linux-kernel - [PATCH 5.9 213/252] iio/adc: ingenic: Fix AUX/VBAT readings when touchscreen is used

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20201123121845.849146781@linuxfoundation.org>
Date:   Mon, 23 Nov 2020 13:22:43 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, Paul Cercueil <paul@...pouillou.net>,
        Artur Rojek <contact@...ur-rojek.eu>,
        Jonathan Cameron <Jonathan.Cameron@...wei.com>
Subject: [PATCH 5.9 213/252] iio/adc: ingenic: Fix AUX/VBAT readings when touchscreen is used

From: Paul Cercueil <paul@...pouillou.net>

commit 6d6aa2907d59ddd3c0ebb2b93e1ddc84e474485b upstream.

When the command feature of the ADC is used, it is possible to program
the ADC, and specify at each step what input should be processed, and in
comparison to what reference.

This broke the AUX and battery readings when the touchscreen was
enabled, most likely because the CMD feature would change the VREF all
the time.

Now, when AUX or battery are read, we temporarily disable the CMD
feature, which means that we won't get touchscreen readings in that time
frame. But it now gives correct values for AUX / battery, and the
touchscreen isn't disabled for long enough to be an actual issue.

Fixes: b96952f498db ("IIO: Ingenic JZ47xx: Add touchscreen mode.")
Signed-off-by: Paul Cercueil <paul@...pouillou.net>
Acked-by: Artur Rojek <contact@...ur-rojek.eu>
Cc: <stable@...r.kernel.org>
Link: https://lore.kernel.org/r/20201103201238.161083-1-paul@crapouillou.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@...wei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>

---
 drivers/iio/adc/ingenic-adc.c |   32 ++++++++++++++++++++++++++------
 1 file changed, 26 insertions(+), 6 deletions(-)

--- a/drivers/iio/adc/ingenic-adc.c
+++ b/drivers/iio/adc/ingenic-adc.c
@@ -177,13 +177,12 @@ static void ingenic_adc_set_config(struc
 	mutex_unlock(&adc->lock);
 }
 
-static void ingenic_adc_enable(struct ingenic_adc *adc,
-			       int engine,
-			       bool enabled)
+static void ingenic_adc_enable_unlocked(struct ingenic_adc *adc,
+					int engine,
+					bool enabled)
 {
 	u8 val;
 
-	mutex_lock(&adc->lock);
 	val = readb(adc->base + JZ_ADC_REG_ENABLE);
 
 	if (enabled)
@@ -192,20 +191,41 @@ static void ingenic_adc_enable(struct in
 		val &= ~BIT(engine);
 
 	writeb(val, adc->base + JZ_ADC_REG_ENABLE);
+}
+
+static void ingenic_adc_enable(struct ingenic_adc *adc,
+			       int engine,
+			       bool enabled)
+{
+	mutex_lock(&adc->lock);
+	ingenic_adc_enable_unlocked(adc, engine, enabled);
 	mutex_unlock(&adc->lock);
 }
 
 static int ingenic_adc_capture(struct ingenic_adc *adc,
 			       int engine)
 {
+	u32 cfg;
 	u8 val;
 	int ret;
 
-	ingenic_adc_enable(adc, engine, true);
+	/*
+	 * Disable CMD_SEL temporarily, because it causes wrong VBAT readings,
+	 * probably due to the switch of VREF. We must keep the lock here to
+	 * avoid races with the buffer enable/disable functions.
+	 */
+	mutex_lock(&adc->lock);
+	cfg = readl(adc->base + JZ_ADC_REG_CFG);
+	writel(cfg & ~JZ_ADC_REG_CFG_CMD_SEL, adc->base + JZ_ADC_REG_CFG);
+
+	ingenic_adc_enable_unlocked(adc, engine, true);
 	ret = readb_poll_timeout(adc->base + JZ_ADC_REG_ENABLE, val,
 				 !(val & BIT(engine)), 250, 1000);
 	if (ret)
-		ingenic_adc_enable(adc, engine, false);
+		ingenic_adc_enable_unlocked(adc, engine, false);
+
+	writel(cfg, adc->base + JZ_ADC_REG_CFG);
+	mutex_unlock(&adc->lock);
 
 	return ret;
 }