --git a/libdwfl/dwfl_segment_report_module.c b/libdwfl/dwfl_segment_report_module.c index f2f866c2c70f..5ad318c78b09 100644 --- a/libdwfl/dwfl_segment_report_module.c +++ b/libdwfl/dwfl_segment_report_module.c @@ -1072,6 +1072,8 @@ dwfl_segment_report_module (Dwfl *dwfl, int ndx, const char *name, elf->flags |= ELF_F_MALLOCED; } + /* Elf comes through a core file, so cannot be an ET_REL. Don't call + __libdwfl_reset_sh_addr. */ if (elf != NULL && mod->main.elf == NULL) { /* Install the file in the module. */ diff --git a/libdwfl/libdwflP.h b/libdwfl/libdwflP.h index 6e394c262010..7a181029c529 100644 --- a/libdwfl/libdwflP.h +++ b/libdwfl/libdwflP.h @@ -482,6 +482,10 @@ extern void __libdwfl_module_free (Dwfl_Module *mod) internal_function; /* Find the main ELF file, update MOD->elferr and/or MOD->main.elf. */ extern void __libdwfl_getelf (Dwfl_Module *mod) internal_function; +/* Reset all sh_addr fields to zero for SHF_ALLOC sections if the Elf + handle is an ET_REL. Called from libdw_open_elf. */ +extern void __libdwfl_reset_sh_addr (Elf *elf) internal_function; + /* Process relocations in debugging sections in an ET_REL file. FILE must be opened with ELF_C_READ_MMAP_PRIVATE or ELF_C_READ, to make it possible to relocate the data in place (or ELF_C_RDWR or diff --git a/libdwfl/open.c b/libdwfl/open.c index e66595fba097..86a3540f2af3 100644 --- a/libdwfl/open.c +++ b/libdwfl/open.c @@ -190,6 +190,8 @@ libdw_open_elf (int *fdp, Elf **elfp, bool close_on_fail, bool archive_ok, *fdp = -1; } + __libdwfl_reset_sh_addr (elf); + *elfp = elf; return error; } diff --git a/libdwfl/relocate.c b/libdwfl/relocate.c index 0497bd4fe1f3..dc9f45e64f3f 100644 --- a/libdwfl/relocate.c +++ b/libdwfl/relocate.c @@ -1,5 +1,6 @@ /* Relocate debug information. Copyright (C) 2005-2011, 2014, 2016, 2018 Red Hat, Inc. + Copyright (C) 2026 Mark J. Wielaard This file is part of elfutils. This file is free software; you can redistribute it and/or modify @@ -37,6 +38,37 @@ typedef uint8_t GElf_Byte; +/* Reset all sh_addr fields to zero for SHF_ALLOC sections if the Elf + handle is an ET_REL. Called from libdw_open_elf. */ + +void +internal_function +__libdwfl_reset_sh_addr (Elf *elf) +{ + if (elf == NULL || elf->kind != ELF_K_ELF) + return; + + GElf_Ehdr ehdr_mem; + GElf_Ehdr *ehdr = gelf_getehdr (elf, &ehdr_mem); + if (ehdr == NULL || ehdr->e_type != ET_REL) + return; + + Elf_Scn *scn = NULL; + while ((scn = elf_nextscn (elf, scn)) != NULL) + { + GElf_Shdr shdr_mem; + GElf_Shdr *shdr = gelf_getshdr (scn, &shdr_mem); + if (shdr != NULL) + { + if ((shdr->sh_flags & SHF_ALLOC) != 0 && shdr->sh_addr != 0) + { + shdr_mem.sh_addr = 0; + gelf_update_shdr (scn, &shdr_mem); + } + } + } +} + /* Adjust *VALUE to add the load address of the SHNDX section. We update the section header in place to cache the result. */