Re: [PATCH 26/33] x86/kvm: Add stack frame dependency to test_cc() inline asm

[Paolo Bonzini]

On 21/01/2016 23:49, Josh Poimboeuf wrote:
> With some configs, gcc doesn't inline test_cc().  When that happens, it
> doesn't create a stack frame before inserting the call instruction.
> This breaks frame pointer convention if CONFIG_FRAME_POINTER is enabled
> and can result in a bad stack trace.
> 
> Force a stack frame to be created if CONFIG_FRAME_POINTER is enabled by
> listing the stack pointer as an output operand for the inline asm
> statement.

If an __always_inline allocation works, that would be better.

Paolo

> Signed-off-by: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> Cc: Gleb Natapov <gleb@xxxxxxxxxx>
> Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> Cc: kvm@xxxxxxxxxxxxxxx
> ---
>  arch/x86/kvm/emulate.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
> index aa4d726..7dba65a 100644
> --- a/arch/x86/kvm/emulate.c
> +++ b/arch/x86/kvm/emulate.c
> @@ -972,11 +972,13 @@ static int em_bsr_c(struct x86_emulate_ctxt *ctxt)
>  static u8 test_cc(unsigned int condition, unsigned long flags)
>  {
>  	u8 rc;
> +	register void *__sp asm(_ASM_SP);
>  	void (*fop)(void) = (void *)em_setcc + 4 * (condition & 0xf);
>  
>  	flags = (flags & EFLAGS_MASK) | X86_EFLAGS_IF;
>  	asm("push %[flags]; popf; call *%[fastop]"
> -	    : "=a"(rc) : [fastop]"r"(fop), [flags]"r"(flags));
> +	    : "=a"(rc), "+r"(__sp)
> +	    : [fastop]"r"(fop), [flags]"r"(flags));
>  	return rc;
>  }
>  
>