Re: [PATCH ath-next v11 08/13] wifi: ath12k: add AHB driver support for IPQ5332

[Jeff Johnson <jeff.johnson@oss.qualcomm.com>]

On 3/18/2025 12:06 PM, Krzysztof Kozlowski wrote:
> On 18/03/2025 19:53, Jeff Johnson wrote:
>> On 3/18/2025 11:19 AM, Krzysztof Kozlowski wrote:
>>> On 18/03/2025 19:16, Krzysztof Kozlowski wrote:
>>>> On 18/03/2025 18:55, Jeff Johnson wrote:
>>>>> On 3/18/2025 8:50 AM, Krzysztof Kozlowski wrote:
>>>>>> On 18/03/2025 16:44, Jeff Johnson wrote:
>>>>>>> On 3/17/2025 1:46 PM, Raj Kumar Bhagat wrote:
>>>>>>>> +	hw_rev = (enum ath12k_hw_rev)of_device_get_match_data(&pdev->dev);
>>>>>>>
>>>>>>> kernel test robot warns:
>>>>>>> cast to smaller integer type 'enum ath12k_hw_rev' from 'const void *'
>>>>>>>
>>>>>>> looks like others have fixed this by first casting to (uintptr_t)
>>>>>>> a few examples:
>>>>>>>
>>>>>> Cast via (kernel_ulong_t)
>>>>>>
>>>>>> But another point is that this patch at stage v11 should not have
>>>>>> compiler warnings and it's not our tools who should point it out. Except
>>>>>> W=1, all standard static analyzers (sparse, smatch and coccinelle) are
>>>>>> expected to be run.
>>>>>
>>>>> I ran what I thought was a reasonable cross-section of builds and did not see
>>>>> this issue. Seems this issue is only flagged with config: um-allmodconfig ??
>>>>>
>>>>> Guess I need to add that configuration to my builds...
>>>>
>>>> This should be visible on every build on 32 bit archs.
>>
>> Yes, I'm seeing it now on my i386 builds
>>
>>> Patchset does not apply neither on next nor on on mainline, so cannot
>>> verify... Probably another problem to solve here - some sort of dependency?
>>
>> T:      git git://git.kernel.org/pub/scm/linux/kernel/git/ath/ath.git
>>
>> The series should apply cleanly on ath.git ath-next branch
> 
> 
> Yeah, but next is fed with for-next branch, not ath-next.... and
> for-next is 10 days behind ath-next. I assume for-next is what you send
> to the next merge window and ath-next is rather next-next-merge window,
> just like DRM people are doing?

we have for-next which feeds linux-next.
normally for-next is a symbolic ref to ath-next.
but when we issue our last PR to wireless-next before the merge window, after
we are pulled we fast-forward and hard branch for-next.
so from that point our contributions to linux-next stop while we continue to
contribute to ath-next.
when the merge window closes we remove the hard branch and reestablish the
symbolic ref.

/jeff