Re: [PATCH] PCI/ASPM: Fix a typo in ASPM restoring logic

From: Bjorn Helgaas
Date: Mon May 06 2024 - 15:24:11 EST


[+cc David]

On Mon, May 06, 2024 at 01:16:02PM +0800, Kai-Heng Feng wrote:
> There's a typo that makes parent device uses child LNKCTL value and vice
> versa. This causes Micron NVMe to trigger a reboot upon system resume.
>
> Correct the typo to fix the issue.
>
> Fixes: 64dbb2d70744 ("PCI/ASPM: Disable L1 before configuring L1 Substates")
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@xxxxxxxxxxxxx>

This is something David did correctly in his original posting
(https://lore.kernel.org/all/20240128233212.1139663-4-david.e.box@xxxxxxxxxxxxxxx/)
and I broke it while reorganizing things
(https://lore.kernel.org/all/20240223213733.GA115410@bhelgaas/).
Thanks for finding this!

Since 64dbb2d70744 was merged for v6.9-rc1, I queued this to for-linus
for v6.9 with this subject:

PCI/ASPM: Restore parent state to parent, child state to child

since it's more than just an innocuous typo.

> ---
> drivers/pci/pcie/aspm.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/pci/pcie/aspm.c b/drivers/pci/pcie/aspm.c
> index 2428d278e015..47761c7ef267 100644
> --- a/drivers/pci/pcie/aspm.c
> +++ b/drivers/pci/pcie/aspm.c
> @@ -177,8 +177,8 @@ void pci_restore_aspm_l1ss_state(struct pci_dev *pdev)
> /* Restore L0s/L1 if they were enabled */
> if (FIELD_GET(PCI_EXP_LNKCTL_ASPMC, clnkctl) ||
> FIELD_GET(PCI_EXP_LNKCTL_ASPMC, plnkctl)) {
> - pcie_capability_write_word(parent, PCI_EXP_LNKCTL, clnkctl);
> - pcie_capability_write_word(pdev, PCI_EXP_LNKCTL, plnkctl);
> + pcie_capability_write_word(parent, PCI_EXP_LNKCTL, plnkctl);
> + pcie_capability_write_word(pdev, PCI_EXP_LNKCTL, clnkctl);
> }
> }
>
> --
> 2.34.1
>