On Tue, 20 Jun 2023, Maciej W. Rozycki wrote:
> Call pci_parent_bus_reset() rather than duplicating it in trigger_sbr().
> There are extra preparatory checks made by the former function, but they
> are supposed to be neutral to the HFI1 device.
>
> Signed-off-by: Maciej W. Rozycki <macro@xxxxxxxxxxx>
> ---
> drivers/infiniband/hw/hfi1/pcie.c | 30 ++++--------------------------
> 1 file changed, 4 insertions(+), 26 deletions(-)
>
> linux-ib-hfi1-pcie-sbr-parent-bus-reset.diff
> Index: linux-macro/drivers/infiniband/hw/hfi1/pcie.c
> ===================================================================
> --- linux-macro.orig/drivers/infiniband/hw/hfi1/pcie.c
> +++ linux-macro/drivers/infiniband/hw/hfi1/pcie.c
> @@ -796,35 +796,13 @@ static void pcie_post_steps(struct hfi1_
> /*
> * Trigger a secondary bus reset (SBR) on ourselves using our parent.
> *
> - * Based on pci_parent_bus_reset() which is not exported by the
> - * kernel core.
> + * This is an end around to do an SBR during probe time. A new API
> + * needs to be implemented to have cleaner interface but this fixes
> + * the current brokenness.
> */
> static int trigger_sbr(struct hfi1_devdata *dd)
> {
> - struct pci_dev *dev = dd->pcidev;
> - struct pci_dev *pdev;
> -
> - /* need a parent */
> - if (!dev->bus->self) {
> - dd_dev_err(dd, "%s: no parent device\n", __func__);
> - return -ENOTTY;
> - }
> -
> - /* should not be anyone else on the bus */
> - list_for_each_entry(pdev, &dev->bus->devices, bus_list)
> - if (pdev != dev) {
> - dd_dev_err(dd,
> - "%s: another device is on the same bus\n",
> - __func__);
> - return -ENOTTY;
> - }
> -
> - /*
> - * This is an end around to do an SBR during probe time. A new API needs
> - * to be implemented to have cleaner interface but this fixes the
> - * current brokenness
> - */
> - return pci_bridge_secondary_bus_reset(dev->bus->self);
> + return pci_parent_bus_reset(dd->pcidev, PCI_RESET_DO_RESET);
> }
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
--
i.
