Re: [PATCH V2] pci: exynos: split into two parts such as Synopsys part and Exynos part

[Kishon Vijay Abraham I <kishon@ti.com>]

Hi,

On Thursday 11 July 2013 12:35 PM, Jingoo Han wrote:
> On Thursday, July 11, 2013 3:40 PM, Kishon Vijay Abraham I wrote:
>> On Thursday 11 July 2013 11:19 AM, Jingoo Han wrote:
>>>
>>>  drivers/pci/host/Makefile          |    1 +
>>>  drivers/pci/host/pcie-designware.c |  963 +++++++++---------------------------
>>>  drivers/pci/host/pcie-designware.h |   71 +++
>>>  drivers/pci/host/pcie-exynos.c     |  523 ++++++++++++++++++++
>>>  4 files changed, 822 insertions(+), 736 deletions(-)
>>>  create mode 100644 drivers/pci/host/pcie-designware.h
>>>  create mode 100644 drivers/pci/host/pcie-exynos.c
>>>
>>> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
>>> index 086d850..7e59864 100644
>>> --- a/drivers/pci/host/Makefile
>>> +++ b/drivers/pci/host/Makefile
>>> @@ -1,2 +1,3 @@
>>>  obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
>>>  obj-$(CONFIG_PCIE_DW) += pcie-designware.o
>>> +obj-$(CONFIG_PCI_EXYNOS) += pcie-exynos.o
>>> diff --git a/drivers/pci/host/pcie-designware.c b/drivers/pci/host/pcie-designware.c
> [...]
>>
>> How about making this a separate driver in itself that does all the
>> configurations for designware core? By this I mean we can have a separate dt
>> node (child node of soc specific wrapper), that will have all the configuration
>> space/IO space and memory space. pci_common_init/dw_pcie_host_init should be
>> done in this driver.
>> We just need to think about a way of passing the ops (since that looks like
>> very much needed because of the sideband bits you have to enable before
>> reading/writing).
>>
> 
> CC'ed Seungwon Jeon(DW-MMC Maintainer), Yulgon Kim (DW-USB Developer)
> 
> 
> Um, maybe you mean dwc3 usb driver (./drivers/usb/dwc3/)'?
> 
> But, I referenced dw mmc driver (./drivers/mmc/host/dw_mmc*.c).
> 
> Now, Exynos PCIe driver, Spear PCIe driver[1], and i.MX PCIe driver[2]
> are submitted to PCIe mailing-list, these are using designware PCIe core.
> 
> There are many differences between Exynos PCIe and Spear PCIe.
> Also, for Exynos PCIe, platform specific part will be changed more.
> 
> Thus, the dw mmc driver model looked more suitable.

Alright. I'll go ahead with a similar implementation then.

Thanks
Kishon