On Mon, 21 Sep 2015, Felipe Balbi wrote: > On Mon, Sep 21, 2015 at 10:50:10AM -0400, Alan Stern wrote: > > On Mon, 21 Sep 2015, Felipe Balbi wrote: > > > > > On Mon, Sep 21, 2015 at 10:31:15AM -0400, Alan Stern wrote: > > > > On Mon, 21 Sep 2015, Roger Quadros wrote: > > > > > > > > > Using spin_lock() in hard irq handler is pointless > > > > > and causes a BUG() in RT (real-time) configuration > > > > > so get rid of it. > > > > > > > > Wait a minute. Who says spin_lock is pointless in an IRQ handler? > > > > > > in the top half IRQs are already disabled, how can this race ? > > > > It can race with code running on a different CPU. > > fair point.
In fact, isn't this the main purpose of spinlocks? There's no point using a spinlock to protect against races occurring on a single CPU, because (in non-RT situations) the kernel can't schedule or preempt while a spinlock is held, so no race can occur. The whole idea of spinlocks is to protect against cross-CPU races. Now, maybe the spinlock usage that Roger is removing really is unnecessary in this top-half handler. I don't know; I haven't looked at the code. But in general, spinlocks are highly necessary in IRQ handlers. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
