Re: [PATCH v10 0/7] Add AMD ISP4 driver

[Bin Du <Bin.Du@amd.com>]

Hi Sakari,

I'm not sure which Smatch version is used in Media CI. Here are some 
findings from my local tests for reference.

For v10, I could reproduce the false positive with an older Smatch 
snapshot built from upstream commit 
URL:https://repo.or.cz/smatch.git/commit/bc575c7a0c5c7a2d4fefd0e1d8b0dc9ccacdd1a6, 
which is a pre-1.74 snapshot.

I could not reproduce the same false positive with a newer Smatch 
snapshot based on upstream commit 
URL:https://repo.or.cz/smatch.git/commit/bcc58b9ccf06d28ab6be4f0992bc74f462aa12f8, 
which is a post-1.75 snapshot.

For v11, neither Smatch snapshot reports the warning.

On 5/6/2026 5:41 PM, Bin Du wrote:
> Many thanks for the feedback, Sakari. I've sent v11 addressing the 
> Smatch false positive and the new checkpatch.pl warnings you pointed out:
>
> <URL: https://lore.kernel.org/all/20260506093250.93460-1-Bin.Du@amd.com/>
>
> Could you please take a look when you have a chance?
>
> On 5/6/2026 4:24 AM, sakari.ailus@linux.intel.com wrote:
>> Hi Bin,
>>
>> On Tue, Apr 07, 2026 at 03:19:20PM +0800, Bin Du wrote:
>>> Hi Sakari,
>>>
>>> Thank you very much for picking up the series and for updating 
>>> Patchwork. I
>>> really appreciate your help and support. It means a lot to us.
>>> Yes, that's right. <20260303224433.87242-1-kinncj@gmail.com> is no 
>>> longer
>>> needed.
>> The Media CI indeed does check there are no warnings from e.g. static
>> checkers such as smatch, and this lead to build failure earlier. If 
>> the set
>> would be merged, this same smatch warning would pop up as an error
>> elsewhere. So in practice it doesn't matter that the smatch warning is a
>> false positive, it needs to be addressed. You should check smatch is 
>> happy
>> with the updated code.
>>
>> While you're sending a new version, can you address the new 
>> checkpatch.pl
>> warnings, too? They're trivial.
>>
>> The best summary is here
>> <URL:https://gitlab.freedesktop.org/linux-media/media-committers/-/merge_requests/278> 
>>
>> I believe.
>>