On Mon, 2014-04-28 at 19:06 +0200, Denys Vlasenko wrote:
> Otherwise, instructions such as cmpxchg and div will be mishandled.
>
> Signed-off-by: Denys Vlasenko <[email protected]>
> CC: Jim Keniston <[email protected]>
> CC: Masami Hiramatsu <[email protected]>
> CC: Srikar Dronamraju <[email protected]>
> CC: Ingo Molnar <[email protected]>
> CC: Oleg Nesterov <[email protected]>
> ---
> arch/x86/kernel/uprobes.c | 57 ++++++++++++++++++++++++++++++-----------------
> 1 file changed, 37 insertions(+), 20 deletions(-)
>
> diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
...
> @@ -296,41 +296,58 @@ static void riprel_analyze(struct arch_uprobe *auprobe, struct insn *insn)
> */
> cursor = auprobe->insn + insn_offset_modrm(insn);
> /*
> - * Convert from rip-relative addressing to register-relative addressing
> - * via a scratch register.
> + * Convert from rip-relative addressing
> + * to register-relative addressing via a scratch register.
> */
This comment looks like a regression. :-)
Looks good otherwise (setting aside your later findings about cmpxchg8b
and such -- I guess we need some way to helpfully reject rip-relative
forms of such instructions).
Reviewed-by: Jim Keniston <[email protected]>