Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751718AbcDTWrB (ORCPT ); Wed, 20 Apr 2016 18:47:01 -0400 Received: from mail-pa0-f42.google.com ([209.85.220.42]:36260 "EHLO mail-pa0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751001AbcDTWqg (ORCPT ); Wed, 20 Apr 2016 18:46:36 -0400 From: Kees Cook To: James Morris Cc: Kees Cook , Mimi Zohar , Joe Perches , Andy Shevchenko , Andrew Morton , "Serge E. Hallyn" , Jonathan Corbet , Kalle Valo , Mauro Carvalho Chehab , Guenter Roeck , Jiri Slaby , Paul Moore , Stephen Smalley , Casey Schaufler , Andreas Gruenbacher , Rasmus Villemoes , Ulf Hansson , Vitaly Kuznetsov , linux-security-module@vger.kernel.org, linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org Subject: [PATCH 6/6] LSM: LoadPin for kernel file loading restrictions Date: Wed, 20 Apr 2016 15:46:28 -0700 Message-Id: <1461192388-13900-7-git-send-email-keescook@chromium.org> X-Mailer: git-send-email 2.6.3 In-Reply-To: <1461192388-13900-1-git-send-email-keescook@chromium.org> References: <1461192388-13900-1-git-send-email-keescook@chromium.org> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 11377 Lines: 345 This LSM enforces that kernel-loaded files (modules, firmware, etc) must all come from the same filesystem, with the expectation that such a filesystem is backed by a read-only device such as dm-verity or CDROM. This allows systems that have a verified and/or unchangeable filesystem to enforce module and firmware loading restrictions without needing to sign the files individually. Signed-off-by: Kees Cook Acked-by: Serge Hallyn --- Documentation/security/LoadPin.txt | 17 ++++ MAINTAINERS | 6 ++ include/linux/lsm_hooks.h | 5 + security/Kconfig | 1 + security/Makefile | 2 + security/loadpin/Kconfig | 10 ++ security/loadpin/Makefile | 1 + security/loadpin/loadpin.c | 190 +++++++++++++++++++++++++++++++++++++ security/security.c | 1 + 9 files changed, 233 insertions(+) create mode 100644 Documentation/security/LoadPin.txt create mode 100644 security/loadpin/Kconfig create mode 100644 security/loadpin/Makefile create mode 100644 security/loadpin/loadpin.c diff --git a/Documentation/security/LoadPin.txt b/Documentation/security/LoadPin.txt new file mode 100644 index 000000000000..e11877f5d3d4 --- /dev/null +++ b/Documentation/security/LoadPin.txt @@ -0,0 +1,17 @@ +LoadPin is a Linux Security Module that ensures all kernel-loaded files +(modules, firmware, etc) all originate from the same filesystem, with +the expectation that such a filesystem is backed by a read-only device +such as dm-verity or CDROM. This allows systems that have a verified +and/or unchangeable filesystem to enforce module and firmware loading +restrictions without needing to sign the files individually. + +The LSM is selectable at build-time with CONFIG_SECURITY_LOADPIN, and +can be controlled at boot-time with the kernel command line option +"loadpin.enabled". By default, it is enabled, but can be disabled at +boot ("loadpin.enabled=0"). + +LoadPin starts pinning when it sees the first file loaded. If the +block device backing the filesystem is not read-only, a sysctl is +created to toggle pinning: /proc/sys/kernel/loadpin/enabled. (Having +a mutable filesystem means pinning is mutable too, but having the +sysctl allows for easy testing on systems with a mutable filesystem.) diff --git a/MAINTAINERS b/MAINTAINERS index 1d5b4becab6f..6042f6318c8f 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -9971,6 +9971,12 @@ T: git git://git.kernel.org/pub/scm/linux/kernel/git/jj/apparmor-dev.git S: Supported F: security/apparmor/ +LOADPIN SECURITY MODULE +M: Kees Cook +T: git git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git lsm/loadpin +S: Supported +F: security/loadpin/ + YAMA SECURITY MODULE M: Kees Cook T: git git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git yama/tip diff --git a/include/linux/lsm_hooks.h b/include/linux/lsm_hooks.h index cdee11cbcdf1..f3402aab1927 100644 --- a/include/linux/lsm_hooks.h +++ b/include/linux/lsm_hooks.h @@ -1893,5 +1893,10 @@ extern void __init yama_add_hooks(void); #else static inline void __init yama_add_hooks(void) { } #endif +#ifdef CONFIG_SECURITY_LOADPIN +void __init loadpin_add_hooks(void); +#else +static inline void loadpin_add_hooks(void) { }; +#endif #endif /* ! __LINUX_LSM_HOOKS_H */ diff --git a/security/Kconfig b/security/Kconfig index e45237897b43..176758cdfa57 100644 --- a/security/Kconfig +++ b/security/Kconfig @@ -122,6 +122,7 @@ source security/selinux/Kconfig source security/smack/Kconfig source security/tomoyo/Kconfig source security/apparmor/Kconfig +source security/loadpin/Kconfig source security/yama/Kconfig source security/integrity/Kconfig diff --git a/security/Makefile b/security/Makefile index c9bfbc84ff50..f2d71cdb8e19 100644 --- a/security/Makefile +++ b/security/Makefile @@ -8,6 +8,7 @@ subdir-$(CONFIG_SECURITY_SMACK) += smack subdir-$(CONFIG_SECURITY_TOMOYO) += tomoyo subdir-$(CONFIG_SECURITY_APPARMOR) += apparmor subdir-$(CONFIG_SECURITY_YAMA) += yama +subdir-$(CONFIG_SECURITY_LOADPIN) += loadpin # always enable default capabilities obj-y += commoncap.o @@ -22,6 +23,7 @@ obj-$(CONFIG_AUDIT) += lsm_audit.o obj-$(CONFIG_SECURITY_TOMOYO) += tomoyo/ obj-$(CONFIG_SECURITY_APPARMOR) += apparmor/ obj-$(CONFIG_SECURITY_YAMA) += yama/ +obj-$(CONFIG_SECURITY_LOADPIN) += loadpin/ obj-$(CONFIG_CGROUP_DEVICE) += device_cgroup.o # Object integrity file lists diff --git a/security/loadpin/Kconfig b/security/loadpin/Kconfig new file mode 100644 index 000000000000..c668ac4eda65 --- /dev/null +++ b/security/loadpin/Kconfig @@ -0,0 +1,10 @@ +config SECURITY_LOADPIN + bool "Pin load of kernel files (modules, fw, etc) to one filesystem" + depends on SECURITY && BLOCK + help + Any files read through the kernel file reading interface + (kernel modules, firmware, kexec images, security policy) will + be pinned to the first filesystem used for loading. Any files + that come from other filesystems will be rejected. This is best + used on systems without an initrd that have a root filesystem + backed by a read-only device such as dm-verity or a CDROM. diff --git a/security/loadpin/Makefile b/security/loadpin/Makefile new file mode 100644 index 000000000000..c2d77f83037b --- /dev/null +++ b/security/loadpin/Makefile @@ -0,0 +1 @@ +obj-$(CONFIG_SECURITY_LOADPIN) += loadpin.o diff --git a/security/loadpin/loadpin.c b/security/loadpin/loadpin.c new file mode 100644 index 000000000000..e4debae3c4d6 --- /dev/null +++ b/security/loadpin/loadpin.c @@ -0,0 +1,190 @@ +/* + * Module and Firmware Pinning Security Module + * + * Copyright 2011-2016 Google Inc. + * + * Author: Kees Cook + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#define pr_fmt(fmt) "LoadPin: " fmt + +#include +#include +#include +#include +#include +#include +#include /* current */ +#include + +static void report_load(const char *origin, struct file *file, char *operation) +{ + char *cmdline, *pathname; + + pathname = kstrdup_quotable_file(file, GFP_KERNEL); + cmdline = kstrdup_quotable_cmdline(current, GFP_KERNEL); + + pr_notice("%s %s obj=%s%s%s pid=%d cmdline=%s%s%s\n", + origin, operation, + (pathname && pathname[0] != '<') ? "\"" : "", + pathname, + (pathname && pathname[0] != '<') ? "\"" : "", + task_pid_nr(current), + cmdline ? "\"" : "", cmdline, cmdline ? "\"" : ""); + + kfree(cmdline); + kfree(pathname); +} + +static int enabled = 1; +static struct super_block *pinned_root; +static DEFINE_SPINLOCK(pinned_root_spinlock); + +#ifdef CONFIG_SYSCTL +static int zero; +static int one = 1; + +static struct ctl_path loadpin_sysctl_path[] = { + { .procname = "kernel", }, + { .procname = "loadpin", }, + { } +}; + +static struct ctl_table loadpin_sysctl_table[] = { + { + .procname = "enabled", + .data = &enabled, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &zero, + .extra2 = &one, + }, + { } +}; + +/* + * This must be called after early kernel init, since then the rootdev + * is available. + */ +static void check_pinning_enforcement(struct super_block *mnt_sb) +{ + bool ro = false; + + /* + * If load pinning is not enforced via a read-only block + * device, allow sysctl to change modes for testing. + */ + if (mnt_sb->s_bdev) { + ro = bdev_read_only(mnt_sb->s_bdev); + pr_info("dev(%u,%u): %s\n", + MAJOR(mnt_sb->s_bdev->bd_dev), + MINOR(mnt_sb->s_bdev->bd_dev), + ro ? "read-only" : "writable"); + } else + pr_info("mnt_sb lacks block device, treating as: writable\n"); + + if (!ro) { + if (!register_sysctl_paths(loadpin_sysctl_path, + loadpin_sysctl_table)) + pr_notice("sysctl registration failed!\n"); + else + pr_info("load pinning can be disabled.\n"); + } else + pr_info("load pinning engaged.\n"); +} +#else +static void check_pinning_enforcement(struct super_block *mnt_sb) +{ + pr_info("load pinning engaged.\n"); +} +#endif + +static void loadpin_sb_free_security(struct super_block *mnt_sb) +{ + /* + * When unmounting the filesystem we were using for load + * pinning, we acknowledge the superblock release, but make sure + * no other modules or firmware can be loaded. + */ + if (!IS_ERR_OR_NULL(pinned_root) && mnt_sb == pinned_root) { + pinned_root = ERR_PTR(-EIO); + pr_info("umount pinned fs: refusing further loads\n"); + } +} + +static int loadpin_read_file(struct file *file, enum kernel_read_file_id id) +{ + struct super_block *load_root; + const char *origin = kernel_read_file_id_str(id); + + /* This handles the older init_module API that has a NULL file. */ + if (!file) { + if (!enabled) { + report_load(origin, NULL, "old-api-pinning-ignored"); + return 0; + } + + report_load(origin, NULL, "old-api-denied"); + return -EPERM; + } + + load_root = file->f_path.mnt->mnt_sb; + + /* First loaded module/firmware defines the root for all others. */ + spin_lock(&pinned_root_spinlock); + /* + * pinned_root is only NULL at startup. Otherwise, it is either + * a valid reference, or an ERR_PTR. + */ + if (!pinned_root) { + pinned_root = load_root; + /* + * Unlock now since it's only pinned_root we care about. + * In the worst case, we will (correctly) report pinning + * failures before we have announced that pinning is + * enabled. This would be purely cosmetic. + */ + spin_unlock(&pinned_root_spinlock); + check_pinning_enforcement(pinned_root); + report_load(origin, file, "pinned"); + } else { + spin_unlock(&pinned_root_spinlock); + } + + if (IS_ERR_OR_NULL(pinned_root) || load_root != pinned_root) { + if (unlikely(!enabled)) { + report_load(origin, file, "pinning-ignored"); + return 0; + } + + report_load(origin, file, "denied"); + return -EPERM; + } + + return 0; +} + +static struct security_hook_list loadpin_hooks[] = { + LSM_HOOK_INIT(sb_free_security, loadpin_sb_free_security), + LSM_HOOK_INIT(kernel_read_file, loadpin_read_file), +}; + +void __init loadpin_add_hooks(void) +{ + pr_info("ready to pin (currently %sabled)", enabled ? "en" : "dis"); + security_add_hooks(loadpin_hooks, ARRAY_SIZE(loadpin_hooks)); +} + +/* Should not be mutable after boot, so not listed in sysfs (perm == 0). */ +module_param(enabled, int, 0); +MODULE_PARM_DESC(enabled, "Pin module/firmware loading (default: true)"); diff --git a/security/security.c b/security/security.c index 3644b0344d29..ce02178c892f 100644 --- a/security/security.c +++ b/security/security.c @@ -60,6 +60,7 @@ int __init security_init(void) */ capability_add_hooks(); yama_add_hooks(); + loadpin_add_hooks(); /* * Load all the remaining security modules. -- 2.6.3