Skip to contentSkip to conversation

Add trusted_for(2) (was O_MAYEXEC)

PATCH Superseded Mickaël Salaün 5 participants 12 messages Oct 8, 2021 – Oct 11, 2021

Subsystems (22)
Documentation Arm (32-bit) x86 VFS +18 more subsystems

v14 of 18 Superseded Latest is v18 →

lore
Share & tools
lore
#0 lore raw

v14 cover Add trusted_for(2) (was O_MAYEXEC)

Hi,

This patch series is mainly a rebase on v5.15-rc4 with some cosmetic changes suggested by Kees Cook. Andrew, can you please consider to merge this into your tree?

Overview
========

The final goal of this patch series is to enable the kernel to be a global policy manager by entrusting processes with access control at their level. To reach this goal, two complementary parts are required:
* user space needs to be able to know if it can trust some file descriptor content for a specific usage;
* and the kernel needs to make available some part of the policy configured by the system administrator.

Primary goal of trusted_for(2)
==============================

This new syscall enables user space to ask the kernel: is this file descriptor's content trusted to be used for this purpose? The set of usage currently only contains "execution", but other may follow (e.g. "configuration", "sensitive_data"). If the kernel identifies the file descriptor as trustworthy for this usage, user space should then take this information into account. The "execution" usage means that the content of the file descriptor is trusted according to the system policy to be executed by user space, which means that it interprets the content or (try to) maps it as executable memory.

A simple system-wide security policy can be enforced by the system administrator through a sysctl configuration consistent with the mount points or the file access rights. The documentation patch explains the prerequisites.

It is important to note that this can only enable to extend access control managed by the kernel. Hence it enables current access control mechanism to be extended and become a superset of what they can currently control. Indeed, the security policy could also be delegated to an LSM, either a MAC system or an integrity system. For instance, this is required to close a major IMA measurement/appraisal interpreter integrity gap by bringing the ability to check the use of scripts [1]. Other uses are expected, such as for magic-links [2], SGX integration [3], bpffs [4].

Complementary W^X protections can be brought by SELinux, IPE [5] and trampfd [6].

Prerequisite of its use
=======================

User space needs to adapt to take advantage of this new feature. For example, the PEP 578 [7] (Runtime Audit Hooks) enables Python 3.8 to be extended with policy enforcement points related to code interpretation, which can be used to align with the PowerShell audit features. Additional Python security improvements (e.g. a limited interpreter without -c, stdin piping of code) are on their way [8].

Examples
========

The initial idea comes from CLIP OS 4 and the original implementation has been used for more than 13 years:
https://github.com/clipos-archive/clipos4_doc
Chrome OS has a similar approach:
https://chromium.googlesource.com/chromiumos/docs/+/master/security/noexec_shell_scripts.md

Userland patches can be found here:
https://github.com/clipos-archive/clipos4_portage-overlay/search?q=O_MAYEXEC
Actually, there is more than the O_MAYEXEC changes (which matches this search) e.g., to prevent Python interactive execution. There are patches for Bash, Wine, Java (Icedtea), Busybox's ash, Perl and Python. There are also some related patches which do not directly rely on O_MAYEXEC but which restrict the use of browser plugins and extensions, which may be seen as scripts too:
https://github.com/clipos-archive/clipos4_portage-overlay/tree/master/www-client

An introduction to O_MAYEXEC was given at the Linux Security Summit Europe 2018 - Linux Kernel Security Contributions by ANSSI:
https://www.youtube.com/watch?v=chNjCRtPKQY&t=17m15s
The "write xor execute" principle was explained at Kernel Recipes 2018 - CLIP OS: a defense-in-depth OS:
https://www.youtube.com/watch?v=PjRE0uBtkHU&t=11m14s
See also a first LWN article about O_MAYEXEC and a new one about trusted_for(2) and its background:
* https://lwn.net/Articles/820000/
* https://lwn.net/Articles/832959/

This patch series can be applied on top of v5.10-rc6 . This can be tested with CONFIG_SYSCTL. I would really appreciate constructive comments on this patch series.

Previous series:
https://lore.kernel.org/r/[email protected]/

[1] https://lore.kernel.org/lkml/[email protected]/
[2] https://lore.kernel.org/lkml/[email protected]/
[3] https://lore.kernel.org/lkml/CALCETrVovr8XNZSroey7pHF46O=kj_c5D9K8h=z2T_cNrpvMig@mail.gmail.com/
[4] https://lore.kernel.org/lkml/CALCETrVeZ0eufFXwfhtaG_j+AdvbzEWE0M3wjXMWVEO7pj+xkw@mail.gmail.com/
[5] https://lore.kernel.org/lkml/[email protected]/
[6] https://lore.kernel.org/lkml/[email protected]/
[7] https://www.python.org/dev/peps/pep-0578/
[8] https://lore.kernel.org/lkml/[email protected]/

Regards,

Mickaël Salaün (3):
fs: Add trusted_for(2) syscall implementation and related sysctl
arch: Wire up trusted_for(2)
selftest/interpreter: Add tests for trusted_for(2) policies

 Documentation/admin-guide/sysctl/fs.rst       |  50 +++
 arch/alpha/kernel/syscalls/syscall.tbl        |   1 +
 arch/arm/tools/syscall.tbl                    |   1 +
 arch/arm64/include/asm/unistd.h               |   2 +-
 arch/arm64/include/asm/unistd32.h             |   2 +
 arch/ia64/kernel/syscalls/syscall.tbl         |   1 +
 arch/m68k/kernel/syscalls/syscall.tbl         |   1 +
 arch/microblaze/kernel/syscalls/syscall.tbl   |   1 +
 arch/mips/kernel/syscalls/syscall_n32.tbl     |   1 +
 arch/mips/kernel/syscalls/syscall_n64.tbl     |   1 +
 arch/mips/kernel/syscalls/syscall_o32.tbl     |   1 +
 arch/parisc/kernel/syscalls/syscall.tbl       |   1 +
 arch/powerpc/kernel/syscalls/syscall.tbl      |   1 +
 arch/s390/kernel/syscalls/syscall.tbl         |   1 +
 arch/sh/kernel/syscalls/syscall.tbl           |   1 +
 arch/sparc/kernel/syscalls/syscall.tbl        |   1 +
 arch/x86/entry/syscalls/syscall_32.tbl        |   1 +
 arch/x86/entry/syscalls/syscall_64.tbl        |   1 +
 arch/xtensa/kernel/syscalls/syscall.tbl       |   1 +
 fs/open.c                                     |  78 ++++
 include/linux/fs.h                            |   1 +
 include/linux/syscalls.h                      |   2 +
 include/uapi/asm-generic/unistd.h             |   4 +-
 include/uapi/linux/trusted-for.h              |  18 +
 kernel/sysctl.c                               |  12 +-
 tools/testing/selftests/Makefile              |   1 +
 .../testing/selftests/interpreter/.gitignore  |   2 +
 tools/testing/selftests/interpreter/Makefile  |  21 +
 tools/testing/selftests/interpreter/config    |   1 +
 .../selftests/interpreter/trust_policy_test.c | 362 ++++++++++++++++++
 30 files changed, 568 insertions(+), 4 deletions(-)
 create mode 100644 include/uapi/linux/trusted-for.h
 create mode 100644 tools/testing/selftests/interpreter/.gitignore
 create mode 100644 tools/testing/selftests/interpreter/Makefile
 create mode 100644 tools/testing/selftests/interpreter/config
 create mode 100644 tools/testing/selftests/interpreter/trust_policy_test.c

base-commit: 9e1ff307c779ce1f0f810c7ecce3d95bbae40896

Signature · 1 line
2.32.0
#1 lore raw

PATCH v14 1/3 fs: Add trusted_for(2) syscall implementation and related sysctl

From: Mickaël Salaün <[email protected]>

The trusted_for() syscall enables user space tasks to check that files are trusted to be executed or interpreted by user space. This may allow script interpreters to check execution permission before reading commands from a file, or dynamic linkers to allow shared object loading. This may be seen as a way for a trusted task (e.g. interpreter) to check the trustworthiness of files (e.g. scripts) before extending its control flow graph with new ones originating from these files.

The security policy is consistently managed by the kernel through the new sysctl: fs.trusted_for_policy . This enables system administrators to enforce two complementary security policies according to the installed system: enforce the noexec mount option, and enforce executable file permission. Indeed, because of compatibility with installed systems, only system administrators are able to check that this new enforcement is in line with the system mount points and file permissions.

For this to be possible, script interpreters must use trusted_for(2) with the TRUSTED_FOR_EXECUTION usage. To be fully effective, these interpreters also need to handle the other ways to execute code: command line parameters (e.g., option -e for Perl), module loading (e.g., option -m for Python), stdin, file sourcing, environment variables, configuration files, etc. According to the threat model, it may be acceptable to allow some script interpreters (e.g. Bash) to interpret commands from stdin, may it be a TTY or a pipe, because it may not be enough to (directly) perform syscalls.

Even without enforced security policy, user space interpreters can use this syscall to try as much as possible to enforce the system policy at their level, knowing that it will not break anything on running systems which do not care about this feature. However, on systems which want this feature enforced, there will be knowledgeable people (i.e. system administrator who configured fs.trusted_for_policy deliberately) to manage it.

Because trusted_for(2) is a mean to enforce a system-wide security policy (but not application-centric policies), it does not make sense for user space to check the sysctl value. Indeed, this new flag only enables to extend the system ability to enforce a policy thanks to (some trusted) user space collaboration. Moreover, additional security policies could be managed by LSMs. This is a best-effort approach from the application developer point of view:
https://lore.kernel.org/lkml/[email protected]/

trusted_for(2) with TRUSTED_FOR_EXECUTION should not be confused with the O_EXEC flag (for open) which is intended for execute-only, which obviously doesn't work for scripts. However, a similar behavior could be implemented in user space with O_PATH:
https://lore.kernel.org/lkml/[email protected]/

Being able to restrict execution also enables to protect the kernel by restricting arbitrary syscalls that an attacker could perform with a crafted binary or certain script languages. It also improves multilevel isolation by reducing the ability of an attacker to use side channels with specific code. These restrictions can natively be enforced for ELF binaries (with the noexec mount option) but require this kernel extension to properly handle scripts (e.g. Python, Perl). To get a consistent execution policy, additional memory restrictions should also be enforced (e.g. thanks to SELinux).

This is a new implementation of a patch initially written by Vincent Strubel for CLIP OS 4:
https://github.com/clipos-archive/src_platform_clip-patches/blob/f5cb330d6b684752e403b4e41b39f7004d88e561/1901_open_mayexec.patch
This patch has been used for more than 13 years with customized script interpreters. Some examples (with the original O_MAYEXEC) can be found here:
https://github.com/clipos-archive/clipos4_portage-overlay/search?q=O_MAYEXEC

---

Changes since v13:
* Rename sysctl from "trust_policy" to "trusted_for_policy" (suggested by Kees Cook).
* Add Acked-by Kees Cook.

Changes since v12:
* Update inode_permission() call to allign with commit 47291baa8ddf ("namei: make permission helpers idmapped mount aware").
* Switch from d_backing_inode(f.file->f_path.dentry) to file_inode(f.file).

Changes since v10:
* Add enum definition to syscalls.h .

Changes since v9:
* Rename the syscall to trusted_for(2) and the sysctl to fs.trust_policy
* Add a dedicated enum trusted_for_usage with include/uapi/linux/trusted-for.h
* Remove the extra MAY_INTROSPECTION_EXEC bit. LSMs can still implement this feature themselves.

Changes since v8:
* Add a dedicated syscall introspect_access() (requested by Al Viro).
* Rename MAY_INTERPRETED_EXEC to MAY_INTROSPECTION_EXEC .
* Rename the sysctl fs.interpreted_access to fs.introspection_policy .
* Update documentation.

Changes since v7:
* Replaces openat2/O_MAYEXEC with faccessat2/X_OK/AT_INTERPRETED . Switching to an FD-based syscall was suggested by Al Viro and Jann Horn.
* Handle special file descriptors.
* Add a compatibility mode for execute/read check.
* Move the sysctl policy from fs/namei.c to fs/open.c for the new faccessat2/AT_INTERPRETED.
* Rename the sysctl from fs.open_mayexec_enforce to fs.interpreted_access .
* Update documentation accordingly.

Changes since v6:
* Allow opening pipes, block devices and character devices with O_MAYEXEC when there is no enforced policy, but forbid any non-regular file opened with O_MAYEXEC otherwise (i.e. for any enforced policy).
* Add a paragraph about the non-regular files policy.
* Move path_noexec() calls out of the fast-path (suggested by Kees Cook).
* Do not set __FMODE_EXEC for now because of inconsistent behavior:
https://lore.kernel.org/lkml/202007160822.CCDB5478@keescook/
* Returns EISDIR when opening a directory with O_MAYEXEC.
* Removed Deven Bowers and Kees Cook Reviewed-by tags because of the current update.

Changes since v5:
* Remove the static enforcement configuration through Kconfig because it makes the code more simple like this, and because the current sysctl configuration can only be set with CAP_SYS_ADMIN, the same way mount options (i.e. noexec) can be set. If an harden distro wants to enforce a configuration, it should restrict capabilities or sysctl configuration. Furthermore, an LSM can easily leverage O_MAYEXEC to fit its need.
* Move checks from inode_permission() to may_open() and make the error codes more consistent according to file types (in line with a previous commit): opening a directory with O_MAYEXEC returns EISDIR and other non-regular file types may return EACCES.
* In may_open(), when OMAYEXEC_ENFORCE_FILE is set, replace explicit call to generic_permission() with an artificial MAY_EXEC to avoid double calls. This makes sense especially when an LSM policy forbids execution of a file.
* Replace the custom proc_omayexec() with proc_dointvec_minmax_sysadmin(), and then replace the CAP_MAC_ADMIN check with a CAP_SYS_ADMIN one (suggested by Kees Cook and Stephen Smalley).
* Use BIT() (suggested by Kees Cook).
* Rename variables (suggested by Kees Cook).
* Reword the kconfig help.
* Import the documentation patch (suggested by Kees Cook):
https://lore.kernel.org/lkml/[email protected]/
* Update documentation and add LWN.net article.

Changes since v4:
* Add kernel configuration options to enforce O_MAYEXEC at build time, and disable the sysctl in such case (requested by James Morris).
* Reword commit message.

Changes since v3:
* Switch back to O_MAYEXEC, but only handle it with openat2(2) which checks unknown flags (suggested by Aleksa Sarai). Cf.
https://lore.kernel.org/lkml/[email protected]/

Changes since v2:
* Replace O_MAYEXEC with RESOLVE_MAYEXEC from openat2(2). This change enables to not break existing application using bogus O_* flags that may be ignored by current kernels by using a new dedicated flag, only usable through openat2(2) (suggested by Jeff Layton). Using this flag will results in an error if the running kernel does not support it. User space needs to manage this case, as with other RESOLVE_* flags. The best effort approach to security (for most common distros) will simply consists of ignoring such an error and retry without RESOLVE_MAYEXEC. However, a fully controlled system may which to error out if such an inconsistency is detected.
* Cosmetic changes.

Changes since v1:
* Set __FMODE_EXEC when using O_MAYEXEC to make this information available through the new fanotify/FAN_OPEN_EXEC event (suggested by Jan Kara and Matthew Bobrowski):
https://lore.kernel.org/lkml/[email protected]/
* Move code from Yama to the FS subsystem (suggested by Kees Cook).
* Make omayexec_inode_permission() static (suggested by Jann Horn).
* Use mode 0600 for the sysctl.
* Only match regular files (not directories nor other types), which follows the same semantic as commit 73601ea5b7b1 ("fs/open.c: allow opening only regular files during execve()").

Changes since v13:
* Rename sysctl to trusted_for_policy (suggested by Kees Cook).
* Add Acked-by Kees Cook

---
 Documentation/admin-guide/sysctl/fs.rst | 50 ++++++++++++++++
 fs/open.c                               | 78 +++++++++++++++++++++++++
 include/linux/fs.h                      |  1 +
 include/linux/syscalls.h                |  2 +
 include/uapi/linux/trusted-for.h        | 18 ++++++
 kernel/sysctl.c                         | 12 +++-
 6 files changed, 159 insertions(+), 2 deletions(-)
 create mode 100644 include/uapi/linux/trusted-for.h
Documentation/admin-guide/sysctl/fs.rst +50 -0
index 2a501c9ddc55..e364d6c45790 100644
@@ -48,6 +48,7 @@ Currently, these files are in /proc/sys/fs:
 - suid_dumpable
 - super-max
 - super-nr
+- trusted_for_policy
 
 
 aio-nr & aio-max-nr
@@ -382,3 +383,52 @@ Each "watch" costs roughly 90 bytes on a 32bit kernel, and roughly 160 bytes
 on a 64bit one.
 The current default value for  max_user_watches  is the 1/25 (4%) of the
 available low memory, divided for the "watch" cost in bytes.
+
+
+trusted_for_policy
+------------------
+
+An interpreter can call :manpage:`trusted_for(2)` with a
+``TRUSTED_FOR_EXECUTION`` usage to check that opened regular files are expected
+to be executable.  If the file is not identified as executable, then the
+syscall returns -EACCES.  This may allow a script interpreter to check
+executable permission before reading commands from a file, or a dynamic linker
+to only load executable shared objects.  One interesting use case is to enforce
+a "write xor execute" policy through interpreters.
+
+The ability to restrict code execution must be thought as a system-wide policy,
+which first starts by restricting mount points with the ``noexec`` option.
+This option is also automatically applied to special filesystems such as /proc .
+This prevents files on such mount points to be directly executed by the kernel
+or mapped as executable memory (e.g. libraries).  With script interpreters
+using :manpage:`trusted_for(2)`, the executable permission can then be checked
+before reading commands from files.  This makes it possible to enforce the
+``noexec`` at the interpreter level, and thus propagates this security policy
+to scripts.  To be fully effective, these interpreters also need to handle the
+other ways to execute code: command line parameters (e.g., option ``-e`` for
+Perl), module loading (e.g., option ``-m`` for Python), stdin, file sourcing,
+environment variables, configuration files, etc.  According to the threat
+model, it may be acceptable to allow some script interpreters (e.g.  Bash) to
+interpret commands from stdin, may it be a TTY or a pipe, because it may not be
+enough to (directly) perform syscalls.
+
+There are two complementary security policies: enforce the ``noexec`` mount
+option, and enforce executable file permission.  These policies are handled by
+the ``fs.trusted_for_policy`` sysctl (writable only with ``CAP_SYS_ADMIN``) as
+a bitmask:
+
+1 - Mount restriction: checks that the mount options for the underlying VFS
+    mount do not prevent execution.
+
+2 - File permission restriction: checks that the file is marked as
+    executable for the current process (e.g., POSIX permissions, ACLs).
+
+Note that as long as a policy is enforced, checking any non-regular file with
+:manpage:`trusted_for(2)` returns -EACCES (e.g. TTYs, pipe), even when such a
+file is marked as executable or is on an executable mount point.
+
+Code samples can be found in
+tools/testing/selftests/interpreter/trust_policy_test.c and interpreter patches
+(for the original O_MAYEXEC) are available at
+https://github.com/clipos-archive/clipos4_portage-overlay/search?q=O_MAYEXEC .
+See also an overview article: https://lwn.net/Articles/820000/ .
fs/open.c +78 -0

90 lines not on this page · raw message

include/linux/fs.h +1 -0
index e7a633353fd2..9689b8a22ec5 100644
@@ -86,6 +86,7 @@ extern int sysctl_protected_symlinks;
 extern int sysctl_protected_hardlinks;
 extern int sysctl_protected_fifos;
 extern int sysctl_protected_regular;
+extern int sysctl_trusted_for_policy;
 
 typedef __kernel_rwf_t rwf_t;
 
include/linux/syscalls.h +2 -0
index 252243c7783d..8a69a6b1c1ef 100644
@@ -71,6 +71,7 @@ struct open_how;
 struct mount_attr;
 struct landlock_ruleset_attr;
 enum landlock_rule_type;
+enum trusted_for_usage;
 
 #include <linux/types.h>
 #include <linux/aio_abi.h>
@@ -461,6 +462,7 @@ asmlinkage long sys_fallocate(int fd, int mode, loff_t offset, loff_t len);
 asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode);
 asmlinkage long sys_faccessat2(int dfd, const char __user *filename, int mode,
 			       int flags);
+asmlinkage long sys_trusted_for(int fd, enum trusted_for_usage usage, u32 flags);
 asmlinkage long sys_chdir(const char __user *filename);
 asmlinkage long sys_fchdir(unsigned int fd);
 asmlinkage long sys_chroot(const char __user *filename);
include/uapi/linux/trusted-for.h new file +18 -0
new file mode 100644
index 000000000000..cc4f030c5103
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+#ifndef _UAPI_LINUX_TRUSTED_FOR_H
+#define _UAPI_LINUX_TRUSTED_FOR_H
+
+/**
+ * enum trusted_for_usage - Usage for which a file descriptor is trusted
+ *
+ * Argument of trusted_for(2).
+ */
+enum trusted_for_usage {
+	/**
+	 * @TRUSTED_FOR_EXECUTION: Check that the data read from a file
+	 * descriptor is trusted to be executed or interpreted (e.g. scripts).
+	 */
+	TRUSTED_FOR_EXECUTION = 1,
+};
+
+#endif /* _UAPI_LINUX_TRUSTED_FOR_H */
kernel/sysctl.c +10 -2
index 083be6af29d7..002dc830c165 100644
@@ -115,6 +115,7 @@ static int sixty = 60;
 
 static int __maybe_unused neg_one = -1;
 static int __maybe_unused two = 2;
+static int __maybe_unused three = 3;
 static int __maybe_unused four = 4;
 static unsigned long zero_ul;
 static unsigned long one_ul = 1;
@@ -936,7 +937,6 @@ static int proc_taint(struct ctl_table *table, int write,
 	return err;
 }
 
-#ifdef CONFIG_PRINTK
 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
 				void *buffer, size_t *lenp, loff_t *ppos)
 {
@@ -945,7 +945,6 @@ static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
 
 	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 }
-#endif
 
 /**
  * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
@@ -3357,6 +3356,15 @@ static struct ctl_table fs_table[] = {
 		.extra1		= SYSCTL_ZERO,
 		.extra2		= &two,
 	},
+	{
+		.procname       = "trusted_for_policy",
+		.data           = &sysctl_trusted_for_policy,
+		.maxlen         = sizeof(int),
+		.mode           = 0600,
+		.proc_handler	= proc_dointvec_minmax_sysadmin,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= &three,
+	},
 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
 	{
 		.procname	= "binfmt_misc",
Signature · 1 line
2.32.0
#2 lore raw

Re: [PATCH v14 1/3] fs: Add trusted_for(2) syscall implementation and related sysctl

* Mickaël Salaün:

Quoted 9 lines

Being able to restrict execution also enables to protect the kernel by restricting arbitrary syscalls that an attacker could perform with a crafted binary or certain script languages. It also improves multilevel isolation by reducing the ability of an attacker to use side channels with specific code. These restrictions can natively be enforced for ELF binaries (with the noexec mount option) but require this kernel extension to properly handle scripts (e.g. Python, Perl). To get a consistent execution policy, additional memory restrictions should also be enforced (e.g. thanks to SELinux).

One example I have come across recently is that code which can be safely loaded as a Perl module is definitely not a no-op as a shell
script: it downloads code and executes it, apparently over an untrusted network connection and without signature checking.

Maybe in the IMA world, the expectation is that such ambiguous code would not be signed in the first place, but general-purpose distributions are heading in a different direction with across-the-board signing:

Signed RPM Contents
<https://fedoraproject.org/wiki/Changes/Signed_RPM_Contents>

So I wonder if we need additional context information for a potential LSM to identify the intended use case.

#3 lore raw

Re: [PATCH v14 1/3] fs: Add trusted_for(2) syscall implementation and related sysctl

Quoted 22 lines · Florian Weimer

* Mickaël Salaün:

Quoted 9 lines

Being able to restrict execution also enables to protect the kernel by restricting arbitrary syscalls that an attacker could perform with a crafted binary or certain script languages. It also improves multilevel isolation by reducing the ability of an attacker to use side channels with specific code. These restrictions can natively be enforced for ELF binaries (with the noexec mount option) but require this kernel extension to properly handle scripts (e.g. Python, Perl). To get a consistent execution policy, additional memory restrictions should also be enforced (e.g. thanks to SELinux).

One example I have come across recently is that code which can be safely loaded as a Perl module is definitely not a no-op as a shell
script: it downloads code and executes it, apparently over an untrusted network connection and without signature checking.

Maybe in the IMA world, the expectation is that such ambiguous code would not be signed in the first place, but general-purpose distributions are heading in a different direction with across-the-board signing:

Quoted 6 lines

Signed RPM Contents
<https://fedoraproject.org/wiki/Changes/Signed_RPM_Contents>

So I wonder if we need additional context information for a potential LSM to identify the intended use case.

This is an interesting use case. I think such policy enforcement could be done either with an existing LSM (e.g. IMA) or a new one (e.g. IPE), but it could also partially be enforced by the script interpreter. The kernel should have enough context: interpreter process (which could be dedicated to a specific usage) and the opened script file, or we could add a new usage flag to the trusted_for syscall if that makes sense. Either way, this doesn't seem to be an issue for the current patch series.

maintainer
#4 lore raw

Re: [PATCH v14 1/3] fs: Add trusted_for(2) syscall implementation and related sysctl

In reply to Florian Weimer

Hi Florian,

Quoted 17 lines · Florian Weimer

* Mickaël Salaün:

Quoted 9 lines

Being able to restrict execution also enables to protect the kernel by restricting arbitrary syscalls that an attacker could perform with a crafted binary or certain script languages. It also improves multilevel isolation by reducing the ability of an attacker to use side channels with specific code. These restrictions can natively be enforced for ELF binaries (with the noexec mount option) but require this kernel extension to properly handle scripts (e.g. Python, Perl). To get a consistent execution policy, additional memory restrictions should also be enforced (e.g. thanks to SELinux).

One example I have come across recently is that code which can be safely loaded as a Perl module is definitely not a no-op as a shell
script: it downloads code and executes it, apparently over an untrusted network connection and without signature checking.

Quoted 4 lines

Maybe in the IMA world, the expectation is that such ambiguous code would not be signed in the first place, but general-purpose distributions are heading in a different direction with across-the-board signing:

Automatically signing code is at least the first step in the right direction of only executing code with known provenance. Perhaps future work would address the code signing granularity.

Quoted 6 lines

Signed RPM Contents
<https://fedoraproject.org/wiki/Changes/Signed_RPM_Contents>

So I wonder if we need additional context information for a potential LSM to identify the intended use case.

My first thoughts were an enumeration UNSIGNED_DOWNLOADED_CODE or maybe even UNTRUSTED_DOWNLOADED_CODE, but that doesn't seem very helpful. What type of context information were you thinking about?

Mimi

#5 lore raw

PATCH v14 2/3 arch: Wire up trusted_for(2)

In reply to Mickaël Salaün

From: Mickaël Salaün <[email protected]>

Wire up trusted_for(2) for all architectures.

---

Changes since v13:
* Add Reviewed-by Kees Cook.

Changes since v12:
* Update syscall IDs to align with the new ones.

Changes since v11:
* Add Acked-by: Geert Uytterhoeven <[email protected]>
* Rebase and leave space for watch_mount(2) and epoll_pwait2(2) from -next.

Changes since v9:
* Rename introspect_access(2) to trusted_for(2).
* Increase syscall number to leave space for memfd_secret(2) in -next.

Changes since v7:
* New patch for the new syscall.
* Increase syscall numbers by 2 to leave space for new ones (in linux-next): watch_mount(2) and process_madvise(2).

---
 arch/alpha/kernel/syscalls/syscall.tbl      | 1 +
 arch/arm/tools/syscall.tbl                  | 1 +
 arch/arm64/include/asm/unistd.h             | 2 +-
 arch/arm64/include/asm/unistd32.h           | 2 ++
 arch/ia64/kernel/syscalls/syscall.tbl       | 1 +
 arch/m68k/kernel/syscalls/syscall.tbl       | 1 +
 arch/microblaze/kernel/syscalls/syscall.tbl | 1 +
 arch/mips/kernel/syscalls/syscall_n32.tbl   | 1 +
 arch/mips/kernel/syscalls/syscall_n64.tbl   | 1 +
 arch/mips/kernel/syscalls/syscall_o32.tbl   | 1 +
 arch/parisc/kernel/syscalls/syscall.tbl     | 1 +
 arch/powerpc/kernel/syscalls/syscall.tbl    | 1 +
 arch/s390/kernel/syscalls/syscall.tbl       | 1 +
 arch/sh/kernel/syscalls/syscall.tbl         | 1 +
 arch/sparc/kernel/syscalls/syscall.tbl      | 1 +
 arch/x86/entry/syscalls/syscall_32.tbl      | 1 +
 arch/x86/entry/syscalls/syscall_64.tbl      | 1 +
 arch/xtensa/kernel/syscalls/syscall.tbl     | 1 +
 include/uapi/asm-generic/unistd.h           | 4 +++-
 19 files changed, 22 insertions(+), 2 deletions(-)
arch/alpha/kernel/syscalls/syscall.tbl +1 -0
index e4a041cd5715..ee33f1631f6d 100644
@@ -488,3 +488,4 @@
 556	common	landlock_restrict_self		sys_landlock_restrict_self
 # 557 reserved for memfd_secret
 558	common	process_mrelease		sys_process_mrelease
+559	common	trusted_for			sys_trusted_for
arch/arm/tools/syscall.tbl +1 -0
index e842209e135d..93edcdeda698 100644
@@ -462,3 +462,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/arm64/include/asm/unistd.h +1 -1
index 3cb206aea3db..6bdb5f5db438 100644
@@ -38,7 +38,7 @@
 #define __ARM_NR_compat_set_tls		(__ARM_NR_COMPAT_BASE + 5)
 #define __ARM_NR_COMPAT_END		(__ARM_NR_COMPAT_BASE + 0x800)
 
-#define __NR_compat_syscalls		449
+#define __NR_compat_syscalls		450
 #endif
 
 #define __ARCH_WANT_SYS_CLONE
arch/arm64/include/asm/unistd32.h +2 -0
index 844f6ae58662..3cb7df3a441c 100644
@@ -903,6 +903,8 @@ __SYSCALL(__NR_landlock_add_rule, sys_landlock_add_rule)
 __SYSCALL(__NR_landlock_restrict_self, sys_landlock_restrict_self)
 #define __NR_process_mrelease 448
 __SYSCALL(__NR_process_mrelease, sys_process_mrelease)
+#define __NR_trusted_for 449
+__SYSCALL(__NR_trusted_for, sys_trusted_for)
 
 /*
  * Please add new compat syscalls above this comment and update
arch/ia64/kernel/syscalls/syscall.tbl +1 -0
index 6fea1844fb95..50970d778005 100644
@@ -369,3 +369,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/m68k/kernel/syscalls/syscall.tbl +1 -0
index 7976dff8f879..caee4759d9f4 100644
@@ -448,3 +448,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/microblaze/kernel/syscalls/syscall.tbl +1 -0
index 6b0e11362bd2..25761d27aa9e 100644
@@ -454,3 +454,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/mips/kernel/syscalls/syscall_n32.tbl +1 -0
index 70e32de2bcaa..5e61c18fa3df 100644
@@ -387,3 +387,4 @@
 446	n32	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	n32	process_mrelease		sys_process_mrelease
+449	n32	trusted_for			sys_trusted_for
arch/mips/kernel/syscalls/syscall_n64.tbl +1 -0
index 1ca7bc337932..9a4928032039 100644
@@ -363,3 +363,4 @@
 446	n64	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	n64	process_mrelease		sys_process_mrelease
+449	n64	trusted_for			sys_trusted_for
arch/mips/kernel/syscalls/syscall_o32.tbl +1 -0
index a61c35edaa74..7fd966e720c2 100644
@@ -436,3 +436,4 @@
 446	o32	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	o32	process_mrelease		sys_process_mrelease
+449	o32	trusted_for			sys_trusted_for
arch/parisc/kernel/syscalls/syscall.tbl +1 -0
index bf751e0732b7..b43fe303922c 100644
@@ -446,3 +446,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/powerpc/kernel/syscalls/syscall.tbl +1 -0
index 7bef917cc84e..0de4795a4f85 100644
@@ -528,3 +528,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/s390/kernel/syscalls/syscall.tbl +1 -0
index df5261e5cfe1..b524343bb67d 100644
@@ -451,3 +451,4 @@
 446  common	landlock_restrict_self	sys_landlock_restrict_self	sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448  common	process_mrelease	sys_process_mrelease		sys_process_mrelease
+449  common	trusted_for		sys_trusted_for			sys_trusted_for
arch/sh/kernel/syscalls/syscall.tbl +1 -0
index 208f131659c5..d4c925d83129 100644
@@ -451,3 +451,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/sparc/kernel/syscalls/syscall.tbl +1 -0
index c37764dc764d..158bf1b028ac 100644
@@ -494,3 +494,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
arch/x86/entry/syscalls/syscall_32.tbl +1 -0
index 960a021d543e..a9eda13c86cd 100644
@@ -453,3 +453,4 @@
 446	i386	landlock_restrict_self	sys_landlock_restrict_self
 447	i386	memfd_secret		sys_memfd_secret
 448	i386	process_mrelease	sys_process_mrelease
+449	i386	trusted_for		sys_trusted_for
arch/x86/entry/syscalls/syscall_64.tbl +1 -0
index 18b5500ea8bf..c352f1b200ee 100644
@@ -370,6 +370,7 @@
 446	common	landlock_restrict_self	sys_landlock_restrict_self
 447	common	memfd_secret		sys_memfd_secret
 448	common	process_mrelease	sys_process_mrelease
+449	common	trusted_for		sys_trusted_for
 
 #
 # Due to a historical design error, certain syscalls are numbered differently
arch/xtensa/kernel/syscalls/syscall.tbl +1 -0
index 104b327f8ac9..f26a1313a055 100644
@@ -419,3 +419,4 @@
 446	common	landlock_restrict_self		sys_landlock_restrict_self
 # 447 reserved for memfd_secret
 448	common	process_mrelease		sys_process_mrelease
+449	common	trusted_for			sys_trusted_for
include/uapi/asm-generic/unistd.h +3 -1
index 1c5fb86d455a..67b4bebe7c87 100644
@@ -879,9 +879,11 @@ __SYSCALL(__NR_memfd_secret, sys_memfd_secret)
 #endif
 #define __NR_process_mrelease 448
 __SYSCALL(__NR_process_mrelease, sys_process_mrelease)
+#define __NR_trusted_for 449
+__SYSCALL(__NR_trusted_for, sys_trusted_for)
 
 #undef __NR_syscalls
-#define __NR_syscalls 449
+#define __NR_syscalls 450
 
 /*
  * 32 bit systems traditionally used different
Signature · 1 line
2.32.0
#6 lore raw

PATCH v14 3/3 selftest/interpreter: Add tests for trusted_for(2) policies

In reply to Mickaël Salaün

From: Mickaël Salaün <[email protected]>

Test that checks performed by trusted_for(2) on file descriptors are consistent with noexec mount points and file execute permissions, according to the policy configured with the fs.trust_policy sysctl.

---

Changes since v13:
* Move -I to CFLAGS (suggested by Kees Cook).
* Update sysctl name.

Changes since v12:
* Fix Makefile's license.

Changes since v10:
* Update selftest Makefile.

Changes since v9:
* Rename the syscall and the sysctl.
* Update tests for enum trusted_for_usage

Changes since v8:
* Update with the dedicated syscall introspect_access(2) and the renamed fs.introspection_policy sysctl.
* Remove check symlink which can't be use as is anymore.
* Use socketpair(2) to test UNIX socket.

Changes since v7:
* Update tests with faccessat2/AT_INTERPRETED, including new ones to check that setting R_OK or W_OK returns EINVAL.
* Add tests for memfd, pipefs and nsfs.
* Rename and move back tests to a standalone directory.

Changes since v6:
* Add full combination tests for all file types, including block devices, character devices, fifos, sockets and symlinks.
* Properly save and restore initial sysctl value for all tests.

Changes since v5:
* Refactor with FIXTURE_VARIANT, which make the tests much more easy to read and maintain.
* Save and restore initial sysctl value (suggested by Kees Cook).
* Test with a sysctl value of 0.
* Check errno in sysctl_access_write test.
* Update tests for the CAP_SYS_ADMIN switch.
* Update tests to check -EISDIR (replacing -EACCES).
* Replace FIXTURE_DATA() with FIXTURE() (spotted by Kees Cook).
* Use global const strings.

Changes since v3:
* Replace RESOLVE_MAYEXEC with O_MAYEXEC.
* Add tests to check that O_MAYEXEC is ignored by open(2) and openat(2).

Changes since v2:
* Move tests from exec/ to openat2/ .
* Replace O_MAYEXEC with RESOLVE_MAYEXEC from openat2(2).
* Cleanup tests.

Changes since v1:
* Move tests from yama/ to exec/ .
* Fix _GNU_SOURCE in kselftest_harness.h .
* Add a new test sysctl_access_write to check if CAP_MAC_ADMIN is taken into account.
* Test directory execution which is always forbidden since commit 73601ea5b7b1 ("fs/open.c: allow opening only regular files during execve()"), and also check that even the root user can not bypass file execution checks.
* Make sure delete_workspace() always as enough right to succeed.
* Cosmetic cleanup.

---
 tools/testing/selftests/Makefile              |   1 +
 .../testing/selftests/interpreter/.gitignore  |   2 +
 tools/testing/selftests/interpreter/Makefile  |  21 +
 tools/testing/selftests/interpreter/config    |   1 +
 .../selftests/interpreter/trust_policy_test.c | 362 ++++++++++++++++++
 5 files changed, 387 insertions(+)
 create mode 100644 tools/testing/selftests/interpreter/.gitignore
 create mode 100644 tools/testing/selftests/interpreter/Makefile
 create mode 100644 tools/testing/selftests/interpreter/config
 create mode 100644 tools/testing/selftests/interpreter/trust_policy_test.c
tools/testing/selftests/Makefile +1 -0
index c852eb40c4f7..3a032a545f74 100644
@@ -20,6 +20,7 @@ TARGETS += ftrace
 TARGETS += futex
 TARGETS += gpio
 TARGETS += intel_pstate
+TARGETS += interpreter
 TARGETS += ipc
 TARGETS += ir
 TARGETS += kcmp
tools/testing/selftests/interpreter/.gitignore new file +2 -0
new file mode 100644
index 000000000000..82a4846cbc4b
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+/*_test
tools/testing/selftests/interpreter/Makefile new file +21 -0
new file mode 100644
index 000000000000..7402fdb6533f
@@ -0,0 +1,21 @@
+# SPDX-License-Identifier: GPL-2.0
+
+CFLAGS += -Wall -O2 -I$(khdr_dir)
+LDLIBS += -lcap
+
+src_test := $(wildcard *_test.c)
+TEST_GEN_PROGS := $(src_test:.c=)
+
+KSFT_KHDR_INSTALL := 1
+include ../lib.mk
+
+khdr_dir = $(top_srcdir)/usr/include
+
+$(khdr_dir)/asm-generic/unistd.h: khdr
+	@:
+
+$(khdr_dir)/linux/trusted-for.h: khdr
+	@:
+
+$(OUTPUT)/%_test: %_test.c $(khdr_dir)/asm-generic/unistd.h $(khdr_dir)/linux/trusted-for.h ../kselftest_harness.h
+	$(LINK.c) $< $(LDLIBS) -o $@
tools/testing/selftests/interpreter/config new file +1 -0
new file mode 100644
index 000000000000..dd53c266bf52
@@ -0,0 +1 @@
+CONFIG_SYSCTL=y
tools/testing/selftests/interpreter/trust_policy_test.c new file +362 -0

362 lines not on this page · raw message

Signature · 1 line
2.32.0
Reviewed-by on v14 3/3
#7 lore raw

Re: [PATCH v14 3/3] selftest/interpreter: Add tests for trusted_for(2) policies

Quoted 6 lines · Mickaël Salaün

From: Mickaël Salaün <[email protected]>

Test that checks performed by trusted_for(2) on file descriptors are consistent with noexec mount points and file execute permissions, according to the policy configured with the fs.trust_policy sysctl.

Quoted 6 lines
Cc Al Viro <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Kees Cook <[email protected]>
Cc Shuah Khan <[email protected]>

Thanks for the adjustments!

Signature · 1 line
Kees Cook
#8 lore raw
In reply to Mickaël Salaün
Quoted 3 lines · Mickaël Salaün

This patch series is mainly a rebase on v5.15-rc4 with some cosmetic changes suggested by Kees Cook. Andrew, can you please consider to merge this into your tree?

Thanks for staying on this series! This is a good step in the right direction for finally plugging the "interpreter" noexec hole. I'm pretty sure Chrome OS will immediately use this as they've been carrying similar functionality for a long time.

Signature · 1 line
Kees Cook
#9 lore raw
In reply to Mickaël Salaün
Quoted 7 lines · On Fri, 8 Oct 2021 12:48:37 +0200 Mickaël Salaün <[email protected]> wrote:

The final goal of this patch series is to enable the kernel to be a global policy manager by entrusting processes with access control at their level. To reach this goal, two complementary parts are required:
* user space needs to be able to know if it can trust some file descriptor content for a specific usage;
* and the kernel needs to make available some part of the policy configured by the system administrator.

Apologies if I missed this...

It would be nice to see a description of the proposed syscall interface in these changelogs! Then a few questions I have will be answered...

long trusted_for(const int fd,
const enum trusted_for_usage usage, const u32 flags)

- `usage' must be equal to TRUSTED_FOR_EXECUTION, so why does it exist? Some future modes are planned? Please expand on this.

- `flags' is unused (must be zero). So why does it exist? What are the plans here?

- what values does the syscall return and what do they mean?

#10 lore raw
Quoted 10 lines · Andrew Morton
Quoted 7 lines · On Fri, 8 Oct 2021 12:48:37 +0200 Mickaël Salaün <[email protected]> wrote:

The final goal of this patch series is to enable the kernel to be a global policy manager by entrusting processes with access control at their level. To reach this goal, two complementary parts are required:
* user space needs to be able to know if it can trust some file descriptor content for a specific usage;
* and the kernel needs to make available some part of the policy configured by the system administrator.

Quoted 4 lines

Apologies if I missed this...

It would be nice to see a description of the proposed syscall interface in these changelogs! Then a few questions I have will be answered...

I described this syscall and it's semantic in the first patch in Documentation/admin-guide/sysctl/fs.rst
Do you want me to copy-paste this content in the cover letter?

Quoted 7 lines

long trusted_for(const int fd,
const enum trusted_for_usage usage, const u32 flags)

- `usage' must be equal to TRUSTED_FOR_EXECUTION, so why does it exist? Some future modes are planned? Please expand on this.

Indeed, the current use case is to check if the kernel would allow execution of a file. But as Florian pointed out, we may want to add more context in the future, e.g. to enforce signature verification, to check if this is a legitimate (system) library, to check if the file is allowed to be used as (trusted) configuration…

Quoted 3 lines

- `flags' is unused (must be zero). So why does it exist? What are the plans here?

This is mostly to follow syscall good practices for extensibility. It could be used in combination with the usage argument (which defines the user space semantic), e.g. to check for extra properties such as cryptographic or integrity requirements, origin of the file…

Quoted 3 lines

- what values does the syscall return and what do they mean?

It returns 0 on success, or -EACCES if the kernel policy denies the specified usage.

#11 lore raw
Quoted 12 lines · On Mon, 11 Oct 2021 10:47:04 +0200 Mickaël Salaün <[email protected]> wrote:
Quoted 10 lines · Andrew Morton
Quoted 7 lines · On Fri, 8 Oct 2021 12:48:37 +0200 Mickaël Salaün <[email protected]> wrote:

The final goal of this patch series is to enable the kernel to be a global policy manager by entrusting processes with access control at their level. To reach this goal, two complementary parts are required:
* user space needs to be able to know if it can trust some file descriptor content for a specific usage;
* and the kernel needs to make available some part of the policy configured by the system administrator.

Quoted 7 lines
Quoted 4 lines

Apologies if I missed this...

It would be nice to see a description of the proposed syscall interface in these changelogs! Then a few questions I have will be answered...

I described this syscall and it's semantic in the first patch in Documentation/admin-guide/sysctl/fs.rst

Well, kinda. It didn't explain why the usage' and flags' arguments exist and what are the plans for them.

Quoted 1 line

Do you want me to copy-paste this content in the cover letter?

That would be best please. It's basically the most important thing when reviewing the implementation.

Quoted 24 lines
Quoted 7 lines

long trusted_for(const int fd,
const enum trusted_for_usage usage, const u32 flags)

- `usage' must be equal to TRUSTED_FOR_EXECUTION, so why does it exist? Some future modes are planned? Please expand on this.

Indeed, the current use case is to check if the kernel would allow execution of a file. But as Florian pointed out, we may want to add more context in the future, e.g. to enforce signature verification, to check if this is a legitimate (system) library, to check if the file is allowed to be used as (trusted) configuration…

Quoted 3 lines

- `flags' is unused (must be zero). So why does it exist? What are the plans here?

This is mostly to follow syscall good practices for extensibility. It could be used in combination with the usage argument (which defines the user space semantic), e.g. to check for extra properties such as cryptographic or integrity requirements, origin of the file…

Quoted 1 line
Quoted 5 lines
Quoted 2 lines

- what values does the syscall return and what do they mean?

It returns 0 on success, or -EACCES if the kernel policy denies the specified usage.

And please document all of this in the changelog also.