[PATCH 4.9 63/94] regulator: fix crash caused by null driver data

From: Greg Kroah-Hartman
Date: Tue Oct 02 2018 - 09:36:38 EST


4.9-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yu Zhao <yuzhao@xxxxxxxxxx>

commit fb6de923ca3358a91525552b4907d4cb38730bdd upstream.

dev_set_drvdata() needs to be called before device_register()
exposes device to userspace. Otherwise kernel crashes after it
gets null pointer from dev_get_drvdata() when userspace tries
to access sysfs entries.

[Removed backtrace for length -- broonie]

Signed-off-by: Yu Zhao <yuzhao@xxxxxxxxxx>
Signed-off-by: Mark Brown <broonie@xxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
drivers/regulator/core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -4054,13 +4054,13 @@ regulator_register(const struct regulato
!rdev->desc->fixed_uV)
rdev->is_switch = true;

+ dev_set_drvdata(&rdev->dev, rdev);
ret = device_register(&rdev->dev);
if (ret != 0) {
put_device(&rdev->dev);
goto unset_supplies;
}

- dev_set_drvdata(&rdev->dev, rdev);
rdev_init_debugfs(rdev);

/* try to resolve regulators supply since a new one was registered */