Merge branch 'core-objtool-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull objtool updates from Ingo Molnar:
 "This is a series from Peter Zijlstra that adds x86 build-time uaccess
  validation of SMAP to objtool, which will detect and warn about the
  following uaccess API usage bugs and weirdnesses:

   - call to %s() with UACCESS enabled
   - return with UACCESS enabled
   - return with UACCESS disabled from a UACCESS-safe function
   - recursive UACCESS enable
   - redundant UACCESS disable
   - UACCESS-safe disables UACCESS

  As it turns out not leaking uaccess permissions outside the intended
  uaccess functionality is hard when the interfaces are complex and when
  such bugs are mostly dormant.

  As a bonus we now also check the DF flag. We had at least one
  high-profile bug in that area in the early days of Linux, and the
  checking is fairly simple. The checks performed and warnings emitted
  are:

   - call to %s() with DF set
   - return with DF set
   - return with modified stack frame
   - recursive STD
   - redundant CLD

  It's all x86-only for now, but later on this can also be used for PAN
  on ARM and objtool is fairly cross-platform in principle.

  While all warnings emitted by this new checking facility that got
  reported to us were fixed, there might be GCC version dependent
  warnings that were not reported yet - which we'll address, should they
  trigger.

  The warnings are non-fatal build warnings"

* 'core-objtool-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (27 commits)
  mm/uaccess: Use 'unsigned long' to placate UBSAN warnings on older GCC versions
  x86/uaccess: Dont leak the AC flag into __put_user() argument evaluation
  sched/x86_64: Don't save flags on context switch
  objtool: Add Direction Flag validation
  objtool: Add UACCESS validation
  objtool: Fix sibling call detection
  objtool: Rewrite alt->skip_orig
  objtool: Add --backtrace support
  objtool: Rewrite add_ignores()
  objtool: Handle function aliases
  objtool: Set insn->func for alternatives
  x86/uaccess, kcov: Disable stack protector
  x86/uaccess, ftrace: Fix ftrace_likely_update() vs. SMAP
  x86/uaccess, ubsan: Fix UBSAN vs. SMAP
  x86/uaccess, kasan: Fix KASAN vs SMAP
  x86/smap: Ditch __stringify()
  x86/uaccess: Introduce user_access_{save,restore}()
  x86/uaccess, signal: Fix AC=1 bloat
  x86/uaccess: Always inline user_access_begin()
  x86/uaccess, xen: Suppress SMAP warnings
  ...

3 files changed, 50 insertions, 21 deletions

diff --git a/arch/x86/lib/copy_user_64.S b/arch/x86/lib/copy_user_64.S
index db4e5aa0858b..b2f1822084ae 100644
--- a/arch/x86/lib/copy_user_64.S
+++ b/arch/x86/lib/copy_user_64.S
@@ -16,6 +16,30 @@
 #include <asm/smap.h>
 #include <asm/export.h>
 
+.macro ALIGN_DESTINATION
+	/* check for bad alignment of destination */
+	movl %edi,%ecx
+	andl $7,%ecx
+	jz 102f				/* already aligned */
+	subl $8,%ecx
+	negl %ecx
+	subl %ecx,%edx
+100:	movb (%rsi),%al
+101:	movb %al,(%rdi)
+	incq %rsi
+	incq %rdi
+	decl %ecx
+	jnz 100b
+102:
+	.section .fixup,"ax"
+103:	addl %ecx,%edx			/* ecx is zerorest also */
+	jmp copy_user_handle_tail
+	.previous
+
+	_ASM_EXTABLE_UA(100b, 103b)
+	_ASM_EXTABLE_UA(101b, 103b)
+	.endm
+
 /*
  * copy_user_generic_unrolled - memory copy with exception handling.
  * This version is for CPUs like P4 that don't have efficient micro
@@ -194,6 +218,30 @@ ENDPROC(copy_user_enhanced_fast_string)
 EXPORT_SYMBOL(copy_user_enhanced_fast_string)
 
 /*
+ * Try to copy last bytes and clear the rest if needed.
+ * Since protection fault in copy_from/to_user is not a normal situation,
+ * it is not necessary to optimize tail handling.
+ *
+ * Input:
+ * rdi destination
+ * rsi source
+ * rdx count
+ *
+ * Output:
+ * eax uncopied bytes or 0 if successful.
+ */
+ALIGN;
+copy_user_handle_tail:
+	movl %edx,%ecx
+1:	rep movsb
+2:	mov %ecx,%eax
+	ASM_CLAC
+	ret
+
+	_ASM_EXTABLE_UA(1b, 2b)
+ENDPROC(copy_user_handle_tail)
+
+/*
  * copy_user_nocache - Uncached memory copy with exception handling
  * This will force destination out of cache for more performance.
  *
diff --git a/arch/x86/lib/memcpy_64.S b/arch/x86/lib/memcpy_64.S
index 3b24dc05251c..9d05572370ed 100644
--- a/arch/x86/lib/memcpy_64.S
+++ b/arch/x86/lib/memcpy_64.S
@@ -257,6 +257,7 @@ ENTRY(__memcpy_mcsafe)
 	/* Copy successful. Return zero */
 .L_done_memcpy_trap:
 	xorl %eax, %eax
+.L_done:
 	ret
 ENDPROC(__memcpy_mcsafe)
 EXPORT_SYMBOL_GPL(__memcpy_mcsafe)
@@ -273,7 +274,7 @@ EXPORT_SYMBOL_GPL(__memcpy_mcsafe)
 	addl	%edx, %ecx
 .E_trailing_bytes:
 	mov	%ecx, %eax
-	ret
+	jmp	.L_done
 
 	/*
 	 * For write fault handling, given the destination is unaligned,
diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c
index ee42bb0cbeb3..9952a01cad24 100644
--- a/arch/x86/lib/usercopy_64.c
+++ b/arch/x86/lib/usercopy_64.c
@@ -55,26 +55,6 @@ unsigned long clear_user(void __user *to, unsigned long n)
 EXPORT_SYMBOL(clear_user);
 
 /*
- * Try to copy last bytes and clear the rest if needed.
- * Since protection fault in copy_from/to_user is not a normal situation,
- * it is not necessary to optimize tail handling.
- */
-__visible unsigned long
-copy_user_handle_tail(char *to, char *from, unsigned len)
-{
-	for (; len; --len, to++) {
-		char c;
-
-		if (__get_user_nocheck(c, from++, sizeof(char)))
-			break;
-		if (__put_user_nocheck(c, to, sizeof(char)))
-			break;
-	}
-	clac();
-	return len;
-}
-
-/*
  * Similar to copy_user_handle_tail, probe for the write fault point,
  * but reuse __memcpy_mcsafe in case a new read error is encountered.
  * clac() is handled in _copy_to_iter_mcsafe().