Re: [PATCH] x86/asm/entry/64: Add forgotten CFI annotation

[Ingo Molnar <mingo@kernel.org>]

* Denys Vlasenko <dvlasenk@redhat.com> wrote:

> Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
> CC: Linus Torvalds <torvalds@linux-foundation.org>
> CC: Steven Rostedt <rostedt@goodmis.org>
> CC: Ingo Molnar <mingo@kernel.org>
> CC: Borislav Petkov <bp@alien8.de>
> CC: "H. Peter Anvin" <hpa@zytor.com>
> CC: Andy Lutomirski <luto@amacapital.net>
> CC: Oleg Nesterov <oleg@redhat.com>
> CC: Frederic Weisbecker <fweisbec@gmail.com>
> CC: Alexei Starovoitov <ast@plumgrid.com>
> CC: Will Drewry <wad@chromium.org>
> CC: Kees Cook <keescook@chromium.org>
> CC: x86@kernel.org
> CC: linux-kernel@vger.kernel.org
> ---
>  arch/x86/kernel/entry_64.S | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
> index 65485b3..15261ba 100644
> --- a/arch/x86/kernel/entry_64.S
> +++ b/arch/x86/kernel/entry_64.S
> @@ -429,6 +429,7 @@ return_from_execve:
>  1:
>  	/* must use IRET code path (pt_regs->cs may have changed) */
>  	addq	$8, %rsp
> +	CFI_ADJUST_CFA_OFFSET -8
>  	ZERO_EXTRA_REGS
>  	movq	%rax,RAX(%rsp)
>  	jmp	int_ret_from_sys_call

So given that this is a pretty common pattern:

triton:~/tip> git grep -A 1 'addq.*%rsp' arch/x86/kernel/entry_64.S
arch/x86/kernel/entry_64.S:     addq    $8, %rsp
arch/x86/kernel/entry_64.S-     CFI_ADJUST_CFA_OFFSET -8
--
arch/x86/kernel/entry_64.S:     addq $8, %rsp
arch/x86/kernel/entry_64.S-     DEFAULT_FRAME 0
--
arch/x86/kernel/entry_64.S:     addq $8, %rsp
arch/x86/kernel/entry_64.S-     DEFAULT_FRAME 0
--
arch/x86/kernel/entry_64.S:     addq $-0x80,(%rsp)              /* Adjust vector to [-256,-1] range */
arch/x86/kernel/entry_64.S-     interrupt do_IRQ
--
arch/x86/kernel/entry_64.S:     addq $0x30,%rsp
arch/x86/kernel/entry_64.S-     CFI_ADJUST_CFA_OFFSET -0x30
--
arch/x86/kernel/entry_64.S:     addq $0x30,%rsp
arch/x86/kernel/entry_64.S-     CFI_ADJUST_CFA_OFFSET -0x30
--
arch/x86/kernel/entry_64.S:     addq $(6*8), %rsp
arch/x86/kernel/entry_64.S-     CFI_ADJUST_CFA_OFFSET -6*8
--
arch/x86/kernel/entry_64.S:     addq $(10*8), %rsp
arch/x86/kernel/entry_64.S-     CFI_ADJUST_CFA_OFFSET -10*8


it might make sense to introduce addq_cfi, to further hide the dwarf 
uglies?

btw., doesn't:

arch/x86/kernel/entry_64.S:     addq $-0x80,(%rsp)              /* Adjust vector to [-256,-1] range */

before the do_IRQ() call need a CFI adjustment as well?

Thanks,

	Ingo