Re: linux-next: Tree for Apr 16 (objtool: warnings)

[Peter Zijlstra]

On Mon, Jun 21, 2021 at 03:34:32PM +0200, Peter Zijlstra wrote:
> On Fri, Apr 16, 2021 at 10:50:49AM -0700, Randy Dunlap wrote:
> > On 4/16/21 4:36 AM, Stephen Rothwell wrote:
> > > Hi all,
> > > 
> > > Changes since 20210415:
> > 
> > 
> > on x86_64, objtool is not liking retpoline.o:
> > 
> > $ gcc --version
> > gcc (SUSE Linux) 7.5.0
> > 
> > 
> > Is there already a patch for these?
> > 
> > 
> > ===== build-r8840.out =====
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rax()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rbx()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rcx()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rdx()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rsi()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rdi()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_rbp()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r8()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r9()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r10()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r11()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r12()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r13()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r14()+0x0: call without frame pointer save/setup
> > arch/x86/lib/retpoline.o: warning: objtool: __x86_indirect_alt_call_r15()+0x0: call without frame pointer save/setup
> 
> Damn, sorry I missed this. I know what the problem is, but I've yet to
> find a solution that's not terrible... hold on.

---
Subject: objtool/x86: Ignore __x86_indirect_alt_* symbols
From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Date: Mon Jun 21 16:13:55 CEST 2021

Because the __x86_indirect_alt* symbols are just that, objtool will
try and validate them as regular symbols, instead of the alternative
replacements that they are.

This goes sideways for FRAME_POINTER=y builds; which generate a fair
amount of warnings.

Fixes: 9bc0bb50727c ("objtool/x86: Rewrite retpoline thunk calls")
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
---
 arch/x86/lib/retpoline.S |    4 ++++
 1 file changed, 4 insertions(+)

--- a/arch/x86/lib/retpoline.S
+++ b/arch/x86/lib/retpoline.S
@@ -58,12 +58,16 @@ SYM_FUNC_START_NOALIGN(__x86_indirect_al
 2:	.skip	5-(2b-1b), 0x90
 SYM_FUNC_END(__x86_indirect_alt_call_\reg)
 
+STACK_FRAME_NON_STANDARD(__x86_indirect_alt_call_\reg)
+
 SYM_FUNC_START_NOALIGN(__x86_indirect_alt_jmp_\reg)
 	ANNOTATE_RETPOLINE_SAFE
 1:	jmp	*%\reg
 2:	.skip	5-(2b-1b), 0x90
 SYM_FUNC_END(__x86_indirect_alt_jmp_\reg)
 
+STACK_FRAME_NON_STANDARD(__x86_indirect_alt_jmp_\reg)
+
 .endm
 
 /*