[PATCH 1/2] PCI: Add new PCIe Fabric End Node flag, PCI_DEV_FLAGS_NO_RELAXED_ORDERING

[Alexander Duyck]

On Mon, May 1, 2017 at 4:13 PM, Casey Leedom <leedom at chelsio.com> wrote:
> The new flag PCI_DEV_FLAGS_NO_RELAXED_ORDERING indicates that the Relaxed
> Ordering Attribute should not be used on Transaction Layer Packets destined
> for the PCIe End Node so flagged.  Initially flagged this way are Intel
> E5-26xx Root Complex Ports which suffer from a Flow Control Credit
> Performance Problem and AMD A1100 ARM ("SEATTLE") Root Complex Ports which
> don't obey PCIe 3.0 ordering rules which can lead to Data Corruption.

So this is a good first step though might I suggest one other change.

We may want to add logic to the core PCIe code that clears the "Enable
Relaxed Ordering" bit in the device control register for all devices
hanging off of this root complex. Assuming the devices conform to the
PCIe spec doing that should disable relaxed ordering in a device
agnostic way that then enables us at a driver level to just enable the
feature always without having to perform any checks for your flag. We
could probably do that as a part of probing the PCIe interfaces
hanging off of these devices.

> ---
>  drivers/pci/quirks.c | 38 ++++++++++++++++++++++++++++++++++++++
>  include/linux/pci.h  |  2 ++
>  2 files changed, 40 insertions(+)
>
> diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
> index f754453..4ae78b3 100644
> --- a/drivers/pci/quirks.c
> +++ b/drivers/pci/quirks.c
> @@ -3979,6 +3979,44 @@ static void quirk_tw686x_class(struct pci_dev *pdev)
>                               quirk_tw686x_class);
>
>  /*
> + * Some devices have problems with Transaction Layer Packets with the Relaxed
> + * Ordering Attribute set.  Such devices should mark themselves and other
> + * Device Drivers should check before sending TLPs with RO set.
> + */
> +static void quirk_relaxedordering_disable(struct pci_dev *dev)
> +{
> +       dev->dev_flags |= PCI_DEV_FLAGS_NO_RELAXED_ORDERING;
> +}
> +
> +/*
> + * Intel E5-26xx Root Complex has a Flow Control Credit issue which can
> + * cause performance problems with Upstream Transaction Layer Packets with
> + * Relaxed Ordering set.
> + */
> +DECLARE_PCI_FIXUP_CLASS_EARLY(0x8086, 0x6f02, PCI_CLASS_NOT_DEFINED, 8,
> +                             quirk_relaxedordering_disable);
> +DECLARE_PCI_FIXUP_CLASS_EARLY(0x8086, 0x6f04, PCI_CLASS_NOT_DEFINED, 8,
> +                             quirk_relaxedordering_disable);
> +DECLARE_PCI_FIXUP_CLASS_EARLY(0x8086, 0x6f08, PCI_CLASS_NOT_DEFINED, 8,
> +                             quirk_relaxedordering_disable);
> +
> +/*
> + * The AMD ARM A1100 (AKA "SEATTLE") SoC has a bug in its PCIe Root Complex
> + * where Upstream Transaction Layer Packets with the Relaxed Ordering
> + * Attribute clear are allowed to bypass earlier TLPs with Relaxed Ordering
> + * set.  This is a violation of the PCIe 3.0 Transaction Ordering Rules
> + * outlined in Section 2.4.1 (PCI Express(r) Base Specification Revision 3.0
> + * November 10, 2010).  As a result, on this platform we can't use Relaxed
> + * Ordering for Upstream TLPs.
> + */
> +DECLARE_PCI_FIXUP_CLASS_EARLY(0x1022, 0x1a00, PCI_CLASS_NOT_DEFINED, 8,
> +                             quirk_relaxedordering_disable);
> +DECLARE_PCI_FIXUP_CLASS_EARLY(0x1022, 0x1a01, PCI_CLASS_NOT_DEFINED, 8,
> +                             quirk_relaxedordering_disable);
> +DECLARE_PCI_FIXUP_CLASS_EARLY(0x1022, 0x1a02, PCI_CLASS_NOT_DEFINED, 8,
> +                             quirk_relaxedordering_disable);
> +
> +/*
>   * Per PCIe r3.0, sec 2.2.9, "Completion headers must supply the same
>   * values for the Attribute as were supplied in the header of the
>   * corresponding Request, except as explicitly allowed when IDO is used."
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index eb3da1a..6764f66 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -178,6 +178,8 @@ enum pci_dev_flags {
>         PCI_DEV_FLAGS_NO_PM_RESET = (__force pci_dev_flags_t) (1 << 7),
>         /* Get VPD from function 0 VPD */
>         PCI_DEV_FLAGS_VPD_REF_F0 = (__force pci_dev_flags_t) (1 << 8),
> +       /* Don't use Relaxed Ordering for TLPs directed at this device */
> +       PCI_DEV_FLAGS_NO_RELAXED_ORDERING = (__force pci_dev_flags_t) (1 << 9),
>  };
>
>  enum pci_irq_reroute_variant {
> --
> 1.9.1
>