Re: [patch V2 11/24] ARM/hw_breakpoint: Use cpuhp_setup_state_cpuslocked()

[Thomas Gleixner]

On Wed, 19 Apr 2017, Mark Rutland wrote:

> Hi,
> 
> On Tue, Apr 18, 2017 at 07:04:53PM +0200, Thomas Gleixner wrote:
> > From: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> > 
> > arch_hw_breakpoint_init() holds get_online_cpus() while registerring the
> > hotplug callbacks.
> > 
> > cpuhp_setup_state() invokes get_online_cpus() as well. This is correct, but
> > prevents the conversion of the hotplug locking to a percpu rwsem.
> > 
> > Use cpuhp_setup_state_cpuslocked() to avoid the nested call.
> > 
> > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> > Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> > Cc: Will Deacon <will.deacon@xxxxxxx>
> > Cc: Mark Rutland <mark.rutland@xxxxxxx>
> > Cc: Russell King <linux@xxxxxxxxxxxxxxx>
> > Cc: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> > 
> > ---
> >  arch/arm/kernel/hw_breakpoint.c |    5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> > 
> > --- a/arch/arm/kernel/hw_breakpoint.c
> > +++ b/arch/arm/kernel/hw_breakpoint.c
> > @@ -1098,8 +1098,9 @@ static int __init arch_hw_breakpoint_ini
> >  	 * assume that a halting debugger will leave the world in a nice state
> >  	 * for us.
> >  	 */
> > -	ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "arm/hw_breakpoint:online",
> > -				dbg_reset_online, NULL);
> > +	ret = cpuhp_setup_state_cpuslocked(CPUHP_AP_ONLINE_DYN,
> > +					   "arm/hw_breakpoint:online",
> > +					   dbg_reset_online, NULL);
> 
> Given the callsite, this particular change looks ok to me. So FWIW:
> 
> Acked-by: Mark Rutland <mark.rutland@xxxxxxx>
> 
> However, as a more general note, the changes make the API feel odd. per
> their current names, {get,put}_online_cpus() sound/feel like refcounting
> ops, which should be able to nest.
> 
> Is there any chance these could get a better name, e.g.
> {lock,unlock}_online_cpus(), so as to align with _cpuslocked?

Yes, that's a follow up cleanup patch treewide once this hit Linus tree.

Thanks,

	tglx