Re: [PATCH v2 6/8] wifi: ath12k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states

On 8/26/2025 9:00 AM, Ilpo Järvinen wrote:
> On Tue, 26 Aug 2025, Jeff Johnson wrote:
> 
>> On 8/25/2025 10:44 AM, Manivannan Sadhasivam via B4 Relay wrote:
>>> --- a/drivers/net/wireless/ath/ath12k/Kconfig
>>> +++ b/drivers/net/wireless/ath/ath12k/Kconfig
>>> @@ -1,7 +1,7 @@
>>>  # SPDX-License-Identifier: BSD-3-Clause-Clear
>>>  config ATH12K
>>>  	tristate "Qualcomm Technologies Wi-Fi 7 support (ath12k)"
>>> -	depends on MAC80211 && HAS_DMA && PCI
>>> +	depends on MAC80211 && HAS_DMA && PCI && PCIEASPM
>>
>> As you point out in patch 1/8, PCIEASPM is protected by EXPERT.
>>
>> Won't this prevent the driver from being built (or even showing up in
>> menuconfig) if EXPERT is not enabled?
> 
> It doesn't work that way, PCIEASPM defaults to y:
> 
> $ sed -i -e 's|CONFIG_PCIEASPM=y|CONFIG_PCIEASPM=n|g' .config && make oldconfig && grep -e 'CONFIG_EXPERT ' -e 'CONFIG_PCIEASPM=' .config
> #
> # configuration written to .config
> #
> # CONFIG_EXPERT is not set
> CONFIG_PCIEASPM=y
> 
>> Should we consider having a separate CONFIG item that is used to protect just
>> the PCI ASPM interfaces? And then we could split out the ath12k_pci_aspm
>> functions into a separate file that is conditionally built based upon that
>> config item?
>>
>> Or am I too paranoid since everyone enables EXPERT?
> 
> One just cannot control PCIEASPM value if EXPERT is not set. ASPM is 
> expected to be enabled, or "experts" get to keep the pieces.
> 

Thanks for the clarification. I now have no issues with the ath driver patches.

/jeff