Merge branch 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 pti fixes from Thomas Gleixner:
 "A small set of fixes for the meltdown/spectre mitigations:

   - Make kprobes aware of retpolines to prevent probes in the retpoline
     thunks.

   - Make the machine check exception speculation protected. MCE used to
     issue an indirect call directly from the ASM entry code. Convert
     that to a direct call into a C-function and issue the indirect call
     from there so the compiler can add the retpoline protection,

   - Make the vmexit_fill_RSB() assembly less stupid

   - Fix a typo in the PTI documentation"

* 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/retpoline: Optimize inline assembler for vmexit_fill_RSB
  x86/pti: Document fix wrong index
  kprobes/x86: Disable optimizing on the function jumps to indirect thunk
  kprobes/x86: Blacklist indirect thunk functions for kprobes
  retpoline: Introduce start/end markers of indirect thunk
  x86/mce: Make machine check speculation protected

Documentation/x86/pti.txt

@@ -78,7 +78,7 @@ this protection comes at a cost:
      non-PTI SYSCALL entry code, so requires mapping fewer
      things into the userspace page tables.  The downside is
      that stacks must be switched at entry time.
-  d. Global pages are disabled for all kernel structures not
+  c. Global pages are disabled for all kernel structures not
      mapped into both kernel and userspace page tables.  This
      feature of the MMU allows different processes to share TLB
      entries mapping the kernel.  Losing the feature means more

arch/x86/entry/entry_64.S

@@ -1264,7 +1264,7 @@ idtentry async_page_fault	do_async_page_fault	has_error_code=1
 #endif
 
 #ifdef CONFIG_X86_MCE
-idtentry machine_check					has_error_code=0	paranoid=1 do_sym=*machine_check_vector(%rip)
+idtentry machine_check		do_mce			has_error_code=0	paranoid=1
 #endif
 
 /*

arch/x86/include/asm/nospec-branch.h

@@ -194,6 +194,9 @@ enum spectre_v2_mitigation {
 	SPECTRE_V2_IBRS,
 };
 
+extern char __indirect_thunk_start[];
+extern char __indirect_thunk_end[];
+
 /*
  * On VMEXIT we must ensure that no RSB predictions learned in the guest
  * can be followed in the host, by overwriting the RSB completely. Both
@@ -203,16 +206,17 @@ enum spectre_v2_mitigation {
 static inline void vmexit_fill_RSB(void)
 {
 #ifdef CONFIG_RETPOLINE
-	unsigned long loops = RSB_CLEAR_LOOPS / 2;
+	unsigned long loops;
 
 	asm volatile (ANNOTATE_NOSPEC_ALTERNATIVE
 		      ALTERNATIVE("jmp 910f",
 				  __stringify(__FILL_RETURN_BUFFER(%0, RSB_CLEAR_LOOPS, %1)),
 				  X86_FEATURE_RETPOLINE)
 		      "910:"
-		      : "=&r" (loops), ASM_CALL_CONSTRAINT
-		      : "r" (loops) : "memory" );
+		      : "=r" (loops), ASM_CALL_CONSTRAINT
+		      : : "memory" );
 #endif
 }
+
 #endif /* __ASSEMBLY__ */
 #endif /* __NOSPEC_BRANCH_H__ */

arch/x86/include/asm/traps.h

@@ -88,6 +88,7 @@ dotraplinkage void do_simd_coprocessor_error(struct pt_regs *, long);
 #ifdef CONFIG_X86_32
 dotraplinkage void do_iret_error(struct pt_regs *, long);
 #endif
+dotraplinkage void do_mce(struct pt_regs *, long);
 
 static inline int get_si_code(unsigned long condition)
 {

arch/x86/kernel/cpu/mcheck/mce.c

@@ -1785,6 +1785,11 @@ static void unexpected_machine_check(struct pt_regs *regs, long error_code)
 void (*machine_check_vector)(struct pt_regs *, long error_code) =
 						unexpected_machine_check;
 
+dotraplinkage void do_mce(struct pt_regs *regs, long error_code)
+{
+	machine_check_vector(regs, error_code);
+}
+
 /*
  * Called for each booted CPU to set up machine checks.
  * Must be called with preempt off:

arch/x86/kernel/kprobes/opt.c

@@ -40,6 +40,7 @@
 #include <asm/debugreg.h>
 #include <asm/set_memory.h>
 #include <asm/sections.h>
+#include <asm/nospec-branch.h>
 
 #include "common.h"
 
@@ -203,7 +204,7 @@ static int copy_optimized_instructions(u8 *dest, u8 *src, u8 *real)
 }
 
 /* Check whether insn is indirect jump */
-static int insn_is_indirect_jump(struct insn *insn)
+static int __insn_is_indirect_jump(struct insn *insn)
 {
 	return ((insn->opcode.bytes[0] == 0xff &&
 		(X86_MODRM_REG(insn->modrm.value) & 6) == 4) || /* Jump */
@@ -237,6 +238,26 @@ static int insn_jump_into_range(struct insn *insn, unsigned long start, int len)
 	return (start <= target && target <= start + len);
 }
 
+static int insn_is_indirect_jump(struct insn *insn)
+{
+	int ret = __insn_is_indirect_jump(insn);
+
+#ifdef CONFIG_RETPOLINE
+	/*
+	 * Jump to x86_indirect_thunk_* is treated as an indirect jump.
+	 * Note that even with CONFIG_RETPOLINE=y, the kernel compiled with
+	 * older gcc may use indirect jump. So we add this check instead of
+	 * replace indirect-jump check.
+	 */
+	if (!ret)
+		ret = insn_jump_into_range(insn,
+				(unsigned long)__indirect_thunk_start,
+				(unsigned long)__indirect_thunk_end -
+				(unsigned long)__indirect_thunk_start);
+#endif
+	return ret;
+}
+
 /* Decode whole function to ensure any instructions don't jump into target */
 static int can_optimize(unsigned long paddr)
 {

arch/x86/kernel/vmlinux.lds.S

@@ -124,6 +124,12 @@ SECTIONS
 		ASSERT(. - _entry_trampoline == PAGE_SIZE, "entry trampoline is too big");
 #endif
 
+#ifdef CONFIG_RETPOLINE
+		__indirect_thunk_start = .;
+		*(.text.__x86.indirect_thunk)
+		__indirect_thunk_end = .;
+#endif
+
 		/* End of text section */
 		_etext = .;
 	} :text = 0x9090

arch/x86/lib/retpoline.S

@@ -9,7 +9,7 @@
 #include <asm/nospec-branch.h>
 
 .macro THUNK reg
-	.section .text.__x86.indirect_thunk.\reg
+	.section .text.__x86.indirect_thunk
 
 ENTRY(__x86_indirect_thunk_\reg)
 	CFI_STARTPROC
@@ -25,7 +25,8 @@ ENDPROC(__x86_indirect_thunk_\reg)
  * than one per register with the correct names. So we do it
  * the simple and nasty way...
  */
-#define EXPORT_THUNK(reg) EXPORT_SYMBOL(__x86_indirect_thunk_ ## reg)
+#define __EXPORT_THUNK(sym) _ASM_NOKPROBE(sym); EXPORT_SYMBOL(sym)
+#define EXPORT_THUNK(reg) __EXPORT_THUNK(__x86_indirect_thunk_ ## reg)
 #define GENERATE_THUNK(reg) THUNK reg ; EXPORT_THUNK(reg)
 
 GENERATE_THUNK(_ASM_AX)