Tick devices

Proxy tick device

The proxy tick device is a synthetic clock event device for handing over the control of the hardware tick device to a high-precision, out-of-band timing logic, which cannot be delayed by the in-band kernel code. With this proxy in place, any out-of-band code can gain control over the timer hardware for carrying out its own timing duties. In the same move, it is required to honor the timing requests received from the in-band timer layer (i.e. hrtimers) since the latter won’t be able to program timer events directly into the hardware while the proxy is active.

In other words, the proxy tick device shares the functionality of the actual device between the in-band and out-of-band contexts, with only the latter actually programming the hardware.

Adapting clock chip devices for proxying

The proxy tick device borrows a real clock chip device from the in-band kernel, controlling it under the hood while substituting for the current tick device. Clock chips which may be controlled by the proxy tick device need their drivers to be specifically adapted for such use, as follows:

  • clockevents_handle_event() must be substituted to any open-coded invocation of the event handler in the interrupt handler.

  • struct clock_event_device::irq must be properly set to the actual IRQ number signaling an event from this device.

  • struct clock_event_device::features must include CLOCK_EVT_FEAT_PIPELINE.

  • __IRQF_TIMER must be set for the action handler of the timer device interrupt.

Adapting the ARM architected timer driver to out-of-band timing

--- a/drivers/clocksource/arm_arch_timer.c
+++ b/drivers/clocksource/arm_arch_timer.c
@@ -585,7 +585,7 @@ static __always_inline irqreturn_t timer_handler(const int access,
 	if (ctrl & ARCH_TIMER_CTRL_IT_STAT) {
 		arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl, evt);
-		evt->event_handler(evt);
+		clockevents_handle_event(evt);
 		return IRQ_HANDLED;
@@ -704,7 +704,7 @@ static int arch_timer_set_next_event_phys_mem(unsigned long evt,
 static void __arch_timer_setup(unsigned type,
 			       struct clock_event_device *clk)
-	clk->features = CLOCK_EVT_FEAT_ONESHOT;
 	if (type == ARCH_TIMER_TYPE_CP15) {
 		if (arch_timer_c3stop)

Only oneshot-capable clock event devices can be shared via the proxy tick device.

Adapting the ARM Global timer driver to out-of-band timing

--- a/drivers/clocksource/arm_global_timer.c
+++ b/drivers/clocksource/arm_global_timer.c
@@ -156,11 +156,11 @@ static irqreturn_t gt_clockevent_interrupt(int irq, void *dev_id)
 	 *	the Global Timer flag _after_ having incremented
 	 *	the Comparator register	value to a higher value.
-	if (clockevent_state_oneshot(evt))
+	if (clockevent_is_oob(evt) || clockevent_state_oneshot(evt))
 		gt_compare_set(ULONG_MAX, 0);
 	writel_relaxed(GT_INT_STATUS_EVENT_FLAG, gt_base + GT_INT_STATUS);
-	evt->event_handler(evt);
+	clockevents_handle_event(evt);
 	return IRQ_HANDLED;
@@ -171,7 +171,7 @@ static int gt_starting_cpu(unsigned int cpu)
 	clk->name = "arm_global_timer";
 	clk->set_state_shutdown = gt_clockevent_shutdown;
 	clk->set_state_periodic = gt_clockevent_set_periodic;
 	clk->set_state_oneshot = gt_clockevent_shutdown;
@@ -195,11 +195,6 @@ static int gt_dying_cpu(unsigned int cpu)
 	return 0;
@@ -302,8 +307,8 @@ static int __init global_timer_of_register(struct device_node *np)
 		goto out_clk;
-	err = request_percpu_irq(gt_ppi, gt_clockevent_interrupt,
-				 "gt", gt_evt);
+	err = __request_percpu_irq(gt_ppi, gt_clockevent_interrupt,
+				   IRQF_TIMER, "gt", gt_evt);
 	if (err) {
 		pr_warn("global-timer: can't register interrupt %d (%d)\n",
 			gt_ppi, err);

This is another example of adapting an existing clock chip driver for serving out-of-band timing requests, with a subtle change in the way we should test for the current state of the clock device in the interrupt handler:

  • A real/original device (such as the ARM global timer in this example) is switched to reserved mode when the proxy tick driver hands it over to the autonomous core, which is similar to the detached mode in effect. Therefore, the ARM global timer state is always reserved from the standpoint of the kernel when proxied, never oneshot. For this reason, clockevent_state_oneshot() would always lead to false in this case.

  • However, since a real device controlled by the proxy for receiving out-of-band events has to be driven in one-shot mode under the hood, testing for clockevent_state_oob() in addition to clockevent_state_oneshot() guarantees that we do take the branch, setting the comparator register to ULONG_MAX when proxied too.

Failing to fix up the way we test for the clock device state would certainly lead to an interrupt storm with any ARM global timer suffering erratum 740657, quickly locking up the board.

Theory of operations

Calling tick_install_proxy() registers an instance of the proxy tick device on each CPU mentioned in the cpumask it receives. This routine is also passed the address of a routine which should setup the given struct clock_proxy_device descriptor for the current CPU. This routine is called indirectly by tick_install_proxy(), for each CPU marked in cpumask.

Initializing the proxy descriptor

tick_install_proxy() prepares the new proxy device for the current CPU, pre-initializing it with settings compatible with the real device’s it interposes on, then calls the setup_proxy() routine. The descriptor is defined as follows:

struct clock_proxy_device {
	struct clock_event_device proxy_device;
	struct clock_event_device *real_device;
	void (*handle_oob_event)(struct clock_event_device *dev);
	/* Internal data - don't depend on this. */
	void (*__setup_handler)(struct clock_proxy_device *dev);
	void (*__original_handler)(struct clock_event_device *dev);

The user setup call must at least set the .handle_oob_event() handler: this is the address of the routine which should be called each time an out-of-band tick is received from the underlying timer hardware the proxy controls. This is the only information required from the setup handler, the rest may be inherited from the pre-set data if the user does not need any particular setting.

If the user code proxying the tick device prefers dealing with nanoseconds instead of clock ticks directly, CLOCK_EVT_FEAT_KTIME should be added to proxy_device.features, along with a valid proxy_device.set_next_ktime() handler and proper min/max delta values.

A setup_proxy() routine preparing a proxy device

static DEFINE_PER_CPU(struct clock_proxy_device, tick_device);

static void oob_event_handler(struct clock_event_device *dev)
	 * We are running on the out-of-band stage, in NMI-like mode.
	 * Schedule a tick on the proxy device to satisfy the
	 * corresponding timing request asap.

static void setup_proxy(struct clock_proxy_device *dev)
	 * Create a proxy which acts as a transparent device, simply
	 * relaying the timing requests to the in-band code, without
	 * any additional out-of-band processing.
	dev->handle_oob_event = oob_event_handler;

The proxy_device.set_next_event() or proxy_device.set_next_ktime() members can be set with the address of a handler which receives timer requests from the in-band kernel. This handler is normally implemented by the autonomous core which takes control over the timer hardware via the proxy device. Whenever that core determines that a tick is due for an outstanding request received from such handler, it should call tick_notify_proxy() to signal the event to the main kernel.

Once the user-supplied setup_proxy() routine returns, the following events happen in sequence:

  1. the proxy device is registered on the clock event framework.

  2. the real device is detached from the clockevent framework, switched to the CLOCK_EVT_STATE_RESERVED state, which makes it non-eligible for any regular operation from the framework. However, its hardware is left in a functional state. In the same move, the proxy device is picked as the new tick device by the framework. From that point, requests to the proxy device may be indirectly channeled to the real device via the proxy when operations on the hardware should be carried out.

  3. the proxy device now controls the real device under the hood to carry out timing requests from the in-band kernel. When the hrtimer layer from the in-band kernel wants to program the next shot of the current tick device, it invokes the set_next_event() handler of the proxy device, which was defined by the user (which defaults to the real device’s set_next_event() handler). If the autonomous core implements its own timer management, this handler should be scheduling in-band ticks at the requested time based on such scheme.

  4. the timer interrupt triggered by the real device is switched to out-of-band handling. As a result, handle_oob_event() receives tick events sent by the real device hardware directly from the out-of-band stage of the interrupt pipeline. This ensures high-precision timing, which the in-band stage cannot delay via interrupt masking. From that point, the out-of-band code can carry out its own timing duties, in addition to honoring the in-band kernel requests for timing.

Step 3. involves emulating ticks scheduled by the in-band kernel by a software logic controlled by some out-of-band timer management, paced by the real ticks received as described in step 4. When this logic decides than the next in-band tick is due, it should call tick_notify_proxy() to trigger the corresponding event for the in-band kernel, which would honor the pending (hr)timer request.

Under the hood

	     [in-band timing request]
	         <tick event>
	         inband_event_handler() [out-of-band stage]
	                     ...(in-band tick emulation)...
	         proxy_irq_handler(proxy_dev) [in-band stage]

Last modified: Sun, 21 May 2023 18:02:07 +0200