Hi,

This series fixes the behavior of the pci_enable_link_state() and
pci_enable_link_state_locked() APIs to be in symmetry with
pci_disable_link_state*() couterparts.

First 3 patches fixes and cleans up the ASPM code and the last 3 patches
modifies the atheros drivers to use the pci{enable/disable}_link_state() APIs
instead of modifying the LNKCTL register directly for enabling ASPM.

NOTE: The current callers of the pci_enable_link_state_locked() APIs (vmd and
pcie-qcom) drivers doesn't look like depending on the old behavior of the API. I
can atleast assure that for pcie-qcom. For VMD, it would be great if VMD folks
CCed could provide their review tags for patch 1/6.

Testing
=======

I've tested this series on Lenovo Thinkpad T14s with WCN7850 chipset (so that's
just ath12k driver). Rest of the drivers are compile tested only.

Merging Strategy
================

Even though there is no build dependency between PCI core and atheros patches,
there is a functional dependency. So I'd recommend creating an immutable branch
with PCI patches and merging that branch into both PCI and linux-wireless trees
and finally merging the atheros patches into linux-wireless tree.

If immutable branch seems like a hassle, then PCI core patches could get merged
for 6.17 and atheros patches can wait for 6.18.

- Mani

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>
---
Manivannan Sadhasivam (6):
      PCI/ASPM: Fix the behavior of pci_enable_link_state*() APIs
      PCI/ASPM: Transition the device to D0 (if required) inside pci_enable_link_state_locked() API
      PCI/ASPM: Improve the kernel-doc for pci_disable_link_state*() APIs
      wifi: ath12k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states
      wifi: ath11k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states
      wifi: ath10k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states

 drivers/net/wireless/ath/ath.h         | 14 ++++++
 drivers/net/wireless/ath/ath10k/pci.c  |  7 +--
 drivers/net/wireless/ath/ath11k/pci.c  | 10 ++---
 drivers/net/wireless/ath/ath12k/pci.c  | 10 ++---
 drivers/pci/controller/dwc/pcie-qcom.c |  5 ---
 drivers/pci/controller/vmd.c           |  5 ---
 drivers/pci/pcie/aspm.c                | 78 ++++++++++++++++++++++++----------
 7 files changed, 79 insertions(+), 50 deletions(-)
---
base-commit: 19272b37aa4f83ca52bdf9c16d5d81bdd1354494
change-id: 20250711-ath-aspm-fix-c17442a5a9ae

Best regards,
-- 
Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>

On 7/16/2025 5:56 AM, Manivannan Sadhasivam via B4 Relay wrote:
> Merging Strategy
> ================
> 
> Even though there is no build dependency between PCI core and atheros patches,
> there is a functional dependency. So I'd recommend creating an immutable branch
> with PCI patches and merging that branch into both PCI and linux-wireless trees
> and finally merging the atheros patches into linux-wireless tree.
> 
> If immutable branch seems like a hassle, then PCI core patches could get merged
> for 6.17 and atheros patches can wait for 6.18.

I'm fine with either strategy. In the first case I'd merge the immutable
branch into the ath tree. Note I plan to issue my final PR to linux-wireless
for the 6.17 merge window on Monday, so we should close on this decision soon.

/jeff

On Wed, Jul 16, 2025 at 10:11:27AM GMT, Jeff Johnson wrote:
> On 7/16/2025 5:56 AM, Manivannan Sadhasivam via B4 Relay wrote:
> > Merging Strategy
> > ================
> > 
> > Even though there is no build dependency between PCI core and atheros patches,
> > there is a functional dependency. So I'd recommend creating an immutable branch
> > with PCI patches and merging that branch into both PCI and linux-wireless trees
> > and finally merging the atheros patches into linux-wireless tree.
> > 
> > If immutable branch seems like a hassle, then PCI core patches could get merged
> > for 6.17 and atheros patches can wait for 6.18.
> 
> I'm fine with either strategy. In the first case I'd merge the immutable
> branch into the ath tree. Note I plan to issue my final PR to linux-wireless
> for the 6.17 merge window on Monday, so we should close on this decision soon.
> 

Looks like there are a couple of things that need to be fixed before this series
gets merged. So we can defer the whole series for 6.18.

- Mani

-- 
மணிவண்ணன் சதாசிவம்