x86/retbleed: Fix return thunk alignment

SYM_FUNC_START_LOCAL_NOALIGN() adds an endbr leading to this layout
(leaving only the last 2 bytes of the address):

  3bff <zen_untrain_ret>:
  3bff:       f3 0f 1e fa             endbr64
  3c03:       f6                      test   $0xcc,%bl

  3c04 <__x86_return_thunk>:
  3c04:       c3                      ret
  3c05:       cc                      int3
  3c06:       0f ae e8                lfence

However, "the RET at __x86_return_thunk must be on a 64 byte boundary,
for alignment within the BTB."

Use SYM_START instead.

Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Cc: <stable@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/arch/x86/lib/retpoline.S b/arch/x86/lib/retpoline.S
index 27ef53f..b3b1e37 100644 (file)
--- a/arch/x86/lib/retpoline.S
+++ b/arch/x86/lib/retpoline.S
@@ -144,8 +144,8 @@ SYM_CODE_END(__x86_indirect_jump_thunk_array)
  */
        .align 64
        .skip 63, 0xcc
  */
        .align 64
        .skip 63, 0xcc

-SYM_FUNC_START_NOALIGN(zen_untrain_ret);
-
+SYM_START(zen_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
+       ANNOTATE_NOENDBR

        /*
         * As executed from zen_untrain_ret, this is:
         *
        /*
         * As executed from zen_untrain_ret, this is:
         *