Hi Arnd, I do not directly add arm64 support to pcie-designware.c, the main reason is that it is hard to port other 5 platform PCIe driver based on pcie-designware to the new implementation. However, not changing those PCIe driver means the code for arm32 must be remained. I add a new file pcie-designware-base.c. Compared with pcie-designware.c, there are main 3 differences. 1. It is a new implementation followed the latest PCIe architecture-independent API without any code for specific architecture. 2. There is no MSI chip code, which will be implemented in a separated file. 3. Improve ATU use - support more than 2 ATUs and use an exclusive ATU to translate MEM transaction. After other platform PCIe owner port the driver to pcie-designware-base.c, The pcie-designware.c can be removed. Thanks, Minghuan > -----Original Message----- > From: Arnd Bergmann [mailto:arnd@xxxxxxxx] > Sent: Wednesday, April 15, 2015 6:37 PM > To: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx > Cc: Lian Minghuan-B31939; linux-pci@xxxxxxxxxxxxxxx; Jingoo Han; Hu > Mingkai-B21284; Zang Roy-R61911; Yoder Stuart-B08248; Bjorn Helgaas; > Wood Scott-B07421 > Subject: Re: [PATCH v2 2/3] pci/designware: Add base driver for Designware > PCIe > > On Wednesday 15 April 2015 17:48:34 Minghuan Lian wrote: > > The Synopsys Designware IP is shared with couples of platforms under > > multiple architectures. The patch is to provide basic > > architecture-independent Designware PCIe host driver including ATU > > initialization and PCI OPS. Currently, which supports arm and arm64 > > simultaneously. > > > > Signed-off-by: Minghuan Lian <Minghuan.Lian@xxxxxxxxxxxxx> > > --- > > change log: > > v1-v2: > > 1. Get MSI chip according to dts property 'msi-parent' > > 2. Simplify platform_get_resource_byname() failure checking > > > > drivers/pci/host/Kconfig | 3 + > > drivers/pci/host/Makefile | 1 + > > drivers/pci/host/pcie-designware-base.c | 293 > > ++++++++++++++++++++++++++++++++ > > drivers/pci/host/pcie-designware-base.h | 63 +++++++ > > > > I don't understand this patch at all. What is the relation between this driver > and the existing pcie-designware.c? From the description above, they would > be the same. > > Arnd -- 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
