Re: [PATCH v3.1 09/10] x86/unwind: add ORC unwinder

From: Josh Poimboeuf
Date: Thu Jul 20 2017 - 17:16:32 EST


On Thu, Jul 20, 2017 at 09:12:16AM +0200, Jiri Slaby wrote:
> On 07/14/2017, 07:22 PM, Josh Poimboeuf wrote:
> > +void __unwind_start(struct unwind_state *state, struct task_struct *task,
> > + struct pt_regs *regs, unsigned long *first_frame)
> > +{
> > + memset(state, 0, sizeof(*state));
> > + state->task = task;
> > +
> > + /*
> > + * Refuse to unwind the stack of a task while it's executing on another
> > + * CPU. This check is racy, but that's ok: the unwinder has other
> > + * checks to prevent it from going off the rails.
> > + */
> > + if (task_on_another_cpu(task))
> > + goto done;
> > +
> > + if (regs) {
> > + if (user_mode(regs))
> > + goto done;
> > +
> > + state->ip = regs->ip;
> > + state->sp = kernel_stack_pointer(regs);
> > + state->bp = regs->bp;
> > + state->regs = regs;
> > + state->full_regs = true;
> > + state->signal = true;
> > +
> > + } else if (task == current) {
> > + asm volatile("lea (%%rip), %0\n\t"
> > + "mov %%rsp, %1\n\t"
> > + "mov %%rbp, %2\n\t"
> > + : "=r" (state->ip), "=r" (state->sp),
> > + "=r" (state->bp));
> > +
> > + } else {
> > + struct inactive_task_frame *frame = (void *)task->thread.sp;
> > +
> > + state->ip = frame->ret_addr;
> > + state->sp = task->thread.sp;
> > + state->bp = frame->bp;
>
> I wonder, if the reads from 'frame' should have READ_ONCE_NOCHECK for
> the same reason as in:
> commit 84936118bdf37bda513d4a361c38181a216427e0
> Author: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> Date: Mon Jan 9 12:00:23 2017 -0600
>
> x86/unwind: Disable KASAN checks for non-current tasks
> ?

Yeah, maybe so. Since the task_on_another_cpu() check above is racy,
here it's remotely possible that the task has since starting executing
and has poisoned the stack memory we're about to read.

I don't know how realistic that scenario is, but it wouldn't hurt to add
a couple of READ_ONCE_NOCHECKs here for the 'frame' dereferences.

--
Josh