[Patch] ARM/MXS/i.MX28: distinguish POR and wake-up event correctly

[Juergen Beisert]

When the built-in RTC in the i.MX28 is programmed to wake-up the SoC, the RTC
reports two events: RST and WAKE. RST is okay in this case, because the PMIC
was really powered down. But the real event is the WAKE from the RTC and
should have precedence over the RST event. So, report the WAKE event for this
special case.

Signed-off-by: Juergen Beisert <jbe at pengutronix.de>

diff --git a/drivers/watchdog/im28wd.c b/drivers/watchdog/im28wd.c
index ca32a72..bc19369 100644
--- a/drivers/watchdog/im28wd.c
+++ b/drivers/watchdog/im28wd.c
@@ -82,6 +82,17 @@ static void __maybe_unused imx28_detect_reset_source(const struct imx28_wd *p)
 	if (reg & MXS_RTC_PERSISTENT0_EXT_RST) {
 		writel(MXS_RTC_PERSISTENT0_EXT_RST,
 			p->regs + MXS_RTC_PERSISTENT0 + MXS_RTC_CLR_ADDR);
+		/*
+		 * if the RTC has woken up the SoC, additionally the ALARM_WAKE
+		 * bit is set. This bit should have precedence, because it
+		 * reports the real event, why we are here.
+		 */
+		if (reg & MXS_RTC_PERSISTENT0_ALARM_WAKE) {
+			writel(MXS_RTC_PERSISTENT0_ALARM_WAKE,
+				p->regs + MXS_RTC_PERSISTENT0 + MXS_RTC_CLR_ADDR);
+			set_reset_source(RESET_WKE);
+			return;
+		}
 		set_reset_source(RESET_POR);
 		return;
 	}


-- 
Pengutronix e.K.                              | Juergen Beisert             |
Linux Solutions for Science and Industry      | http://www.pengutronix.de/  |