[PATCH 01/10] regulator: irq_helpers: Clear caller handle in regulator_irq_helper_cancel()

[PATCH 01/10] regulator: irq_helpers: Clear caller handle in regulator_irq_helper_cancel()

[Lee Jones]

regulator_irq_helper_cancel() is documented as clearing the caller's
handle pointer upon return, and takes a void **handle parameter for
that purpose.  Instead, it assigns NULL to the local pointer h before
returning, leaving the caller with a dangling pointer to freed memory.

Dereference handle to clear the caller's pointer as intended.

Fixes: 7111c6d1b31b ("regulator: IRQ based event/error notification helpers")
Closes: https://sashiko.dev/#/bug/linux-5c3d3b21-3823-4654-bf5d-85407c5fd05f
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/irq_helpers.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/regulator/irq_helpers.c b/drivers/regulator/irq_helpers.c
index cdbd170d2b15..7b25eae667e3 100644
--- a/drivers/regulator/irq_helpers.c
+++ b/drivers/regulator/irq_helpers.c
@@ -403,7 +403,7 @@ void regulator_irq_helper_cancel(void **handle)
 				destroy_workqueue(h->wq);
 		}
 
-		h = NULL;
+		*handle = NULL;
 	}
 }
 EXPORT_SYMBOL_GPL(regulator_irq_helper_cancel);
-- 
2.56.0.385.gac671c819a-goog

[PATCH 02/10] regulator: arizona-ldo1: Avoid double gpiod_put() on enable GPIO

[Lee Jones]

Passing config.ena_gpiod to devm_regulator_register() transfers ownership
of the GPIO descriptor to the regulator core, which releases it via
gpiod_put() in regulator_ena_gpio_free() upon registration failure or
when devres unregisters the regulator.

Calling gpiod_put(ldo1->ena_gpiod) in arizona_ldo1_remove() prematurely
frees the enable GPIO while the regulator is still registered and causes
a second gpiod_put() when devm_rdev_release() runs.  Drop the redundant
remove callback and ena_gpiod field so the regulator core manages the
descriptor lifetime exclusively.

Fixes: a9191579ba10 ("regulator: arizona-ldo1: Use correct device to get enable GPIO")
Closes: https://sashiko.dev/#/bug/linux-7988402e-ca08-4bd3-9488-5779d6da0450
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/arizona-ldo1.c | 16 ----------------
 1 file changed, 16 deletions(-)

diff --git a/drivers/regulator/arizona-ldo1.c b/drivers/regulator/arizona-ldo1.c
index 501843996faa..981e047167ea 100644
--- a/drivers/regulator/arizona-ldo1.c
+++ b/drivers/regulator/arizona-ldo1.c
@@ -35,8 +35,6 @@ struct arizona_ldo1 {
 
 	struct regulator_consumer_supply supply;
 	struct regulator_init_data init_data;
-
-	struct gpio_desc *ena_gpiod;
 };
 
 static int arizona_ldo1_hc_set_voltage_sel(struct regulator_dev *rdev,
@@ -262,8 +260,6 @@ static int arizona_ldo1_common_init(struct platform_device *pdev,
 	if (IS_ERR(config.ena_gpiod))
 		return PTR_ERR(config.ena_gpiod);
 
-	ldo1->ena_gpiod = config.ena_gpiod;
-
 	if (pdata->init_data)
 		config.init_data = pdata->init_data;
 	else
@@ -287,8 +283,6 @@ static int arizona_ldo1_common_init(struct platform_device *pdev,
 		return ret;
 	}
 
-	platform_set_drvdata(pdev, ldo1);
-
 	return 0;
 }
 
@@ -339,14 +333,6 @@ static int arizona_ldo1_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static void arizona_ldo1_remove(struct platform_device *pdev)
-{
-	struct arizona_ldo1 *ldo1 = platform_get_drvdata(pdev);
-
-	if (ldo1->ena_gpiod)
-		gpiod_put(ldo1->ena_gpiod);
-}
-
 static int madera_ldo1_probe(struct platform_device *pdev)
 {
 	struct madera *madera = dev_get_drvdata(pdev->dev.parent);
@@ -375,7 +361,6 @@ static int madera_ldo1_probe(struct platform_device *pdev)
 
 static struct platform_driver arizona_ldo1_driver = {
 	.probe = arizona_ldo1_probe,
-	.remove = arizona_ldo1_remove,
 	.driver = {
 		.name = "arizona-ldo1",
 		.probe_type = PROBE_FORCE_SYNCHRONOUS,
@@ -384,7 +369,6 @@ static struct platform_driver arizona_ldo1_driver = {
 
 static struct platform_driver madera_ldo1_driver = {
 	.probe = madera_ldo1_probe,
-	.remove = arizona_ldo1_remove,
 	.driver = {
 		.name = "madera-ldo1",
 		.probe_type = PROBE_FORCE_SYNCHRONOUS,
-- 
2.56.0.385.gac671c819a-goog

[PATCH 03/10] regulator: core: Flush delayed disable_work before releasing supply in regulator_unregister()

[Lee Jones]

regulator_unregister() currently calls flush_work() on
rdev->disable_work.work after tearing down and freeing rdev->supply.
This has two bugs:

1. Because rdev->disable_work is a delayed_work, flush_work() only waits
   for work that has already been queued and ignores an armed timer,
   allowing delayed_work_timer_fn() to fire after device_unregister()
   has freed rdev.
2. regulator_disable_work() locks and disables rdev->supply via
   regulator_lock_dependent() and _regulator_disable(), so releasing
   rdev->supply before flushing disable_work races with any pending or
   running deferred disable work item.

Use flush_delayed_work() at the start of regulator_unregister() so any
deferred disable work completes before rdev->supply is torn down.

Fixes: da07ecd93b19 ("regulator: Implement deferred disable support")
Fixes: e032b376551a ("regulator: Fix deadlock on removal of regulators with supplies")
Closes: https://sashiko.dev/#/bug/linux-6ee97926-52f2-4842-ba55-b554f9b39d53
Closes: https://sashiko.dev/#/bug/linux-aac296d1-789a-469b-8bde-8e110c4447ea
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 14422cfe80e2..935fee27f126 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -6299,6 +6299,8 @@ void regulator_unregister(struct regulator_dev *rdev)
 	if (rdev == NULL)
 		return;
 
+	flush_delayed_work(&rdev->disable_work);
+
 	if (rdev->supply) {
 		regulator_unregister_notifier(rdev->supply,
 					      &rdev->supply_fwd_nb);
@@ -6308,8 +6310,6 @@ void regulator_unregister(struct regulator_dev *rdev)
 		regulator_put(rdev->supply);
 	}
 
-	flush_work(&rdev->disable_work.work);
-
 	mutex_lock(&regulator_list_mutex);
 
 	WARN_ON(rdev->open_count);
-- 
2.56.0.385.gac671c819a-goog

[PATCH 04/10] regulator: arizona-micsupp: Cancel check_cp_work on device teardown

[Lee Jones]

micsupp->check_cp_work is initialised with raw INIT_WORK() and is never
cancelled when the regulator device is unbound or when registration
fails after set_machine_constraints() has run, allowing a pending work
item to dereference micsupp after devres frees it.

Initialise check_cp_work with devm_work_autocancel() before registering
the regulator so the work item is cancelled after regulator teardown and
before micsupp is freed.

Fixes: e6ed90581bfa ("regulator: arizona-micsupp: Enable SYSCLK for charge pump")
Closes: https://sashiko.dev/#/bug/linux-3f85eadb-7493-4aea-9e4d-ce319b6e0688
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/arizona-micsupp.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/regulator/arizona-micsupp.c b/drivers/regulator/arizona-micsupp.c
index be7208bc7409..70b939d3e704 100644
--- a/drivers/regulator/arizona-micsupp.c
+++ b/drivers/regulator/arizona-micsupp.c
@@ -10,6 +10,7 @@
 #include <linux/moduleparam.h>
 #include <linux/init.h>
 #include <linux/bitops.h>
+#include <linux/devm-helpers.h>
 #include <linux/err.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
@@ -255,7 +256,10 @@ static int arizona_micsupp_common_init(struct platform_device *pdev,
 	struct regulator_config config = { };
 	int ret;
 
-	INIT_WORK(&micsupp->check_cp_work, arizona_micsupp_check_cp);
+	ret = devm_work_autocancel(&pdev->dev, &micsupp->check_cp_work,
+				   arizona_micsupp_check_cp);
+	if (ret)
+		return ret;
 
 	micsupp->init_data.consumer_supplies = &micsupp->supply;
 	micsupp->supply.dev_name = dev_name(micsupp->dev);
-- 
2.56.0.385.gac671c819a-goog

[PATCH 05/10] regulator: core: Hold supply device reference until constraints are resolved in regulator_resolve_supply()

[Lee Jones]

In regulator_resolve_supply(), set_supply() previously consumed the
caller's lookup reference on &r->dev without taking its own reference
for rdev->supply, while the concurrent path (when rdev->supply was
already set) dropped put_device(&r->dev) immediately after unlocking
and before retrying set_machine_constraints().  In both paths, if
set_machine_constraints() fails (which may drop rdev->supply inside
set_machine_constraints() or race with a concurrent task in
unset_supply), the error path subsequently passes r to
regulator_lock_two(rdev, r, &ww_ctx) without holding a guaranteed local
reference on &r->dev.

Take a dedicated device reference for rdev->supply in set_supply() to
match _regulator_put(), and retain the local lookup reference on &r->dev
throughout regulator_resolve_supply() until returning.

Fixes: 8d38423d9dea ("regulator: core: don't fail regulator_register() with missing required supply")
Closes: https://sashiko.dev/#/bug/linux-3a22ff58-1a66-4a6a-9e19-83135ef603e0
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/core.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 935fee27f126..6d522731e4d9 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -1842,6 +1842,7 @@ static int set_supply(struct regulator_dev *rdev,
 		err = -ENOMEM;
 		return err;
 	}
+	get_device(&supply_rdev->dev);
 	supply_rdev->open_count++;
 
 	return 0;
@@ -2290,22 +2291,20 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 		supply = rdev->supply;
 
 		regulator_unlock_two(rdev, r, &ww_ctx);
-		put_device(&r->dev);
 
 		/*
 		 * Supply resolved by concurrent task, and constraints set as
 		 * well (or not required): fast path.
 		 */
 		if (done)
-			goto out;
+			goto out_put;
 
 		do_final_setup = false;
 	} else {
 		ret = set_supply(rdev, r);
 		if (ret < 0) {
 			regulator_unlock_two(rdev, r, &ww_ctx);
-			put_device(&r->dev);
-			goto out;
+			goto out_put;
 		}
 
 		supply = rdev->supply;
@@ -2342,7 +2341,7 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 			 */
 			regulator_unlock_dependent(rdev, &ww_ctx);
 			ret = -EPROBE_DEFER;
-			goto out;
+			goto out_put;
 		}
 
 		ret = set_machine_constraints(rdev, true);
@@ -2364,7 +2363,7 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 				 * that again via unset_supply.
 				 */
 				regulator_unlock_two(rdev, r, &ww_ctx);
-				goto out;
+				goto out_put;
 			}
 
 			unregister_regulator_event_forwarding(rdev);
@@ -2376,11 +2375,13 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 	regulator_unlock_dependent(rdev, &ww_ctx);
 
 	if (!do_final_setup)
-		goto out;
+		goto out_put;
 
 	/* rdev->supply was created in set_supply() */
 	link_and_create_debugfs(rdev->supply, rdev->supply->rdev, &rdev->dev);
 
+out_put:
+	put_device(&r->dev);
 out:
 	return ret;
 
@@ -2391,6 +2392,7 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 	regulator_unlock_two(rdev, supply->rdev, &ww_ctx);
 
 	regulator_put(supply);
+	put_device(&r->dev);
 
 	return ret;
 }
-- 
2.56.0.385.gac671c819a-goog

[PATCH 06/10] regulator: core: Use ktime_get_boottime() in set_machine_constraints()

[Lee Jones]

rdev->last_off is tracked against CLOCK_BOOTTIME in both
_regulator_do_enable() and _regulator_do_disable(), whereas
set_machine_constraints() initialises rdev->last_off using ktime_get()
(CLOCK_MONOTONIC).  If the system suspends before a regulator with
off_on_delay is enabled, comparing a CLOCK_BOOTTIME timestamp against a
CLOCK_MONOTONIC timestamp produces an incorrect elapsed off-time.

Use ktime_get_boottime() in set_machine_constraints() so rdev->last_off
uses the same clock base everywhere.

Fixes: 691c1fcda535 ("regulator: core: Shorten off-on-delay-us for always-on/boot-on by time since booted")
Closes: https://sashiko.dev/#/bug/linux-4afd38c3-0589-4544-961d-4e3cc5ed8639
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 6d522731e4d9..6b5fae3cd486 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -1698,7 +1698,7 @@ static int set_machine_constraints(struct regulator_dev *rdev,
 		if (rdev->constraints->always_on)
 			rdev->use_count++;
 	} else if (rdev->desc->off_on_delay) {
-		rdev->last_off = ktime_get();
+		rdev->last_off = ktime_get_boottime();
 	}
 
 	if (!rdev->constraints->pw_budget_mW)
-- 
2.56.0.385.gac671c819a-goog

[PATCH 07/10] regulator: bq257xx: Propagate GPIO error in bq257xx_reg_dt_parse_gpio()

[Lee Jones]

When devm_fwnode_gpiod_get_index() fails with an error other than
-ENOENT (such as -EPROBE_DEFER), bq257xx_reg_dt_parse_gpio() logs the
error and returns void, leaving the error pointer stored in
pdata->otg_en_gpio.  bq257xx_regulator_probe() then registers the
regulator and bq25703_vbus_enable() / bq25703_vbus_disable() pass the
ERR_PTR to gpiod_set_value_cansleep().

Return the error code from bq257xx_reg_dt_parse_gpio() via
dev_err_probe() and propagate it from bq257xx_regulator_probe().

Fixes: 981dd162b635 ("regulator: bq257xx: Add bq257xx boost regulator driver")
Closes: https://sashiko.dev/#/bug/linux-0d1fc150-1400-4ba5-ad27-3d7376fb2d7d
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/bq257xx-regulator.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/drivers/regulator/bq257xx-regulator.c b/drivers/regulator/bq257xx-regulator.c
index 577b277efd7f..0db99a74e9e8 100644
--- a/drivers/regulator/bq257xx-regulator.c
+++ b/drivers/regulator/bq257xx-regulator.c
@@ -103,7 +103,7 @@ static const struct regulator_desc bq25703_vbus_desc = {
 };
 
 /* Get optional GPIO for OTG regulator enable. */
-static void bq257xx_reg_dt_parse_gpio(struct platform_device *pdev)
+static int bq257xx_reg_dt_parse_gpio(struct platform_device *pdev)
 {
 	struct device_node *child, *subchild;
 	struct bq257xx_reg_data *pdata = platform_get_drvdata(pdev);
@@ -111,12 +111,12 @@ static void bq257xx_reg_dt_parse_gpio(struct platform_device *pdev)
 	child = of_get_child_by_name(pdev->dev.of_node,
 				     pdata->desc.regulators_node);
 	if (!child)
-		return;
+		return 0;
 
 	subchild = of_get_child_by_name(child, pdata->desc.of_match);
 	of_node_put(child);
 	if (!subchild)
-		return;
+		return 0;
 
 	pdata->otg_en_gpio = devm_fwnode_gpiod_get_index(&pdev->dev,
 							 of_fwnode_handle(subchild),
@@ -130,13 +130,14 @@ static void bq257xx_reg_dt_parse_gpio(struct platform_device *pdev)
 		if (PTR_ERR(pdata->otg_en_gpio) == -ENOENT) {
 			/* No GPIO, will only use register writes for OTG */
 			pdata->otg_en_gpio = NULL;
-			return;
+			return 0;
 		}
 
-		dev_err(&pdev->dev, "Error getting enable gpio: %ld\n",
-			PTR_ERR(pdata->otg_en_gpio));
-		return;
+		return dev_err_probe(&pdev->dev, PTR_ERR(pdata->otg_en_gpio),
+				     "Error getting enable gpio\n");
 	}
+
+	return 0;
 }
 
 static int bq257xx_regulator_probe(struct platform_device *pdev)
@@ -144,6 +145,7 @@ static int bq257xx_regulator_probe(struct platform_device *pdev)
 	struct device *dev = &pdev->dev;
 	struct bq257xx_reg_data *pdata;
 	struct regulator_config cfg = {};
+	int ret;
 
 	device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
 
@@ -154,7 +156,9 @@ static int bq257xx_regulator_probe(struct platform_device *pdev)
 	pdata->desc = bq25703_vbus_desc;
 
 	platform_set_drvdata(pdev, pdata);
-	bq257xx_reg_dt_parse_gpio(pdev);
+	ret = bq257xx_reg_dt_parse_gpio(pdev);
+	if (ret)
+		return ret;
 
 	cfg.dev = &pdev->dev;
 	cfg.driver_data = pdata;
-- 
2.56.0.385.gac671c819a-goog

[PATCH 08/10] regulator: tps6594-regulator: Fix LDO4 n_voltages in tps6594_ldo_regs

[Lee Jones]

Unlike TPS6594_MASK_LDO123_VSET (GENMASK(6, 1)), TPS6594_MASK_LDO4_VSET
is GENMASK(6, 0) (0x7f) and uses an unshifted 7-bit selector field,
with ldos_4_ranges defining valid voltage selectors from 0x20 up to
0x74 (116) in 25mV steps up to 3.3V.  In tps6594_ldo_regs, LDO4
erroneously passes TPS6594_MASK_LDO4_VSET >> 1 (63) for n_voltages, so
the regulator core rejects any selector >= 63 and leaves selectors 0x3f
through 0x74 (1.975V to 3.3V) unreachable.

Pass TPS6594_MASK_LDO4_VSET for LDO4 n_voltages to allow the full LDO4
voltage range.

Fixes: f17ccc5deb4d ("regulator: tps6594-regulator: Add driver for TI TPS6594 regulators")
Closes: https://sashiko.dev/#/bug/linux-c680c361-7684-48a1-96f7-0eae5c8477d5
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/tps6594-regulator.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/regulator/tps6594-regulator.c b/drivers/regulator/tps6594-regulator.c
index 503b89a2f5f8..fa4d700252e6 100644
--- a/drivers/regulator/tps6594-regulator.c
+++ b/drivers/regulator/tps6594-regulator.c
@@ -457,7 +457,7 @@ static const struct regulator_desc tps6594_ldo_regs[] = {
 			  TPS6594_BIT_LDO_EN, 0, 0, ldos_1_2_3_ranges,
 			  1, 0, 0, NULL, 0, TPS6594_BIT_LDO_BYPASS),
 	TPS6594_REGULATOR("LDO4", "ldo4", TPS6594_LDO_4,
-			  REGULATOR_VOLTAGE, tps6594_ldos_4_ops, TPS6594_MASK_LDO4_VSET >> 1,
+			  REGULATOR_VOLTAGE, tps6594_ldos_4_ops, TPS6594_MASK_LDO4_VSET,
 			  TPS6594_REG_LDOX_VOUT(3),
 			  TPS6594_MASK_LDO4_VSET,
 			  TPS6594_REG_LDOX_CTRL(3),
-- 
2.56.0.385.gac671c819a-goog

[PATCH 09/10] regulator: max8907: Avoid NULL idata dereference in max8907_regulator_probe()

[Lee Jones]

When a board does not provide platform data or a Device Tree node for
BBAT, SDBY, or VRTC, idata is NULL in max8907_regulator_probe().  Unlike
the MAX8907_MBATT case above it, the BBAT/SDBY/VRTC case unconditionally
dereferences idata to assign idata->supply_regulator, causing a NULL
pointer dereference during probe.

Check that idata is non-NULL before dereferencing it.

Fixes: ffee19091a4e ("regulator: add MAX8907 driver")
Fixes: 5fc72f57e62e ("regulator: max8907: fix use of possibly NULL idata")
Closes: https://sashiko.dev/#/bug/linux-372cd4b4-65df-47bd-8b31-5a7b416b2f91
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/max8907-regulator.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/regulator/max8907-regulator.c b/drivers/regulator/max8907-regulator.c
index e59aa7a3ee54..ee7d9b7ef7fc 100644
--- a/drivers/regulator/max8907-regulator.c
+++ b/drivers/regulator/max8907-regulator.c
@@ -331,7 +331,8 @@ static int max8907_regulator_probe(struct platform_device *pdev)
 		case MAX8907_BBAT:
 		case MAX8907_SDBY:
 		case MAX8907_VRTC:
-			idata->supply_regulator = mbatt_rail_name;
+			if (idata)
+				idata->supply_regulator = mbatt_rail_name;
 			break;
 		}
 
-- 
2.56.0.385.gac671c819a-goog

[PATCH 10/10] regulator: rt6190: Check pm_runtime_resume_and_get() return value in rt6190_out_enable()

[Lee Jones]

rt6190_out_enable() ignores the return value of pm_runtime_get_sync()
and proceeds with register accesses even if runtime resume fails.

Use pm_runtime_resume_and_get() and return the error code on failure so
the usage counter is not left elevated and registers are not accessed
while the device is suspended.

Fixes: e6999e7cca7e ("regulator: rt6190: Add support for Richtek RT6190 regulator")
Closes: https://sashiko.dev/#/bug/linux-dcfab9b2-550f-4319-87a5-682c3f6cef75
Cc: stable@vger.kernel.org
Signed-off-by: Lee Jones <lee@kernel.org>
---
 drivers/regulator/rt6190-regulator.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/regulator/rt6190-regulator.c b/drivers/regulator/rt6190-regulator.c
index 21ced348a2cf..973335a9178d 100644
--- a/drivers/regulator/rt6190-regulator.c
+++ b/drivers/regulator/rt6190-regulator.c
@@ -99,7 +99,9 @@ static int rt6190_out_enable(struct regulator_dev *rdev)
 	u8 out_cfg[4];
 	int ret;
 
-	pm_runtime_get_sync(data->dev);
+	ret = pm_runtime_resume_and_get(data->dev);
+	if (ret)
+		return ret;
 
 	/*
 	 * From off to on, vout config will restore to IC default.
-- 
2.56.0.385.gac671c819a-goog

Re: [PATCH 01/10] regulator: irq_helpers: Clear caller handle in regulator_irq_helper_cancel()

[Matti Vaittinen]

On 04/11/2026 20:20, Lee Jones wrote:
> regulator_irq_helper_cancel() is documented as clearing the caller's
> handle pointer upon return, and takes a void **handle parameter for
> that purpose.  Instead, it assigns NULL to the local pointer h before
> returning, leaving the caller with a dangling pointer to freed memory.
> 
> Dereference handle to clear the caller's pointer as intended.
> 
> Fixes: 7111c6d1b31b ("regulator: IRQ based event/error notification helpers")
> Closes: https://sashiko.dev/#/bug/linux-5c3d3b21-3823-4654-bf5d-85407c5fd05f
> Cc: stable@vger.kernel.org
> Signed-off-by: Lee Jones <lee@kernel.org>

Good catch! Thanks a lot :)

Reviewed-by: Matti Vaittinen <mazziesaccount@gmail.com>

> ---
>   drivers/regulator/irq_helpers.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/regulator/irq_helpers.c b/drivers/regulator/irq_helpers.c
> index cdbd170d2b15..7b25eae667e3 100644
> --- a/drivers/regulator/irq_helpers.c
> +++ b/drivers/regulator/irq_helpers.c
> @@ -403,7 +403,7 @@ void regulator_irq_helper_cancel(void **handle)
>   				destroy_workqueue(h->wq);
>   		}
>   
> -		h = NULL;
> +		*handle = NULL;
>   	}
>   }
>   EXPORT_SYMBOL_GPL(regulator_irq_helper_cancel);


-- 
Matti Vaittinen
Linux kernel developer at ROHM Semiconductors
Oulu Finland

~~ When things go utterly wrong vim users can always type :help! ~~

Re: [PATCH 02/10] regulator: arizona-ldo1: Avoid double gpiod_put() on enable GPIO

[Charles Keepax]

On Wed, Nov 04, 2026 at 06:20:52PM +0000, Lee Jones wrote:
> Passing config.ena_gpiod to devm_regulator_register() transfers ownership
> of the GPIO descriptor to the regulator core, which releases it via
> gpiod_put() in regulator_ena_gpio_free() upon registration failure or
> when devres unregisters the regulator.
> 
> Calling gpiod_put(ldo1->ena_gpiod) in arizona_ldo1_remove() prematurely
> frees the enable GPIO while the regulator is still registered and causes
> a second gpiod_put() when devm_rdev_release() runs.  Drop the redundant
> remove callback and ena_gpiod field so the regulator core manages the
> descriptor lifetime exclusively.
> 
> Fixes: a9191579ba10 ("regulator: arizona-ldo1: Use correct device to get enable GPIO")
> Closes: https://sashiko.dev/#/bug/linux-7988402e-ca08-4bd3-9488-5779d6da0450
> Cc: stable@vger.kernel.org
> Signed-off-by: Lee Jones <lee@kernel.org>
> ---

Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>

Thanks,
Charles

Re: [PATCH 04/10] regulator: arizona-micsupp: Cancel check_cp_work on device teardown

[Charles Keepax]

On Wed, Nov 04, 2026 at 06:20:54PM +0000, Lee Jones wrote:
> micsupp->check_cp_work is initialised with raw INIT_WORK() and is never
> cancelled when the regulator device is unbound or when registration
> fails after set_machine_constraints() has run, allowing a pending work
> item to dereference micsupp after devres frees it.
> 
> Initialise check_cp_work with devm_work_autocancel() before registering
> the regulator so the work item is cancelled after regulator teardown and
> before micsupp is freed.
> 
> Fixes: e6ed90581bfa ("regulator: arizona-micsupp: Enable SYSCLK for charge pump")
> Closes: https://sashiko.dev/#/bug/linux-3f85eadb-7493-4aea-9e4d-ce319b6e0688
> Cc: stable@vger.kernel.org
> Signed-off-by: Lee Jones <lee@kernel.org>
> ---

Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>

Thanks,
Charles

Re: [PATCH 05/10] regulator: core: Hold supply device reference until constraints are resolved in regulator_resolve_supply()

[André Draszik]

Hi Lee,

Overall looks good, thanks for catching this. A few small comments
below:

On Wed, 2026-11-04 at 18:20 +0000, Lee Jones wrote:
> In regulator_resolve_supply(), set_supply() previously consumed the
> caller's lookup reference on &r->dev without taking its own reference
> for rdev->supply, while the concurrent path (when rdev->supply was
> already set) dropped put_device(&r->dev) immediately after unlocking
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

I think
"called put_device(&r->dev) immediately" or
"dropped its reference immediately" reads a little bit better.


> diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
> index 935fee27f126..6d522731e4d9 100644
> --- a/drivers/regulator/core.c
> +++ b/drivers/regulator/core.c
[...]
> @@ -2290,22 +2291,20 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
>  		supply = rdev->supply;
>  
>  		regulator_unlock_two(rdev, r, &ww_ctx);
> -		put_device(&r->dev);
>  
>  		/*
>  		 * Supply resolved by concurrent task, and constraints set as
>  		 * well (or not required): fast path.
>  		 */
>  		if (done)
> -			goto out;
> +			goto out_put;
>  
>  		do_final_setup = false;
>  	} else {
>  		ret = set_supply(rdev, r);
>  		if (ret < 0) {
>  			regulator_unlock_two(rdev, r, &ww_ctx);
> -			put_device(&r->dev);
> -			goto out;
> +			goto out_put;
>  		}

There are three locations in this file that still do
'put_device(&r->dev); goto out;' in this function and could be updated
as well:

* the '!device_is_bound(r->dev.parent) exit' around line 2264
* the exit after regulator_resolve_supply(r) fails around line 2275
* the 'if (r == rdev) {' branch around line 2242 (with a little massaging)

Then there'd be just one place that releases the lookup reference.

Cheers,
Andre'