On Thu, May 5, 2011 at 1:33 PM, Jesse Barnes <jbarnes@xxxxxxxxxxxxxxxx> wrote:
> Add support to allow drivers to enable/disable ID-based ordering. Where
> supported, ID-based ordering can significantly improve the latency of
> individual requests by preventing them from queueing up behind unrelated
> traffic.
>
> Signed-off-by: Jesse Barnes <jbarnes@xxxxxxxxxxxxxxxx>
> ---
> drivers/pci/pci.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++
> include/linux/pci.h | 13 +++++++++++
> include/linux/pci_regs.h | 2 +
> 3 files changed, 69 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 44d1c7c..44bf8f0 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -1834,6 +1834,60 @@ void pci_enable_ari(struct pci_dev *dev)
> bridge->ari_enabled = 1;
> }
>
> +/**
> + * pci_enable_ido - enable ID-based ordering on a device
> + * @dev: the PCI device
> + * @type: which types of IDO to enable
> + *
> + * Enable ID-based ordering on @dev. @type can contain the bits
> + * %PCI_EXP_IDO_REQUEST and/or %PCI_EXP_IDO_COMPLETION to indicate
> + * which types of transactions are allowed to be re-ordered.
> + */
> +void pci_enable_ido(struct pci_dev *dev, unsigned long type)
> +{
> + int pos;
> + u16 ctrl;
> +
> + if (!pci_is_pcie(dev))
> + return;
> +
> + pos = pci_pcie_cap(dev);
> + if (!pos)
> + return;
Do we gain something by testing !pci_is_pcie() before testing
pci_pcie_cap()? The current pci_is_pcie() implementation is merely:
static inline bool pci_is_pcie(struct pci_dev *dev)
{
return !!pci_pcie_cap(dev);
}
so we could just drop the pci_is_pcie() test. But maybe there's some
abstraction or other benefit to keeping it.
Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
