Re: [PATCH 1/3] tee: qcomtee: call: Fix confusing cleanup.h syntax

[Amirreza Zarrabi <amirreza.zarrabi@oss.qualcomm.com>]

Hi,

On 12/12/2025 12:39 PM, Sumit Garg wrote:
> On Fri, Dec 12, 2025 at 02:07:40AM +0100, Krzysztof Kozlowski wrote:
>> On 12/12/2025 01:55, Sumit Garg wrote:
>>> On Mon, Dec 08, 2025 at 03:08:45AM +0100, Krzysztof Kozlowski wrote:
>>>> Initializing automatic __free variables to NULL without need (e.g.
>>>> branches with different allocations), followed by actual allocation is
>>>> in contrary to explicit coding rules guiding cleanup.h:
>>>>
>>>> "Given that the "__free(...) = NULL" pattern for variables defined at
>>>> the top of the function poses this potential interdependency problem the
>>>> recommendation is to always define and assign variables in one statement
>>>> and not group variable definitions at the top of the function when
>>>> __free() is used."
>>>>
>>>> Code does not have a bug, but is less readable and uses discouraged
>>>> coding practice, so fix that by moving declaration to the place of
>>>> assignment.
>>>
>>> Okay I see but..
>>>
>>>>
>>>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
>>>> ---
>>>>  drivers/tee/qcomtee/call.c | 17 ++++++++---------
>>>>  1 file changed, 8 insertions(+), 9 deletions(-)
>>>>
>>>> diff --git a/drivers/tee/qcomtee/call.c b/drivers/tee/qcomtee/call.c
>>>> index 65f9140d4e1f..8f8830f0df26 100644
>>>> --- a/drivers/tee/qcomtee/call.c
>>>> +++ b/drivers/tee/qcomtee/call.c
>>>> @@ -395,9 +395,7 @@ static int qcomtee_object_invoke(struct tee_context *ctx,
>>>>  				 struct tee_ioctl_object_invoke_arg *arg,
>>>>  				 struct tee_param *params)
>>>>  {
>>>> -	struct qcomtee_object_invoke_ctx *oic __free(kfree) = NULL;
>>>>  	struct qcomtee_context_data *ctxdata = ctx->data;
>>>> -	struct qcomtee_arg *u __free(kfree) = NULL;
>>>>  	struct qcomtee_object *object;
>>>>  	int i, ret, result;
>>>>  
>>>> @@ -412,12 +410,14 @@ static int qcomtee_object_invoke(struct tee_context *ctx,
>>>>  	}
>>>>  
>>>>  	/* Otherwise, invoke a QTEE object: */
>>>> -	oic = qcomtee_object_invoke_ctx_alloc(ctx);
>>>> +	struct qcomtee_object_invoke_ctx *oic __free(kfree) =
>>>> +		qcomtee_object_invoke_ctx_alloc(ctx);
>>>>  	if (!oic)
>>>>  		return -ENOMEM;
>>>>  
>>>>  	/* +1 for ending QCOMTEE_ARG_TYPE_INV. */
>>>> -	u = kcalloc(arg->num_params + 1, sizeof(*u), GFP_KERNEL);
>>>> +	struct qcomtee_arg *u __free(kfree) = kcalloc(arg->num_params + 1, sizeof(*u),
>>>> +						      GFP_KERNEL);
>>>
>>> ..this makes the code less readable with variable declarations floating
>>
>> Which is intentional.
>>
>>> within the function. I would rather favor to not use the cleanup.h construct
>>> but use explicit kfree() invocations instead like it's done in all other
>>> allocations in the TEE subsystem.
>>
>> Sure, fair. I just don't get why introducing cleanup.h without actually
>> accepting its explicitly documented style...
>>
> 
> TBH, it is likely overlooked during review of the QTEE driver. Having a
> builtin warning for the undesired syntax would help the reviewers here.
> 
> -Sumit

While the style may seem unusual -- as stated in cleanup.h, using cleanup helpers
makes the code more readable overall compared to relying on multiple goto statements.
Also, it’s not just about the "__free(...) = NULL" use cases -- there are locks
involved as well. Switching to direct free() would require reverting those locks,
since mixing cleanup helpers with manual cleanup is not acceptable.

If this behavior is explicitly documented in cleanup.h, there is no reason not
to use it as intended. I also support Krzysztof’s suggestion.

Best regards,
Amir