Re: [RFC PATCH 6/6] uprobes/x86: Emulate rip-relative conditional "near" jmp's

[Jim Keniston]

On Mon, 2014-04-07 at 16:28 +0200, Oleg Nesterov wrote:
> On 04/06, Oleg Nesterov wrote:
> >
> > But I'll try to cleanup this patch...
> 
> See v2 below.
> 
> -------------------------------------------------------------------------------
> Subject: [RFC PATCH v2 6/6] uprobes/x86: Emulate rip-relative conditional "near" jmp's
> 
> It seems that 16bit condi jmp is just 0x0f + short_jump_opc_plus_0x10.

Yes, but the code could use a comment to that effect.  See below.

Searching for "jump" in the AMD manual, I see that there are 3 other
instructions that are essentially conditional branches: loop,
loope/loopz, and loopne/loopnz.  They decrement ecx/rcx and then
conditionally branch.  The offset is always 8 bits.

BTW, patches 2 and 3 look fine to me.

> 
> Reported-by: Jonathan Lebon <jlebon@xxxxxxxxxx>
> Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
> ---
>  arch/x86/kernel/uprobes.c |    4 ++++
>  1 files changed, 4 insertions(+), 0 deletions(-)
> 
> diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
> index 3865d8b..dae02f9 100644
> --- a/arch/x86/kernel/uprobes.c
> +++ b/arch/x86/kernel/uprobes.c
> @@ -591,6 +591,10 @@ static int ttt_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
>  		auprobe->ttt.opc1 = opc1;
>  		break;
> 
> +	case 0x0f:
> +		if (insn->opcode.nbytes != 2)
> +			return -ENOSYS;
		/*
		 * Map 0f 8x (Jcc with 32-bit displacement) to 7x
		 * (Jcc with 8-bit displacement).  insn lib maps both
		 * to 32 bits.
		 */
> +		opc1 = OPCODE2(insn) - 0x10;
>  	default:
>  		if (!is_cond_jmp_opcode(opc1))
>  			return -ENOSYS;

Jim

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/