On Mon, Aug 03, 2015 at 12:59:50AM -0600, Sagar Dharia wrote:

+	if (cur_clk_state == SLIM_CLK_ENTERING_PAUSE) {
+		if (txn->mc != SLIM_MSG_MC_BEGIN_RECONFIGURATION &&
+		    txn->mc != SLIM_MSG_MC_RECONFIGURE_NOW &&
+		    txn->mc != SLIM_MSG_MC_NEXT_PAUSE_CLOCK)
+			return -EBUSY;
+	} else {
+		int ret = pm_runtime_get_sync(dev->dev);
+
+		if (ret < 0) {

This is taking a runtime PM reference whenever we're about to start a
transfer - this is the sort of thing I was talking about expecting the
framework to do rather than individual drivers.

+			pm_runtime_set_suspended(dev->dev);
+			dev_err(dev->dev, "runtime-pm vote failed:%d\n", ret);
+			return ret;

If the get failed there should be no need to suspend the device, the
get() should leave us in a sensible state...

+static int msm_slim_suspend(struct device *dev)
+{
+	int ret = 0;
+
+	if (!pm_runtime_enabled(dev) ||
+		(!pm_runtime_suspended(dev))) {
+		dev_dbg(dev, "system suspend");
+		ret = msm_slim_runtime_suspend(dev);
+	}
+	if (ret == -EISCONN) {
+		/**
+		* If the clock pause failed due to active channels, there is
+		* a possibility that some audio stream is active during suspend
+		* We dont want to return suspend failure in that case so that
+		* display and relevant components can still go to suspend.
+		* If there is some other error, then it should prevent
+		* system level suspend
+		*/

Best mention that this is for the modem use case for Android's suspend
model since it's really surprising to see.