Re: [PATCH v5 3.1.0-rc4-tip 17/26] x86: arch specific hooks forpre/post singlestep handling.

[Peter Zijlstra]

On Mon, 2011-09-26 at 22:04 +0530, Srikar Dronamraju wrote:
> * Peter Zijlstra <peterz@xxxxxxxxxxxxx> [2011-09-26 16:23:53]:
> 
> > On Tue, 2011-09-20 at 17:33 +0530, Srikar Dronamraju wrote:
> > > +fail:
> > > +       pr_warn_once("uprobes: Failed to adjust return address after"
> > > +               " single-stepping call instruction;"
> > > +               " pid=%d, sp=%#lx\n", current->pid, sp);
> > > +       return -EFAULT; 
> > 
> > So how can that happen? Single-Step while someone unmapped the stack?
> 
> We do a copy_to_user, copy_from_user just above this,

I saw that,

>  Now if either of
> them fail, we have no choice but to Bail out.

Agreed,

>  What caused this EFault may not be under Uprobes's Control.

I never said it was.. All I asked is what (outside of uprobe) was done
to cause this, and why is this particular error important enough to
warrant a warn.
--
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/