From: Topi Miettinen <toiwoton@gmail.com>
Date: Sun, 6 Nov 2022 19:12:45 +0000 (+0200)
Subject: execute: use prctl(PR_SET_MDWE) for MemoryDenyWriteExecute=yes
X-Git-Tag: v254-rc1~1045
X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=7a114ed4b39e9670f6a511f3eecb6fd58274d27b;p=thirdparty%2Fsystemd.git

execute: use prctl(PR_SET_MDWE) for MemoryDenyWriteExecute=yes

On some ARM platforms, the dynamic linker could use PROT_BTI memory protection
flag with `mprotect(..., PROT_BTI | PROT_EXEC)` to enable additional memory
protection for executable pages. But `MemoryDenyWriteExecute=yes` blocks this
with seccomp filter denying all `mprotect(..., x | PROT_EXEC)`.

Newly preferred method is to use prctl(PR_SET_MDWE) on supported kernels. Then
in-kernel implementation can allow PROT_BTI as necessary, without weakening
MDWE. In-kernel version may also be extended to more sophisticated protections
in the future.
---

diff --git a/man/systemd.exec.xml b/man/systemd.exec.xml
index 795e2f0671c..653aa0d7921 100644
--- a/man/systemd.exec.xml
+++ b/man/systemd.exec.xml
@@ -2080,9 +2080,11 @@ RestrictNamespaces=~cgroup net</programlisting>
 
         <listitem><para>Takes a boolean argument. If set, attempts to create memory mappings that are writable and
         executable at the same time, or to change existing memory mappings to become executable, or mapping shared
-        memory segments as executable, are prohibited. Specifically, a system call filter is added that rejects
-        <citerefentry><refentrytitle>mmap</refentrytitle><manvolnum>2</manvolnum></citerefentry> system calls with both
-        <constant>PROT_EXEC</constant> and <constant>PROT_WRITE</constant> set,
+        memory segments as executable, are prohibited. Specifically, a system call filter is added (or
+        preferably, an equivalent kernel check is enabled with
+        <citerefentry><refentrytitle>prctl</refentrytitle><manvolnum>2</manvolnum></citerefentry>) that
+        rejects <citerefentry><refentrytitle>mmap</refentrytitle><manvolnum>2</manvolnum></citerefentry>
+        system calls with both <constant>PROT_EXEC</constant> and <constant>PROT_WRITE</constant> set,
         <citerefentry><refentrytitle>mprotect</refentrytitle><manvolnum>2</manvolnum></citerefentry> or
         <citerefentry><refentrytitle>pkey_mprotect</refentrytitle><manvolnum>2</manvolnum></citerefentry> system calls
         with <constant>PROT_EXEC</constant> set and
diff --git a/src/basic/missing_prctl.h b/src/basic/missing_prctl.h
index ab851306bac..016085bb02d 100644
--- a/src/basic/missing_prctl.h
+++ b/src/basic/missing_prctl.h
@@ -12,3 +12,11 @@
 #define PR_CAP_AMBIENT_LOWER     3
 #define PR_CAP_AMBIENT_CLEAR_ALL 4
 #endif
+
+/* b507808ebce23561d4ff8c2aa1fb949fe402bc61 (6.3) */
+#ifndef PR_SET_MDWE
+#define PR_SET_MDWE 65
+#endif
+#ifndef PR_MDWE_REFUSE_EXEC_GAIN
+#define PR_MDWE_REFUSE_EXEC_GAIN 1
+#endif
diff --git a/src/core/execute.c b/src/core/execute.c
index 3800c7a38bf..857b0b0070b 100644
--- a/src/core/execute.c
+++ b/src/core/execute.c
@@ -73,6 +73,7 @@
 #include "memory-util.h"
 #include "missing_fs.h"
 #include "missing_ioprio.h"
+#include "missing_prctl.h"
 #include "mkdir-label.h"
 #include "mount-util.h"
 #include "mountpoint-util.h"
@@ -1571,12 +1572,25 @@ static int apply_address_families(const Unit* u, const ExecContext *c) {
 }
 
 static int apply_memory_deny_write_execute(const Unit* u, const ExecContext *c) {
+        int r;
+
         assert(u);
         assert(c);
 
         if (!c->memory_deny_write_execute)
                 return 0;
 
+        /* use prctl() if kernel supports it (6.3) */
+        r = prctl(PR_SET_MDWE, PR_MDWE_REFUSE_EXEC_GAIN, 0, 0, 0);
+        if (r == 0) {
+                log_unit_debug(u, "Enabled MemoryDenyWriteExecute= with PR_SET_MDWE");
+                return 0;
+        }
+        if (r < 0 && errno != EINVAL)
+                return log_unit_debug_errno(u, errno, "Failed to enable MemoryDenyWriteExecute= with PR_SET_MDWE: %m");
+        /* else use seccomp */
+        log_unit_debug(u, "Kernel doesn't support PR_SET_MDWE: falling back to seccomp");
+
         if (skip_seccomp_unavailable(u, "MemoryDenyWriteExecute="))
                 return 0;