[PATCH v2 3/8] riscv: Add PLATFORM_SUPPORTS_RISCV_ISA_SVNAPOT Kconfig option

[Charlie Jenkins]

On Wed, May 08, 2024 at 10:00:48AM +0100, Ben Dooks wrote:
> On 08/05/2024 02:36, Charlie Jenkins wrote:
> > The existing "RISCV_ISA_SVNAPOT" option is repurposed to be used to by
> > kernel code to determine if either
> > PLATFORM_MAY_SUPPORT_RISCV_ISA_SVNAPOT or
> > PLATFORM_SUPPORTS_RISCV_ISA_SVNAPOT has been set.
> > 
> > PLATFORM_MAY_SUPPORT_RISCV_ISA_SVNAPOT will check if the hardware
> > supports Svnapot before using it, while
> > PLATFORM_SUPPORTS_RISCV_ISA_SVNAPOT will assume that the hardware
> > supports Svnapot.
> > 
> > Signed-off-by: Charlie Jenkins <charlie at rivosinc.com>
> > ---
> >   arch/riscv/Kconfig               | 19 -----------------
> >   arch/riscv/Kconfig.isa           | 44 ++++++++++++++++++++++++++++++++++++++++
> >   arch/riscv/include/asm/pgtable.h |  3 ++-
> >   3 files changed, 46 insertions(+), 20 deletions(-)
> > 
> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> > index c2e9eded0a7d..3c1960e8cd7c 100644
> > --- a/arch/riscv/Kconfig
> > +++ b/arch/riscv/Kconfig
> > @@ -484,25 +484,6 @@ config RISCV_ALTERNATIVE_EARLY
> >   	help
> >   	  Allows early patching of the kernel for special errata
> > -config RISCV_ISA_SVNAPOT
> > -	bool "Svnapot extension support for supervisor mode NAPOT pages"
> > -	depends on 64BIT && MMU
> > -	depends on RISCV_ALTERNATIVE
> > -	default y
> > -	help
> > -	  Add support for the Svnapot ISA-extension in the kernel when it
> > -	  is detected at boot.
> > -
> > -	  The Svnapot extension is used to mark contiguous PTEs as a range
> > -	  of contiguous virtual-to-physical translations for a naturally
> > -	  aligned power-of-2 (NAPOT) granularity larger than the base 4KB page
> > -	  size. When HUGETLBFS is also selected this option unconditionally
> > -	  allocates some memory for each NAPOT page size supported by the kernel.
> > -	  When optimizing for low memory consumption and for platforms without
> > -	  the Svnapot extension, it may be better to say N here.
> > -
> > -	  If you don't know what to do here, say Y.
> > -
> >   config RISCV_ISA_SVPBMT
> >   	bool "Svpbmt extension support for supervisor mode page-based memory types"
> >   	depends on 64BIT && MMU
> > diff --git a/arch/riscv/Kconfig.isa b/arch/riscv/Kconfig.isa
> > index 0663c98b5b17..37585bcd763e 100644
> > --- a/arch/riscv/Kconfig.isa
> > +++ b/arch/riscv/Kconfig.isa
> > @@ -124,3 +124,47 @@ config RISCV_ISA_V_PREEMPTIVE
> >   	  This config allows kernel to run SIMD without explicitly disable
> >   	  preemption. Enabling this config will result in higher memory
> >   	  consumption due to the allocation of per-task's kernel Vector context.
> > +
> > +config RISCV_ISA_SVNAPOT
> > +	bool
> > +
> > +choice
> > +	prompt "Svnapot extension support for supervisor mode NAPOT pages"
> > +	default PLATFORM_MAY_SUPPORT_RISCV_ISA_SVNAPOT
> > +	help
> > +	  This selects the level of support for Svnapot in the Linux Kernel.
> > +
> > +	  The Svnapot extension is used to mark contiguous PTEs as a range
> > +	  of contiguous virtual-to-physical translations for a naturally
> > +	  aligned power-of-2 (NAPOT) granularity larger than the base 4KB page
> > +	  size. When HUGETLBFS is also selected this option unconditionally
> > +	  allocates some memory for each NAPOT page size supported by the kernel.
> > +	  When optimizing for low memory consumption and for platforms without
> > +	  the Svnapot extension, it may be better to prohibit Svnapot.
> > +
> > +config PROHIBIT_RISCV_ISA_SVNAPOT
> > +	bool "Prohibit Svnapot extension"
> > +	help
> > +	  Regardless of if the platform supports Svnapot, prohibit the kernel
> > +	  from using Svnapot.
> > +
> > +config PLATFORM_MAY_SUPPORT_RISCV_ISA_SVNAPOT
> > +	bool "Allow Svnapot extension if supported"
> > +	depends on 64BIT && MMU
> > +	depends on RISCV_ALTERNATIVE
> > +	select RISCV_ISA_SVNAPOT
> > +	help
> > +	  Add support for the Svnapot ISA-extension in the kernel when it
> > +	  is detected at boot.
> > +
> > +config PLATFORM_SUPPORTS_RISCV_ISA_SVNAPOT
> > +	bool "Emit Svnapot mappings when building Linux"
> > +	depends on 64BIT && MMU
> > +	depends on NONPORTABLE
> > +	select RISCV_ISA_SVNAPOT
> > +	help
> > +	  Compile a kernel that assumes that the platform supports Svnapot.
> > +	  This option produces a kernel that will not run on systems that do
> > +	  not support Svnapot.
> > +
> > +endchoice
> > diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
> > index 6afd6bb4882e..432be9691b78 100644
> > --- a/arch/riscv/include/asm/pgtable.h
> > +++ b/arch/riscv/include/asm/pgtable.h
> > @@ -289,7 +289,8 @@ static inline pte_t pud_pte(pud_t pud)
> >   static __always_inline bool has_svnapot(void)
> >   {
> > -	return riscv_has_extension_likely(RISCV_ISA_EXT_SVNAPOT);
> > +	return IS_ENABLED(CONFIG_PLATFORM_SUPPORTS_RISCV_ISA_SVNAPOT) ||
> > +	       riscv_has_extension_likely(RISCV_ISA_EXT_SVNAPOT);
> 
> could you add the IS_ENABLED(*) check into riscv_has_extension_likely
> and other such functions?

I wasn't sure how to support that. An option I was debating about this
was fixing up riscv_has_extension_likely() so that it's a macro and
SVNAPOT could be expanded to both
CONFIG_PLATFORM_SUPPORTS_RISCV_ISA_SVNAPOT and RISCV_ISA_EXT_SVNAPOT.

- Charlie

> 
> 
> -- 
> Ben Dooks				http://www.codethink.co.uk/
> Senior Engineer				Codethink - Providing Genius
> 
> https://www.codethink.co.uk/privacy.html
>