2022-05-23 08:11:12

by Jisheng Zhang

[permalink] [raw]
Subject: [PATCH v4 0/2] use static key to optimize pgtable_l4_enabled

The pgtable_l4|[l5]_enabled check sits at hot code path, performance
is impacted a lot. Since pgtable_l4|[l5]_enabled isn't changed after
boot, so static key can be used to solve the performance issue[1].

An unified way static key was introduced in [2], but it only targets
riscv isa extension. We dunno whether SV48 and SV57 will be considered
as isa extension, so the unified solution isn't used for
pgtable_l4[l5]_enabled now.

patch1 fixes a NULL pointer deference if static key is used a bit earlier.
patch2 uses the static key to optimize pgtable_l4|[l5]_enabled.

[1] http://lists.infradead.org/pipermail/linux-riscv/2021-December/011164.html
[2] https://lore.kernel.org/linux-riscv/[email protected]/T/#t

Since v3:
- fix W=1 call to undeclared function 'static_branch_likely' error

Since v2:
- move the W=1 warning fix to a separate patch
- move the unified way to use static key to a new patch series.

Since v1:
- Add a W=1 warning fix
- Fix W=1 error
- Based on v5.18-rcN, since SV57 support is added, so convert
pgtable_l5_enabled as well.



Jisheng Zhang (2):
riscv: move sbi_init() earlier before jump_label_init()
riscv: turn pgtable_l4|[l5]_enabled to static key for RV64

arch/riscv/include/asm/pgalloc.h | 16 ++++----
arch/riscv/include/asm/pgtable-32.h | 3 ++
arch/riscv/include/asm/pgtable-64.h | 60 ++++++++++++++++++---------
arch/riscv/include/asm/pgtable.h | 5 +--
arch/riscv/kernel/cpu.c | 4 +-
arch/riscv/kernel/setup.c | 2 +-
arch/riscv/mm/init.c | 64 ++++++++++++++++++-----------
arch/riscv/mm/kasan_init.c | 16 ++++----
8 files changed, 104 insertions(+), 66 deletions(-)

--
2.34.1



2022-06-26 05:07:04

by Anup Patel

[permalink] [raw]
Subject: Re: [PATCH v4 0/2] use static key to optimize pgtable_l4_enabled

On Sat, May 21, 2022 at 8:13 PM Jisheng Zhang <[email protected]> wrote:
>
> The pgtable_l4|[l5]_enabled check sits at hot code path, performance
> is impacted a lot. Since pgtable_l4|[l5]_enabled isn't changed after
> boot, so static key can be used to solve the performance issue[1].
>
> An unified way static key was introduced in [2], but it only targets
> riscv isa extension. We dunno whether SV48 and SV57 will be considered
> as isa extension, so the unified solution isn't used for
> pgtable_l4[l5]_enabled now.
>
> patch1 fixes a NULL pointer deference if static key is used a bit earlier.
> patch2 uses the static key to optimize pgtable_l4|[l5]_enabled.
>
> [1] http://lists.infradead.org/pipermail/linux-riscv/2021-December/011164.html
> [2] https://lore.kernel.org/linux-riscv/[email protected]/T/#t
>
> Since v3:
> - fix W=1 call to undeclared function 'static_branch_likely' error
>
> Since v2:
> - move the W=1 warning fix to a separate patch
> - move the unified way to use static key to a new patch series.
>
> Since v1:
> - Add a W=1 warning fix
> - Fix W=1 error
> - Based on v5.18-rcN, since SV57 support is added, so convert
> pgtable_l5_enabled as well.
>
>
>
> Jisheng Zhang (2):
> riscv: move sbi_init() earlier before jump_label_init()
> riscv: turn pgtable_l4|[l5]_enabled to static key for RV64

I have tested both these patches on QEMU RV64 and RV32.

Tested-by: Anup Patel <[email protected]>

Thanks,
Anup

>
> arch/riscv/include/asm/pgalloc.h | 16 ++++----
> arch/riscv/include/asm/pgtable-32.h | 3 ++
> arch/riscv/include/asm/pgtable-64.h | 60 ++++++++++++++++++---------
> arch/riscv/include/asm/pgtable.h | 5 +--
> arch/riscv/kernel/cpu.c | 4 +-
> arch/riscv/kernel/setup.c | 2 +-
> arch/riscv/mm/init.c | 64 ++++++++++++++++++-----------
> arch/riscv/mm/kasan_init.c | 16 ++++----
> 8 files changed, 104 insertions(+), 66 deletions(-)
>
> --
> 2.34.1
>

2022-07-02 04:03:35

by Palmer Dabbelt

[permalink] [raw]
Subject: Re: [PATCH v4 0/2] use static key to optimize pgtable_l4_enabled

On Sat, 25 Jun 2022 21:33:07 PDT (-0700), [email protected] wrote:
> On Sat, May 21, 2022 at 8:13 PM Jisheng Zhang <[email protected]> wrote:
>>
>> The pgtable_l4|[l5]_enabled check sits at hot code path, performance
>> is impacted a lot. Since pgtable_l4|[l5]_enabled isn't changed after
>> boot, so static key can be used to solve the performance issue[1].
>>
>> An unified way static key was introduced in [2], but it only targets
>> riscv isa extension. We dunno whether SV48 and SV57 will be considered
>> as isa extension, so the unified solution isn't used for
>> pgtable_l4[l5]_enabled now.
>>
>> patch1 fixes a NULL pointer deference if static key is used a bit earlier.
>> patch2 uses the static key to optimize pgtable_l4|[l5]_enabled.
>>
>> [1] http://lists.infradead.org/pipermail/linux-riscv/2021-December/011164.html
>> [2] https://lore.kernel.org/linux-riscv/[email protected]/T/#t
>>
>> Since v3:
>> - fix W=1 call to undeclared function 'static_branch_likely' error
>>
>> Since v2:
>> - move the W=1 warning fix to a separate patch
>> - move the unified way to use static key to a new patch series.
>>
>> Since v1:
>> - Add a W=1 warning fix
>> - Fix W=1 error
>> - Based on v5.18-rcN, since SV57 support is added, so convert
>> pgtable_l5_enabled as well.
>>
>>
>>
>> Jisheng Zhang (2):
>> riscv: move sbi_init() earlier before jump_label_init()
>> riscv: turn pgtable_l4|[l5]_enabled to static key for RV64
>
> I have tested both these patches on QEMU RV64 and RV32.
>
> Tested-by: Anup Patel <[email protected]>
>
> Thanks,
> Anup

Thanks for testing these. Unfortunatly they're failing for me under my
kasan+sparsemem-vmemmap config, which looks like a defconfig with

CONFIG_KASAN=y
# CONFIG_FLATMEM_MANUAL is not set
CONFIG_SPARSEMEM_MANUAL=y
CONFIG_SPARSEMEM=y
# CONFIG_SPARSEMEM_VMEMMAP is not set

Nothing's really jumping out and I'm not sure that's a super compelling
configuration, but IIRC it's found a handful of issues before so I'm not
sure it's sane to just toss it.

I've put this all on the riscv-pgtable_static_key branch of
kernel.org/palmer/linux . If nobody has the time to look then I'll try
and give it another shot, but I'm pretty buried right now so happy to
have the help.

>
>>
>> arch/riscv/include/asm/pgalloc.h | 16 ++++----
>> arch/riscv/include/asm/pgtable-32.h | 3 ++
>> arch/riscv/include/asm/pgtable-64.h | 60 ++++++++++++++++++---------
>> arch/riscv/include/asm/pgtable.h | 5 +--
>> arch/riscv/kernel/cpu.c | 4 +-
>> arch/riscv/kernel/setup.c | 2 +-
>> arch/riscv/mm/init.c | 64 ++++++++++++++++++-----------
>> arch/riscv/mm/kasan_init.c | 16 ++++----
>> 8 files changed, 104 insertions(+), 66 deletions(-)
>>
>> --
>> 2.34.1
>>

2022-07-15 13:49:25

by Jisheng Zhang

[permalink] [raw]
Subject: Re: [PATCH v4 0/2] use static key to optimize pgtable_l4_enabled

On Fri, Jul 01, 2022 at 08:48:25PM -0700, Palmer Dabbelt wrote:
> On Sat, 25 Jun 2022 21:33:07 PDT (-0700), [email protected] wrote:
> > On Sat, May 21, 2022 at 8:13 PM Jisheng Zhang <[email protected]> wrote:
> > >
> > > The pgtable_l4|[l5]_enabled check sits at hot code path, performance
> > > is impacted a lot. Since pgtable_l4|[l5]_enabled isn't changed after
> > > boot, so static key can be used to solve the performance issue[1].
> > >
> > > An unified way static key was introduced in [2], but it only targets
> > > riscv isa extension. We dunno whether SV48 and SV57 will be considered
> > > as isa extension, so the unified solution isn't used for
> > > pgtable_l4[l5]_enabled now.
> > >
> > > patch1 fixes a NULL pointer deference if static key is used a bit earlier.
> > > patch2 uses the static key to optimize pgtable_l4|[l5]_enabled.
> > >
> > > [1] http://lists.infradead.org/pipermail/linux-riscv/2021-December/011164.html
> > > [2] https://lore.kernel.org/linux-riscv/[email protected]/T/#t
> > >
> > > Since v3:
> > > - fix W=1 call to undeclared function 'static_branch_likely' error
> > >
> > > Since v2:
> > > - move the W=1 warning fix to a separate patch
> > > - move the unified way to use static key to a new patch series.
> > >
> > > Since v1:
> > > - Add a W=1 warning fix
> > > - Fix W=1 error
> > > - Based on v5.18-rcN, since SV57 support is added, so convert
> > > pgtable_l5_enabled as well.
> > >
> > >
> > >
> > > Jisheng Zhang (2):
> > > riscv: move sbi_init() earlier before jump_label_init()
> > > riscv: turn pgtable_l4|[l5]_enabled to static key for RV64
> >
> > I have tested both these patches on QEMU RV64 and RV32.
> >
> > Tested-by: Anup Patel <[email protected]>
> >
> > Thanks,
> > Anup
>
> Thanks for testing these. Unfortunatly they're failing for me under my
> kasan+sparsemem-vmemmap config, which looks like a defconfig with
>
> CONFIG_KASAN=y
> # CONFIG_FLATMEM_MANUAL is not set
> CONFIG_SPARSEMEM_MANUAL=y
> CONFIG_SPARSEMEM=y
> # CONFIG_SPARSEMEM_VMEMMAP is not set

Hi Palmer,

Thank you for the hint, I find the reason: SPARSEMEM is the key, KASAN
doesn't matter. To fix this issue, we need to move
riscv_finalise_pgtable_lx() after sparse_init(). I will send out a
newer version soon.

>
> Nothing's really jumping out and I'm not sure that's a super compelling
> configuration, but IIRC it's found a handful of issues before so I'm not
> sure it's sane to just toss it.
>
> I've put this all on the riscv-pgtable_static_key branch of
> kernel.org/palmer/linux . If nobody has the time to look then I'll try and
> give it another shot, but I'm pretty buried right now so happy to have the
> help.

Let me know if you want a seperate patch against
riscv-pgtable_static_key branch.

Thanks

>
> >
> > >
> > > arch/riscv/include/asm/pgalloc.h | 16 ++++----
> > > arch/riscv/include/asm/pgtable-32.h | 3 ++
> > > arch/riscv/include/asm/pgtable-64.h | 60 ++++++++++++++++++---------
> > > arch/riscv/include/asm/pgtable.h | 5 +--
> > > arch/riscv/kernel/cpu.c | 4 +-
> > > arch/riscv/kernel/setup.c | 2 +-
> > > arch/riscv/mm/init.c | 64 ++++++++++++++++++-----------
> > > arch/riscv/mm/kasan_init.c | 16 ++++----
> > > 8 files changed, 104 insertions(+), 66 deletions(-)
> > >
> > > --
> > > 2.34.1
> > >