Re: [PATCH] coresight: etm4x: Ensure valid drvdata and clock before clk_put()

[Anshuman Khandual]

On 8/11/23 14:39, Suzuki K Poulose wrote:
> On 11/08/2023 09:39, James Clark wrote:
>>
>>
>> On 11/08/2023 07:27, Anshuman Khandual wrote:
>>> This validates 'drvdata' and 'drvdata->pclk' clock before calling clk_put()
>>> in etm4_remove_platform_dev(). The problem was detected using Smatch static
>>> checker as reported.
>>>
>>> Cc: Suzuki K Poulose <suzuki.poulose@xxxxxxx>
>>> Cc: Mike Leach <mike.leach@xxxxxxxxxx>
>>> Cc: James Clark <james.clark@xxxxxxx>
>>> Cc: coresight@xxxxxxxxxxxxxxxx
>>> Cc: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
>>> Cc: linux-kernel@xxxxxxxxxxxxxxx
>>> Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
>>> Closes: https://lists.linaro.org/archives/list/coresight@xxxxxxxxxxxxxxxx/thread/G4N6P4OXELPLLQSNU3GU2MR4LOLRXRMJ/
>>> Signed-off-by: Anshuman Khandual <anshuman.khandual@xxxxxxx>
>>> ---
>>> This applies on coresight-next
>>>
>>>   drivers/hwtracing/coresight/coresight-etm4x-core.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> index 703b6fcbb6a5..eb412ce302cc 100644
>>> --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c
>>> @@ -2269,7 +2269,7 @@ static int __exit etm4_remove_platform_dev(struct platform_device *pdev)
>>>           etm4_remove_dev(drvdata);
>>>       pm_runtime_disable(&pdev->dev);
>>>   -    if (drvdata->pclk)
>>> +    if (drvdata && drvdata->pclk && !IS_ERR(drvdata->pclk))
>>>           clk_put(drvdata->pclk);
>>>         return 0;
>>
>> It could be !IS_ERR_OR_NULL(drvdata->pclk), but I wouldn't bother
>> changing it at this point.
> 
> +1, please could we have that. Someone else will run a code scanner and
> send a patch later. Given this is straight and easy change, lets do it
> in the first place.

But we already have a drvdata->pclk validation check before IS_ERR().
Would not _OR_NULL be redundant ?