Re: [PATCH 08/15] powerpc/watchpoint: Disable all available watchpoints when !dawr_force_enable

[Ravi Bangoria]

On 3/17/20 4:05 PM, Christophe Leroy wrote:
> Le 09/03/2020 Ã 09:57, Ravi Bangoria a ÃcritÂ:
> > Instead of disabling only first watchpoint, disable all available
> > watchpoints while clearing dawr_force_enable.
> >
> > Signed-off-by: Ravi Bangoria <ravi.bangoria@xxxxxxxxxxxxx>
> > ---
> > Â arch/powerpc/kernel/dawr.c | 10 +++++++---
> > Â 1 file changed, 7 insertions(+), 3 deletions(-)
> >
> > diff --git a/arch/powerpc/kernel/dawr.c b/arch/powerpc/kernel/dawr.c
> > index 311e51ee09f4..5c882f07ac7d 100644
> > --- a/arch/powerpc/kernel/dawr.c
> > +++ b/arch/powerpc/kernel/dawr.c
> > @@ -50,9 +50,13 @@ int set_dawr(struct arch_hw_breakpoint *brk, int nr)
> > ÂÂÂÂÂ return 0;
> > Â }
> > -static void set_dawr_cb(void *info)
> > +static void disable_dawrs(void *info)
>
> Can you explain a bit more what you do exactly ? Why do you change the name of the function and why the parameter becomes NULL ? And why it doens't take into account the parameter anymore ?

Before:
  static void set_dawr_cb(void *info)
  {
          set_dawr(info);
  }
  
  static ssize_t dawr_write_file_bool(...)
  {
  	...
          /* If we are clearing, make sure all CPUs have the DAWR cleared */
          if (!dawr_force_enable)
                  smp_call_function(set_dawr_cb, &null_brk, 0);
  }

After:
  static void disable_dawrs(void *info)
  {
          struct arch_hw_breakpoint null_brk = {0};
          int i;
  
          for (i = 0; i < nr_wp_slots(); i++)
                  set_dawr(&null_brk, i);
  }
  
  static ssize_t dawr_write_file_bool(...)
  {
  	...
          /* If we are clearing, make sure all CPUs have the DAWR cleared */
          if (!dawr_force_enable)
                  smp_call_function(disable_dawrs, NULL, 0);
  }

We use callback function only for disabling watchpoint. Thus I renamed
it to disable_dawrs(). And we are passing null_brk as parameter which
is not really required while disabling watchpoint. So removed it.

Thanks,
Ravi