[PATCH v8] RO/NX protection for loadable kernel modules

From: Siarhei Liakh
Date: Sun Jan 31 2010 - 18:22:44 EST


This patch is a logical extension of the protection provided by
CONFIG_DEBUG_RODATA to LKMs. The protection is provided by splitting
module_core and module_init into three logical parts each and setting
appropriate page access permissions for each individual section:

1. Code: RO+X
2. RO data: RO+NX
3. RW data: RW+NX

In order to achieve proper protection, layout_sections() have been
modified to align each of the three parts mentioned above onto page
boundary. Next, the corresponding page access permissions are set
right before successful exit from load_module(). Further, free_module()
and sys_init_module have been modified to set module_core and
module_init as RW+NX right before calling module_free().

By default, the original section layout and access flags are preserved.
When compiled with CONFIG_DEBUG_SET_MODULE_RONX=y, the patch
will page-align each group of sections to ensure that each page contains
only one type of content and will enforce RO/NX for each group of pages.

v1: Initial proof-of-concept patch.

v2: The patch have been re-written to reduce the number of #ifdefs and
to make it architecture-agnostic. Code formatting have been corrected also.

v3: Opportunistic RO/NX protectiuon is now unconditional. Section
page-alignment is enabled when CONFIG_DEBUG_RODATA=y.

v4: Removed most macros and improved coding style.

v5: Changed page-alignment and RO/NX section size calculation

v6: Fixed comments. Restricted RO/NX enforcement to x86 only

v7: Introduced CONFIG_DEBUG_SET_MODULE_RONX, added calls to
set_all_modules_text_rw() and set_all_modules_text_ro() in ftrace

v8: updated for compatibility with linux 2.6.33-rc5

The patch have been developed for Linux 2.6.33-rc5 by Siarhei Liakh
<sliakh.lkml@xxxxxxxxx> and Xuxian Jiang <jiang@xxxxxxxxxxx>.

---

Signed-off-by: Siarhei Liakh <sliakh.lkml@xxxxxxxxx>
Signed-off-by: Xuxian Jiang <jiang@xxxxxxxxxxx>
Acked-by: Arjan van de Ven <arjan@xxxxxxxxxxxxxxx>

diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug
index bc01e3e..7c80e28 100644
--- a/arch/x86/Kconfig.debug
+++ b/arch/x86/Kconfig.debug
@@ -115,6 +115,17 @@ config DEBUG_RODATA_TEST
feature as well as for the change_page_attr() infrastructure.
If in doubt, say "N"

+config DEBUG_SET_MODULE_RONX
+ bool "Set loadable kernel module data as NX and text as RO"
+ default n
+ depends on X86 && MODULES
+ ---help---
+ This option helps to catch unintended modifications to loadable
+ kernel module's text and read-only data. It also prevents execution
+ of LKM's data. Such protection may interfere with run-time code
+ patching and dynamic kernel tracing.
+ If in doubt, say "N".
+
config DEBUG_NX_TEST
tristate "Testcase for the NX non-executable stack feature"
depends on DEBUG_KERNEL && m
diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c
index 3096892..b47a154 100644
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -19,6 +19,7 @@
#include <linux/sched.h>
#include <linux/init.h>
#include <linux/list.h>
+#include <linux/module.h>

#include <trace/syscall.h>

@@ -29,16 +30,17 @@


#ifdef CONFIG_DYNAMIC_FTRACE
-
int ftrace_arch_code_modify_prepare(void)
{
set_kernel_text_rw();
+ set_all_modules_text_rw();
return 0;
}

int ftrace_arch_code_modify_post_process(void)
{
set_kernel_text_ro();
+ set_all_modules_text_ro();
return 0;
}

diff --git a/include/linux/module.h b/include/linux/module.h
index 6cb1a3c..c6040ce 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -300,6 +300,9 @@ struct module
/* The size of the executable code in each section. */
unsigned int init_text_size, core_text_size;

+ /* Size of RO sections of the module (text+rodata) */
+ unsigned int init_ro_size, core_ro_size;
+
/* Arch-specific module values */
struct mod_arch_specific arch;

@@ -542,6 +545,9 @@ extern void print_modules(void);
extern void module_update_tracepoints(void);
extern int module_get_iter_tracepoints(struct tracepoint_iter *iter);

+void set_all_modules_text_rw(void);
+void set_all_modules_text_ro(void);
+
#else /* !CONFIG_MODULES... */
#define EXPORT_SYMBOL(sym)
#define EXPORT_SYMBOL_GPL(sym)
@@ -662,6 +668,13 @@ static inline int
module_get_iter_tracepoints(struct tracepoint_iter *iter)
return 0;
}

+static inline void set_all_modules_text_rw()
+{
+}
+
+static inline void set_all_modules_text_ro()
+{
+}
#endif /* CONFIG_MODULES */

struct device_driver;
diff --git a/kernel/module.c b/kernel/module.c
index f82386b..d382b59 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -55,6 +55,7 @@
#include <linux/async.h>
#include <linux/percpu.h>
#include <linux/kmemleak.h>
+#include <linux/pfn.h>

#define CREATE_TRACE_POINTS
#include <trace/events/module.h>
@@ -71,6 +72,26 @@ EXPORT_TRACEPOINT_SYMBOL(module_get);
#define ARCH_SHF_SMALL 0
#endif

+/*
+ * Modules' sections will be aligned on page boundaries
+ * to ensure complete separation of code and data, but
+ * only when CONFIG_DEBUG_SET_MODULE_RONX=y
+ */
+#ifdef CONFIG_DEBUG_SET_MODULE_RONX
+#define debug_align(X) ALIGN(X, PAGE_SIZE)
+#else
+#define debug_align(X) (X)
+#endif
+
+/*
+ * Given BASE and SIZE this macro calculates the number of pages the
+ * memory regions occupies
+ */
+#define NUMBER_OF_PAGES(BASE, SIZE) ((SIZE > 0) ? \
+ (PFN_DOWN((unsigned long)BASE + SIZE - 1) - \
+ PFN_DOWN((unsigned long)BASE) + 1) \
+ : (0UL))
+
/* If this is set, the section belongs in the init part of the module */
#define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG-1))

@@ -1372,6 +1393,134 @@ static int __unlink_module(void *_mod)
return 0;
}

+#ifdef CONFIG_DEBUG_SET_MODULE_RONX
+/*
+ * LKM RO/NX protection: protect module's text/ro-data
+ * from modification and any data from execution.
+ */
+static void set_section_ro_nx(void *base,
+ unsigned long text_size,
+ unsigned long ro_size,
+ unsigned long total_size)
+{
+ /* begin and end PFNs of the current subsection */
+ unsigned long begin_pfn;
+ unsigned long end_pfn;
+
+ /*
+ * Set RO for module text and RO-data:
+ * - Always protect first page.
+ * - Do not protect last partial page.
+ */
+ if (ro_size > 0) {
+ begin_pfn = PFN_DOWN((unsigned long)base);
+ end_pfn = PFN_DOWN((unsigned long)base + ro_size);
+ if (end_pfn > begin_pfn)
+ set_memory_ro(begin_pfn << PAGE_SHIFT,
+ end_pfn - begin_pfn);
+ }
+
+ /*
+ * Set NX permissions for module data:
+ * - Do not protect first partial page.
+ * - Always protect last page.
+ */
+ if (total_size > text_size) {
+ begin_pfn = PFN_UP((unsigned long)base + text_size);
+ end_pfn = PFN_UP((unsigned long)base + total_size);
+ if (end_pfn > begin_pfn)
+ set_memory_nx(begin_pfn << PAGE_SHIFT,
+ end_pfn - begin_pfn);
+ }
+}
+
+/* Setting memory back to RW+NX before releasing it */
+void unset_section_ro_nx(struct module *mod, void *module_region)
+{
+ unsigned long total_pages;
+
+ if (mod->module_core == module_region) {
+ /* Set core as NX+RW */
+ total_pages = NUMBER_OF_PAGES(mod->module_core, mod->core_size);
+ set_memory_nx((unsigned long)mod->module_core, total_pages);
+ set_memory_rw((unsigned long)mod->module_core, total_pages);
+
+ } else if (mod->module_init == module_region) {
+ /* Set init as NX+RW */
+ total_pages = NUMBER_OF_PAGES(mod->module_init, mod->init_size);
+ set_memory_nx((unsigned long)mod->module_init, total_pages);
+ set_memory_rw((unsigned long)mod->module_init, total_pages);
+ }
+}
+
+/* Iterate through all modules and set each module's text as RW */
+void set_all_modules_text_rw()
+{
+ struct module *mod;
+ unsigned long begin_pfn;
+ unsigned long end_pfn;
+
+ mutex_lock(&module_mutex);
+ list_for_each_entry_rcu(mod, &modules, list) {
+ if ((mod->module_core) && (mod->core_text_size > 0)) {
+ begin_pfn = PFN_DOWN((unsigned long)mod->module_core);
+ end_pfn = PFN_DOWN((unsigned long)mod->module_core +
+ mod->core_text_size);
+ if (end_pfn > begin_pfn)
+ set_memory_rw(begin_pfn << PAGE_SHIFT,
+ end_pfn - begin_pfn);
+ }
+ if ((mod->module_init) && (mod->init_text_size > 0)) {
+ begin_pfn = PFN_DOWN((unsigned long)mod->module_init);
+ end_pfn = PFN_DOWN((unsigned long)mod->module_init +
+ mod->init_text_size);
+ if (end_pfn > begin_pfn)
+ set_memory_rw(begin_pfn << PAGE_SHIFT,
+ end_pfn - begin_pfn);
+ }
+ }
+ mutex_unlock(&module_mutex);
+}
+
+/* Iterate through all modules and set each module's text as RO */
+void set_all_modules_text_ro()
+{
+ struct module *mod;
+ unsigned long begin_pfn;
+ unsigned long end_pfn;
+
+ mutex_lock(&module_mutex);
+ list_for_each_entry_rcu(mod, &modules, list) {
+ if ((mod->module_core) && (mod->core_text_size > 0)) {
+ begin_pfn = PFN_DOWN((unsigned long)mod->module_core);
+ end_pfn = PFN_DOWN((unsigned long)mod->module_core +
+ mod->core_text_size);
+ if (end_pfn > begin_pfn)
+ set_memory_ro(begin_pfn << PAGE_SHIFT,
+ end_pfn - begin_pfn);
+ }
+ if ((mod->module_init) && (mod->init_text_size > 0)) {
+ begin_pfn = PFN_DOWN((unsigned long)mod->module_init);
+ end_pfn = PFN_DOWN((unsigned long)mod->module_init +
+ mod->init_text_size);
+ if (end_pfn > begin_pfn)
+ set_memory_ro(begin_pfn << PAGE_SHIFT,
+ end_pfn - begin_pfn);
+ }
+ }
+ mutex_unlock(&module_mutex);
+}
+#else
+static void set_section_ro_nx(void *base,
+ unsigned long text_size,
+ unsigned long ro_size,
+ unsigned long total_size) { }
+
+void unset_section_ro_nx(struct module *mod, void *module_region) { }
+void set_all_modules_text_rw() { }
+void set_all_modules_text_ro() { }
+#endif
+
/* Free a module, remove from lists, etc (must hold module_mutex). */
static void free_module(struct module *mod)
{
@@ -1393,6 +1542,7 @@ static void free_module(struct module *mod)
destroy_params(mod->kp, mod->num_kp);

/* This may be NULL, but that's OK */
+ unset_section_ro_nx(mod, mod->module_init);
module_free(mod, mod->module_init);
kfree(mod->args);
if (mod->percpu)
@@ -1405,6 +1555,7 @@ static void free_module(struct module *mod)
lockdep_free_key_range(mod->module_core, mod->core_size);

/* Finally, free the core (containing the module structure) */
+ unset_section_ro_nx(mod, mod->module_core);
module_free(mod, mod->module_core);

#ifdef CONFIG_MPU
@@ -1582,8 +1733,19 @@ static void layout_sections(struct module *mod,
s->sh_entsize = get_offset(mod, &mod->core_size, s, i);
DEBUGP("\t%s\n", secstrings + s->sh_name);
}
- if (m == 0)
+ switch (m) {
+ case 0: /* executable */
+ mod->core_size = debug_align(mod->core_size);
mod->core_text_size = mod->core_size;
+ break;
+ case 1: /* RO: text and ro-data */
+ mod->core_size = debug_align(mod->core_size);
+ mod->core_ro_size = mod->core_size;
+ break;
+ case 3: /* whole core */
+ mod->core_size = debug_align(mod->core_size);
+ break;
+ }
}

DEBUGP("Init section allocation order:\n");
@@ -1600,8 +1762,19 @@ static void layout_sections(struct module *mod,
| INIT_OFFSET_MASK);
DEBUGP("\t%s\n", secstrings + s->sh_name);
}
- if (m == 0)
+ switch (m) {
+ case 0: /* executable */
+ mod->init_size = debug_align(mod->init_size);
mod->init_text_size = mod->init_size;
+ break;
+ case 1: /* RO: text and ro-data */
+ mod->init_size = debug_align(mod->init_size);
+ mod->init_ro_size = mod->init_size;
+ break;
+ case 3: /* whole init */
+ mod->init_size = debug_align(mod->init_size);
+ break;
+ }
}
}

@@ -2382,6 +2555,18 @@ static noinline struct module *load_module(void
__user *umod,

trace_module_load(mod);

+ /* Set RO and NX regions for core */
+ set_section_ro_nx(mod->module_core,
+ mod->core_text_size,
+ mod->core_ro_size,
+ mod->core_size);
+
+ /* Set RO and NX regions for init */
+ set_section_ro_nx(mod->module_init,
+ mod->init_text_size,
+ mod->init_ro_size,
+ mod->init_size);
+
/* Done! */
return mod;

@@ -2504,6 +2689,7 @@ SYSCALL_DEFINE3(init_module, void __user *, umod,
mod->symtab = mod->core_symtab;
mod->strtab = mod->core_strtab;
#endif
+ unset_section_ro_nx(mod, mod->module_init);
module_free(mod, mod->module_init);
mod->module_init = NULL;
mod->init_size = 0;
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/