Merge tag 'hardening-v5.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux

Pull kernel hardening updates from Kees Cook:

 - Add arm64 Shadow Call Stack support for GCC 12 (Dan Li)

 - Avoid memset with stack offset randomization under Clang (Marco
   Elver)

 - Clean up stackleak plugin to play nice with .noinstr (Kees Cook)

 - Check stack depth for greater usercopy hardening coverage (Kees Cook)

* tag 'hardening-v5.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  arm64: Add gcc Shadow Call Stack support
  m68k: Implement "current_stack_pointer"
  xtensa: Implement "current_stack_pointer"
  usercopy: Check valid lifetime via stack depth
  stack: Constrain and fix stack offset randomization with Clang builds
  stack: Introduce CONFIG_RANDOMIZE_KSTACK_OFFSET
  gcc-plugins/stackleak: Ignore .noinstr.text and .entry.text
  gcc-plugins/stackleak: Exactly match strings instead of prefixes
  gcc-plugins/stackleak: Provide verbose mode

arch/Kconfig

@@ -599,21 +599,22 @@ config STACKPROTECTOR_STRONG
 config ARCH_SUPPORTS_SHADOW_CALL_STACK
 	bool
 	help
-	  An architecture should select this if it supports Clang's Shadow
-	  Call Stack and implements runtime support for shadow stack
+	  An architecture should select this if it supports the compiler's
+	  Shadow Call Stack and implements runtime support for shadow stack
 	  switching.
 
 config SHADOW_CALL_STACK
-	bool "Clang Shadow Call Stack"
-	depends on CC_IS_CLANG && ARCH_SUPPORTS_SHADOW_CALL_STACK
+	bool "Shadow Call Stack"
+	depends on ARCH_SUPPORTS_SHADOW_CALL_STACK
 	depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
 	help
-	  This option enables Clang's Shadow Call Stack, which uses a
-	  shadow stack to protect function return addresses from being
-	  overwritten by an attacker. More information can be found in
-	  Clang's documentation:
+	  This option enables the compiler's Shadow Call Stack, which
+	  uses a shadow stack to protect function return addresses from
+	  being overwritten by an attacker. More information can be found
+	  in the compiler's documentation:
 
-	    https://clang.llvm.org/docs/ShadowCallStack.html
+	  - Clang: https://clang.llvm.org/docs/ShadowCallStack.html
+	  - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options
 
 	  Note that security guarantees in the kernel differ from the
 	  ones documented for user space. The kernel must store addresses
@@ -1159,16 +1160,30 @@ config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
 	  to the compiler, so it will attempt to add canary checks regardless
 	  of the static branch state.
 
-config RANDOMIZE_KSTACK_OFFSET_DEFAULT
-	bool "Randomize kernel stack offset on syscall entry"
+config RANDOMIZE_KSTACK_OFFSET
+	bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT
+	default y
 	depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
+	depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000
 	help
 	  The kernel stack offset can be randomized (after pt_regs) by
 	  roughly 5 bits of entropy, frustrating memory corruption
 	  attacks that depend on stack address determinism or
-	  cross-syscall address exposures. This feature is controlled
-	  by kernel boot param "randomize_kstack_offset=on/off", and this
-	  config chooses the default boot state.
+	  cross-syscall address exposures.
+
+	  The feature is controlled via the "randomize_kstack_offset=on/off"
+	  kernel boot param, and if turned off has zero overhead due to its use
+	  of static branches (see JUMP_LABEL).
+
+	  If unsure, say Y.
+
+config RANDOMIZE_KSTACK_OFFSET_DEFAULT
+	bool "Default state of kernel stack offset randomization"
+	depends on RANDOMIZE_KSTACK_OFFSET
+	help
+	  Kernel stack offset randomization is controlled by kernel boot param
+	  "randomize_kstack_offset=on/off", and this config chooses the default
+	  boot state.
 
 config ARCH_OPTIONAL_KERNEL_RWX
 	def_bool n

arch/arm/Kconfig

@@ -5,6 +5,7 @@ config ARM
 	select ARCH_32BIT_OFF_T
 	select ARCH_CORRECT_STACKTRACE_ON_KRETPROBE if HAVE_KRETPROBES && FRAME_POINTER && !ARM_UNWIND
 	select ARCH_HAS_BINFMT_FLAT
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DEBUG_VIRTUAL if MMU
 	select ARCH_HAS_DMA_WRITE_COMBINE if !ARM_DMA_MEM_BUFFERABLE
 	select ARCH_HAS_ELF_RANDOMIZE

arch/arm64/Kconfig

@@ -19,6 +19,7 @@ config ARM64
 	select ARCH_ENABLE_SPLIT_PMD_PTLOCK if PGTABLE_LEVELS > 2
 	select ARCH_ENABLE_THP_MIGRATION if TRANSPARENT_HUGEPAGE
 	select ARCH_HAS_CACHE_LINE_SIZE
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DEBUG_VIRTUAL
 	select ARCH_HAS_DEBUG_VM_PGTABLE
 	select ARCH_HAS_DMA_PREP_COHERENT
@@ -1257,7 +1258,7 @@ config HW_PERF_EVENTS
 	def_bool y
 	depends on ARM_PMU
 
-# Supported by clang >= 7.0
+# Supported by clang >= 7.0 or GCC >= 12.0.0
 config CC_HAVE_SHADOW_CALL_STACK
 	def_bool $(cc-option, -fsanitize=shadow-call-stack -ffixed-x18)
 

arch/m68k/Kconfig

@@ -4,6 +4,7 @@ config M68K
 	default y
 	select ARCH_32BIT_OFF_T
 	select ARCH_HAS_BINFMT_FLAT
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DMA_PREP_COHERENT if HAS_DMA && MMU && !COLDFIRE
 	select ARCH_HAS_SYNC_DMA_FOR_DEVICE if HAS_DMA
 	select ARCH_HAVE_NMI_SAFE_CMPXCHG if RMW_INSNS

arch/m68k/include/asm/current.h

@@ -24,6 +24,8 @@ static inline struct task_struct *get_current(void)
 
 #define	current	get_current()
 
-#endif /* CONFNIG_MMU */
+#endif /* CONFIG_MMU */
+
+register unsigned long current_stack_pointer __asm__("sp");
 
 #endif /* !(_M68K_CURRENT_H) */

arch/powerpc/Kconfig

@@ -108,6 +108,7 @@ config PPC
 	select ARCH_ENABLE_MEMORY_HOTPLUG
 	select ARCH_ENABLE_MEMORY_HOTREMOVE
 	select ARCH_HAS_COPY_MC			if PPC64
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DEBUG_VIRTUAL
 	select ARCH_HAS_DEBUG_VM_PGTABLE
 	select ARCH_HAS_DEBUG_WX		if STRICT_KERNEL_RWX

arch/s390/Kconfig

@@ -60,6 +60,7 @@ config S390
 	select ARCH_ENABLE_MEMORY_HOTPLUG if SPARSEMEM
 	select ARCH_ENABLE_MEMORY_HOTREMOVE
 	select ARCH_ENABLE_SPLIT_PMD_PTLOCK if PGTABLE_LEVELS > 2
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DEBUG_VM_PGTABLE
 	select ARCH_HAS_DEBUG_WX
 	select ARCH_HAS_DEVMEM_IS_ALLOWED

arch/sh/Kconfig

@@ -7,6 +7,7 @@ config SUPERH
 	select ARCH_HAVE_CUSTOM_GPIO_H
 	select ARCH_HAVE_NMI_SAFE_CMPXCHG if (GUSA_RB || CPU_SH4A)
 	select ARCH_HAS_BINFMT_FLAT if !MMU
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_GIGANTIC_PAGE
 	select ARCH_HAS_GCOV_PROFILE_ALL
 	select ARCH_HAS_PTE_SPECIAL

arch/x86/Kconfig

@@ -69,6 +69,7 @@ config X86
 	select ARCH_ENABLE_THP_MIGRATION if X86_64 && TRANSPARENT_HUGEPAGE
 	select ARCH_HAS_ACPI_TABLE_UPGRADE	if ACPI
 	select ARCH_HAS_CACHE_LINE_SIZE
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DEBUG_VIRTUAL
 	select ARCH_HAS_DEBUG_VM_PGTABLE	if !X86_PAE
 	select ARCH_HAS_DEVMEM_IS_ALLOWED

arch/xtensa/Kconfig

@@ -3,6 +3,7 @@ config XTENSA
 	def_bool y
 	select ARCH_32BIT_OFF_T
 	select ARCH_HAS_BINFMT_FLAT if !MMU
+	select ARCH_HAS_CURRENT_STACK_POINTER
 	select ARCH_HAS_DMA_PREP_COHERENT if MMU
 	select ARCH_HAS_SYNC_DMA_FOR_CPU if MMU
 	select ARCH_HAS_SYNC_DMA_FOR_DEVICE if MMU

arch/xtensa/include/asm/current.h

@@ -26,6 +26,8 @@ static inline struct task_struct *get_current(void)
 
 #define current get_current()
 
+register unsigned long current_stack_pointer __asm__("a1");
+
 #else
 
 #define GET_CURRENT(reg,sp)		\

arch/xtensa/include/asm/stacktrace.h

@@ -19,14 +19,14 @@ struct stackframe {
 
 static __always_inline unsigned long *stack_pointer(struct task_struct *task)
 {
-	unsigned long *sp;
+	unsigned long sp;
 
 	if (!task || task == current)
-		__asm__ __volatile__ ("mov %0, a1\n" : "=a"(sp));
+		sp = current_stack_pointer;
 	else
-		sp = (unsigned long *)task->thread.sp;
+		sp = task->thread.sp;
 
-	return sp;
+	return (unsigned long *)sp;
 }
 
 void walk_stackframe(unsigned long *sp,

arch/xtensa/kernel/irq.c

@@ -36,9 +36,8 @@ asmlinkage void do_IRQ(int hwirq, struct pt_regs *regs)
 #ifdef CONFIG_DEBUG_STACKOVERFLOW
 	/* Debugging check for stack overflow: is there less than 1KB free? */
 	{
-		unsigned long sp;
+		unsigned long sp = current_stack_pointer;
 
-		__asm__ __volatile__ ("mov %0, a1\n" : "=a" (sp));
 		sp &= THREAD_SIZE - 1;
 
 		if (unlikely(sp < (sizeof(thread_info) + 1024)))

include/linux/compiler-gcc.h

@@ -97,6 +97,10 @@
 #define KASAN_ABI_VERSION 4
 #endif
 
+#ifdef CONFIG_SHADOW_CALL_STACK
+#define __noscs __attribute__((__no_sanitize__("shadow-call-stack")))
+#endif
+
 #if __has_attribute(__no_sanitize_address__)
 #define __no_sanitize_address __attribute__((no_sanitize_address))
 #else

include/linux/randomize_kstack.h

@@ -2,6 +2,7 @@
 #ifndef _LINUX_RANDOMIZE_KSTACK_H
 #define _LINUX_RANDOMIZE_KSTACK_H
 
+#ifdef CONFIG_RANDOMIZE_KSTACK_OFFSET
 #include <linux/kernel.h>
 #include <linux/jump_label.h>
 #include <linux/percpu-defs.h>
@@ -16,8 +17,20 @@ DECLARE_PER_CPU(u32, kstack_offset);
  * alignment. Also, since this use is being explicitly masked to a max of
  * 10 bits, stack-clash style attacks are unlikely. For more details see
  * "VLAs" in Documentation/process/deprecated.rst
+ *
+ * The normal __builtin_alloca() is initialized with INIT_STACK_ALL (currently
+ * only with Clang and not GCC). Initializing the unused area on each syscall
+ * entry is expensive, and generating an implicit call to memset() may also be
+ * problematic (such as in noinstr functions). Therefore, if the compiler
+ * supports it (which it should if it initializes allocas), always use the
+ * "uninitialized" variant of the builtin.
  */
-void *__builtin_alloca(size_t size);
+#if __has_builtin(__builtin_alloca_uninitialized)
+#define __kstack_alloca __builtin_alloca_uninitialized
+#else
+#define __kstack_alloca __builtin_alloca
+#endif
+
 /*
  * Use, at most, 10 bits of entropy. We explicitly cap this to keep the
  * "VLA" from being unbounded (see above). 10 bits leaves enough room for
@@ -36,7 +49,7 @@ void *__builtin_alloca(size_t size);
 	if (static_branch_maybe(CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT,	\
 				&randomize_kstack_offset)) {		\
 		u32 offset = raw_cpu_read(kstack_offset);		\
-		u8 *ptr = __builtin_alloca(KSTACK_OFFSET_MAX(offset));	\
+		u8 *ptr = __kstack_alloca(KSTACK_OFFSET_MAX(offset));	\
 		/* Keep allocation even after "ptr" loses scope. */	\
 		asm volatile("" :: "r"(ptr) : "memory");		\
 	}								\
@@ -50,5 +63,9 @@ void *__builtin_alloca(size_t size);
 		raw_cpu_write(kstack_offset, offset);			\
 	}								\
 } while (0)
+#else /* CONFIG_RANDOMIZE_KSTACK_OFFSET */
+#define add_random_kstack_offset()		do { } while (0)
+#define choose_random_kstack_offset(rand)	do { } while (0)
+#endif /* CONFIG_RANDOMIZE_KSTACK_OFFSET */
 
 #endif

init/main.c

@@ -853,7 +853,7 @@ static void __init mm_init(void)
 	pti_init();
 }
 
-#ifdef CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
+#ifdef CONFIG_RANDOMIZE_KSTACK_OFFSET
 DEFINE_STATIC_KEY_MAYBE_RO(CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT,
 			   randomize_kstack_offset);
 DEFINE_PER_CPU(u32, kstack_offset);

mm/Kconfig

@@ -744,6 +744,15 @@ config IDLE_PAGE_TRACKING
 config ARCH_HAS_CACHE_LINE_SIZE
 	bool
 
+config ARCH_HAS_CURRENT_STACK_POINTER
+	bool
+	help
+	  In support of HARDENED_USERCOPY performing stack variable lifetime
+	  checking, an architecture-agnostic way to find the stack pointer
+	  is needed. Once an architecture defines an unsigned long global
+	  register alias named "current_stack_pointer", this config can be
+	  selected.
+
 config ARCH_HAS_PTE_DEVMAP
 	bool
 

mm/usercopy.c

@@ -29,7 +29,7 @@
  * Returns:
  *	NOT_STACK: not at all on the stack
  *	GOOD_FRAME: fully within a valid stack frame
- *	GOOD_STACK: fully on the stack (when can't do frame-checking)
+ *	GOOD_STACK: within the current stack (when can't frame-check exactly)
  *	BAD_STACK: error condition (invalid stack position or bad stack frame)
  */
 static noinline int check_stack_object(const void *obj, unsigned long len)
@@ -55,6 +55,17 @@ static noinline int check_stack_object(const void *obj, unsigned long len)
 	if (ret)
 		return ret;
 
+	/* Finally, check stack depth if possible. */
+#ifdef CONFIG_ARCH_HAS_CURRENT_STACK_POINTER
+	if (IS_ENABLED(CONFIG_STACK_GROWSUP)) {
+		if ((void *)current_stack_pointer < obj + len)
+			return BAD_STACK;
+	} else {
+		if (obj < (void *)current_stack_pointer)
+			return BAD_STACK;
+	}
+#endif
+
 	return GOOD_STACK;
 }
 
@@ -280,7 +291,15 @@ void __check_object_size(const void *ptr, unsigned long n, bool to_user)
 		 */
 		return;
 	default:
-		usercopy_abort("process stack", NULL, to_user, 0, n);
+		usercopy_abort("process stack", NULL, to_user,
+#ifdef CONFIG_ARCH_HAS_CURRENT_STACK_POINTER
+			IS_ENABLED(CONFIG_STACK_GROWSUP) ?
+				ptr - (void *)current_stack_pointer :
+				(void *)current_stack_pointer - ptr,
+#else
+			0,
+#endif
+			n);
 	}
 
 	/* Check for bad heap object. */

scripts/Makefile.gcc-plugins

@@ -37,6 +37,8 @@ gcc-plugin-cflags-$(CONFIG_GCC_PLUGIN_STACKLEAK)		\
 		+= -fplugin-arg-stackleak_plugin-track-min-size=$(CONFIG_STACKLEAK_TRACK_MIN_SIZE)
 gcc-plugin-cflags-$(CONFIG_GCC_PLUGIN_STACKLEAK)		\
 		+= -fplugin-arg-stackleak_plugin-arch=$(SRCARCH)
+gcc-plugin-cflags-$(CONFIG_GCC_PLUGIN_STACKLEAK_VERBOSE)	\
+		+= -fplugin-arg-stackleak_plugin-verbose
 ifdef CONFIG_GCC_PLUGIN_STACKLEAK
     DISABLE_STACKLEAK_PLUGIN += -fplugin-arg-stackleak_plugin-disable
 endif

scripts/gcc-plugins/stackleak_plugin.c

@@ -429,6 +429,23 @@ static unsigned int stackleak_cleanup_execute(void)
 	return 0;
 }
 
+/*
+ * STRING_CST may or may not be NUL terminated:
+ * https://gcc.gnu.org/onlinedocs/gccint/Constant-expressions.html
+ */
+static inline bool string_equal(tree node, const char *string, int length)
+{
+	if (TREE_STRING_LENGTH(node) < length)
+		return false;
+	if (TREE_STRING_LENGTH(node) > length + 1)
+		return false;
+	if (TREE_STRING_LENGTH(node) == length + 1 &&
+	    TREE_STRING_POINTER(node)[length] != '\0')
+		return false;
+	return !memcmp(TREE_STRING_POINTER(node), string, length);
+}
+#define STRING_EQUAL(node, str)	string_equal(node, str, strlen(str))
+
 static bool stackleak_gate(void)
 {
 	tree section;
@@ -438,13 +455,17 @@ static bool stackleak_gate(void)
 	if (section && TREE_VALUE(section)) {
 		section = TREE_VALUE(TREE_VALUE(section));
 
-		if (!strncmp(TREE_STRING_POINTER(section), ".init.text", 10))
+		if (STRING_EQUAL(section, ".init.text"))
+			return false;
+		if (STRING_EQUAL(section, ".devinit.text"))
+			return false;
+		if (STRING_EQUAL(section, ".cpuinit.text"))
 			return false;
-		if (!strncmp(TREE_STRING_POINTER(section), ".devinit.text", 13))
+		if (STRING_EQUAL(section, ".meminit.text"))
 			return false;
-		if (!strncmp(TREE_STRING_POINTER(section), ".cpuinit.text", 13))
+		if (STRING_EQUAL(section, ".noinstr.text"))
 			return false;
-		if (!strncmp(TREE_STRING_POINTER(section), ".meminit.text", 13))
+		if (STRING_EQUAL(section, ".entry.text"))
 			return false;
 	}
 

security/Kconfig.hardening

@@ -174,6 +174,16 @@ config GCC_PLUGIN_STACKLEAK
 	   * https://grsecurity.net/
 	   * https://pax.grsecurity.net/
 
+config GCC_PLUGIN_STACKLEAK_VERBOSE
+	bool "Report stack depth analysis instrumentation" if EXPERT
+	depends on GCC_PLUGIN_STACKLEAK
+	depends on !COMPILE_TEST	# too noisy
+	help
+	  This option will cause a warning to be printed each time the
+	  stackleak plugin finds a function it thinks needs to be
+	  instrumented. This is useful for comparing coverage between
+	  builds.
+
 config STACKLEAK_TRACK_MIN_SIZE
 	int "Minimum stack frame size of functions tracked by STACKLEAK"
 	default 100