Re: [RFC] [PATCH 2.6.37-rc5-tip 13/20] 13: x86: x86 specific probe handling

[Peter Zijlstra <peterz@infradead.org>]

On Thu, 2011-01-27 at 15:10 +0530, Srikar Dronamraju wrote:
> * Peter Zijlstra <peterz@infradead.org> [2011-01-25 14:56:22]:
> 
> > On Thu, 2010-12-16 at 15:29 +0530, Srikar Dronamraju wrote:
> > > 
> > > +void arch_uprobe_enable_sstep(struct pt_regs *regs)
> > > +{
> > > +       /*
> > > +        * Enable single-stepping by
> > > +        * - Set TF on stack
> > > +        * - Set TIF_SINGLESTEP: Guarantees that TF is set when
> > > +        *      returning to user mode.
> > > +        *  - Indicate that TF is set by us.
> > > +        */
> > > +       regs->flags |= X86_EFLAGS_TF;
> > > +       set_thread_flag(TIF_SINGLESTEP);
> > > +       set_thread_flag(TIF_FORCED_TF);
> > > +}
> > > +
> > > +void arch_uprobe_disable_sstep(struct pt_regs *regs)
> > > +{
> > > +       /* Disable single-stepping by clearing what we set */
> > > +       clear_thread_flag(TIF_SINGLESTEP);
> > > +       clear_thread_flag(TIF_FORCED_TF);
> > > +       regs->flags &= ~X86_EFLAGS_TF;
> > > +} 
> > 
> > Why not use the code from arch/x86/kernel/step.c?
> 
> user_enable_single_step and user_disable_single_step that are
> defined in arch/x86/kernel/step.c cant be called in interrupt context.

Right, because of is_setting_trap_flag()..

> Initially we were looking at enabling/disabling singlestep in
> interrupt context. Even now we disable singlestep in post notifier in
> interrupt context.
> 
> Though  arch/x86/kernel/step.c has a static function
> enable_single_step which is identical to arch_uprobe_enable_sstep;
> there is no equivalent function for arch_uprobe_disable_sstep.

Its not even close to identical, its very careful to deal with user-mode
already doing single step.

But I'll leave this to the x86 people who actually know the intricacies
of the single step cruft, I was just wondering why you weren't using (or
extending) the existing code.