diff mbox series

i2c: busses: fix uninit-value in pci1xxxx_i2c_xfer

Message ID 20241022190845.23536-1-surajsonawane0215@gmail.com
State New
Headers show
Series i2c: busses: fix uninit-value in pci1xxxx_i2c_xfer | expand

Commit Message

Suraj Sonawane Oct. 22, 2024, 7:08 p.m. UTC
Fix an issue reported by the smatch static analysis tool:
drivers/i2c/busses/i2c-mchp-pci1xxxx.c:1030 pci1xxxx_i2c_xfer() error:
uninitialized symbol 'retval'.

The error occurs because retval may be used without being set if the
transfer loop does not execute (e.g., when num is 0). This could cause
the function to return an undefined value, leading to unpredictable
behavior.

Initialize retval to 0 before the transfer loop to ensure that the
function returns a valid value even if no transfers are processed. This
change also preserves proper error handling within the loop.

Signed-off-by: Suraj Sonawane <surajsonawane0215@gmail.com>
---
 drivers/i2c/busses/i2c-mchp-pci1xxxx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Andi Shyti Oct. 23, 2024, 11:02 p.m. UTC | #1
Hi Suraj,

Thanks for your patch, but...

On Wed, Oct 23, 2024 at 12:38:45AM +0530, Suraj Sonawane wrote:
> Fix an issue reported by the smatch static analysis tool:
> drivers/i2c/busses/i2c-mchp-pci1xxxx.c:1030 pci1xxxx_i2c_xfer() error:
> uninitialized symbol 'retval'.
> 
> The error occurs because retval may be used without being set if the
> transfer loop does not execute (e.g., when num is 0). This could cause
> the function to return an undefined value, leading to unpredictable
> behavior.

... if num is '0', then we would never reach here, check
__i2c_transfer().

Thanks,
Andi

> Initialize retval to 0 before the transfer loop to ensure that the
> function returns a valid value even if no transfers are processed. This
> change also preserves proper error handling within the loop.
> 
> Signed-off-by: Suraj Sonawane <surajsonawane0215@gmail.com>
Suraj Sonawane Oct. 26, 2024, 7:39 a.m. UTC | #2
On 24/10/24 04:32, Andi Shyti wrote:
> Hi Suraj,
> 
> Thanks for your patch, but...
> 
> On Wed, Oct 23, 2024 at 12:38:45AM +0530, Suraj Sonawane wrote:
>> Fix an issue reported by the smatch static analysis tool:
>> drivers/i2c/busses/i2c-mchp-pci1xxxx.c:1030 pci1xxxx_i2c_xfer() error:
>> uninitialized symbol 'retval'.
>>
>> The error occurs because retval may be used without being set if the
>> transfer loop does not execute (e.g., when num is 0). This could cause
>> the function to return an undefined value, leading to unpredictable
>> behavior.
> 
> ... if num is '0', then we would never reach here, check
> __i2c_transfer().
> 
> Thanks,
> Andi
> 
>> Initialize retval to 0 before the transfer loop to ensure that the
>> function returns a valid value even if no transfers are processed. This
>> change also preserves proper error handling within the loop.
>>
>> Signed-off-by: Suraj Sonawane <surajsonawane0215@gmail.com>
Thank you for reviewing and clarifying.

Best regards,
Suraj Sonawane
diff mbox series

Patch

diff --git a/drivers/i2c/busses/i2c-mchp-pci1xxxx.c b/drivers/i2c/busses/i2c-mchp-pci1xxxx.c
index 5ef136c3e..4dfa11650 100644
--- a/drivers/i2c/busses/i2c-mchp-pci1xxxx.c
+++ b/drivers/i2c/busses/i2c-mchp-pci1xxxx.c
@@ -994,7 +994,7 @@  static int pci1xxxx_i2c_xfer(struct i2c_adapter *adap,
 {
 	struct pci1xxxx_i2c *i2c = i2c_get_adapdata(adap);
 	u8 slaveaddr;
-	int retval;
+	int retval = 0;
 	u32 i;
 
 	i2c->i2c_xfer_in_progress = true;