| Message ID | 1541728209-3224-1-git-send-email-atish.patra@wdc.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Unify CPU topology across ARM64 & RISC-V | expand |
On 11/8/2018 6:50 PM, Atish Patra wrote: > The cpu-map DT entry in ARM64 can describe the CPU topology in > much better way compared to other existing approaches. RISC-V can > easily adopt this binding to represent it's own CPU topology. > Thus, both cpu-map DT binding and topology parsing code can be > moved to a common location so that RISC-V or any other > architecture can leverage that. > > The relevant discussion regarding unifying cpu topology can be > found in [1]. > > arch_topology seems to be a perfect place to move the common > code. I have not introduced any functional changes in the moved > to code. The only downside in this approach is that the capacity > code will be executed for RISC-V as well. But, it will exit > immediately after not able to find the appropriate DT node. If > the overhead is considered too much, we can always compile out > capacity related functions under a different config for the > architectures that do not support them. > > The patches have been tested for RISC-V and compile tested for > ARM64. > > The socket changes[2] can be merged on top of this series or vice > versa. > > [1] https://lkml.org/lkml/2018/11/6/19 > [2] https://lkml.org/lkml/2018/11/7/918 > > Atish Patra (3): > dt-binding: cpu-topology: Move cpu-map to a common binding. > cpu-topology: Move cpu topology code to common code. > RISC-V: Parse cpu topology during boot. > > Documentation/devicetree/bindings/arm/topology.txt | 475 ------------------- > .../devicetree/bindings/cpu/cpu-topology.txt | 526 +++++++++++++++++++++ > arch/arm64/include/asm/topology.h | 23 +- > arch/arm64/kernel/topology.c | 305 +----------- > arch/riscv/Kconfig | 1 + > arch/riscv/kernel/smpboot.c | 6 +- > drivers/base/arch_topology.c | 303 ++++++++++++ > include/linux/arch_topology.h | 23 + > include/linux/topology.h | 1 + > 9 files changed, 864 insertions(+), 799 deletions(-) > delete mode 100644 Documentation/devicetree/bindings/arm/topology.txt > create mode 100644 Documentation/devicetree/bindings/cpu/cpu-topology.txt > I was interested in testing these on QDF2400, an ARM64 platform, since this series touches core ARM64 code and I'd hate to see a regression. However, I can't figure out what baseline to use to apply these. Different patches cause different conflicts of a variety of baselines I attempted. What are these intended to apply to? Also, you might want to run them through checkpatch next time. There are several whitespace errors.
On 11/15/18 10:31 AM, Jeffrey Hugo wrote: > On 11/8/2018 6:50 PM, Atish Patra wrote: >> The cpu-map DT entry in ARM64 can describe the CPU topology in >> much better way compared to other existing approaches. RISC-V can >> easily adopt this binding to represent it's own CPU topology. >> Thus, both cpu-map DT binding and topology parsing code can be >> moved to a common location so that RISC-V or any other >> architecture can leverage that. >> >> The relevant discussion regarding unifying cpu topology can be >> found in [1]. >> >> arch_topology seems to be a perfect place to move the common >> code. I have not introduced any functional changes in the moved >> to code. The only downside in this approach is that the capacity >> code will be executed for RISC-V as well. But, it will exit >> immediately after not able to find the appropriate DT node. If >> the overhead is considered too much, we can always compile out >> capacity related functions under a different config for the >> architectures that do not support them. >> >> The patches have been tested for RISC-V and compile tested for >> ARM64. >> >> The socket changes[2] can be merged on top of this series or vice >> versa. >> >> [1] https://lkml.org/lkml/2018/11/6/19 >> [2] https://lkml.org/lkml/2018/11/7/918 >> >> Atish Patra (3): >> dt-binding: cpu-topology: Move cpu-map to a common binding. >> cpu-topology: Move cpu topology code to common code. >> RISC-V: Parse cpu topology during boot. >> >> Documentation/devicetree/bindings/arm/topology.txt | 475 ------------------- >> .../devicetree/bindings/cpu/cpu-topology.txt | 526 +++++++++++++++++++++ >> arch/arm64/include/asm/topology.h | 23 +- >> arch/arm64/kernel/topology.c | 305 +----------- >> arch/riscv/Kconfig | 1 + >> arch/riscv/kernel/smpboot.c | 6 +- >> drivers/base/arch_topology.c | 303 ++++++++++++ >> include/linux/arch_topology.h | 23 + >> include/linux/topology.h | 1 + >> 9 files changed, 864 insertions(+), 799 deletions(-) >> delete mode 100644 Documentation/devicetree/bindings/arm/topology.txt >> create mode 100644 Documentation/devicetree/bindings/cpu/cpu-topology.txt >> > > I was interested in testing these on QDF2400, an ARM64 platform, since > this series touches core ARM64 code and I'd hate to see a regression. > However, I can't figure out what baseline to use to apply these. > Different patches cause different conflicts of a variety of baselines I > attempted. > > What are these intended to apply to? > I had rebased them on top of 4.20-rc1. > Also, you might want to run them through checkpatch next time. There > are several whitespace errors. > Sorry I missed couple of them. Thanks for trying to test the patches. I will send a next version as Rob suggested. Please test that. Regards, Atish
On Thu, Nov 15, 2018 at 11:31:33AM -0700, Jeffrey Hugo wrote: [...] > > I was interested in testing these on QDF2400, an ARM64 platform, since this > series touches core ARM64 code and I'd hate to see a regression. However, I > can't figure out what baseline to use to apply these. Different patches > cause different conflicts of a variety of baselines I attempted. > Good to know that we can test DT configuration on QDF2400. I always assumed it's ACPI only. > What are these intended to apply to? > The series alone may not get the package/socket ids correct on QDF2400. I have not yet added support for the same as I wanted to get the initial feedback on DT bindings. The movement of DT binding and corresponding code should not regress and you should be able to validate only that part. -- Regards, Sudeep
On 11/20/2018 4:11 AM, Sudeep Holla wrote: > On Thu, Nov 15, 2018 at 11:31:33AM -0700, Jeffrey Hugo wrote: > > [...] > >> >> I was interested in testing these on QDF2400, an ARM64 platform, since this >> series touches core ARM64 code and I'd hate to see a regression. However, I >> can't figure out what baseline to use to apply these. Different patches >> cause different conflicts of a variety of baselines I attempted. >> > > Good to know that we can test DT configuration on QDF2400. I always assumed > it's ACPI only. It is ACPI only in the production configuration. I suppose we could hack things up to do basic DT sanity, but I expect it would be nasty and non-trivial. > >> What are these intended to apply to? >> > > The series alone may not get the package/socket ids correct on QDF2400. > I have not yet added support for the same as I wanted to get the initial > feedback on DT bindings. The movement of DT binding and corresponding > code should not regress and you should be able to validate only that > part. > On a cursory glance, it looks like some of the reorganized code would also be used in the ACPI path (things that are common between DT and ACPI). I do not expect problems, but I still feel its prudent to do a sanity check on actual hardware.