[PATCH v2 4/5] riscv: RISCV_ISA_SV48 and RISCV_ISA_SV57 under NONPORTABLE
Jisheng Zhang
jszhang at kernel.org
Mon Sep 7 20:49:10 PDT 2026
On Mon, Sep 07, 2026 at 05:21:11PM +0100, Conor Dooley wrote:
> On Mon, Sep 07, 2026 at 11:14:36PM +0800, Jisheng Zhang wrote:
> > In last commit, we have optimized the pgtable_l4|l5_enabled with isa
> > extension alternative mechanism, thus a typical access_ok(addr, 1);
> > is compiled as:
> >
> > ffffffff8001dd30: 1141 addi sp,sp,-16
> > ffffffff8001dd32: e022 sd s0,0(sp)
> > ffffffff8001dd34: e406 sd ra,8(sp)
> > ffffffff8001dd36: 0800 addi s0,sp,16
> > ffffffff8001dd38: 00000013 nop
> > ffffffff8001dd3c: 02a0006f j ffffffff8001dd66 <foo+0x3e>
> > ffffffff8001dd40: 60a2 ld ra,8(sp)
> > ffffffff8001dd42: 6402 ld s0,0(sp)
> > ffffffff8001dd44: 57fd li a5,-1
> > ffffffff8001dd46: 83a1 srli a5,a5,0x8
> > ffffffff8001dd48: 00a7b533 sltu a0,a5,a0
> > ffffffff8001dd4c: 00154513 xori a0,a0,1
> > ffffffff8001dd50: 0141 addi sp,sp,16
> > ffffffff8001dd52: 8082 ret
> > ...
> > ffffffff8001dd64: bfe1 j ffffffff8001dd3c <foo+0x14>
> > ffffffff8001dd66: 0180006f j ffffffff8001dd7e <foo+0x56>
> > ffffffff8001dd6a: 60a2 ld ra,8(sp)
> > ffffffff8001dd6c: 6402 ld s0,0(sp)
> > ffffffff8001dd6e: 57fd li a5,-1
> > ffffffff8001dd70: 83c5 srli a5,a5,0x11
> > ffffffff8001dd72: 00a7b533 sltu a0,a5,a0
> > ffffffff8001dd76: 00154513 xori a0,a0,1
> > ffffffff8001dd7a: 0141 addi sp,sp,16
> > ffffffff8001dd7c: 8082 ret
> > ffffffff8001dd7e: 60a2 ld ra,8(sp)
> > ffffffff8001dd80: 6402 ld s0,0(sp)
> > ffffffff8001dd82: 57fd li a5,-1
> > ffffffff8001dd84: 83e9 srli a5,a5,0x1a
> > ffffffff8001dd86: 00a7b533 sltu a0,a5,a0
> > ffffffff8001dd8a: 00154513 xori a0,a0,1
> > ffffffff8001dd8e: 0141 addi sp,sp,16
> > ffffffff8001dd90: 8082 ret
> >
> > As can be seen, different branches for SV39/SV48/SV57 are still there,
> > since access_ok() sits at hot code path, why not reduce the code size
> > to optimize to make the instruction cache happy?
> >
> > Introduce RISCV_ISA_SV48 and RISCV_ISA_SV57 under NONPORTABLE, so that
> > the embedded platforms can choose the best option themselves, while
> > still keep the feature of unified one kernel Image for all SV39, SV48
> > and SV57.
> >
> > Before the patch, vmlinux built with pure RISCV64 defconfig
> > text data bss dec hex filename
> > 13439437 6790446 483221 20713104 13c0e90 /tmp/old/vmlinux
> >
> > After the patch, vmlinux built with pure RISCV64 defconfig:
> > text data bss dec hex filename
> > 13314445 6787934 466837 20569216 139dc80 /tmp/new/vmlinux
> >
> > .text section is reduced by 122KB!
> >
> > And after the patch, typical access_ok(addr, 1) is compiled as:
> > ffffffff8001809a: 1141 addi sp,sp,-16
> > ffffffff8001809c: e022 sd s0,0(sp)
> > ffffffff8001809e: e406 sd ra,8(sp)
> > ffffffff800180a0: 0800 addi s0,sp,16
> > ffffffff800180a2: 00000013 nop
> > ffffffff800180a6: 60a2 ld ra,8(sp)
> > ffffffff800180a8: 6402 ld s0,0(sp)
> > ffffffff800180aa: 4785 li a5,1
> > ffffffff800180ac: 179a slli a5,a5,0x26
> > ffffffff800180ae: 00f53533 sltu a0,a0,a5
> > ffffffff800180b2: 0141 addi sp,sp,16
> > ffffffff800180b4: 8082 ret
> >
> > only 12 instruction!! For refernece, before the patch, the access(addr,
> > 1) needs 32 instructions.
> >
> > Signed-off-by: Jisheng Zhang <jszhang at kernel.org>
> > ---
> > arch/riscv/Kconfig | 24 ++++++++++++++++++++++++
> > arch/riscv/include/asm/pgtable-64.h | 12 ++++++++++++
> > arch/riscv/mm/init.c | 10 ++++++++++
> > 3 files changed, 46 insertions(+)
> >
> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> > index e9476b8cbeb0..4cdb1e3f34be 100644
> > --- a/arch/riscv/Kconfig
> > +++ b/arch/riscv/Kconfig
> > @@ -602,6 +602,30 @@ config RISCV_ISA_SSQOSID
> > cache partitioning scheme and use the MCID to track how much
> > cache a process, or a group of processes, is using.
> >
> > +config RISCV_ISA_SV48
> > + bool "Page-Based 48-bit Virtual-Memory System"
> > + depends on 64BIT && NONPORTABLE
> > + default y
>
> I don't understand how you've got this set up here.
> If this is genuinely non-portable, then you should not set the default to
> enabled. However, I'm not really sure that this is actually non-portable
> at all....
>
> > + help
> > + Add support for the Page-Based 48-bit Virtual-Memory System.
> > +
> > + Enable this option for systems for which a 39-bit virtual address
> > + space is insufficient. But if 39-bit is enough, disabling this
> > + option will result in smaller kernel size and slighly better
> > + performance.
> > +
> > +config RISCV_ISA_SV57
> > + bool "Page-Based 57-bit Virtual-Memory System"
> > + depends on 64BIT && NONPORTABLE
> > + default y
> > + help
> > + Add support for the Page-Based 57-bit Virtual-Memory System.
> > +
> > + Enable this option for systems for which a 48-bit virtual address
> > + space is insufficient. But if 48-bit is enough, disabling this
> > + option will result in smaller kernel size and slighly better
> > + performance.
> > +
> > config RISCV_ISA_SVPBMT
> > bool "Svpbmt extension support for supervisor mode page-based memory types"
> > depends on 64BIT && MMU
> > diff --git a/arch/riscv/include/asm/pgtable-64.h b/arch/riscv/include/asm/pgtable-64.h
> > index 7e30ec0e9f6c..f2d29da93eea 100644
> > --- a/arch/riscv/include/asm/pgtable-64.h
> > +++ b/arch/riscv/include/asm/pgtable-64.h
> > @@ -16,21 +16,33 @@ extern bool _pgtable_l5_enabled;
> > #ifdef USE_EARLY_PGTABLE_LEVELS
> > static __always_inline bool pgtable_l5_enabled(void)
> > {
> > + if (IS_ENABLED(CONFIG_NONPORTABLE) && !IS_ENABLED(CONFIG_RISCV_ISA_SV57))
> > + return false;
> > +
> > return _pgtable_l5_enabled;
> > }
> >
> > static __always_inline bool pgtable_l4_enabled(void)
> > {
> > + if (IS_ENABLED(CONFIG_NONPORTABLE) && !IS_ENABLED(CONFIG_RISCV_ISA_SV48))
> > + return false;
> > +
> > return _pgtable_l4_enabled;
> > }
> > #else
> > static __always_inline bool pgtable_l4_enabled(void)
> > {
> > + if (IS_ENABLED(CONFIG_NONPORTABLE) && !IS_ENABLED(CONFIG_RISCV_ISA_SV48))
> > + return false;
> > +
> > return riscv_has_extension_likely(RISCV_ISA_EXT_SV48);
> > }
> >
> > static __always_inline bool pgtable_l5_enabled(void)
> > {
> > + if (IS_ENABLED(CONFIG_NONPORTABLE) && !IS_ENABLED(CONFIG_RISCV_ISA_SV57))
> > + return false;
> > +
> > return riscv_has_extension_likely(RISCV_ISA_EXT_SV57);
> > }
> > #endif
>
> ...because the code here won't enable sv57 etc if not supported...
>
> > diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
> > index 2a2402d79bc4..90bb85c5f167 100644
> > --- a/arch/riscv/mm/init.c
> > +++ b/arch/riscv/mm/init.c
> > @@ -844,6 +844,16 @@ static __init void set_satp_mode(uintptr_t dtb_pa)
> >
> > kernel_map.page_offset = PAGE_OFFSET_L5;
> >
> > + /* Fall back to Sv48 if Sv57 is unsupported on non-portable systems */
> > + if (IS_ENABLED(CONFIG_NONPORTABLE) && !IS_ENABLED(CONFIG_RISCV_ISA_SV57) &&
> > + satp_mode_limit == 0)
> > + satp_mode_limit = SATP_MODE_48;
>
> ...and neither does the code here.
>
> > Introduce RISCV_ISA_SV48 and RISCV_ISA_SV57 under NONPORTABLE, so that
> > the embedded platforms can choose the best option themselves, while
> > still keep the feature of unified one kernel Image for all SV39, SV48
> > and SV57.
>
> NONPORTABLE is what you're meant to depend on when the kernel that gets
> built cannot run on platforms without the feature.
The reason I put these two Kconfig options under NONPORTABLE is: all
distributions won't enable NONPORTABLE for their defconfig, and I guess
the distributions want to kept unified one kernel Image for all SV39,
SV48 and SV57.
I have no preference whether move them out of NONPORTABLE or not.
>
> > +
> > + /* Fall back to Sv39 if Sv48 is unsupported on non-portable systems */
> > + if (IS_ENABLED(CONFIG_NONPORTABLE) && !IS_ENABLED(CONFIG_RISCV_ISA_SV48) &&
> > + satp_mode_limit == SATP_MODE_48)
> > + satp_mode_limit = SATP_MODE_39;
> > +
> > if (satp_mode_limit == SATP_MODE_48) {
> > disable_pgtable_l5();
> > } else if (satp_mode_limit == SATP_MODE_39) {
> > --
> > 2.53.0
> >
More information about the linux-riscv
mailing list