On 06/10/2021 18.24, Philipp Zabel wrote:

quoted

+static int apple_pmgr_reset_reset(struct reset_controller_dev *rcdev, unsigned long id)
+{
+	int ret;
+
+	ret = apple_pmgr_reset_assert(rcdev, id);
+	if (ret)
+		return ret;
+
+	usleep_range(APPLE_PMGR_RESET_TIME, 2 * APPLE_PMGR_RESET_TIME);

Is this delay known to be long enough for all consumers using the
reset_control_reset() functionality? Are there any users at all?

I tested this for UART and ANS outside of Linux, and found that even 
back to back register writes worked, so the 1us thing is already 
conservative. I have no idea if we'll run into some weirdo block that 
needs more time, though. If we do then this will have to be bumped or 
turned into a property.

Is it ok for a genpd transition to happen during this sleep?

I expect consumers to call reset while the device is active; it won't 
even work without that, as the reset is synchronous and just doesn't 
take effect while clock gated (at least for UART). See the dev_err()s 
that fire when that happens.

-- 
Hector Martin (marcan@marcan.st)
Public Key: https://mrcn.st/pub