[PATCH v3 0/4] Add support for tsens controller reinit via trustzone

[PATCH v3 0/4] Add support for tsens controller reinit via trustzone

[Bhupesh Sharma]

Changes since v2:
-----------------
- v2 can be viewed here: https://lore.kernel.org/linux-arm-msm/20220724122424.2509021-1-bhupesh.sharma@linaro.org/
- Dropped sm6375 specific patch from v3, as suggested by Konrad.
- Rebased on latest linux-next (master branch) tip.

Changes since v1:
-----------------
- v1 can be viewed here: https://lore.kernel.org/linux-arm-msm/20220701145815.2037993-1-bhupesh.sharma@linaro.org/
- Addressed several comments from Bjorn regarding locking, serialization
  etc received on v1.
- Addressed Konrad's concerns about the tsens controller found on sm6375
  SoC which seems to start in a bad state or is disabled when entering
  the linux world.
- This series would depend on sm6375 tsens controller changes being
  added by Konrad. It is based on linux-next (master branch) tip.

Some versions of Qualcomm tsens controller might enter a
'bad state' causing sensor temperatures/interrupts status
to be in an 'invalid' state.

It is recommended to re-initialize the tsens controller
via trustzone (secure registers) using scm call(s) when that
happens.

This patchset adds the support for the same.

Cc: bjorn.andersson@linaro.org
Cc: Amit Kucheria <amitk@kernel.org>
Cc: Thara Gopinath <thara.gopinath@gmail.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-msm@vger.kernel.org

Bhupesh Sharma (4):
  firmware: qcom: scm: Add support for tsens reinit workaround
  thermal: qcom: tsens: Add hooks for supplying platform specific reinit
    quirks
  thermal: qcom: tsens: Add driver support for re-initialization quirk
  thermal: qcom: tsens: Add reinit quirk support for tsens v2
    controllers

 drivers/firmware/qcom_scm.c     |  15 +++
 drivers/firmware/qcom_scm.h     |   4 +
 drivers/thermal/qcom/tsens-v2.c |  15 +++
 drivers/thermal/qcom/tsens.c    | 200 ++++++++++++++++++++++++++++++++
 drivers/thermal/qcom/tsens.h    |  18 ++-
 include/linux/qcom_scm.h        |   2 +
 6 files changed, 253 insertions(+), 1 deletion(-)

-- 
2.35.3

[PATCH v3 1/4] firmware: qcom: scm: Add support for tsens reinit workaround

[Bhupesh Sharma]

Some versions of Qualcomm tsens controller might enter a
'bad state' while running stability tests causing sensor
temperatures/interrupts status to be in an 'invalid' state.

It is recommended to re-initialize the tsens controller
via trustzone (secure registers) using scm call(s) when that
happens.

Add support for the same in the qcom_scm driver.

Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Amit Kucheria <amitk@kernel.org>
Cc: Thara Gopinath <thara.gopinath@gmail.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-msm@vger.kernel.org
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
---
 drivers/firmware/qcom_scm.c | 15 +++++++++++++++
 drivers/firmware/qcom_scm.h |  4 ++++
 include/linux/qcom_scm.h    |  2 ++
 3 files changed, 21 insertions(+)

diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c
index cdbfe54c8146..93adcc046a62 100644
--- a/drivers/firmware/qcom_scm.c
+++ b/drivers/firmware/qcom_scm.c
@@ -858,6 +858,21 @@ int qcom_scm_mem_protect_video_var(u32 cp_start, u32 cp_size,
 }
 EXPORT_SYMBOL(qcom_scm_mem_protect_video_var);
 
+int qcom_scm_tsens_reinit(void)
+{
+	int ret;
+	const struct qcom_scm_desc desc = {
+		.svc = QCOM_SCM_SVC_TSENS,
+		.cmd = QCOM_SCM_TSENS_INIT_ID,
+	};
+	struct qcom_scm_res res;
+
+	ret = qcom_scm_call(__scm->dev, &desc, &res);
+
+	return ret ? : res.result[0];
+}
+EXPORT_SYMBOL(qcom_scm_tsens_reinit);
+
 static int __qcom_scm_assign_mem(struct device *dev, phys_addr_t mem_region,
 				 size_t mem_sz, phys_addr_t src, size_t src_sz,
 				 phys_addr_t dest, size_t dest_sz)
diff --git a/drivers/firmware/qcom_scm.h b/drivers/firmware/qcom_scm.h
index 0d51eef2472f..495fa00230c7 100644
--- a/drivers/firmware/qcom_scm.h
+++ b/drivers/firmware/qcom_scm.h
@@ -94,6 +94,10 @@ extern int scm_legacy_call(struct device *dev, const struct qcom_scm_desc *desc,
 #define QCOM_SCM_PIL_PAS_IS_SUPPORTED	0x07
 #define QCOM_SCM_PIL_PAS_MSS_RESET	0x0a
 
+/* TSENS Services and Function IDs */
+#define QCOM_SCM_SVC_TSENS		0x1E
+#define QCOM_SCM_TSENS_INIT_ID		0x5
+
 #define QCOM_SCM_SVC_IO			0x05
 #define QCOM_SCM_IO_READ		0x01
 #define QCOM_SCM_IO_WRITE		0x02
diff --git a/include/linux/qcom_scm.h b/include/linux/qcom_scm.h
index f8335644a01a..5c37e1658cef 100644
--- a/include/linux/qcom_scm.h
+++ b/include/linux/qcom_scm.h
@@ -124,4 +124,6 @@ extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val,
 extern int qcom_scm_lmh_profile_change(u32 profile_id);
 extern bool qcom_scm_lmh_dcvsh_available(void);
 
+extern int qcom_scm_tsens_reinit(void);
+
 #endif
-- 
2.35.3

[PATCH v3 2/4] thermal: qcom: tsens: Add hooks for supplying platform specific reinit quirks

[Bhupesh Sharma]

Add hooks inside platform specific data which can be
used by Qualcomm tsens controller(s) which might need
reinitialization via trustzone.

Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Amit Kucheria <amitk@kernel.org>
Cc: Thara Gopinath <thara.gopinath@gmail.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-msm@vger.kernel.org
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
---
 drivers/thermal/qcom/tsens.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/thermal/qcom/tsens.h b/drivers/thermal/qcom/tsens.h
index ba05c8233356..92787017c6ab 100644
--- a/drivers/thermal/qcom/tsens.h
+++ b/drivers/thermal/qcom/tsens.h
@@ -513,6 +513,7 @@ struct tsens_features {
  * @num_sensors: Number of sensors supported by platform
  * @ops: operations the tsens instance supports
  * @hw_ids: Subset of sensors ids supported by platform, if not the first n
+ * @needs_reinit_wa: tsens controller might need reinit via trustzone
  * @feat: features of the IP
  * @fields: bitfield locations
  */
@@ -520,6 +521,7 @@ struct tsens_plat_data {
 	const u32		num_sensors;
 	const struct tsens_ops	*ops;
 	unsigned int		*hw_ids;
+	bool			needs_reinit_wa;
 	struct tsens_features	*feat;
 	const struct reg_field		*fields;
 };
@@ -542,6 +544,7 @@ struct tsens_context {
  * @srot_map: pointer to SROT register address space
  * @tm_offset: deal with old device trees that don't address TM and SROT
  *             address space separately
+ * @needs_reinit_wa: tsens controller might need reinit via trustzone
  * @ul_lock: lock while processing upper/lower threshold interrupts
  * @crit_lock: lock while processing critical threshold interrupts
  * @rf: array of regmap_fields used to store value of the field
@@ -559,6 +562,7 @@ struct tsens_priv {
 	struct regmap			*tm_map;
 	struct regmap			*srot_map;
 	u32				tm_offset;
+	bool				needs_reinit_wa;
 
 	/* lock for upper/lower threshold interrupts */
 	spinlock_t			ul_lock;
-- 
2.35.3

[PATCH v3 3/4] thermal: qcom: tsens: Add driver support for re-initialization quirk

[Bhupesh Sharma]

Since for some Qualcomm tsens controllers, its suggested to
monitor the controller health periodically and in case an
issue is detected, to re-initialize the tsens controller
via trustzone, add the support for the same in the
qcom tsens driver.

Note that once the tsens controller is reset using scm call,
all SROT and TM region registers will enter the reset mode.

While all the SROT registers will be re-programmed and
re-enabled in trustzone prior to the scm call exit, the TM
region registers will not re-initialized in trustzone and thus
need to be handled by the tsens driver.

Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Amit Kucheria <amitk@kernel.org>
Cc: Thara Gopinath <thara.gopinath@gmail.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-msm@vger.kernel.org
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
---
 drivers/thermal/qcom/tsens-v2.c |   3 +
 drivers/thermal/qcom/tsens.c    | 197 ++++++++++++++++++++++++++++++++
 drivers/thermal/qcom/tsens.h    |  12 ++
 3 files changed, 212 insertions(+)

diff --git a/drivers/thermal/qcom/tsens-v2.c b/drivers/thermal/qcom/tsens-v2.c
index b293ed32174b..f521e4479cc5 100644
--- a/drivers/thermal/qcom/tsens-v2.c
+++ b/drivers/thermal/qcom/tsens-v2.c
@@ -88,6 +88,9 @@ static const struct reg_field tsens_v2_regfields[MAX_REGFIELDS] = {
 
 	/* TRDY: 1=ready, 0=in progress */
 	[TRDY] = REG_FIELD(TM_TRDY_OFF, 0, 0),
+
+	/* FIRST_ROUND_COMPLETE: 1=complete, 0=not complete */
+	[FIRST_ROUND_COMPLETE] = REG_FIELD(TM_TRDY_OFF, 3, 3),
 };
 
 static const struct tsens_ops ops_generic_v2 = {
diff --git a/drivers/thermal/qcom/tsens.c b/drivers/thermal/qcom/tsens.c
index e49f58e83513..c2d085fb5447 100644
--- a/drivers/thermal/qcom/tsens.c
+++ b/drivers/thermal/qcom/tsens.c
@@ -7,6 +7,7 @@
 #include <linux/debugfs.h>
 #include <linux/err.h>
 #include <linux/io.h>
+#include <linux/qcom_scm.h>
 #include <linux/module.h>
 #include <linux/nvmem-consumer.h>
 #include <linux/of.h>
@@ -594,6 +595,113 @@ static void tsens_disable_irq(struct tsens_priv *priv)
 	regmap_field_write(priv->rf[INT_EN], 0);
 }
 
+static int tsens_reenable_hw_after_scm(struct tsens_priv *priv)
+{
+	/*
+	 * Re-enable watchdog, unmask the bark and
+	 * disable cycle completion monitoring.
+	 */
+	regmap_field_write(priv->rf[WDOG_BARK_CLEAR], 1);
+	regmap_field_write(priv->rf[WDOG_BARK_CLEAR], 0);
+	regmap_field_write(priv->rf[WDOG_BARK_MASK], 0);
+	regmap_field_write(priv->rf[CC_MON_MASK], 1);
+
+	/* Re-enable interrupts */
+	tsens_enable_irq(priv);
+
+	return 0;
+}
+
+static int tsens_health_check_and_reinit(struct tsens_priv *priv,
+					 int hw_id)
+{
+	int ret, trdy, first_round, sw_reg;
+	unsigned long timeout;
+
+	/* First check if TRDY is SET */
+	ret = regmap_field_read(priv->rf[TRDY], &trdy);
+	if (ret)
+		goto err;
+
+	if (!trdy) {
+		ret = regmap_field_read(priv->rf[FIRST_ROUND_COMPLETE], &first_round);
+		if (ret)
+			goto err;
+
+		if (!first_round) {
+			WARN_ON(!mutex_is_locked(&priv->reinit_mutex));
+
+			/* Wait for 2 ms for tsens controller to recover */
+			timeout = jiffies + msecs_to_jiffies(RESET_TIMEOUT_MS);
+			do {
+				ret = regmap_field_read(priv->rf[FIRST_ROUND_COMPLETE],
+						&first_round);
+				if (ret)
+					goto err;
+
+				if (first_round) {
+					dev_dbg(priv->dev, "tsens controller recovered\n");
+					return 0; /* success */
+				}
+			} while (time_before(jiffies, timeout));
+
+			spin_lock(&priv->reinit_lock);
+
+			/*
+			 * Invoke SCM call only if SW register write is
+			 * reflecting in controller. Try it for 2 ms.
+			 * In case that fails mark the tsens controller
+			 * as unrecoverable.
+			 */
+			timeout = jiffies + msecs_to_jiffies(RESET_TIMEOUT_MS);
+			do {
+				ret = regmap_field_write(priv->rf[INT_EN], CRITICAL_INT_EN);
+				if (ret)
+					goto err;
+
+				ret = regmap_field_read(priv->rf[INT_EN], &sw_reg);
+				if (ret)
+					goto err;
+			} while ((sw_reg & CRITICAL_INT_EN) && (time_before(jiffies, timeout)));
+
+			if (!(sw_reg & CRITICAL_INT_EN)) {
+				ret = -ENOTRECOVERABLE;
+				goto err;
+			}
+
+			/*
+			 * tsens controller did not recover,
+			 * proceed with SCM call to re-init it.
+			 */
+			ret = qcom_scm_tsens_reinit();
+			if (ret) {
+				dev_err(priv->dev, "tsens reinit scm call failed (%d)\n", ret);
+				goto err;
+			}
+
+			/*
+			 * After the SCM call, we need to re-enable
+			 * the interrupts and also set active threshold
+			 * for each sensor.
+			 */
+			ret = tsens_reenable_hw_after_scm(priv);
+			if (ret) {
+				dev_err(priv->dev,
+					"tsens re-enable after scm call failed (%d)\n", ret);
+				goto err;
+			}
+
+			/* Notify reinit wa worker */
+			queue_work(system_highpri_wq, &priv->reinit_wa_notify);
+
+			spin_unlock(&priv->reinit_lock);
+		}
+	}
+
+err:
+	return ret;
+}
+
 int get_temp_tsens_valid(const struct tsens_sensor *s, int *temp)
 {
 	struct tsens_priv *priv = s->priv;
@@ -607,6 +715,21 @@ int get_temp_tsens_valid(const struct tsens_sensor *s, int *temp)
 	if (tsens_version(priv) == VER_0)
 		goto get_temp;
 
+	/*
+	 * For some tsens controllers, its suggested to
+	 * monitor the controller health periodically
+	 * and in case an issue is detected to reinit
+	 * tsens controller via trustzone.
+	 */
+	if (priv->needs_reinit_wa) {
+		mutex_lock(&priv->reinit_mutex);
+		ret = tsens_health_check_and_reinit(priv, hw_id);
+		mutex_unlock(&priv->reinit_mutex);
+
+		if (ret)
+			return ret;
+	}
+
 	/* Valid bit is 0 for 6 AHB clock cycles.
 	 * At 19.2MHz, 1 AHB clock is ~60ns.
 	 * We should enter this loop very, very rarely.
@@ -739,6 +862,40 @@ static const struct regmap_config tsens_srot_config = {
 	.reg_stride	= 4,
 };
 
+static void __tsens_reinit_worker(struct tsens_priv *priv)
+{
+	int ret, temp;
+	unsigned int i;
+	struct tsens_irq_data d;
+
+	for (i = 0; i < priv->num_sensors; i++) {
+		const struct tsens_sensor *s = &priv->sensor[i];
+		u32 hw_id = s->hw_id;
+
+		if (!s->tzd)
+			continue;
+		if (!tsens_threshold_violated(priv, hw_id, &d))
+			continue;
+
+		ret = get_temp_tsens_valid(s, &temp);
+		if (ret) {
+			dev_err(priv->dev, "[%u] error reading sensor during reinit\n", hw_id);
+			continue;
+		}
+
+		tsens_read_irq_state(priv, hw_id, s, &d);
+
+		if ((d.up_thresh < temp) || (d.low_thresh > temp)) {
+			dev_dbg(priv->dev, "[%u] TZ update trigger during reinit (%d mC)\n",
+				hw_id, temp);
+			thermal_zone_device_update(s->tzd, THERMAL_EVENT_UNSPECIFIED);
+		} else {
+			dev_dbg(priv->dev, "[%u] no violation during reinit (%d)\n",
+				hw_id, temp);
+		}
+	}
+}
+
 int __init init_common(struct tsens_priv *priv)
 {
 	void __iomem *tm_base, *srot_base;
@@ -860,6 +1017,14 @@ int __init init_common(struct tsens_priv *priv)
 		goto err_put_device;
 	}
 
+	priv->rf[FIRST_ROUND_COMPLETE] = devm_regmap_field_alloc(dev,
+								priv->tm_map,
+								priv->fields[FIRST_ROUND_COMPLETE]);
+	if (IS_ERR(priv->rf[FIRST_ROUND_COMPLETE])) {
+		ret = PTR_ERR(priv->rf[FIRST_ROUND_COMPLETE]);
+		goto err_put_device;
+	}
+
 	/* This loop might need changes if enum regfield_ids is reordered */
 	for (j = LAST_TEMP_0; j <= UP_THRESH_15; j += 16) {
 		for (i = 0; i < priv->feat->max_sensors; i++) {
@@ -1082,6 +1247,14 @@ static int tsens_register(struct tsens_priv *priv)
 	return ret;
 }
 
+static void tsens_reinit_worker_notify(struct work_struct *work)
+{
+	struct tsens_priv *priv = container_of(work, struct tsens_priv,
+					       reinit_wa_notify);
+
+	__tsens_reinit_worker(priv);
+}
+
 static int tsens_probe(struct platform_device *pdev)
 {
 	int ret, i;
@@ -1123,6 +1296,11 @@ static int tsens_probe(struct platform_device *pdev)
 
 	priv->dev = dev;
 	priv->num_sensors = num_sensors;
+	priv->needs_reinit_wa = data->needs_reinit_wa;
+
+	if (priv->needs_reinit_wa && !qcom_scm_is_available())
+		return -EPROBE_DEFER;
+
 	priv->ops = data->ops;
 	for (i = 0;  i < priv->num_sensors; i++) {
 		if (data->hw_ids)
@@ -1138,6 +1316,25 @@ static int tsens_probe(struct platform_device *pdev)
 	if (!priv->ops || !priv->ops->init || !priv->ops->get_temp)
 		return -EINVAL;
 
+	/*
+	 * Reinitialization workaround is currently supported only for
+	 * tsens controller versions v2.
+	 *
+	 * If incorrect platform data is passed to this effect, ignore
+	 * the requested setting and move forward.
+	 */
+	if (priv->needs_reinit_wa && (tsens_version(priv) < VER_2_X)) {
+		dev_warn(dev,
+			 "%s: Reinit quirk available only for tsens v2\n", __func__);
+		priv->needs_reinit_wa = false;
+	}
+
+	mutex_init(&priv->reinit_mutex);
+	spin_lock_init(&priv->reinit_lock);
+
+	if (priv->needs_reinit_wa)
+		INIT_WORK(&priv->reinit_wa_notify, tsens_reinit_worker_notify);
+
 	ret = priv->ops->init(priv);
 	if (ret < 0) {
 		dev_err(dev, "%s: init failed\n", __func__);
diff --git a/drivers/thermal/qcom/tsens.h b/drivers/thermal/qcom/tsens.h
index 92787017c6ab..900d2a74d25e 100644
--- a/drivers/thermal/qcom/tsens.h
+++ b/drivers/thermal/qcom/tsens.h
@@ -14,9 +14,12 @@
 #define SLOPE_FACTOR		1000
 #define SLOPE_DEFAULT		3200
 #define TIMEOUT_US		100
+#define RESET_TIMEOUT_MS	2
 #define THRESHOLD_MAX_ADC_CODE	0x3ff
 #define THRESHOLD_MIN_ADC_CODE	0x0
 
+#define CRITICAL_INT_EN		(BIT(2))
+
 #include <linux/interrupt.h>
 #include <linux/thermal.h>
 #include <linux/regmap.h>
@@ -165,6 +168,7 @@ enum regfield_ids {
 	/* ----- TM ------ */
 	/* TRDY */
 	TRDY,
+	FIRST_ROUND_COMPLETE,
 	/* INTERRUPT ENABLE */
 	INT_EN,	/* v2+ has separate enables for crit, upper and lower irq */
 	/* STATUS */
@@ -564,6 +568,14 @@ struct tsens_priv {
 	u32				tm_offset;
 	bool				needs_reinit_wa;
 
+	struct work_struct		reinit_wa_notify;
+
+	/* protects reinit related serialization */
+	struct mutex			reinit_mutex;
+
+	/* lock for reinit workaround */
+	spinlock_t			reinit_lock;
+
 	/* lock for upper/lower threshold interrupts */
 	spinlock_t			ul_lock;
 
-- 
2.35.3

[PATCH v3 4/4] thermal: qcom: tsens: Add reinit quirk support for tsens v2 controllers

[Bhupesh Sharma]

Some Qualcomm tsens v2 controllers like those present on
sm8150 SoC might require re-initialization via trustzone
[via scm call(s)] when it enters a 'bad state' causing
sensor temperatures/interrupts status to be in an
'invalid' state.

Add hooks for the same in the qcom tsens driver.

Devices requiring the same can pass the relevant
compatible string in dt and the driver hook can
be used accordingly.

Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Amit Kucheria <amitk@kernel.org>
Cc: Thara Gopinath <thara.gopinath@gmail.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-msm@vger.kernel.org
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
---
 drivers/thermal/qcom/tsens-v2.c | 12 ++++++++++++
 drivers/thermal/qcom/tsens.c    |  3 +++
 drivers/thermal/qcom/tsens.h    |  2 +-
 3 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/drivers/thermal/qcom/tsens-v2.c b/drivers/thermal/qcom/tsens-v2.c
index f521e4479cc5..431f17f99d34 100644
--- a/drivers/thermal/qcom/tsens-v2.c
+++ b/drivers/thermal/qcom/tsens-v2.c
@@ -104,6 +104,18 @@ struct tsens_plat_data data_tsens_v2 = {
 	.fields	= tsens_v2_regfields,
 };
 
+/*
+ * For some tsens v2 controllers, its suggested to monitor the
+ * controller health periodically and in case an issue is detected
+ * to reinit tsens controller via trustzone.
+ */
+struct tsens_plat_data data_tsens_v2_reinit = {
+	.ops		= &ops_generic_v2,
+	.feat		= &tsens_v2_feat,
+	.needs_reinit_wa = true,
+	.fields	= tsens_v2_regfields,
+};
+
 /* Kept around for backward compatibility with old msm8996.dtsi */
 struct tsens_plat_data data_8996 = {
 	.num_sensors	= 13,
diff --git a/drivers/thermal/qcom/tsens.c b/drivers/thermal/qcom/tsens.c
index c2d085fb5447..928b6a44fda8 100644
--- a/drivers/thermal/qcom/tsens.c
+++ b/drivers/thermal/qcom/tsens.c
@@ -1145,6 +1145,9 @@ static const struct of_device_id tsens_table[] = {
 	}, {
 		.compatible = "qcom,msm8996-tsens",
 		.data = &data_8996,
+	}, {
+		.compatible = "qcom,sm8150-tsens",
+		.data = &data_tsens_v2_reinit,
 	}, {
 		.compatible = "qcom,tsens-v1",
 		.data = &data_tsens_v1,
diff --git a/drivers/thermal/qcom/tsens.h b/drivers/thermal/qcom/tsens.h
index 900d2a74d25e..03cc3a790972 100644
--- a/drivers/thermal/qcom/tsens.h
+++ b/drivers/thermal/qcom/tsens.h
@@ -607,6 +607,6 @@ extern struct tsens_plat_data data_8916, data_8939, data_8974, data_9607;
 extern struct tsens_plat_data data_tsens_v1, data_8976;
 
 /* TSENS v2 targets */
-extern struct tsens_plat_data data_8996, data_tsens_v2;
+extern struct tsens_plat_data data_8996, data_tsens_v2_reinit, data_tsens_v2;
 
 #endif /* __QCOM_TSENS_H__ */
-- 
2.35.3

Re: [PATCH v3 0/4] Add support for tsens controller reinit via trustzone

[Bhupesh Sharma]

On 8/4/22 11:16 AM, Bhupesh Sharma wrote:
> Changes since v2:
> -----------------
> - v2 can be viewed here: https://lore.kernel.org/linux-arm-msm/20220724122424.2509021-1-bhupesh.sharma@linaro.org/
> - Dropped sm6375 specific patch from v3, as suggested by Konrad.
> - Rebased on latest linux-next (master branch) tip.
> 
> Changes since v1:
> -----------------
> - v1 can be viewed here: https://lore.kernel.org/linux-arm-msm/20220701145815.2037993-1-bhupesh.sharma@linaro.org/
> - Addressed several comments from Bjorn regarding locking, serialization
>    etc received on v1.
> - Addressed Konrad's concerns about the tsens controller found on sm6375
>    SoC which seems to start in a bad state or is disabled when entering
>    the linux world.
> - This series would depend on sm6375 tsens controller changes being
>    added by Konrad. It is based on linux-next (master branch) tip.
> 
> Some versions of Qualcomm tsens controller might enter a
> 'bad state' causing sensor temperatures/interrupts status
> to be in an 'invalid' state.
> 
> It is recommended to re-initialize the tsens controller
> via trustzone (secure registers) using scm call(s) when that
> happens.
> 
> This patchset adds the support for the same.
> 
> Cc: bjorn.andersson@linaro.org
> Cc: Amit Kucheria <amitk@kernel.org>
> Cc: Thara Gopinath <thara.gopinath@gmail.com>
> Cc: linux-pm@vger.kernel.org
> Cc: linux-arm-msm@vger.kernel.org
> 
> Bhupesh Sharma (4):
>    firmware: qcom: scm: Add support for tsens reinit workaround
>    thermal: qcom: tsens: Add hooks for supplying platform specific reinit
>      quirks
>    thermal: qcom: tsens: Add driver support for re-initialization quirk
>    thermal: qcom: tsens: Add reinit quirk support for tsens v2
>      controllers
> 
>   drivers/firmware/qcom_scm.c     |  15 +++
>   drivers/firmware/qcom_scm.h     |   4 +
>   drivers/thermal/qcom/tsens-v2.c |  15 +++
>   drivers/thermal/qcom/tsens.c    | 200 ++++++++++++++++++++++++++++++++
>   drivers/thermal/qcom/tsens.h    |  18 ++-
>   include/linux/qcom_scm.h        |   2 +
>   6 files changed, 253 insertions(+), 1 deletion(-)
> 

Gentle ping.

Thanks,
Bhupesh

Re: [PATCH v3 3/4] thermal: qcom: tsens: Add driver support for re-initialization quirk

[Bjorn Andersson]

On Thu, Aug 04, 2022 at 11:16:37AM +0530, Bhupesh Sharma wrote:
> Since for some Qualcomm tsens controllers, its suggested to
> monitor the controller health periodically and in case an
> issue is detected, to re-initialize the tsens controller
> via trustzone, add the support for the same in the
> qcom tsens driver.
> 
> Note that once the tsens controller is reset using scm call,
> all SROT and TM region registers will enter the reset mode.
> 
> While all the SROT registers will be re-programmed and
> re-enabled in trustzone prior to the scm call exit, the TM
> region registers will not re-initialized in trustzone and thus
> need to be handled by the tsens driver.
> 
> Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
> Cc: Amit Kucheria <amitk@kernel.org>
> Cc: Thara Gopinath <thara.gopinath@gmail.com>
> Cc: linux-pm@vger.kernel.org
> Cc: linux-arm-msm@vger.kernel.org
> Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
> ---
>  drivers/thermal/qcom/tsens-v2.c |   3 +
>  drivers/thermal/qcom/tsens.c    | 197 ++++++++++++++++++++++++++++++++
>  drivers/thermal/qcom/tsens.h    |  12 ++
>  3 files changed, 212 insertions(+)
> 
> diff --git a/drivers/thermal/qcom/tsens-v2.c b/drivers/thermal/qcom/tsens-v2.c
> index b293ed32174b..f521e4479cc5 100644
> --- a/drivers/thermal/qcom/tsens-v2.c
> +++ b/drivers/thermal/qcom/tsens-v2.c
> @@ -88,6 +88,9 @@ static const struct reg_field tsens_v2_regfields[MAX_REGFIELDS] = {
>  
>  	/* TRDY: 1=ready, 0=in progress */
>  	[TRDY] = REG_FIELD(TM_TRDY_OFF, 0, 0),
> +
> +	/* FIRST_ROUND_COMPLETE: 1=complete, 0=not complete */
> +	[FIRST_ROUND_COMPLETE] = REG_FIELD(TM_TRDY_OFF, 3, 3),
>  };
>  
>  static const struct tsens_ops ops_generic_v2 = {
> diff --git a/drivers/thermal/qcom/tsens.c b/drivers/thermal/qcom/tsens.c
> index e49f58e83513..c2d085fb5447 100644
> --- a/drivers/thermal/qcom/tsens.c
> +++ b/drivers/thermal/qcom/tsens.c
> @@ -7,6 +7,7 @@
>  #include <linux/debugfs.h>
>  #include <linux/err.h>
>  #include <linux/io.h>
> +#include <linux/qcom_scm.h>
>  #include <linux/module.h>
>  #include <linux/nvmem-consumer.h>
>  #include <linux/of.h>
> @@ -594,6 +595,113 @@ static void tsens_disable_irq(struct tsens_priv *priv)
>  	regmap_field_write(priv->rf[INT_EN], 0);
>  }
>  
> +static int tsens_reenable_hw_after_scm(struct tsens_priv *priv)

As written, this is a void function.

> +{
> +	/*
> +	 * Re-enable watchdog, unmask the bark and
> +	 * disable cycle completion monitoring.
> +	 */
> +	regmap_field_write(priv->rf[WDOG_BARK_CLEAR], 1);
> +	regmap_field_write(priv->rf[WDOG_BARK_CLEAR], 0);
> +	regmap_field_write(priv->rf[WDOG_BARK_MASK], 0);
> +	regmap_field_write(priv->rf[CC_MON_MASK], 1);
> +
> +	/* Re-enable interrupts */
> +	tsens_enable_irq(priv);
> +
> +	return 0;
> +}
> +
> +static int tsens_health_check_and_reinit(struct tsens_priv *priv,
> +					 int hw_id)
> +{
> +	int ret, trdy, first_round, sw_reg;
> +	unsigned long timeout;
> +
> +	/* First check if TRDY is SET */
> +	ret = regmap_field_read(priv->rf[TRDY], &trdy);
> +	if (ret)
> +		goto err;
> +
> +	if (!trdy) {

if (trdy)
	return 0;

Would save you one level of indentation.

> +		ret = regmap_field_read(priv->rf[FIRST_ROUND_COMPLETE], &first_round);
> +		if (ret)
> +			goto err;
> +
> +		if (!first_round) {

if (first_round)
	return 0;

Would save you another level of indentation.

> +			WARN_ON(!mutex_is_locked(&priv->reinit_mutex));

At least for now the function is only called within a small locked
region, so it's going to be locked here. But I'm wondering if there's
any relationship between the lock state of reinit_mutex and the values
of TRDY and FIRST_ROUND_COMPLETE.

Seems like it's possible to hit this function repeatedly and have it
exit early because of TRDY and FIRST_ROUND_COMPLETE values and then one
day if will reach here and trip.

So how about starting the function with this check, to make it more
likely to be hit in our testing?

> +
> +			/* Wait for 2 ms for tsens controller to recover */
> +			timeout = jiffies + msecs_to_jiffies(RESET_TIMEOUT_MS);
> +			do {
> +				ret = regmap_field_read(priv->rf[FIRST_ROUND_COMPLETE],
> +						&first_round);
> +				if (ret)
> +					goto err;
> +
> +				if (first_round) {
> +					dev_dbg(priv->dev, "tsens controller recovered\n");
> +					return 0; /* success */
> +				}
> +			} while (time_before(jiffies, timeout));

I see no delays in this loop, so we're presumably going to spin here
tightly for 2ms.

I think you could write this loop as:

	ret = regmap_field_read_poll_timeout(priv->rf[FIRST_ROUND_COMPLETE],
					     &first_round, first_round, 100, 2000);
	if (ret == 0) {
		dev_dbg(priv->dev, "tsens controller recovered\n");
		return 0;
	}

> +
> +			spin_lock(&priv->reinit_lock);
> +
> +			/*
> +			 * Invoke SCM call only if SW register write is
> +			 * reflecting in controller. Try it for 2 ms.
> +			 * In case that fails mark the tsens controller
> +			 * as unrecoverable.
> +			 */
> +			timeout = jiffies + msecs_to_jiffies(RESET_TIMEOUT_MS);
> +			do {
> +				ret = regmap_field_write(priv->rf[INT_EN], CRITICAL_INT_EN);
> +				if (ret)
> +					goto err;

You're holding reinit_lock here.

> +
> +				ret = regmap_field_read(priv->rf[INT_EN], &sw_reg);
> +				if (ret)
> +					goto err;

And here.

> +			} while ((sw_reg & CRITICAL_INT_EN) && (time_before(jiffies, timeout)));

And again, this is a tight loop. Please add a usleep_range(100, 1000),
perhaps inbetween the write and read?

> +
> +			if (!(sw_reg & CRITICAL_INT_EN)) {
> +				ret = -ENOTRECOVERABLE;
> +				goto err;

Again, reinit_lock is held here.

> +			}
> +
> +			/*
> +			 * tsens controller did not recover,
> +			 * proceed with SCM call to re-init it.
> +			 */
> +			ret = qcom_scm_tsens_reinit();
> +			if (ret) {
> +				dev_err(priv->dev, "tsens reinit scm call failed (%d)\n", ret);
> +				goto err;

And here.

> +			}
> +
> +			/*
> +			 * After the SCM call, we need to re-enable
> +			 * the interrupts and also set active threshold
> +			 * for each sensor.
> +			 */
> +			ret = tsens_reenable_hw_after_scm(priv);

As written tsens_reenable_hw_after_scm() doesn't return any value, so
skip the error handling.

> +			if (ret) {
> +				dev_err(priv->dev,
> +					"tsens re-enable after scm call failed (%d)\n", ret);
> +				goto err;

And here...

> +			}
> +
> +			/* Notify reinit wa worker */
> +			queue_work(system_highpri_wq, &priv->reinit_wa_notify);
> +
> +			spin_unlock(&priv->reinit_lock);
> +		}
> +	}
> +
> +err:
> +	return ret;
> +}
> +
>  int get_temp_tsens_valid(const struct tsens_sensor *s, int *temp)
>  {
>  	struct tsens_priv *priv = s->priv;
> @@ -607,6 +715,21 @@ int get_temp_tsens_valid(const struct tsens_sensor *s, int *temp)
>  	if (tsens_version(priv) == VER_0)
>  		goto get_temp;
>  
> +	/*
> +	 * For some tsens controllers, its suggested to
> +	 * monitor the controller health periodically
> +	 * and in case an issue is detected to reinit
> +	 * tsens controller via trustzone.

Please use your 80 chars.

Regards,
Bjorn

Re: [PATCH v3 4/4] thermal: qcom: tsens: Add reinit quirk support for tsens v2 controllers

[Bjorn Andersson]

On Thu, Aug 04, 2022 at 11:16:38AM +0530, Bhupesh Sharma wrote:
> Some Qualcomm tsens v2 controllers like those present on
> sm8150 SoC might require re-initialization via trustzone
> [via scm call(s)] when it enters a 'bad state' causing
> sensor temperatures/interrupts status to be in an
> 'invalid' state.
> 
> Add hooks for the same in the qcom tsens driver.
> 
> Devices requiring the same can pass the relevant
> compatible string in dt and the driver hook can
> be used accordingly.
> 
> Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
> Cc: Amit Kucheria <amitk@kernel.org>
> Cc: Thara Gopinath <thara.gopinath@gmail.com>
> Cc: linux-pm@vger.kernel.org
> Cc: linux-arm-msm@vger.kernel.org
> Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>

Reviewed-by: Bjorn Andersson <andersson@kernel.org>

> ---
>  drivers/thermal/qcom/tsens-v2.c | 12 ++++++++++++
>  drivers/thermal/qcom/tsens.c    |  3 +++
>  drivers/thermal/qcom/tsens.h    |  2 +-
>  3 files changed, 16 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/thermal/qcom/tsens-v2.c b/drivers/thermal/qcom/tsens-v2.c
> index f521e4479cc5..431f17f99d34 100644
> --- a/drivers/thermal/qcom/tsens-v2.c
> +++ b/drivers/thermal/qcom/tsens-v2.c
> @@ -104,6 +104,18 @@ struct tsens_plat_data data_tsens_v2 = {
>  	.fields	= tsens_v2_regfields,
>  };
>  
> +/*
> + * For some tsens v2 controllers, its suggested to monitor the
> + * controller health periodically and in case an issue is detected
> + * to reinit tsens controller via trustzone.
> + */
> +struct tsens_plat_data data_tsens_v2_reinit = {
> +	.ops		= &ops_generic_v2,
> +	.feat		= &tsens_v2_feat,
> +	.needs_reinit_wa = true,
> +	.fields	= tsens_v2_regfields,
> +};
> +
>  /* Kept around for backward compatibility with old msm8996.dtsi */
>  struct tsens_plat_data data_8996 = {
>  	.num_sensors	= 13,
> diff --git a/drivers/thermal/qcom/tsens.c b/drivers/thermal/qcom/tsens.c
> index c2d085fb5447..928b6a44fda8 100644
> --- a/drivers/thermal/qcom/tsens.c
> +++ b/drivers/thermal/qcom/tsens.c
> @@ -1145,6 +1145,9 @@ static const struct of_device_id tsens_table[] = {
>  	}, {
>  		.compatible = "qcom,msm8996-tsens",
>  		.data = &data_8996,
> +	}, {
> +		.compatible = "qcom,sm8150-tsens",
> +		.data = &data_tsens_v2_reinit,
>  	}, {
>  		.compatible = "qcom,tsens-v1",
>  		.data = &data_tsens_v1,
> diff --git a/drivers/thermal/qcom/tsens.h b/drivers/thermal/qcom/tsens.h
> index 900d2a74d25e..03cc3a790972 100644
> --- a/drivers/thermal/qcom/tsens.h
> +++ b/drivers/thermal/qcom/tsens.h
> @@ -607,6 +607,6 @@ extern struct tsens_plat_data data_8916, data_8939, data_8974, data_9607;
>  extern struct tsens_plat_data data_tsens_v1, data_8976;
>  
>  /* TSENS v2 targets */
> -extern struct tsens_plat_data data_8996, data_tsens_v2;
> +extern struct tsens_plat_data data_8996, data_tsens_v2_reinit, data_tsens_v2;
>  
>  #endif /* __QCOM_TSENS_H__ */
> -- 
> 2.35.3
> 

Re: [PATCH v3 1/4] firmware: qcom: scm: Add support for tsens reinit workaround

[Bjorn Andersson]

On Thu, Aug 04, 2022 at 11:16:35AM +0530, Bhupesh Sharma wrote:
> Some versions of Qualcomm tsens controller might enter a
> 'bad state' while running stability tests causing sensor
> temperatures/interrupts status to be in an 'invalid' state.
> 
> It is recommended to re-initialize the tsens controller
> via trustzone (secure registers) using scm call(s) when that
> happens.
> 
> Add support for the same in the qcom_scm driver.
> 
> Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
> Cc: Amit Kucheria <amitk@kernel.org>
> Cc: Thara Gopinath <thara.gopinath@gmail.com>
> Cc: linux-pm@vger.kernel.org
> Cc: linux-arm-msm@vger.kernel.org
> Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
> ---
>  drivers/firmware/qcom_scm.c | 15 +++++++++++++++
>  drivers/firmware/qcom_scm.h |  4 ++++
>  include/linux/qcom_scm.h    |  2 ++
>  3 files changed, 21 insertions(+)
> 
> diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c
> index cdbfe54c8146..93adcc046a62 100644
> --- a/drivers/firmware/qcom_scm.c
> +++ b/drivers/firmware/qcom_scm.c
> @@ -858,6 +858,21 @@ int qcom_scm_mem_protect_video_var(u32 cp_start, u32 cp_size,
>  }
>  EXPORT_SYMBOL(qcom_scm_mem_protect_video_var);
>  
> +int qcom_scm_tsens_reinit(void)
> +{
> +	int ret;
> +	const struct qcom_scm_desc desc = {
> +		.svc = QCOM_SCM_SVC_TSENS,
> +		.cmd = QCOM_SCM_TSENS_INIT_ID,
> +	};
> +	struct qcom_scm_res res;
> +
> +	ret = qcom_scm_call(__scm->dev, &desc, &res);
> +
> +	return ret ? : res.result[0];
> +}
> +EXPORT_SYMBOL(qcom_scm_tsens_reinit);
> +
>  static int __qcom_scm_assign_mem(struct device *dev, phys_addr_t mem_region,
>  				 size_t mem_sz, phys_addr_t src, size_t src_sz,
>  				 phys_addr_t dest, size_t dest_sz)
> diff --git a/drivers/firmware/qcom_scm.h b/drivers/firmware/qcom_scm.h
> index 0d51eef2472f..495fa00230c7 100644
> --- a/drivers/firmware/qcom_scm.h
> +++ b/drivers/firmware/qcom_scm.h
> @@ -94,6 +94,10 @@ extern int scm_legacy_call(struct device *dev, const struct qcom_scm_desc *desc,
>  #define QCOM_SCM_PIL_PAS_IS_SUPPORTED	0x07
>  #define QCOM_SCM_PIL_PAS_MSS_RESET	0x0a
>  
> +/* TSENS Services and Function IDs */
> +#define QCOM_SCM_SVC_TSENS		0x1E

It would be nice if this 'E' was lowercase.

Reviewed-by: Bjorn Andersson <andersson@kernel.org>

Thanks,
Bjorn

> +#define QCOM_SCM_TSENS_INIT_ID		0x5
> +
>  #define QCOM_SCM_SVC_IO			0x05
>  #define QCOM_SCM_IO_READ		0x01
>  #define QCOM_SCM_IO_WRITE		0x02
> diff --git a/include/linux/qcom_scm.h b/include/linux/qcom_scm.h
> index f8335644a01a..5c37e1658cef 100644
> --- a/include/linux/qcom_scm.h
> +++ b/include/linux/qcom_scm.h
> @@ -124,4 +124,6 @@ extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val,
>  extern int qcom_scm_lmh_profile_change(u32 profile_id);
>  extern bool qcom_scm_lmh_dcvsh_available(void);
>  
> +extern int qcom_scm_tsens_reinit(void);
> +
>  #endif
> -- 
> 2.35.3
> 

Re: [PATCH v3 1/4] firmware: qcom: scm: Add support for tsens reinit workaround

[Bhupesh Sharma]

On 8/30/22 3:45 AM, Bjorn Andersson wrote:
> On Thu, Aug 04, 2022 at 11:16:35AM +0530, Bhupesh Sharma wrote:
>> Some versions of Qualcomm tsens controller might enter a
>> 'bad state' while running stability tests causing sensor
>> temperatures/interrupts status to be in an 'invalid' state.
>>
>> It is recommended to re-initialize the tsens controller
>> via trustzone (secure registers) using scm call(s) when that
>> happens.
>>
>> Add support for the same in the qcom_scm driver.
>>
>> Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
>> Cc: Amit Kucheria <amitk@kernel.org>
>> Cc: Thara Gopinath <thara.gopinath@gmail.com>
>> Cc: linux-pm@vger.kernel.org
>> Cc: linux-arm-msm@vger.kernel.org
>> Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
>> ---
>>   drivers/firmware/qcom_scm.c | 15 +++++++++++++++
>>   drivers/firmware/qcom_scm.h |  4 ++++
>>   include/linux/qcom_scm.h    |  2 ++
>>   3 files changed, 21 insertions(+)
>>
>> diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c
>> index cdbfe54c8146..93adcc046a62 100644
>> --- a/drivers/firmware/qcom_scm.c
>> +++ b/drivers/firmware/qcom_scm.c
>> @@ -858,6 +858,21 @@ int qcom_scm_mem_protect_video_var(u32 cp_start, u32 cp_size,
>>   }
>>   EXPORT_SYMBOL(qcom_scm_mem_protect_video_var);
>>   
>> +int qcom_scm_tsens_reinit(void)
>> +{
>> +	int ret;
>> +	const struct qcom_scm_desc desc = {
>> +		.svc = QCOM_SCM_SVC_TSENS,
>> +		.cmd = QCOM_SCM_TSENS_INIT_ID,
>> +	};
>> +	struct qcom_scm_res res;
>> +
>> +	ret = qcom_scm_call(__scm->dev, &desc, &res);
>> +
>> +	return ret ? : res.result[0];
>> +}
>> +EXPORT_SYMBOL(qcom_scm_tsens_reinit);
>> +
>>   static int __qcom_scm_assign_mem(struct device *dev, phys_addr_t mem_region,
>>   				 size_t mem_sz, phys_addr_t src, size_t src_sz,
>>   				 phys_addr_t dest, size_t dest_sz)
>> diff --git a/drivers/firmware/qcom_scm.h b/drivers/firmware/qcom_scm.h
>> index 0d51eef2472f..495fa00230c7 100644
>> --- a/drivers/firmware/qcom_scm.h
>> +++ b/drivers/firmware/qcom_scm.h
>> @@ -94,6 +94,10 @@ extern int scm_legacy_call(struct device *dev, const struct qcom_scm_desc *desc,
>>   #define QCOM_SCM_PIL_PAS_IS_SUPPORTED	0x07
>>   #define QCOM_SCM_PIL_PAS_MSS_RESET	0x0a
>>   
>> +/* TSENS Services and Function IDs */
>> +#define QCOM_SCM_SVC_TSENS		0x1E
> 
> It would be nice if this 'E' was lowercase.

Sure, will fix this in v4.

> Reviewed-by: Bjorn Andersson <andersson@kernel.org>

Thanks,
Bhupesh

>> +#define QCOM_SCM_TSENS_INIT_ID		0x5
>> +
>>   #define QCOM_SCM_SVC_IO			0x05
>>   #define QCOM_SCM_IO_READ		0x01
>>   #define QCOM_SCM_IO_WRITE		0x02
>> diff --git a/include/linux/qcom_scm.h b/include/linux/qcom_scm.h
>> index f8335644a01a..5c37e1658cef 100644
>> --- a/include/linux/qcom_scm.h
>> +++ b/include/linux/qcom_scm.h
>> @@ -124,4 +124,6 @@ extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val,
>>   extern int qcom_scm_lmh_profile_change(u32 profile_id);
>>   extern bool qcom_scm_lmh_dcvsh_available(void);
>>   
>> +extern int qcom_scm_tsens_reinit(void);
>> +
>>   #endif
>> -- 
>> 2.35.3
>>

Re: [PATCH v3 3/4] thermal: qcom: tsens: Add driver support for re-initialization quirk

[Bhupesh Sharma]

On 8/30/22 3:44 AM, Bjorn Andersson wrote:
> On Thu, Aug 04, 2022 at 11:16:37AM +0530, Bhupesh Sharma wrote:
>> Since for some Qualcomm tsens controllers, its suggested to
>> monitor the controller health periodically and in case an
>> issue is detected, to re-initialize the tsens controller
>> via trustzone, add the support for the same in the
>> qcom tsens driver.
>>
>> Note that once the tsens controller is reset using scm call,
>> all SROT and TM region registers will enter the reset mode.
>>
>> While all the SROT registers will be re-programmed and
>> re-enabled in trustzone prior to the scm call exit, the TM
>> region registers will not re-initialized in trustzone and thus
>> need to be handled by the tsens driver.
>>
>> Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
>> Cc: Amit Kucheria <amitk@kernel.org>
>> Cc: Thara Gopinath <thara.gopinath@gmail.com>
>> Cc: linux-pm@vger.kernel.org
>> Cc: linux-arm-msm@vger.kernel.org
>> Signed-off-by: Bhupesh Sharma <bhupesh.sharma@linaro.org>
>> ---
>>   drivers/thermal/qcom/tsens-v2.c |   3 +
>>   drivers/thermal/qcom/tsens.c    | 197 ++++++++++++++++++++++++++++++++
>>   drivers/thermal/qcom/tsens.h    |  12 ++
>>   3 files changed, 212 insertions(+)
>>
>> diff --git a/drivers/thermal/qcom/tsens-v2.c b/drivers/thermal/qcom/tsens-v2.c
>> index b293ed32174b..f521e4479cc5 100644
>> --- a/drivers/thermal/qcom/tsens-v2.c
>> +++ b/drivers/thermal/qcom/tsens-v2.c
>> @@ -88,6 +88,9 @@ static const struct reg_field tsens_v2_regfields[MAX_REGFIELDS] = {
>>   
>>   	/* TRDY: 1=ready, 0=in progress */
>>   	[TRDY] = REG_FIELD(TM_TRDY_OFF, 0, 0),
>> +
>> +	/* FIRST_ROUND_COMPLETE: 1=complete, 0=not complete */
>> +	[FIRST_ROUND_COMPLETE] = REG_FIELD(TM_TRDY_OFF, 3, 3),
>>   };
>>   
>>   static const struct tsens_ops ops_generic_v2 = {
>> diff --git a/drivers/thermal/qcom/tsens.c b/drivers/thermal/qcom/tsens.c
>> index e49f58e83513..c2d085fb5447 100644
>> --- a/drivers/thermal/qcom/tsens.c
>> +++ b/drivers/thermal/qcom/tsens.c
>> @@ -7,6 +7,7 @@
>>   #include <linux/debugfs.h>
>>   #include <linux/err.h>
>>   #include <linux/io.h>
>> +#include <linux/qcom_scm.h>
>>   #include <linux/module.h>
>>   #include <linux/nvmem-consumer.h>
>>   #include <linux/of.h>
>> @@ -594,6 +595,113 @@ static void tsens_disable_irq(struct tsens_priv *priv)
>>   	regmap_field_write(priv->rf[INT_EN], 0);
>>   }
>>   
>> +static int tsens_reenable_hw_after_scm(struct tsens_priv *priv)
> 
> As written, this is a void function.

Ok.

>> +{
>> +	/*
>> +	 * Re-enable watchdog, unmask the bark and
>> +	 * disable cycle completion monitoring.
>> +	 */
>> +	regmap_field_write(priv->rf[WDOG_BARK_CLEAR], 1);
>> +	regmap_field_write(priv->rf[WDOG_BARK_CLEAR], 0);
>> +	regmap_field_write(priv->rf[WDOG_BARK_MASK], 0);
>> +	regmap_field_write(priv->rf[CC_MON_MASK], 1);
>> +
>> +	/* Re-enable interrupts */
>> +	tsens_enable_irq(priv);
>> +
>> +	return 0;
>> +}
>> +
>> +static int tsens_health_check_and_reinit(struct tsens_priv *priv,
>> +					 int hw_id)
>> +{
>> +	int ret, trdy, first_round, sw_reg;
>> +	unsigned long timeout;
>> +
>> +	/* First check if TRDY is SET */
>> +	ret = regmap_field_read(priv->rf[TRDY], &trdy);
>> +	if (ret)
>> +		goto err;
>> +
>> +	if (!trdy) {
> 
> if (trdy)
> 	return 0;
> 
> Would save you one level of indentation.
> 
>> +		ret = regmap_field_read(priv->rf[FIRST_ROUND_COMPLETE], &first_round);
>> +		if (ret)
>> +			goto err;
>> +
>> +		if (!first_round) {
> 
> if (first_round)
> 	return 0;
> 
> Would save you another level of indentation.
> 
>> +			WARN_ON(!mutex_is_locked(&priv->reinit_mutex));
> 
> At least for now the function is only called within a small locked
> region, so it's going to be locked here. But I'm wondering if there's
> any relationship between the lock state of reinit_mutex and the values
> of TRDY and FIRST_ROUND_COMPLETE.
> 
> Seems like it's possible to hit this function repeatedly and have it
> exit early because of TRDY and FIRST_ROUND_COMPLETE values and then one
> day if will reach here and trip.
> 
> So how about starting the function with this check, to make it more
> likely to be hit in our testing?
> 
>> +
>> +			/* Wait for 2 ms for tsens controller to recover */
>> +			timeout = jiffies + msecs_to_jiffies(RESET_TIMEOUT_MS);
>> +			do {
>> +				ret = regmap_field_read(priv->rf[FIRST_ROUND_COMPLETE],
>> +						&first_round);
>> +				if (ret)
>> +					goto err;
>> +
>> +				if (first_round) {
>> +					dev_dbg(priv->dev, "tsens controller recovered\n");
>> +					return 0; /* success */
>> +				}
>> +			} while (time_before(jiffies, timeout));
> 
> I see no delays in this loop, so we're presumably going to spin here
> tightly for 2ms.
> 
> I think you could write this loop as:
> 
> 	ret = regmap_field_read_poll_timeout(priv->rf[FIRST_ROUND_COMPLETE],
> 					     &first_round, first_round, 100, 2000);
> 	if (ret == 0) {
> 		dev_dbg(priv->dev, "tsens controller recovered\n");
> 		return 0;
> 	}
> 
>> +
>> +			spin_lock(&priv->reinit_lock);
>> +
>> +			/*
>> +			 * Invoke SCM call only if SW register write is
>> +			 * reflecting in controller. Try it for 2 ms.
>> +			 * In case that fails mark the tsens controller
>> +			 * as unrecoverable.
>> +			 */
>> +			timeout = jiffies + msecs_to_jiffies(RESET_TIMEOUT_MS);
>> +			do {
>> +				ret = regmap_field_write(priv->rf[INT_EN], CRITICAL_INT_EN);
>> +				if (ret)
>> +					goto err;
> 
> You're holding reinit_lock here.
> 
>> +
>> +				ret = regmap_field_read(priv->rf[INT_EN], &sw_reg);
>> +				if (ret)
>> +					goto err;
> 
> And here.
> 
>> +			} while ((sw_reg & CRITICAL_INT_EN) && (time_before(jiffies, timeout)));
> 
> And again, this is a tight loop. Please add a usleep_range(100, 1000),
> perhaps inbetween the write and read?
> 
>> +
>> +			if (!(sw_reg & CRITICAL_INT_EN)) {
>> +				ret = -ENOTRECOVERABLE;
>> +				goto err;
> 
> Again, reinit_lock is held here.
> 
>> +			}
>> +
>> +			/*
>> +			 * tsens controller did not recover,
>> +			 * proceed with SCM call to re-init it.
>> +			 */
>> +			ret = qcom_scm_tsens_reinit();
>> +			if (ret) {
>> +				dev_err(priv->dev, "tsens reinit scm call failed (%d)\n", ret);
>> +				goto err;
> 
> And here.
> 
>> +			}
>> +
>> +			/*
>> +			 * After the SCM call, we need to re-enable
>> +			 * the interrupts and also set active threshold
>> +			 * for each sensor.
>> +			 */
>> +			ret = tsens_reenable_hw_after_scm(priv);
> 
> As written tsens_reenable_hw_after_scm() doesn't return any value, so
> skip the error handling.
> 
>> +			if (ret) {
>> +				dev_err(priv->dev,
>> +					"tsens re-enable after scm call failed (%d)\n", ret);
>> +				goto err;
> 
> And here...
> 
>> +			}
>> +
>> +			/* Notify reinit wa worker */
>> +			queue_work(system_highpri_wq, &priv->reinit_wa_notify);
>> +
>> +			spin_unlock(&priv->reinit_lock);
>> +		}
>> +	}
>> +
>> +err:
>> +	return ret;
>> +}
>> +
>>   int get_temp_tsens_valid(const struct tsens_sensor *s, int *temp)
>>   {
>>   	struct tsens_priv *priv = s->priv;
>> @@ -607,6 +715,21 @@ int get_temp_tsens_valid(const struct tsens_sensor *s, int *temp)
>>   	if (tsens_version(priv) == VER_0)
>>   		goto get_temp;
>>   
>> +	/*
>> +	 * For some tsens controllers, its suggested to
>> +	 * monitor the controller health periodically
>> +	 * and in case an issue is detected to reinit
>> +	 * tsens controller via trustzone.
> 
> Please use your 80 chars.

Ack to all the above. Will send a fixed v4 shortly.

Thanks.