[PATCH] pci: host: xgene: fix incorrectly returned address by map_bus

[Bjorn Helgaas]

On Tue, Feb 17, 2015 at 03:14:00PM -0800, Feng Kan wrote:
> The generic accessor functions for pci-xgene uses map_bus
> call that returns the base address but did not add the additional
> offset.
> 
> Signed-off-by: Feng Kan <fkan at apm.com>
> ---
>  drivers/pci/host/pci-xgene.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/host/pci-xgene.c b/drivers/pci/host/pci-xgene.c
> index aab5547..ee082c0 100644
> --- a/drivers/pci/host/pci-xgene.c
> +++ b/drivers/pci/host/pci-xgene.c
> @@ -127,7 +127,7 @@ static bool xgene_pcie_hide_rc_bars(struct pci_bus *bus, int offset)
>  	return false;
>  }
>  
> -static int xgene_pcie_map_bus(struct pci_bus *bus, unsigned int devfn,
> +static void __iomem *xgene_pcie_map_bus(struct pci_bus *bus, unsigned int devfn,
>  			      int offset)
>  {
>  	struct xgene_pcie_port *port = bus->sysdata;
> @@ -137,7 +137,7 @@ static int xgene_pcie_map_bus(struct pci_bus *bus, unsigned int devfn,
>  		return NULL;
>  
>  	xgene_pcie_set_rtdid_reg(bus, devfn);
> -	return xgene_pcie_get_cfg_base(bus);
> +	return xgene_pcie_get_cfg_base(bus) + offset;

Where's the locking here?  ECAM doesn't need locking because the
bus/dev/fn/offset is all encoded in the MMIO address.  But it looks
like X-Gene doesn't work that way and bus/dev/fn is in the RTDID register.

So it seems like X-Gene needs locking that not everybody needs.  Are you
relying on higher-level locking somewhere?

>  }
>  
>  static struct pci_ops xgene_pcie_ops = {
> -- 
> 1.9.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo at vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html