| Message ID | 7eb4d1a03031b4d873ce08e8912900d7936a3ca8.1395832156.git.viresh.kumar@linaro.org |
|---|---|
| State | New |
| Headers | show |
On 03/26/2014 04:51 PM, Viresh Kumar wrote: > In hrtimers_init() we need to call init_hrtimers_cpu() for boot CPU. For this, > currently we are emulating a call to hotplug notifier. Probably this was done > initially to get rid of code redundancy. But this sequence always called a > single routine, i.e. init_hrtimers_cpu(), and so calling that routine directly > would be better. This would get rid of emulating a notifier call, few typecasts > and the extra steps we are doing in notifier callback. > > So, this patch calls init_hrtimers_cpu() directly from hrtimers_init(). > I don't think this is such a good idea. Open-coding a part of that callback in the init routine can lead to loop-holes down the road: what if someone changes or adds something to the CPU_UP_PREPARE switch-case, and forgets to do the same in the init-routine? It is more comforting to know that there is just one single place where CPU hotplug operations are handled (hrtimer_cpu_notify). That, in turn is good for reliability because it makes it easier to write bug-free code. Regards, Srivatsa S. Bhat > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> > --- > kernel/hrtimer.c | 3 +-- > 1 file changed, 1 insertion(+), 2 deletions(-) > > diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c > index f14d861..39dbdbd 100644 > --- a/kernel/hrtimer.c > +++ b/kernel/hrtimer.c > @@ -1756,8 +1756,7 @@ static struct notifier_block hrtimers_nb = { > > void __init hrtimers_init(void) > { > - hrtimer_cpu_notify(&hrtimers_nb, (unsigned long)CPU_UP_PREPARE, > - (void *)(long)smp_processor_id()); > + init_hrtimers_cpu(smp_processor_id()); > register_cpu_notifier(&hrtimers_nb); > #ifdef CONFIG_HIGH_RES_TIMERS > open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq); > -- 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/
On 26 March 2014 18:10, Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com> wrote: > I don't think this is such a good idea. Open-coding a part of that callback > in the init routine can lead to loop-holes down the road: We think that we are open-coding part of that callback here because it is implemented that way on the first design. Rather, we should have a common routine which should do all the work required when a CPU comes up. And any modification should be done to that code. > what if someone > changes or adds something to the CPU_UP_PREPARE switch-case, and forgets to > do the same in the init-routine? This is not a driver which only 2-3 people use. This part is so well reviewed by so many highly smart people that this should never happen. And if it happens than its nothing but a review mistake. > It is more comforting to know that there is just one single place where CPU > hotplug operations are handled (hrtimer_cpu_notify). That, in turn is good > for reliability because it makes it easier to write bug-free code. And for me that single place is: init_hrtimers_cpu() :) -- 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/
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index f14d861..39dbdbd 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -1756,8 +1756,7 @@ static struct notifier_block hrtimers_nb = { void __init hrtimers_init(void) { - hrtimer_cpu_notify(&hrtimers_nb, (unsigned long)CPU_UP_PREPARE, - (void *)(long)smp_processor_id()); + init_hrtimers_cpu(smp_processor_id()); register_cpu_notifier(&hrtimers_nb); #ifdef CONFIG_HIGH_RES_TIMERS open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq);
In hrtimers_init() we need to call init_hrtimers_cpu() for boot CPU. For this, currently we are emulating a call to hotplug notifier. Probably this was done initially to get rid of code redundancy. But this sequence always called a single routine, i.e. init_hrtimers_cpu(), and so calling that routine directly would be better. This would get rid of emulating a notifier call, few typecasts and the extra steps we are doing in notifier callback. So, this patch calls init_hrtimers_cpu() directly from hrtimers_init(). Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> --- kernel/hrtimer.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)