Skip to contentSkip to conversation

platform/x86: surface3_power: i2c_acpi_new_device() returns a PTR_ERR

PATCH In review Hans de Goede 2 participants 3 messages Apr 20, 2020 – Apr 21, 2020 x86 platform drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 platform/x86: surface3_power: i2c_acpi_new_device() returns a PTR_ERR

i2c_acpi_new_device() never returns NULL, it either returns an i2c_client or a PTR_ERR. Adjust the mshw0011_probe() error handling to take this into account.

Note the goto out_err will cause i2c_unregister_device() to get called even though the i2c_acpi_new_device() fails, this is ok as it accepts a NULL pointer argument (and treats it as a no-op).

---
 drivers/platform/x86/surface3_power.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/platform/x86/surface3_power.c +4 -2
index 946ac2dc08ae..32e6e86e27dd 100644
@@ -522,8 +522,10 @@ static int mshw0011_probe(struct i2c_client *client)
 	strlcpy(board_info.type, "MSHW0011-bat0", I2C_NAME_SIZE);
 
 	bat0 = i2c_acpi_new_device(dev, 1, &board_info);
-	if (!bat0)
-		return -ENOMEM;
+	if (IS_ERR(bat0)) {
+		error = PTR_ERR(bat0);
+		goto out_err;
+	}
 
 	data->bat0 = bat0;
 	i2c_set_clientdata(bat0, data);
Signature · 1 line
2.26.0
#1 lore raw
Quoted 9 lines · Hans de Goede

i2c_acpi_new_device() never returns NULL, it either returns an i2c_client or a PTR_ERR. Adjust the mshw0011_probe() error handling to take this into account.

Note the goto out_err will cause i2c_unregister_device() to get called even though the i2c_acpi_new_device() fails, this is ok as it accepts a NULL pointer argument (and treats it as a no-op).

Thanks, I guess it repeats [1].

[1]: http://git.infradead.org/linux-platform-drivers-x86.git/commitdiff/4dbccb873f2b35ad1b26419ff88c80509e2d4cbb

Quoted 25 lines
---
 drivers/platform/x86/surface3_power.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/platform/x86/surface3_power.c +4 -3
index 946ac2dc08ae..32e6e86e27dd 100644
@@ -522,8 +522,10 @@ static int mshw0011_probe(struct i2c_client *client)
        strlcpy(board_info.type, "MSHW0011-bat0", I2C_NAME_SIZE);
 
        bat0 = i2c_acpi_new_device(dev, 1, &board_info);
-       if (!bat0)
-               return -ENOMEM;
+       if (IS_ERR(bat0)) {
+               error = PTR_ERR(bat0);
+               goto out_err;
+       }
 
        data->bat0 = bat0;
        i2c_set_clientdata(bat0, data);
--

2.26.0

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#2 lore raw

Hi,

Quoted 6 lines · Andy Shevchenko
Quoted 5 lines · Hans de Goede

i2c_acpi_new_device() never returns NULL, it either returns an i2c_client or a PTR_ERR. Adjust the mshw0011_probe() error handling to take this into account.

Quoted 8 lines
Quoted 4 lines

Note the goto out_err will cause i2c_unregister_device() to get called even though the i2c_acpi_new_device() fails, this is ok as it accepts a NULL pointer argument (and treats it as a no-op).

Thanks, I guess it repeats [1].

[1]: http://git.infradead.org/linux-platform-drivers-x86.git/commitdiff/4dbccb873f2b35ad1b26419ff88c80509e2d4cbb

I guess it does and Dan's solution for the error handling is claner then mine, a direct return indeed is the proper thing to do here.

Regards,

Hans

Quoted 28 lines
Quoted 26 lines
---
  drivers/platform/x86/surface3_power.c | 6 ++++--
  1 file changed, 4 insertions(+), 2 deletions(-)
drivers/platform/x86/surface3_power.c +4 -3
index 946ac2dc08ae..32e6e86e27dd 100644
@@ -522,8 +522,10 @@ static int mshw0011_probe(struct i2c_client *client)
         strlcpy(board_info.type, "MSHW0011-bat0", I2C_NAME_SIZE);
 
         bat0 = i2c_acpi_new_device(dev, 1, &board_info);
-       if (!bat0)
-               return -ENOMEM;
+       if (IS_ERR(bat0)) {
+               error = PTR_ERR(bat0);
+               goto out_err;
+       }
 
         data->bat0 = bat0;
         i2c_set_clientdata(bat0, data);
--

2.26.0