On Tue, Mar 26, 2019 at 10:13:06PM +0100, Thomas Gleixner wrote: > On Wed, 27 Feb 2019, Ricardo Neri wrote: > > +#ifdef CONFIG_X86_HARDLOCKUP_DETECTOR_HPET > > +struct hpet_hld_data *hpet_hardlockup_detector_assign_timer(void) > > +{ > > + struct hpet_hld_data *hdata; > > + unsigned int cfg; > > + > > + cfg = hpet_readl(HPET_Tn_CFG(HPET_WD_TIMER_NR)); > > + > > + if (!(cfg & HPET_TN_FSB_CAP)) > > + return NULL; > > + > > + hdata = kzalloc(sizeof(*hdata), GFP_KERNEL); > > + if (!hdata) > > + return NULL; > > + > > + hdata->flags = HPET_DEV_FSB_CAP; > > Pointless.
Agreed. Only if the timer is FSB-capable is hdata is initialized. > > > + > > + if (cfg & HPET_TN_PERIODIC_CAP) > > + hdata->flags |= HPET_DEV_PERI_CAP; > > This can be expressed by a simple: > > hdata->has_periodic = 1; > > And no flag shuffling required at all. Sure. I'll implement this change. Thanks and BR, Ricardo