[PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[Paul Burton]

If we built the kernel targeting the microMIPS ISA then the very fact
that the kernel is running implies that the CPU supports microMIPS. Thus
we can hardcode cpu_has_mmips to 1 allowing the compiler greater scope
for optimisation due to the compile-time constant.

Signed-off-by: Paul Burton <paul.burton@mips.com>
---

 arch/mips/include/asm/cpu-features.h | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/mips/include/asm/cpu-features.h b/arch/mips/include/asm/cpu-features.h
index 0edba3e75747..8669fdb503a5 100644
--- a/arch/mips/include/asm/cpu-features.h
+++ b/arch/mips/include/asm/cpu-features.h
@@ -195,7 +195,9 @@
 #endif
 
 #ifndef cpu_has_mmips
-# ifdef CONFIG_SYS_SUPPORTS_MICROMIPS
+# if defined(__mips_micromips)
+#  define cpu_has_mmips		1
+# elif defined(CONFIG_SYS_SUPPORTS_MICROMIPS)
 #  define cpu_has_mmips		__opt(MIPS_CPU_MICROMIPS)
 # else
 #  define cpu_has_mmips		0
-- 
2.19.1

Re: [PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[Paul Burton]

Hello,

Paul Burton wrote:
> If we built the kernel targeting the microMIPS ISA then the very fact
> that the kernel is running implies that the CPU supports microMIPS. Thus
> we can hardcode cpu_has_mmips to 1 allowing the compiler greater scope
> for optimisation due to the compile-time constant.
> 
> Signed-off-by: Paul Burton <paul.burton@mips.com>

Applied to mips-next.

Thanks,
    Paul

[ This message was auto-generated; if you believe anything is incorrect
  then please email paul.burton@mips.com to report it. ]

Re: [PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[Paul Burton]

Hello,

Paul Burton wrote:
> If we built the kernel targeting the microMIPS ISA then the very fact
> that the kernel is running implies that the CPU supports microMIPS. Thus
> we can hardcode cpu_has_mmips to 1 allowing the compiler greater scope
> for optimisation due to the compile-time constant.
> 
> Signed-off-by: Paul Burton <paul.burton@mips.com>

Applied to mips-next.

Thanks,
    Paul

[ This message was auto-generated; if you believe anything is incorrect
  then please email paul.burton@mips.com to report it. ]

Re: [PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[Maciej W. Rozycki]

On Wed, 7 Nov 2018, Paul Burton wrote:

> diff --git a/arch/mips/include/asm/cpu-features.h b/arch/mips/include/asm/cpu-features.h
> index 0edba3e75747..8669fdb503a5 100644
> --- a/arch/mips/include/asm/cpu-features.h
> +++ b/arch/mips/include/asm/cpu-features.h
> @@ -195,7 +195,9 @@
>  #endif
>  
>  #ifndef cpu_has_mmips
> -# ifdef CONFIG_SYS_SUPPORTS_MICROMIPS
> +# if defined(__mips_micromips)

 Wouldn't it be cleaner if it was written:

# if defined(CONFIG_CPU_MICROMIPS)

?

  Maciej

Re: [PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[Paul Burton]

Hi Maciej,

On Wed, Nov 28, 2018 at 03:10:53AM +0000, Maciej W. Rozycki wrote:
> On Wed, 7 Nov 2018, Paul Burton wrote:
> 
> > diff --git a/arch/mips/include/asm/cpu-features.h b/arch/mips/include/asm/cpu-features.h
> > index 0edba3e75747..8669fdb503a5 100644
> > --- a/arch/mips/include/asm/cpu-features.h
> > +++ b/arch/mips/include/asm/cpu-features.h
> > @@ -195,7 +195,9 @@
> >  #endif
> >  
> >  #ifndef cpu_has_mmips
> > -# ifdef CONFIG_SYS_SUPPORTS_MICROMIPS
> > +# if defined(__mips_micromips)
> 
>  Wouldn't it be cleaner if it was written:
> 
> # if defined(CONFIG_CPU_MICROMIPS)

I suppose it's just a matter of preference - in practice both ought to
be defined or undefined at the same times. My personal preference is the
standard macro provided by the compiler, so that's what I used.

Thanks,
    Paul

Re: [PATCH] MIPS: Hardcode cpu_has_mmips=1 for microMIPS kernels

[Maciej W. Rozycki]

Hi Paul,

> > > -# ifdef CONFIG_SYS_SUPPORTS_MICROMIPS
> > > +# if defined(__mips_micromips)
> > 
> >  Wouldn't it be cleaner if it was written:
> > 
> > # if defined(CONFIG_CPU_MICROMIPS)
> 
> I suppose it's just a matter of preference - in practice both ought to
> be defined or undefined at the same times. My personal preference is the
> standard macro provided by the compiler, so that's what I used.

 The minor difference as I see it is that CONFIG_CPU_MICROMIPS is ours 
and the primary knob while `__mips_micromips' is GCC's.  I don't see it as 
a big problem if any at all, it's just my guts feeling.

 Do we have a consistency check between the two macros anywhere?

  Maciej