Re: [tip:x86/mm] x86, mm, kprobes: fault.c, simplify notify_page_fault()

Re: [tip:x86/mm] x86, mm, kprobes: fault.c, simplify notify_page_fault()

[Masami Hiramatsu]

Ingo Molnar wrote:
> Author:     Ingo Molnar <mingo@elte.hu>
> AuthorDate: Fri, 20 Feb 2009 22:42:57 +0100
> Commit:     Ingo Molnar <mingo@elte.hu>
> CommitDate: Sat, 21 Feb 2009 00:09:42 +0100
> 
> x86, mm, kprobes: fault.c, simplify notify_page_fault()
> 
> Impact: cleanup
> 
> Remove an #ifdef from notify_page_fault(). The function still
> compiles to nothing in the !CONFIG_KPROBES case.
> 
> Introduce kprobes_built_in() and kprobe_fault_handler() helpers
> to allow this - they returns 0 if !CONFIG_KPROBES.
> 
> No code changed:
> 
>    text	   data	    bss	    dec	    hex	filename
>    4618	     32	     24	   4674	   1242	fault.o.before
>    4618	     32	     24	   4674	   1242	fault.o.after

It seems good for me. Thank you for cleanup!

Acked-by: Masami Hiramatsu <mhiramat@redhat.com>

> 
> Cc: Masami Hiramatsu <mhiramat@redhat.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> 
> 
> ---
>  arch/x86/mm/fault.c     |    6 +-----
>  include/linux/kprobes.h |   22 +++++++++++++++++++---
>  2 files changed, 20 insertions(+), 8 deletions(-)
> 
> diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
> index fe99af4..379beae 100644
> --- a/arch/x86/mm/fault.c
> +++ b/arch/x86/mm/fault.c
> @@ -68,11 +68,10 @@ static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr)
>  
>  static inline int notify_page_fault(struct pt_regs *regs)
>  {
> -#ifdef CONFIG_KPROBES
>  	int ret = 0;
>  
>  	/* kprobe_running() needs smp_processor_id() */
> -	if (!user_mode_vm(regs)) {
> +	if (kprobes_built_in() && !user_mode_vm(regs)) {
>  		preempt_disable();
>  		if (kprobe_running() && kprobe_fault_handler(regs, 14))
>  			ret = 1;
> @@ -80,9 +79,6 @@ static inline int notify_page_fault(struct pt_regs *regs)
>  	}
>  
>  	return ret;
> -#else
> -	return 0;
> -#endif
>  }
>  
>  /*
> diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
> index 32851ee..2ec6cc1 100644
> --- a/include/linux/kprobes.h
> +++ b/include/linux/kprobes.h
> @@ -182,6 +182,14 @@ struct kprobe_blackpoint {
>  DECLARE_PER_CPU(struct kprobe *, current_kprobe);
>  DECLARE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk);
>  
> +/*
> + * For #ifdef avoidance:
> + */
> +static inline int kprobes_built_in(void)
> +{
> +	return 1;
> +}
> +
>  #ifdef CONFIG_KRETPROBES
>  extern void arch_prepare_kretprobe(struct kretprobe_instance *ri,
>  				   struct pt_regs *regs);
> @@ -271,8 +279,16 @@ void unregister_kretprobes(struct kretprobe **rps, int num);
>  void kprobe_flush_task(struct task_struct *tk);
>  void recycle_rp_inst(struct kretprobe_instance *ri, struct hlist_head *head);
>  
> -#else /* CONFIG_KPROBES */
> +#else /* !CONFIG_KPROBES: */
>  
> +static inline int kprobes_built_in(void)
> +{
> +	return 0;
> +}
> +static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr)
> +{
> +	return 0;
> +}
>  static inline struct kprobe *get_kprobe(void *addr)
>  {
>  	return NULL;
> @@ -329,5 +345,5 @@ static inline void unregister_kretprobes(struct kretprobe **rps, int num)
>  static inline void kprobe_flush_task(struct task_struct *tk)
>  {
>  }
> -#endif				/* CONFIG_KPROBES */
> -#endif				/* _LINUX_KPROBES_H */
> +#endif /* CONFIG_KPROBES */
> +#endif /* _LINUX_KPROBES_H */

-- 
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: mhiramat@redhat.com

Re: [tip:x86/mm] x86, mm, kprobes: fault.c, simplify notify_page_fault()

[Ingo Molnar]

* Masami Hiramatsu <mhiramat@redhat.com> wrote:

> Ingo Molnar wrote:
> > Author:     Ingo Molnar <mingo@elte.hu>
> > AuthorDate: Fri, 20 Feb 2009 22:42:57 +0100
> > Commit:     Ingo Molnar <mingo@elte.hu>
> > CommitDate: Sat, 21 Feb 2009 00:09:42 +0100
> > 
> > x86, mm, kprobes: fault.c, simplify notify_page_fault()
> > 
> > Impact: cleanup
> > 
> > Remove an #ifdef from notify_page_fault(). The function still
> > compiles to nothing in the !CONFIG_KPROBES case.
> > 
> > Introduce kprobes_built_in() and kprobe_fault_handler() helpers
> > to allow this - they returns 0 if !CONFIG_KPROBES.
> > 
> > No code changed:
> > 
> >    text	   data	    bss	    dec	    hex	filename
> >    4618	     32	     24	   4674	   1242	fault.o.before
> >    4618	     32	     24	   4674	   1242	fault.o.after
> 
> It seems good for me. Thank you for cleanup!
> 
> Acked-by: Masami Hiramatsu <mhiramat@redhat.com>

another very small thing, while we are discussing kprobes:

I always found that the __kprobes annotation is very confusingly 
euphemistic: what those annotations really mean is not 
'kprobes', but 'no kprobes'.

So how about renaming __kprobes to __nokprobes, similar to how 
we have the notrace attribute?

We have about 350 __kprobes annotations in the kernel, so 
renaming it now would not be practical - but any objections 
against me sending Linus a rename patch somewhere late in the 
next merge window that just does this rename?

[ likewise, i'll rename notrace to __notrace to make it visually 
  less intrusive to the return value type. There's a lot less 
  such annotations in the kernel. ]

	Ingo

Re: [tip:x86/mm] x86, mm, kprobes: fault.c, simplify notify_page_fault()

[Masami Hiramatsu]

Hi Ingo,

Ingo Molnar wrote:
> another very small thing, while we are discussing kprobes:
> 
> I always found that the __kprobes annotation is very confusingly 
> euphemistic: what those annotations really mean is not 
> 'kprobes', but 'no kprobes'.

As far as I know, __kprobes originally means 'this function will be
called from kprobes'. However, now many functions are tagged as
__kprobes, it might be confusingly.

> So how about renaming __kprobes to __nokprobes, similar to how 
> we have the notrace attribute?

Would you mean that will include changing section name of
'.text.kprobes'? That's what I mind.

> We have about 350 __kprobes annotations in the kernel, so 
> renaming it now would not be practical - but any objections 
> against me sending Linus a rename patch somewhere late in the 
> next merge window that just does this rename?

Just renaming __kprobes to __nokprobes seems good for me.

Thank you!


> [ likewise, i'll rename notrace to __notrace to make it visually 
>   less intrusive to the return value type. There's a lot less 
>   such annotations in the kernel. ]
> 
> 	Ingo
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: mhiramat@redhat.com

Re: [tip:x86/mm] x86, mm, kprobes: fault.c, simplify notify_page_fault()

[Ananth N Mavinakayanahalli]

On Sun, Feb 22, 2009 at 10:31:09AM +0100, Ingo Molnar wrote:
> 
> * Masami Hiramatsu <mhiramat@redhat.com> wrote:
> 
> > Ingo Molnar wrote:
> > > Author:     Ingo Molnar <mingo@elte.hu>
> > > AuthorDate: Fri, 20 Feb 2009 22:42:57 +0100
> > > Commit:     Ingo Molnar <mingo@elte.hu>
> > > CommitDate: Sat, 21 Feb 2009 00:09:42 +0100
> > > 
> > > x86, mm, kprobes: fault.c, simplify notify_page_fault()
> > > 
> > > Impact: cleanup
> > > 
> > > Remove an #ifdef from notify_page_fault(). The function still
> > > compiles to nothing in the !CONFIG_KPROBES case.
> > > 
> > > Introduce kprobes_built_in() and kprobe_fault_handler() helpers
> > > to allow this - they returns 0 if !CONFIG_KPROBES.
> > > 
> > > No code changed:
> > > 
> > >    text	   data	    bss	    dec	    hex	filename
> > >    4618	     32	     24	   4674	   1242	fault.o.before
> > >    4618	     32	     24	   4674	   1242	fault.o.after
> > 
> > It seems good for me. Thank you for cleanup!
> > 
> > Acked-by: Masami Hiramatsu <mhiramat@redhat.com>
> 
> another very small thing, while we are discussing kprobes:
> 
> I always found that the __kprobes annotation is very confusingly 
> euphemistic: what those annotations really mean is not 
> 'kprobes', but 'no kprobes'.

Right!

> So how about renaming __kprobes to __nokprobes, similar to how 
> we have the notrace attribute?
> 
> We have about 350 __kprobes annotations in the kernel, so 
> renaming it now would not be practical - but any objections 
> against me sending Linus a rename patch somewhere late in the 
> next merge window that just does this rename?

No issues with that.

Ananth