Linux PCI subsystem development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Manivannan Sadhasivam" <mani@kernel.org>
Cc: linux-pci@vger.kernel.org
Subject: Re: [PATCH v3 6/8] wifi: ath12k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states
Date: Wed, 08 Jul 2026 14:49:03 +0000	[thread overview]
Message-ID: <20260708144904.0CC4F1F000E9@smtp.kernel.org> (raw)
In-Reply-To: <20260708-pci-aspm-fix-v3-6-6bd72451746e@kernel.org>

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

New issues:
- [Medium] Adding `depends on PCIEASPM` unnecessarily restricts the driver from being built on kernels where ASPM control is disabled.

Pre-existing issues:
- [High] ASPM state is permanently lost (leaked as disabled) if initialization fails before `ath12k_pci_start()`.
--

--- Patch [6]: [PATCH v3 6/8] wifi: ath12k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states ---
commit 6aeb96e7a1d3a96df659d6e1d629f287d946f802
Author: Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>

wifi: ath12k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states

This commit changes the ath12k driver to use the PCI core's
pci_disable_link_state() and pci_force_enable_link_state() APIs for managing
ASPM states instead of directly modifying the LNKCTL register. It also adds a
Kconfig dependency on PCIEASPM.

> diff --git a/drivers/net/wireless/ath/ath12k/Kconfig b/drivers/net/wireless/ath/ath12k/Kconfig
> index 4a2b240f967af..7852ede3eaea8 100644
> --- 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 Wi-Fi 7 support (ath12k)"
> -	depends on MAC80211 && HAS_DMA && PCI
> +	depends on MAC80211 && HAS_DMA && PCI && PCIEASPM

[Severity: Medium]
Is the PCIEASPM dependency necessary here?

The PCI subsystem provides static inline no-ops for pci_disable_link_state(),
pcie_aspm_enabled(), and pci_force_enable_link_state() in linux/pci.h when
CONFIG_PCIEASPM is not set. Adding this strict dependency restricts the
driver from being built on kernels where ASPM control is intentionally
disabled.

>  	select QCOM_QMI_HELPERS
>  	select MHI_BUS
>  	select QRTR

[ ... ]

> diff --git a/drivers/net/wireless/ath/ath12k/pci.c b/drivers/net/wireless/ath/ath12k/pci.c
> index d9a22d6afbb02..6d48fb25690de 100644
> --- a/drivers/net/wireless/ath/ath12k/pci.c
> +++ b/drivers/net/wireless/ath/ath12k/pci.c
> @@ -884,19 +884,9 @@ static void ath12k_pci_free_region(struct ath12k_pci *ab_pci)
>  
>  static void ath12k_pci_aspm_disable(struct ath12k_pci *ab_pci)
>  {
> -	struct ath12k_base *ab = ab_pci->ab;
> -
> -	pcie_capability_read_word(ab_pci->pdev, PCI_EXP_LNKCTL,
> -				  &ab_pci->link_ctl);
> -
> -	ath12k_dbg(ab, ATH12K_DBG_PCI, "pci link_ctl 0x%04x L0s %d L1 %d\n",
> -		   ab_pci->link_ctl,
> -		   u16_get_bits(ab_pci->link_ctl, PCI_EXP_LNKCTL_ASPM_L0S),
> -		   u16_get_bits(ab_pci->link_ctl, PCI_EXP_LNKCTL_ASPM_L1));
> +	ab_pci->aspm_states = pcie_aspm_enabled(ab_pci->pdev);

[Severity: High]
This is a pre-existing issue, but does this logic permanently lose the ASPM
state if initialization fails before ath12k_pci_start()?

If ath12k_mhi_start() fails in ath12k_pci_power_up(), the initialization
is aborted. The current ASPM state is unconditionally saved to
ab_pci->aspm_states here and then disabled. On failure,
ath12k_pci_power_down() is called to clean up, but it returns early if
ATH12K_PCI_FLAG_INIT_DONE is not set:

drivers/net/wireless/ath/ath12k/pci.c:ath12k_pci_power_down() {
	if (!test_bit(ATH12K_PCI_FLAG_INIT_DONE, &ab_pci->flags))
		return;

	/* restore aspm in case firmware bootup fails */
	ath12k_pci_aspm_restore(ab_pci);
	...
}

This early return skips the restore and leaves ASPM disabled. When the
device is later retried, ath12k_pci_power_up() is called again, overwriting
ab_pci->aspm_states with the now-disabled state (0). When initialization
finally succeeds, it restores 0, permanently disabling ASPM.

>  
> -	/* disable L0s and L1 */
> -	pcie_capability_clear_word(ab_pci->pdev, PCI_EXP_LNKCTL,
> -				   PCI_EXP_LNKCTL_ASPMC);
> +	pci_disable_link_state(ab_pci->pdev, PCIE_LINK_STATE_ASPM_ALL);
>  
>  	set_bit(ATH12K_PCI_ASPM_RESTORE, &ab_pci->flags);
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260708-pci-aspm-fix-v3-0-6bd72451746e@kernel.org?part=6

  reply	other threads:[~2026-07-08 14:49 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-08 14:30 [PATCH v3 0/8] PCI/ASPM: Add pci_force_enable_link_state() API and convert Atheros drivers Manivannan Sadhasivam
2026-07-08 14:30 ` [PATCH v3 1/8] PCI/ASPM: Always disable ASPM when driver requests it Manivannan Sadhasivam
2026-07-08 14:44   ` sashiko-bot
2026-07-08 14:30 ` [PATCH v3 2/8] PCI/ASPM: Add pci_force_enable_link_state() API Manivannan Sadhasivam
2026-07-08 14:44   ` sashiko-bot
2026-07-08 14:30 ` [PATCH v3 3/8] PCI/ASPM: Transition the device to D0 (if required) when enabling ASPM link states Manivannan Sadhasivam
2026-07-08 14:50   ` sashiko-bot
2026-07-08 14:30 ` [PATCH v3 4/8] PCI/ASPM: Improve the kernel-doc for pci_{enable,disable}_link_state*() APIs Manivannan Sadhasivam
2026-07-08 14:37   ` sashiko-bot
2026-07-08 14:30 ` [PATCH v3 5/8] PCI/ASPM: Return enabled ASPM states from pcie_aspm_enabled() API Manivannan Sadhasivam
2026-07-08 14:36   ` sashiko-bot
2026-07-08 14:30 ` [PATCH v3 6/8] wifi: ath12k: Use pci_{enable/disable}_link_state() APIs to enable/disable ASPM states Manivannan Sadhasivam
2026-07-08 14:49   ` sashiko-bot [this message]
2026-07-10 20:25   ` Bjorn Helgaas
2026-07-08 14:30 ` [PATCH v3 7/8] wifi: ath11k: " Manivannan Sadhasivam
2026-07-08 14:47   ` sashiko-bot
2026-07-08 14:30 ` [PATCH v3 8/8] wifi: ath10k: " Manivannan Sadhasivam
2026-07-08 14:51   ` sashiko-bot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260708144904.0CC4F1F000E9@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=mani@kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox