[PATCH v4 5/6] PCI: layerscape: add PCIe support for LS1043a and LS2080a

[Bjorn Helgaas]

[+cc Fabio, Lucas]

Hi Minghuan,

On Fri, Oct 16, 2015 at 03:19:19PM +0800, Minghuan Lian wrote:
> Both LS1043a and LS2080a are based on ARMv8 64-bit architecture
> and have similar PCIe implementation. LUT is added to controller.
> The patch removes the necessary fields from struct ls_pcie.
> 
> Signed-off-by: Minghuan Lian <Minghuan.Lian at freescale.com>
> ---
> This patch is based on v4.3-rc4 and [PATCH v10 3/6]
> PCI: designware: Add ARM64 support.
> 
> change log
> v4: 
> 1. split to 6 patches.
> 2. use ARCH_LAYERSCAPE instead of ARM64
> 
> v3:
> 1. Use 8 or 16 bit access function to simplify code
> 2. Add ls_add_pcie_port in accordance with other DesignWare-based drivers
> 
> v2:
> 1. Rename ls2085a to ls2080a
> 2. Add ls_pcie_msi_host_init()
> 
>  drivers/pci/host/Kconfig          |  2 +-
>  drivers/pci/host/pci-layerscape.c | 72 +++++++++++++++++++++++++++++++++++----
>  2 files changed, 67 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> index ae873be..8eb09ea 100644
> --- a/drivers/pci/host/Kconfig
> +++ b/drivers/pci/host/Kconfig
> @@ -105,7 +105,7 @@ config PCI_XGENE_MSI
>  
>  config PCI_LAYERSCAPE
>  	bool "Freescale Layerscape PCIe controller"
> -	depends on OF && ARM
> +	depends on OF && (ARM || ARCH_LAYERSCAPE)
>  	select PCIE_DW
>  	select MFD_SYSCON
>  	help
> diff --git a/drivers/pci/host/pci-layerscape.c b/drivers/pci/host/pci-layerscape.c
> index 891e504..c53692a 100644
> --- a/drivers/pci/host/pci-layerscape.c
> +++ b/drivers/pci/host/pci-layerscape.c
> @@ -31,23 +31,26 @@
>  #define LTSSM_STATE_MASK	0x3f

BTW, not related to *this* patch, but does LTSSM_STATE_MASK really need to
be 0x3f (6 bits), or could it be 0x1f (5 bits)?

I'd like to include Layerscape in the LTSSM_STATE_MASK cleanup done by
Fabio:

https://git.kernel.org/cgit/linux/kernel/git/helgaas/pci.git/commit/?h=pci/host-designware&id=4788fe6ebf4594c9a95b620cbff05147c8504823

I don't have specs for any of these devices, so I don't know if this is
really something that can vary between the different DesignWare-based
devices, or if they all should use a mask of 0x1f.

Bjorn