Hey Mike,
On Mon, Feb 27, 2012 at 6:06 PM, Turquette, Mike mturquette@ti.com wrote:
On Sun, Feb 26, 2012 at 8:47 PM, Robert Lee rob.lee@linaro.org wrote:
+/**
- cpuidle_enter_wrap - performing timekeeping and irq around enter function
- @dev: pointer to a valid cpuidle_device object
- @drv: pointer to a valid cpuidle_driver object
- @index: index of the target cpuidle state.
- */
+static inline int cpuidle_wrap_enter(struct cpuidle_device *dev,
- struct cpuidle_driver *drv, int index,
- int (*enter)(struct cpuidle_device *dev,
- struct cpuidle_driver *drv, int index))
+{
- ktime_t time_start, time_end;
- s64 diff;
- time_start = ktime_get();
- index = enter(dev, drv, index);
- time_end = ktime_get();
- local_irq_enable();
- diff = ktime_to_us(ktime_sub(time_end, time_start));
- if (diff > INT_MAX)
- diff = INT_MAX;
- dev->last_residency = (int) diff;
- return index;
+}
Any reason that this code is in the header? Why not in cpuidle.c?
Not a strong reason. I thought making it an inline would introduce slightly less new execution when adding this code (realizing that there are function calls immediately after, so the only benefit is the reduce popping and pushing). But it does require an extra copy of this code for any platform driver that does not enable en_core_tk_irqen and instead makes calls to it directly (like omap3). For this case, I don't think the inline implementation should add extra code from what exists today as it should simply replace the existing platform time keeping calls to a standard one defined by the core cpuidle.
I don't have a strong preference with using the inline so if you or others can give your opinion on which method to use and why, I'd be glad to read it.
Regards, Mike