Re: [PATCH v4 3.0-rc2-tip 13/22] 13: uprobes: Handing int3 andsinglestep exception.

[Srikar Dronamraju]

> > +
> > +cleanup_ret:
> > +       if (u) {
> > +               down_read(&mm->mmap_sem);
> > +               if (!set_orig_insn(current, u, probept, true))
> 
> we try to undo the probe? That doesn't make any sense. I thought you
> meant to return to userspace, let it re-take the trap and try again
> until you do manage to allocate the user resource.

I meant removing the probe itself
https://lkml.org/lkml/2011/4/21/279

We could try reseting and retrying the trap. Just that we might end up
looping under memory pressure.

> 
> This behaviour makes probes totally unreliable under memory pressure. 

Under memory pressure we could be unreliable.

> 
> > +                       atomic_dec(&mm->uprobes_count);
> > +               up_read(&mm->mmap_sem);
> > +               put_uprobe(u);
> > +       } else {
> > +       /*TODO Return SIGTRAP signal */
> > +       }
> > +       if (utask) {
> > +               utask->active_uprobe = NULL;
> > +               utask->state = UTASK_RUNNING;
> > +       }
> > +       set_instruction_pointer(regs, probept);
> > +} 
> 
> Also, there's a scary amount of TODO in there...

All of those deal with delaying the signals. I am working on it at this
moment. 

-- 
Thanks and Regards
Srikar
--
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/