[PATCH v4] arch/riscv: add Zihintpause support

[Heiko Stuebner]

Am Montag, 11. Juli 2022, 22:44:21 CEST schrieb Conor Dooley:
> On 20/06/2022 21:15, Dao Lu wrote:
> > [PATCH v4] arch/riscv: add Zihintpause support
> 
> A little ornery/pedantic maybe, but the "arch/" isn't needed.
> Guess it can easily be fixed on application or if there's a
> v5 so there's prob no need to resend.

I noticed that the patch prefix on riscv is all over the place
in a lot of cases. There are

- RISC-V:
-riscv:
- now arch/riscv:

and probably even more.

I guess someone should just decide on one prefix that then
gets used (and slightly enforced) all the time.

Even just modifying applied patches to one specific prefix should
already solve the issue in a short time, as I guess most people will
just get the appicable prefix from a "git log" ;-) .


Heiko

> > Implement support for the ZiHintPause extension.
> > 
> > The PAUSE instruction is a HINT that indicates the current hart’s rate
> > of instruction retirement should be temporarily reduced or paused.
> > 
> > Reviewed-by: Heiko Stuebner <heiko at sntech.de>
> > Tested-by: Heiko Stuebner <heiko at sntech.de>
> > Signed-off-by: Dao Lu <daolu at rivosinc.com>
> > ---
> > 
> > v1 -> v2:
> >  Remove the usage of static branch, use PAUSE if toolchain supports it
> > v2 -> v3:
> >  Added the static branch back, cpu_relax() behavior is kept the same for
> > systems that do not support ZiHintPause
> > v3 -> v4:
> >  Adopted the newly added unified static keys for extensions
> > ---
> >  arch/riscv/Makefile                     |  4 ++++
> >  arch/riscv/include/asm/hwcap.h          |  5 +++++
> >  arch/riscv/include/asm/vdso/processor.h | 21 ++++++++++++++++++---
> >  arch/riscv/kernel/cpu.c                 |  1 +
> >  arch/riscv/kernel/cpufeature.c          |  1 +
> >  5 files changed, 29 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/riscv/Makefile b/arch/riscv/Makefile
> > index 34cf8a598617..6ddacc6f44b9 100644
> > --- a/arch/riscv/Makefile
> > +++ b/arch/riscv/Makefile
> > @@ -56,6 +56,10 @@ riscv-march-$(CONFIG_RISCV_ISA_C)	:= $(riscv-march-y)c
> >  toolchain-need-zicsr-zifencei := $(call cc-option-yn, -march=$(riscv-march-y)_zicsr_zifencei)
> >  riscv-march-$(toolchain-need-zicsr-zifencei) := $(riscv-march-y)_zicsr_zifencei
> >  
> > +# Check if the toolchain supports Zihintpause extension
> > +toolchain-supports-zihintpause := $(call cc-option-yn, -march=$(riscv-march-y)_zihintpause)
> > +riscv-march-$(toolchain-supports-zihintpause) := $(riscv-march-y)_zihintpause
> > +
> >  KBUILD_CFLAGS += -march=$(subst fd,,$(riscv-march-y))
> >  KBUILD_AFLAGS += -march=$(riscv-march-y)
> >  
> > diff --git a/arch/riscv/include/asm/hwcap.h b/arch/riscv/include/asm/hwcap.h
> > index e48eebdd2631..dc47019a0b38 100644
> > --- a/arch/riscv/include/asm/hwcap.h
> > +++ b/arch/riscv/include/asm/hwcap.h
> > @@ -8,6 +8,7 @@
> >  #ifndef _ASM_RISCV_HWCAP_H
> >  #define _ASM_RISCV_HWCAP_H
> >  
> > +#include <asm/errno.h>
> >  #include <linux/bits.h>
> >  #include <uapi/asm/hwcap.h>
> >  
> > @@ -54,6 +55,7 @@ extern unsigned long elf_hwcap;
> >  enum riscv_isa_ext_id {
> >  	RISCV_ISA_EXT_SSCOFPMF = RISCV_ISA_EXT_BASE,
> >  	RISCV_ISA_EXT_SVPBMT,
> > +	RISCV_ISA_EXT_ZIHINTPAUSE,
> >  	RISCV_ISA_EXT_ID_MAX = RISCV_ISA_EXT_MAX,
> >  };
> >  
> > @@ -64,6 +66,7 @@ enum riscv_isa_ext_id {
> >   */
> >  enum riscv_isa_ext_key {
> >  	RISCV_ISA_EXT_KEY_FPU,		/* For 'F' and 'D' */
> > +	RISCV_ISA_EXT_KEY_ZIHINTPAUSE,
> >  	RISCV_ISA_EXT_KEY_MAX,
> >  };
> >  
> > @@ -83,6 +86,8 @@ static __always_inline int riscv_isa_ext2key(int num)
> >  		return RISCV_ISA_EXT_KEY_FPU;
> >  	case RISCV_ISA_EXT_d:
> >  		return RISCV_ISA_EXT_KEY_FPU;
> > +	case RISCV_ISA_EXT_ZIHINTPAUSE:
> > +		return RISCV_ISA_EXT_KEY_ZIHINTPAUSE;
> >  	default:
> >  		return -EINVAL;
> >  	}
> > diff --git a/arch/riscv/include/asm/vdso/processor.h b/arch/riscv/include/asm/vdso/processor.h
> > index 134388cbaaa1..1e4f8b4aef79 100644
> > --- a/arch/riscv/include/asm/vdso/processor.h
> > +++ b/arch/riscv/include/asm/vdso/processor.h
> > @@ -4,15 +4,30 @@
> >  
> >  #ifndef __ASSEMBLY__
> >  
> > +#include <linux/jump_label.h>
> >  #include <asm/barrier.h>
> > +#include <asm/hwcap.h>
> >  
> >  static inline void cpu_relax(void)
> >  {
> > +	if (!static_branch_likely(&riscv_isa_ext_keys[RISCV_ISA_EXT_KEY_ZIHINTPAUSE])) {
> >  #ifdef __riscv_muldiv
> > -	int dummy;
> > -	/* In lieu of a halt instruction, induce a long-latency stall. */
> > -	__asm__ __volatile__ ("div %0, %0, zero" : "=r" (dummy));
> > +		int dummy;
> > +		/* In lieu of a halt instruction, induce a long-latency stall. */
> > +		__asm__ __volatile__ ("div %0, %0, zero" : "=r" (dummy));
> >  #endif
> > +	} else {
> > +		/*
> > +		 * Reduce instruction retirement.
> > +		 * This assumes the PC changes.
> > +		 */
> > +#ifdef __riscv_zihintpause
> > +		__asm__ __volatile__ ("pause");
> > +#else
> > +		/* Encoding of the pause instruction */
> > +		__asm__ __volatile__ (".4byte 0x100000F");
> > +#endif
> > +	}
> >  	barrier();
> >  }
> >  
> > diff --git a/arch/riscv/kernel/cpu.c b/arch/riscv/kernel/cpu.c
> > index fba9e9f46a8c..a123e92b14dd 100644
> > --- a/arch/riscv/kernel/cpu.c
> > +++ b/arch/riscv/kernel/cpu.c
> > @@ -89,6 +89,7 @@ int riscv_of_parent_hartid(struct device_node *node)
> >  static struct riscv_isa_ext_data isa_ext_arr[] = {
> >  	__RISCV_ISA_EXT_DATA(sscofpmf, RISCV_ISA_EXT_SSCOFPMF),
> >  	__RISCV_ISA_EXT_DATA(svpbmt, RISCV_ISA_EXT_SVPBMT),
> > +	__RISCV_ISA_EXT_DATA(zihintpause, RISCV_ISA_EXT_ZIHINTPAUSE),
> >  	__RISCV_ISA_EXT_DATA("", RISCV_ISA_EXT_MAX),
> >  };
> >  
> > diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/cpufeature.c
> > index 1b3ec44e25f5..708df2c0bc34 100644
> > --- a/arch/riscv/kernel/cpufeature.c
> > +++ b/arch/riscv/kernel/cpufeature.c
> > @@ -198,6 +198,7 @@ void __init riscv_fill_hwcap(void)
> >  			} else {
> >  				SET_ISA_EXT_MAP("sscofpmf", RISCV_ISA_EXT_SSCOFPMF);
> >  				SET_ISA_EXT_MAP("svpbmt", RISCV_ISA_EXT_SVPBMT);
> > +				SET_ISA_EXT_MAP("zihintpause", RISCV_ISA_EXT_ZIHINTPAUSE);
> >  			}
> >  #undef SET_ISA_EXT_MAP
> >  		}
>