5.15-stable review patch. If anyone has any objections, please let me know.
------------------
From: John Sperbeck <jsperbeck@xxxxxxxxxx>
When dbf460087755 ("objtool/x86: Fixup frame-pointer vs rethunk")
was backported to some stable branches, the check for dest->embedded_insn
in is_special_call() was missed. The result is that the warning it
was intended to suppress still appears. For example on 6.1 (on kernels
before 6.1, the '-s' argument would instead be 'check'):
$ tools/objtool/objtool -s arch/x86/lib/retpoline.o
arch/x86/lib/retpoline.o: warning: objtool: srso_untrain_ret+0xd:
call without frame pointer save/setup
With this patch, the warning is correctly suppressed, and the
kernel still passes the normal Google kernel developer tests.
Signed-off-by: John Sperbeck <jsperbeck@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
tools/objtool/check.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -2202,7 +2202,7 @@ static bool is_special_call(struct instr
if (!dest)
return false;
- if (dest->fentry)
+ if (dest->fentry || dest->embedded_insn)
return true;
}
