Re: [PATCH] PCI: Disable RRS polling for Intel SSDPE2KX020T8 nvme

[Bjorn Helgaas <helgaas@kernel.org>]

On Wed, Jun 11, 2025 at 06:14:42PM +0800, Hui Wang wrote:
> Prior to commit d591f6804e7e ("PCI: Wait for device readiness with
> Configuration RRS"), this Intel nvme [8086:0a54] works well. Since
> that patch is merged to the kernel, this nvme stops working.
> 
> Through debugging, we found that commit introduces the RRS polling in
> the pci_dev_wait(), for this nvme, when polling the PCI_VENDOR_ID, it
> will return ~0 if the config access is not ready yet, but the polling
> expects a return value of 0x0001 or a valid vendor_id, so the RRS
> polling doesn't work for this nvme.
> 
> Here we add a pci quirk to disable the RRS polling for the device.
> 
> Fixes: d591f6804e7e ("PCI: Wait for device readiness with Configuration RRS")
> Link: https://bugs.launchpad.net/bugs/2111521
> Signed-off-by: Hui Wang <hui.wang@canonical.com>

To help keep this top of mind:

#regzbot introduced: d591f6804e7e ("PCI: Wait for device readiness with Configuration RRS")

> ---
>  drivers/pci/probe.c  |  3 ++-
>  drivers/pci/quirks.c | 18 ++++++++++++++++++
>  include/linux/pci.h  |  2 ++
>  3 files changed, 22 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c
> index 4b8693ec9e4c..848fa0e6cf60 100644
> --- a/drivers/pci/probe.c
> +++ b/drivers/pci/probe.c
> @@ -1270,7 +1270,8 @@ static void pci_enable_rrs_sv(struct pci_dev *pdev)
>  	if (root_cap & PCI_EXP_RTCAP_RRS_SV) {
>  		pcie_capability_set_word(pdev, PCI_EXP_RTCTL,
>  					 PCI_EXP_RTCTL_RRS_SVE);
> -		pdev->config_rrs_sv = 1;
> +		if (!(pdev->dev_flags & PCI_DEV_FLAGS_NO_RRS_SV))
> +			pdev->config_rrs_sv = 1;
>  	}
>  }
>  
> diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
> index cf483d82572c..519e48ff6448 100644
> --- a/drivers/pci/quirks.c
> +++ b/drivers/pci/quirks.c
> @@ -6336,3 +6336,21 @@ static void pci_mask_replay_timer_timeout(struct pci_dev *pdev)
>  DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_GLI, 0x9750, pci_mask_replay_timer_timeout);
>  DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_GLI, 0x9755, pci_mask_replay_timer_timeout);
>  #endif
> +
> +/*
> + * Although the root port device claims to support RRS, some devices don't work
> + * with RRS polling, when reading the Vendor ID, they just return ~0 if config
> + * access is not ready, this will break the pci_dev_wait(). Here disable the RRS
> + * forcibly for this type of device.
> + */
> +static void quirk_no_rrs_sv(struct pci_dev *dev)
> +{
> +	struct pci_dev *root;
> +
> +	root = pcie_find_root_port(dev);
> +	if (root) {
> +		root->dev_flags |= PCI_DEV_FLAGS_NO_RRS_SV;
> +		root->config_rrs_sv = 0;
> +	}
> +}
> +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x0a54, quirk_no_rrs_sv);
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index 05e68f35f392..f4dd9ada12e4 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -247,6 +247,8 @@ enum pci_dev_flags {
>  	PCI_DEV_FLAGS_HAS_MSI_MASKING = (__force pci_dev_flags_t) (1 << 12),
>  	/* Device requires write to PCI_MSIX_ENTRY_DATA before any MSIX reads */
>  	PCI_DEV_FLAGS_MSIX_TOUCH_ENTRY_DATA_FIRST = (__force pci_dev_flags_t) (1 << 13),
> +	/* Do not use Configuration Request Retry Status polling in pci_dev_wait() */
> +	PCI_DEV_FLAGS_NO_RRS_SV = (__force pci_dev_flags_t) (1 << 14),
>  };
>  
>  enum pci_irq_reroute_variant {
> -- 
> 2.34.1
>