From patchwork Wed May 23 19:54:17 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thomas Garnier X-Patchwork-Id: 10422299 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id ED3A160224 for ; Wed, 23 May 2018 20:00:51 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id DAFAA291D0 for ; Wed, 23 May 2018 20:00:51 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id CEA01291DC; Wed, 23 May 2018 20:00:51 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-12.8 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,MAILING_LIST_MULTI,RCVD_IN_DNSWL_MED, USER_IN_DEF_DKIM_WL autolearn=ham version=3.3.1 Received: from mother.openwall.net (mother.openwall.net [195.42.179.200]) by mail.wl.linuxfoundation.org (Postfix) with SMTP id 47898291DB for ; Wed, 23 May 2018 20:00:50 +0000 (UTC) Received: (qmail 25623 invoked by uid 550); 23 May 2018 19:56:26 -0000 Mailing-List: contact kernel-hardening-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-ID: Delivered-To: mailing list kernel-hardening@lists.openwall.com Received: (qmail 24376 invoked from network); 23 May 2018 19:56:13 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=6L/OmSajWN9TQKps8HIRaaC4bJLB3PZawt8YKLIVIPE=; b=HlqUFo8LWr4zCV1qgTK25c7vWzT01FM96NWJ/Z/pGO5grv5jT0junmV5JoY8a/cxjs SvwF/GFf/al7vyux01PoV/0uCvN2p1sX1tumCedJWJMfERCCY/xubwzhNyIzYenLzW7M xzRa372b0inFxqyJtPSGzbavSNBM6jx/TVlR/lX9Z5OxOmoegScF4txYwHKhixdzMI6r 8gInH1Kwy7llUkNhNdA1B+KQqVkeFO4cXzu2sbUW9CSeA8EzOr9RyU5FLFbz4RbHxxft 6f7L/cnzDk6g/0xNGdY8H46fWuc6NksstRqFWVk6enXDLirv20ZpXuWUNr6s5v0IYxUH FDQQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references; bh=6L/OmSajWN9TQKps8HIRaaC4bJLB3PZawt8YKLIVIPE=; b=prIC9No3Bw+AquY8Znf/LLecKttoiZ0N/oM6/okQ/6Y3iLoLCtK7yaHXvst7poa0cC 2HVUt2Hz9zkv6iwGwv4H+gwMuSg1pkJoySALqivTIdTDe2N9QhxTnH4gK0iByCCuP816 OkWFTwdIHTH9UylPTFbv9JUdzmuyGcwVzIYqi1PnqwcSlRoKMHfemxxBL+cdgdZcZVCy a4VlZ2m3J3q9eeKVl/kpgguR8BBg/7uElHg26nFiT/02p7al+0+mNU82Ae33lamEv76s XUuo41Vx4msUaDcSVhH+nUtPJqDgxKTXHBnhBVe+sv8p/d2kmd+9QFKwQh4ZshnWYUpW BjxQ== X-Gm-Message-State: ALKqPwdhXZy350Hfzcx4eX+Nlx+m75t/Mnu+BUH1kqHLtt3OeGmoSKky lb4NQ8zHgoGh+XRSFQjl/ZslJA== X-Google-Smtp-Source: AB8JxZo3KVucNg7O9JiulMp33s0N0xSgm3zsAlvM9k0wQLK5mitBLXSKcu6iiHTMaBm64rfWQLGkWA== X-Received: by 2002:a63:a807:: with SMTP id o7-v6mr3304472pgf.83.1527105360984; Wed, 23 May 2018 12:56:00 -0700 (PDT) From: Thomas Garnier To: Herbert Xu , "David S . Miller" , Thomas Gleixner , Ingo Molnar , "H . Peter Anvin" , Peter Zijlstra , Josh Poimboeuf , Greg Kroah-Hartman , Thomas Garnier , Philippe Ombredanne , Kate Stewart , Arnaldo Carvalho de Melo , Yonghong Song , Andrey Ryabinin , Kees Cook , Tom Lendacky , "Kirill A . Shutemov" , Andy Lutomirski , Dominik Brodowski , Borislav Petkov , Borislav Petkov , "Rafael J . Wysocki" , Len Brown , Pavel Machek , Juergen Gross , Alok Kataria , Steven Rostedt , Jan Kiszka , Tejun Heo , Christoph Lameter , Dennis Zhou , Boris Ostrovsky , Alexey Dobriyan , Masami Hiramatsu , Cao jin , Francis Deslauriers , "Paul E . McKenney" , Nicolas Pitre , Andrew Morton , Randy Dunlap , "Luis R . Rodriguez" , Arnd Bergmann , Christopher Li , Jason Baron , Mika Westerberg , Lukas Wunner , Dou Liyang , Sergey Senozhatsky , Petr Mladek , Masahiro Yamada , Ingo Molnar , Nicholas Piggin , "H . J . Lu" , Paolo Bonzini , =?UTF-8?q?Radim=20Kr=C4=8Dm=C3=A1=C5=99?= , Joerg Roedel , David Woodhouse , Dave Hansen , Rik van Riel , Jia Zhang , Ricardo Neri , Jonathan Corbet , Jan Beulich , Matthias Kaehlcke , Baoquan He , =?UTF-8?q?Jan=20H=20=2E=20Sch=C3=B6nherr?= , Daniel Micay Cc: x86@kernel.org, linux-crypto@vger.kernel.org, linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org, virtualization@lists.linux-foundation.org, xen-devel@lists.xenproject.org, linux-arch@vger.kernel.org, linux-sparse@vger.kernel.org, kvm@vger.kernel.org, linux-doc@vger.kernel.org, kernel-hardening@lists.openwall.com Subject: [PATCH v3 23/27] x86/modules: Adapt module loading for PIE support Date: Wed, 23 May 2018 12:54:17 -0700 Message-Id: <20180523195421.180248-24-thgarnie@google.com> X-Mailer: git-send-email 2.17.0.441.gb46fe60e1d-goog In-Reply-To: <20180523195421.180248-1-thgarnie@google.com> References: <20180523195421.180248-1-thgarnie@google.com> X-Virus-Scanned: ClamAV using ClamSMTP Adapt module loading to support PIE relocations. Generate dynamic GOT if a symbol requires it but no entry exist in the kernel GOT. Position Independent Executable (PIE) support will allow to extended the KASLR randomization range below the -2G memory limit. Signed-off-by: Thomas Garnier --- arch/x86/Makefile | 4 + arch/x86/include/asm/module.h | 11 ++ arch/x86/include/asm/sections.h | 4 + arch/x86/kernel/module.c | 181 +++++++++++++++++++++++++++++++- arch/x86/kernel/module.lds | 3 + 5 files changed, 198 insertions(+), 5 deletions(-) create mode 100644 arch/x86/kernel/module.lds diff --git a/arch/x86/Makefile b/arch/x86/Makefile index 277ffc57ae13..20bb6cbd8938 100644 --- a/arch/x86/Makefile +++ b/arch/x86/Makefile @@ -134,7 +134,11 @@ else KBUILD_CFLAGS += $(cflags-y) KBUILD_CFLAGS += -mno-red-zone +ifdef CONFIG_X86_PIE + KBUILD_LDFLAGS_MODULE += -T $(srctree)/arch/x86/kernel/module.lds +else KBUILD_CFLAGS += -mcmodel=kernel +endif # -funit-at-a-time shrinks the kernel .text considerably # unfortunately it makes reading oopses harder. diff --git a/arch/x86/include/asm/module.h b/arch/x86/include/asm/module.h index 7948a17febb4..68ff05e14288 100644 --- a/arch/x86/include/asm/module.h +++ b/arch/x86/include/asm/module.h @@ -5,12 +5,23 @@ #include #include +#ifdef CONFIG_X86_PIE +struct mod_got_sec { + struct elf64_shdr *got; + int got_num_entries; + int got_max_entries; +}; +#endif + struct mod_arch_specific { #ifdef CONFIG_UNWINDER_ORC unsigned int num_orcs; int *orc_unwind_ip; struct orc_entry *orc_unwind; #endif +#ifdef CONFIG_X86_PIE + struct mod_got_sec core; +#endif }; #ifdef CONFIG_X86_64 diff --git a/arch/x86/include/asm/sections.h b/arch/x86/include/asm/sections.h index da3d98bb2bcb..89b3a95c8d11 100644 --- a/arch/x86/include/asm/sections.h +++ b/arch/x86/include/asm/sections.h @@ -17,4 +17,8 @@ extern char __entry_trampoline_start[], __entry_trampoline_end[]; extern char __start_got[], __end_got[]; #endif +#if defined(CONFIG_X86_PIE) +extern char __start_got[], __end_got[]; +#endif + #endif /* _ASM_X86_SECTIONS_H */ diff --git a/arch/x86/kernel/module.c b/arch/x86/kernel/module.c index f58336af095c..88895f3d474b 100644 --- a/arch/x86/kernel/module.c +++ b/arch/x86/kernel/module.c @@ -30,6 +30,7 @@ #include #include #include +#include #include #include @@ -77,6 +78,173 @@ static unsigned long int get_module_load_offset(void) } #endif +#ifdef CONFIG_X86_PIE +static u64 find_got_kernel_entry(Elf64_Sym *sym, const Elf64_Rela *rela) +{ + u64 *pos; + + for (pos = (u64*)__start_got; pos < (u64*)__end_got; pos++) { + if (*pos == sym->st_value) + return (u64)pos + rela->r_addend; + } + + return 0; +} + +static u64 module_emit_got_entry(struct module *mod, void *loc, + const Elf64_Rela *rela, Elf64_Sym *sym) +{ + struct mod_got_sec *gotsec = &mod->arch.core; + u64 *got = (u64*)gotsec->got->sh_addr; + int i = gotsec->got_num_entries; + u64 ret; + + /* Check if we can use the kernel GOT */ + ret = find_got_kernel_entry(sym, rela); + if (ret) + return ret; + + got[i] = sym->st_value; + + /* + * Check if the entry we just created is a duplicate. Given that the + * relocations are sorted, this will be the last entry we allocated. + * (if one exists). + */ + if (i > 0 && got[i] == got[i - 2]) { + ret = (u64)&got[i - 1]; + } else { + gotsec->got_num_entries++; + BUG_ON(gotsec->got_num_entries > gotsec->got_max_entries); + ret = (u64)&got[i]; + } + + return ret + rela->r_addend; +} + +#define cmp_3way(a,b) ((a) < (b) ? -1 : (a) > (b)) + +static int cmp_rela(const void *a, const void *b) +{ + const Elf64_Rela *x = a, *y = b; + int i; + + /* sort by type, symbol index and addend */ + i = cmp_3way(ELF64_R_TYPE(x->r_info), ELF64_R_TYPE(y->r_info)); + if (i == 0) + i = cmp_3way(ELF64_R_SYM(x->r_info), ELF64_R_SYM(y->r_info)); + if (i == 0) + i = cmp_3way(x->r_addend, y->r_addend); + return i; +} + +static bool duplicate_rel(const Elf64_Rela *rela, int num) +{ + /* + * Entries are sorted by type, symbol index and addend. That means + * that, if a duplicate entry exists, it must be in the preceding + * slot. + */ + return num > 0 && cmp_rela(rela + num, rela + num - 1) == 0; +} + +static unsigned int count_gots(Elf64_Sym *syms, Elf64_Rela *rela, int num) +{ + unsigned int ret = 0; + Elf64_Sym *s; + int i; + + for (i = 0; i < num; i++) { + switch (ELF64_R_TYPE(rela[i].r_info)) { + case R_X86_64_GOTPCREL: + s = syms + ELF64_R_SYM(rela[i].r_info); + + /* + * Use the kernel GOT when possible, else reserve a + * custom one for this module. + */ + if (!duplicate_rel(rela, i) && + !find_got_kernel_entry(s, rela + i)) + ret++; + break; + } + } + return ret; +} + +/* + * Generate GOT entries for GOTPCREL relocations that do not exists in the + * kernel GOT. Based on arm64 module-plts implementation. + */ +int module_frob_arch_sections(Elf_Ehdr *ehdr, Elf_Shdr *sechdrs, + char *secstrings, struct module *mod) +{ + unsigned long gots = 0; + Elf_Shdr *symtab = NULL; + Elf64_Sym *syms = NULL; + char *strings, *name; + int i; + + /* + * Find the empty .got section so we can expand it to store the PLT + * entries. Record the symtab address as well. + */ + for (i = 0; i < ehdr->e_shnum; i++) { + if (!strcmp(secstrings + sechdrs[i].sh_name, ".got")) { + mod->arch.core.got = sechdrs + i; + } else if (sechdrs[i].sh_type == SHT_SYMTAB) { + symtab = sechdrs + i; + syms = (Elf64_Sym *)symtab->sh_addr; + } + } + + if (!mod->arch.core.got) { + pr_err("%s: module GOT section missing\n", mod->name); + return -ENOEXEC; + } + if (!syms) { + pr_err("%s: module symtab section missing\n", mod->name); + return -ENOEXEC; + } + + for (i = 0; i < ehdr->e_shnum; i++) { + Elf64_Rela *rels = (void *)ehdr + sechdrs[i].sh_offset; + int numrels = sechdrs[i].sh_size / sizeof(Elf64_Rela); + + if (sechdrs[i].sh_type != SHT_RELA) + continue; + + /* sort by type, symbol index and addend */ + sort(rels, numrels, sizeof(Elf64_Rela), cmp_rela, NULL); + + gots += count_gots(syms, rels, numrels); + } + + mod->arch.core.got->sh_type = SHT_NOBITS; + mod->arch.core.got->sh_flags = SHF_ALLOC; + mod->arch.core.got->sh_addralign = L1_CACHE_BYTES; + mod->arch.core.got->sh_size = (gots + 1) * sizeof(u64); + mod->arch.core.got_num_entries = 0; + mod->arch.core.got_max_entries = gots; + + /* + * If a _GLOBAL_OFFSET_TABLE_ symbol exists, make it absolute for + * modules to correctly reference it. Similar to s390 implementation. + */ + strings = (void *) ehdr + sechdrs[symtab->sh_link].sh_offset; + for (i = 0; i < symtab->sh_size/sizeof(Elf_Sym); i++) { + if (syms[i].st_shndx != SHN_UNDEF) + continue; + name = strings + syms[i].st_name; + if (!strcmp(name, "_GLOBAL_OFFSET_TABLE_")) { + syms[i].st_shndx = SHN_ABS; + break; + } + } + return 0; +} +#endif + void *module_alloc(unsigned long size) { void *p; @@ -190,16 +358,20 @@ int apply_relocate_add(Elf64_Shdr *sechdrs, if ((s64)val != *(s32 *)loc) goto overflow; break; +#ifdef CONFIG_X86_PIE + case R_X86_64_GOTPCREL: + val = module_emit_got_entry(me, loc, rel + i, sym); + /* fallthrough */ +#endif case R_X86_64_PC32: case R_X86_64_PLT32: if (*(u32 *)loc != 0) goto invalid_relocation; val -= (u64)loc; *(u32 *)loc = val; -#if 0 - if ((s64)val != *(s32 *)loc) + if (IS_ENABLED(CONFIG_X86_PIE) && + (s64)val != *(s32 *)loc) goto overflow; -#endif break; default: pr_err("%s: Unknown rela relocation: %llu\n", @@ -217,8 +389,7 @@ int apply_relocate_add(Elf64_Shdr *sechdrs, overflow: pr_err("overflow in relocation type %d val %Lx\n", (int)ELF64_R_TYPE(rel[i].r_info), val); - pr_err("`%s' likely not compiled with -mcmodel=kernel\n", - me->name); + pr_err("`%s' likely too far from the kernel\n", me->name); return -ENOEXEC; } #endif diff --git a/arch/x86/kernel/module.lds b/arch/x86/kernel/module.lds new file mode 100644 index 000000000000..fd6e95a4b454 --- /dev/null +++ b/arch/x86/kernel/module.lds @@ -0,0 +1,3 @@ +SECTIONS { + .got (NOLOAD) : { BYTE(0) } +}