Re: [PATCH -v5 05/17] x86/ftrace: Use text_poke()

[Peter Zijlstra]

On Tue, Nov 12, 2019 at 11:24:13PM +0100, Peter Zijlstra wrote:
> On Tue, Nov 12, 2019 at 01:25:36PM -0500, Steven Rostedt wrote:
> > On Mon, 11 Nov 2019 14:12:57 +0100
> > Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> 
> > >  int ftrace_arch_code_modify_post_process(void)
> > >      __releases(&text_mutex)
> > >  {
> > > -	set_all_modules_text_ro();
> > > -	set_kernel_text_ro();
> > > +	text_poke_finish();
> > 
> > Why is the text_poke_finish() needed here? Can we add a comment about
> > why?
> 
> I think this is because of the text_poke_queue() in
> ftrace_modify_code_direct(). I seem to have forgotten the code-flow
> between the core and arch parts of ftrace again.
> 
> But sure, I can try and dig that out again and write a comment.

These are the two callgraphs:

  ftrace_module_enable()
    ftrace_arch_code_modify_prepare()
      do_for_each_ftrace_rec()
	__ftrace_replace_code()
	  ftrace_make_{call,nop}()
	    ftrace_modify_code_direct()
	      text_poke_queue()

    ftrace_arch_code_modify_post_process()
      text_poke_finish();


  ftrace_run_update_code()
    ftrace_arch_code_modify_prepare()
    arch_ftrace_update_code()
      ftrace_modify_all_code()
	ftrace_replace_code()
	  for_ftrace_rec_iter()
	    text_poke_queue()
	  text_poke_finish()
    ftrace_arch_code_modify_post_process()
      text_poke_finish()


So while it is superfluous for ftrace_run_update_code() it is required
for ftrace_module_enable(), and, as I said, pairs with the
text_poke_queue() in ftrace_modify_code_direct().

I'll stick in a comment.