Skip to contentSkip to conversation

regulator: debugging and fixing supply deps

PATCH RESEND In review Michał Mirosław 3 participants 10 messages Nov 13, 2020 Voltage and current regulator framework

lore
Share & tools
lore
#0 lore raw

cover regulator: debugging and fixing supply deps

It turns out that commit aea6cb99703e ("regulator: resolve supply after creating regulator") exposed a number of issues in regulator initialization and introduced a memory leak of its own. One uncovered problem was already fixed by cf1ad559a20d ("regulator: defer probe when trying to get voltage from unresolved supply"). This series fixes the remaining ones and adds a two debugging aids to help in the future.

The final patch adds a workaround to preexisting problem occurring with regulators that have the same name as its supply_name. This worked before by accident, so might be worth backporting. The error message is left on purpose so that these configurations can be detected and fixed.

(The first two patches are resends from Nov 5).

(Series resent because of wrong arm-kernel ML address.)

Michał Mirosław (4):
regulator: fix memory leak with repeated set_machine_constraints()
regulator: debug early supply resolving
regulator: avoid resolve_supply() infinite recursion
regulator: workaround self-referent regulators

 drivers/regulator/core.c | 40 ++++++++++++++++++++++++----------------
 1 file changed, 24 insertions(+), 16 deletions(-)
Signature · 1 line
2.20.1
#1 lore raw

PATCH RESEND 1/4 regulator: fix memory leak with repeated set_machine_constraints()

Fixed commit introduced a possible second call to set_machine_constraints() and that allocates memory for rdev->constraints. Move the allocation to the caller so it's easier to manage and done once.

Fixes aea6cb99703e ("regulator: resolve supply after creating regulator")
---
 drivers/regulator/core.c | 29 +++++++++++++----------------
 1 file changed, 13 insertions(+), 16 deletions(-)
drivers/regulator/core.c +13 -16
index 2e1ea18221ef..bcd64ba21fb9 100644
@@ -1315,7 +1315,6 @@ static int _regulator_do_enable(struct regulator_dev *rdev);
 /**
  * set_machine_constraints - sets regulator constraints
  * @rdev: regulator source
- * @constraints: constraints to apply
  *
  * Allows platform initialisation code to define and constrain
  * regulator circuits e.g. valid voltage/current ranges, etc.  NOTE:
@@ -1323,21 +1322,11 @@ static int _regulator_do_enable(struct regulator_dev *rdev);
  * regulator operations to proceed i.e. set_voltage, set_current_limit,
  * set_mode.
  */
-static int set_machine_constraints(struct regulator_dev *rdev,
-	const struct regulation_constraints *constraints)
+static int set_machine_constraints(struct regulator_dev *rdev)
 {
 	int ret = 0;
 	const struct regulator_ops *ops = rdev->desc->ops;
 
-	if (constraints)
-		rdev->constraints = kmemdup(constraints, sizeof(*constraints),
-					    GFP_KERNEL);
-	else
-		rdev->constraints = kzalloc(sizeof(*constraints),
-					    GFP_KERNEL);
-	if (!rdev->constraints)
-		return -ENOMEM;
-
 	ret = machine_constraints_voltage(rdev, rdev->constraints);
 	if (ret != 0)
 		return ret;
@@ -5146,7 +5135,6 @@ struct regulator_dev *
 regulator_register(const struct regulator_desc *regulator_desc,
 		   const struct regulator_config *cfg)
 {
-	const struct regulation_constraints *constraints = NULL;
 	const struct regulator_init_data *init_data;
 	struct regulator_config *config = NULL;
 	static atomic_t regulator_no = ATOMIC_INIT(-1);
@@ -5285,14 +5273,23 @@ regulator_register(const struct regulator_desc *regulator_desc,
 
 	/* set regulator constraints */
 	if (init_data)
-		constraints = &init_data->constraints;
+		rdev->constraints = kmemdup(&init_data->constraints,
+					    sizeof(*rdev->constraints),
+					    GFP_KERNEL);
+	else
+		rdev->constraints = kzalloc(sizeof(*rdev->constraints),
+					    GFP_KERNEL);
+	if (!rdev->constraints) {
+		ret = -ENOMEM;
+		goto wash;
+	}
 
 	if (init_data && init_data->supply_regulator)
 		rdev->supply_name = init_data->supply_regulator;
 	else if (regulator_desc->supply_name)
 		rdev->supply_name = regulator_desc->supply_name;
 
-	ret = set_machine_constraints(rdev, constraints);
+	ret = set_machine_constraints(rdev);
 	if (ret == -EPROBE_DEFER) {
 		/* Regulator might be in bypass mode and so needs its supply
 		 * to set the constraints */
@@ -5301,7 +5298,7 @@ regulator_register(const struct regulator_desc *regulator_desc,
 		 * that is just being created */
 		ret = regulator_resolve_supply(rdev);
 		if (!ret)
-			ret = set_machine_constraints(rdev, constraints);
+			ret = set_machine_constraints(rdev);
 		else
 			rdev_dbg(rdev, "unable to resolve supply early: %pe\n",
 				 ERR_PTR(ret));
Signature · 1 line
2.20.1
#2 lore raw

PATCH RESEND 2/4 regulator: debug early supply resolving

In reply to Michał Mirosław

Help debugging the case when set_machine_constraints() needs to be repeated.

---
 drivers/regulator/core.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/regulator/core.c +2 -0
index bcd64ba21fb9..ad36f03d7ee6 100644
@@ -5296,6 +5296,8 @@ regulator_register(const struct regulator_desc *regulator_desc,
 		/* FIXME: this currently triggers a chicken-and-egg problem
 		 * when creating -SUPPLY symlink in sysfs to a regulator
 		 * that is just being created */
+		rdev_dbg(rdev, "will resolve supply early: %s\n",
+			 rdev->supply_name);
 		ret = regulator_resolve_supply(rdev);
 		if (!ret)
 			ret = set_machine_constraints(rdev);
Signature · 1 line
2.20.1
maintainer
#3 lore raw

Re: [PATCH RESEND 2/4] regulator: debug early supply resolving

Quoted 4 lines · Michał Mirosław

Help debugging the case when set_machine_constraints() needs to be repeated.

Please when sending patches containing a mix of different kinds of patches put any fixes at the start of the series before anything else, including new features and trace additions like this one. This makes it much easier to send things as fixes without unneeded dependencies on the new stuff.

#4 lore raw

PATCH RESEND 3/4 regulator: avoid resolve_supply() infinite recursion

In reply to Michał Mirosław

When a regulator's name equals its supply's name the regulator_resolve_supply() recurses indefinitely. Add a check so that debugging the problem is easier. The "fixed" commit just exposed the problem.

Fixes aea6cb99703e ("regulator: resolve supply after creating regulator")
---
 drivers/regulator/core.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/regulator/core.c +6 -0
index ad36f03d7ee6..ab922ed273f3 100644
@@ -1841,6 +1841,12 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 		}
 	}
 
+	if (r == rdev) {
+		dev_err(dev, "Supply for %s (%s) resolved to itself\n",
+			rdev->desc->name, rdev->supply_name);
+		return -EINVAL;
+	}
+
 	/*
 	 * If the supply's parent device is not the same as the
 	 * regulator's parent device, then ensure the parent device
Signature · 1 line
2.20.1
#5 lore raw

PATCH RESEND 4/4 regulator: workaround self-referent regulators

In reply to Michał Mirosław

Workaround regulators whose supply name happens to be the same as its own name. This fixes boards that used to work before the early supply resolving was removed. The error message is left in place so that offending drivers can be detected.

Fixes aea6cb99703e ("regulator: resolve supply after creating regulator")
---
 drivers/regulator/core.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/regulator/core.c +4 -1
index ab922ed273f3..38ba579efe2b 100644
@@ -1844,7 +1844,10 @@ static int regulator_resolve_supply(struct regulator_dev *rdev)
 	if (r == rdev) {
 		dev_err(dev, "Supply for %s (%s) resolved to itself\n",
 			rdev->desc->name, rdev->supply_name);
-		return -EINVAL;
+		if (!have_full_constraints())
+			return -EINVAL;
+		r = dummy_regulator_rdev;
+		get_device(&r->dev);
 	}
 
 	/*
Signature · 1 line
2.20.1
Tested-by on the v1 series
#6 lore raw
In reply to Michał Mirosław

Hello Michał,

Quoted 7 lines · Michał Mirosław

It turns out that commit aea6cb99703e ("regulator: resolve supply after creating regulator") exposed a number of issues in regulator initialization and introduced a memory leak of its own. One uncovered problem was already fixed by cf1ad559a20d ("regulator: defer probe when trying to get voltage from unresolved supply"). This series fixes the remaining ones and adds a two debugging aids to help in the future.

Quoted 8 lines

The final patch adds a workaround to preexisting problem occurring with regulators that have the same name as its supply_name. This worked before by accident, so might be worth backporting. The error message is left on purpose so that these configurations can be detected and fixed.

(The first two patches are resends from Nov 5).

(Series resent because of wrong arm-kernel ML address.)

lxa-mc1 (STM32MP1 board with STPMIC) now manages to boot again with following new warning:

stpmic1-regulator 5c002000.i2c:stpmic@33:regulators: Supply for VREF_DDR

So for the whole series,

Thanks!
Ahmad

Quoted 10 lines

Michał Mirosław (4):
regulator: fix memory leak with repeated set_machine_constraints()
regulator: debug early supply resolving
regulator: avoid resolve_supply() infinite recursion
regulator: workaround self-referent regulators

 drivers/regulator/core.c | 40 ++++++++++++++++++++++++----------------
 1 file changed, 24 insertions(+), 16 deletions(-)
Signature · 4 lines
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
#7 lore raw

CC += linux-stm32 as other STM32MP1 boards are also affected.

Quoted 41 lines · Ahmad Fatoum

Hello Michał,

Quoted 7 lines · Michał Mirosław

It turns out that commit aea6cb99703e ("regulator: resolve supply after creating regulator") exposed a number of issues in regulator initialization and introduced a memory leak of its own. One uncovered problem was already fixed by cf1ad559a20d ("regulator: defer probe when trying to get voltage from unresolved supply"). This series fixes the remaining ones and adds a two debugging aids to help in the future.

Quoted 8 lines

The final patch adds a workaround to preexisting problem occurring with regulators that have the same name as its supply_name. This worked before by accident, so might be worth backporting. The error message is left on purpose so that these configurations can be detected and fixed.

(The first two patches are resends from Nov 5).

(Series resent because of wrong arm-kernel ML address.)

lxa-mc1 (STM32MP1 board with STPMIC) now manages to boot again with following new warning:

stpmic1-regulator 5c002000.i2c:stpmic@33:regulators: Supply for VREF_DDR

So for the whole series,

Thanks!
Ahmad

Quoted 10 lines

Michał Mirosław (4):
regulator: fix memory leak with repeated set_machine_constraints()
regulator: debug early supply resolving
regulator: avoid resolve_supply() infinite recursion
regulator: workaround self-referent regulators

 drivers/regulator/core.c | 40 ++++++++++++++++++++++++----------------
 1 file changed, 24 insertions(+), 16 deletions(-)
Signature · 9 lines
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
maintainer
#8 lore raw
In reply to Michał Mirosław
Quoted 8 lines · Michał Mirosław

It turns out that commit aea6cb99703e ("regulator: resolve supply after creating regulator") exposed a number of issues in regulator initialization and introduced a memory leak of its own. One uncovered problem was already fixed by cf1ad559a20d ("regulator: defer probe when trying to get voltage from unresolved supply"). This series fixes the remaining ones and adds a two debugging aids to help in the future.

[...]

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator.git for-next

Thanks!

[1/3] regulator: fix memory leak with repeated set_machine_constraints()

      commit: 57a6ad482af256b2a13de14194fb8f67c1a65f10
[2/3] regulator: avoid resolve_supply() infinite recursion
      commit: 4b639e254d3d4f15ee4ff2b890a447204cfbeea9
[3/3] regulator: workaround self-referent regulators
      commit: f5c042b23f7429e5c2ac987b01a31c69059a978b

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

maintainer
#9 lore raw
In reply to Michał Mirosław
Quoted 8 lines · Michał Mirosław

It turns out that commit aea6cb99703e ("regulator: resolve supply after creating regulator") exposed a number of issues in regulator initialization and introduced a memory leak of its own. One uncovered problem was already fixed by cf1ad559a20d ("regulator: defer probe when trying to get voltage from unresolved supply"). This series fixes the remaining ones and adds a two debugging aids to help in the future.

[...]

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator.git for-next

Thanks!

[1/1] regulator: debug early supply resolving

      commit: 0917c9db23accb8690d8f1987ada36eb5b1a5ac9

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel