Re: [PATCH v2 1/3] PCI/PM: Honor PCI_DEV_FLAGS_NO_D3 in pci_bridge_d3_possible()
From: Kuppuswamy Sathyanarayanan
Date: Mon Nov 16 2026 - 12:47:41 EST
On 11/16/26 08:51, Aviv Bakal wrote:
> pci_set_power_state() does not put a device with PCI_DEV_FLAGS_NO_D3
> into D3, but pci_bridge_d3_possible() ignores the flag.
> pcie_portdrv_probe() therefore enables runtime PM on such a port, and
> on runtime suspend pciehp masks the hotplug interrupt while the port
> stays in D0.
>
> Return false from pci_bridge_d3_possible() when the flag is set.
> pci_bridge_d3_possible() must return the same value for the lifetime
> of the device, and FINAL fixups may set the flag after portdrv has
> probed, so record the flag in pci_pm_init() and check that copy.
>
> Fixes: d4c79b63d82d ("PCI: Allow D3 for native hotplug-capable Root Ports on non-x86 platforms")
> Cc: stable@xxxxxxxxxxxxxxx # 7.3.x
> Co-developed-by: Zeev Zilberman <zeev@xxxxxxxxxx>
> Signed-off-by: Zeev Zilberman <zeev@xxxxxxxxxx>
> Signed-off-by: Aviv Bakal <avivb@xxxxxxxxxx>
> ---
> drivers/pci/pci.c | 10 ++++++++++
> include/linux/pci.h | 1 +
> 2 files changed, 11 insertions(+)
>
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 7c6f2aba8144..eb7f2f03d9ae 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -3055,6 +3055,10 @@ bool pci_bridge_d3_possible(struct pci_dev *bridge)
> if (pci_bridge_d3_disable)
> return false;
>
> + /* The port is known not to work correctly in D3 */
> + if (bridge->no_d3)
> + return false;
> +
> /*
> * Hotplug ports handled by platform firmware may not be put
> * into D3 by the OS, e.g. ACPI slots ...
> @@ -3227,6 +3231,12 @@ void pci_pm_init(struct pci_dev *dev)
> dev->pm_cap = 0;
> dev->pme_support = 0;
>
> + /*
> + * pci_bridge_d3_possible() must not change after this point, so
> + * ignore NO_D3 set by later fixups.
> + */
> + dev->no_d3 = !!(dev->dev_flags & PCI_DEV_FLAGS_NO_D3);
This ignores quirk_plx_pex_no_d3(). It is the only other quirk that
sets NO_D3 on a PCIe port, and it is a FINAL fixup. So
pci_bridge_d3_possible() still returns true for those PLX downstream
ports, and portdrv still enables runtime PM on them.
> +
> /* find PCI PM capability in list */
> pm = pci_find_capability(dev, PCI_CAP_ID_PM);
> if (!pm)
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index aa3765a20a96..1238667d643c 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -420,6 +420,7 @@ struct pci_dev {
> unsigned int no_d1d2:1; /* D1 and D2 are forbidden */
> unsigned int no_d3cold:1; /* D3cold is forbidden */
> unsigned int bridge_d3:1; /* Allow D3 for bridge */
> + unsigned int no_d3:1; /* NO_D3 set before PM init */
> unsigned int d3cold_allowed:1; /* D3cold is allowed by user */
> unsigned int mmio_always_on:1; /* Disallow turning off io/mem
> decoding during BAR sizing */
--
Sathyanarayanan Kuppuswamy
Linux Kernel Developer