Re: [RFC PATCH] riscv: export cpu/freq invariant to scheduler

[Song Shuai]

Andrew Jones <ajones@xxxxxxxxxxxxxxxx> 于2023年3月22日周三 08:03写道：
>
> On Wed, Mar 22, 2023 at 02:18:56PM +0800, Song Shuai wrote:
> > RISC-V now manages CPU topology using arch_topology which provides
> > CPU capacity and frequency related interfaces to access the cpu/freq
> > invariant in possible heterogeneous or DVFS-enabled platforms.
> >
> > Here adds topology.h file to export the arch_topology interfaces for
> > replacing the scheduler's constant-based cpu/freq invariant accounting.
> >
> > Signed-off-by: Song Shuai <suagrfillet@xxxxxxxxx>
> > ---
> >  arch/riscv/include/asm/topology.h | 19 +++++++++++++++++++
> >  1 file changed, 19 insertions(+)
> >  create mode 100644 arch/riscv/include/asm/topology.h
> >
> > diff --git a/arch/riscv/include/asm/topology.h b/arch/riscv/include/asm/topology.h
> > new file mode 100644
> > index 000000000000..14bbd2472af6
> > --- /dev/null
> > +++ b/arch/riscv/include/asm/topology.h
> > @@ -0,0 +1,19 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +#ifndef __ASM_RISCV_TOPOLOGY_H
> > +#define __ASM_RISCV_TOPOLOGY_H
>
> riscv uses a single leading underscore.
ok.
>
> > +
> > +#include <linux/arch_topology.h>
> > +
> > +/* Replace task scheduler's default frequency-invariant accounting */
> > +#define arch_scale_freq_tick topology_scale_freq_tick
> > +#define arch_set_freq_scale topology_set_freq_scale
> > +#define arch_scale_freq_capacity topology_get_freq_scale
> > +#define arch_scale_freq_invariant topology_scale_freq_invariant
> > +
> > +/* Replace task scheduler's default cpu-invariant accounting */
> > +#define arch_scale_cpu_capacity topology_get_cpu_scale
> > +#define arch_update_cpu_topology topology_update_cpu_topology
> > +
> > +
> > +#include <asm-generic/topology.h>
> > +#endif /* __ASM_RISCV_TOPOLOGY_H */
> > --
> > 2.20.1
> >
>
> This looks reasonable, at least from a "do what arm64 does" type of
> perspective. Why the RFC? Is it not tested?
I only tested it in the qemu sifive_u machine with a customed dtb
not sure if it works in real hardware, so I posted it with RFC.
>
> Also, if you repost, could please neaten it up a bit? Aligning all
> the defines would help and removing the extra blank line. Finally,
> why wasn't the "/* Enable topology flag updates */" comment also
> lifted from arm64 like the others?
I'll add the comment back and tidy it up in the next post.
>
> Thanks,
> drew



-- 
Thanks,
Song