On Fri, Jun 15, 2012 at 08:59:37AM -0700, David Daney wrote:
> On 06/15/2012 01:09 AM, Huacai Chen wrote:
> >Loongson-3 is a multi-core MIPS family CPU, it support MIPS64R2
> >fully. Loongson-3 has the same IMP field (0x6300) as Loongson-2.
> >Loongson-3 has a hardware-maintained cache, system software doesn't
> >need to maintain coherency.
> >Loongson-3A is the first revision of Loongson-3, and it is the quad-
> >core version of Loongson-2G. Loongson-3A has a simplified version named
> >Loongson-2Gq, the main difference between Loongson-3A/2Gq is 3A has two
> >HyperTransport controller but 2Gq has only one. HT0 is used for cross-
> >chip interconnection and HT1 is used to link PCI bus. Therefore, 2Gq
> >cannot support NUMA but 3A can.
> >Exsisting Loongson family CPUs:
> >Loongson-1: Loongson-1A, Loongson-1B, they are 32-bit MIPS CPUs.
> >Loongson-2: Loongson-2E, Loongson-2F, Loongson-2G(including Loongson-
> > 2Gq), they are 64-bit MIPS CPUs.
> >Loongson-3: Loongson-3A, it is a 64-bit MIPS CPU.
> >Signed-off-by: Huacai Chen<firstname.lastname@example.org>
> >Signed-off-by: Hongliang Tao<email@example.com>
> >Signed-off-by: Hua Yan<firstname.lastname@example.org>
> > arch/mips/Kconfig | 13 ++++
> > arch/mips/include/asm/addrspace.h | 6 ++
> > arch/mips/include/asm/cpu.h | 6 +-
> > arch/mips/include/asm/mach-loongson/spaces.h | 15 +++++
> > arch/mips/include/asm/module.h | 2 +
> > arch/mips/include/asm/pgtable-bits.h | 7 ++
> > arch/mips/kernel/Makefile | 1 +
> > arch/mips/kernel/cpu-probe.c | 12 +++-
> > arch/mips/lib/Makefile | 1 +
> > arch/mips/loongson/Kconfig | 4 +
> > arch/mips/loongson/Platform | 1 +
> > arch/mips/loongson/common/env.c | 3 +
> > arch/mips/loongson/common/setup.c | 6 +-
> > arch/mips/mm/Makefile | 1 +
> > arch/mips/mm/c-r4k.c | 84
> > ++++++++++++++++++++++++++
> > arch/mips/mm/tlb-r4k.c | 2 +-
> > arch/mips/mm/tlbex.c | 1 +
> > 17 files changed, 156 insertions(+), 9 deletions(-)
> This patch is too big. It should be split up into smaller but
> related parts.
> For example, the parts that add new identifier constants should be
> first. Then a separate patch for cpu-probe.c where they are used.
> > create mode 100644 arch/mips/include/asm/mach-loongson/spaces.h
> >diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig
> >index c179461..38e460b 100644
> >--- a/arch/mips/Kconfig
> >+++ b/arch/mips/Kconfig
> >@@ -1544,6 +1544,16 @@ config CPU_LOONGSON2
> > select CPU_SUPPORTS_64BIT_KERNEL
> > select CPU_SUPPORTS_HIGHMEM
> >+config CPU_LOONGSON3
> >+ bool "Loongson 3 CPU"
> >+ depends on SYS_HAS_CPU_LOONGSON3
> >+ select CPU_SUPPORTS_32BIT_KERNEL
> >+ select CPU_SUPPORTS_64BIT_KERNEL
> >+ select CPU_SUPPORTS_HIGHMEM
> >+ help
> >+ The Loongson 3 processor implements the MIPS III instruction set
> >+ with many extensions.
> This bit must be the very last patch of the entire set, not the first.
> Ask yourself what would happen if someone did a build selecting
> CPU_LOONGSON3 after this patch were applied? Would a runnable
> kernel result? If the answer is no, then you did it in the wrong
Just to clarify and also because I only recently had the discussion with
somebody. What is very important is that a patch series never ever
break something when it gets only partially applied. This is because
all the automated testing and debugging that is being used today. The
most common example is git-bisect but lately aiaiai  is becoming
fashionable, too. So I rather spend some extra time on reviewing a big,
complex patch now than living with the consequences of improper splitting
later. But of course properlz split patches are always preferred :-)