Re: [PATCH v2 1/9] riscv: add __init section marker to some functions

[Jisheng Zhang]

On Fri, 2 Apr 2021 09:38:02 +0530
Anup Patel <anup@xxxxxxxxxxxxxx> wrote:


> 
> 
> On Wed, Mar 31, 2021 at 10:00 PM Jisheng Zhang
> <jszhang3@xxxxxxxxxxxxxxxx> wrote:
> >
> > From: Jisheng Zhang <jszhang@xxxxxxxxxx>
> >
> > They are not needed after booting, so mark them as __init to move them
> > to the __init section.
> >
> > Signed-off-by: Jisheng Zhang <jszhang@xxxxxxxxxx>
> > ---
> >  arch/riscv/kernel/traps.c  | 2 +-
> >  arch/riscv/mm/init.c       | 6 +++---
> >  arch/riscv/mm/kasan_init.c | 6 +++---
> >  arch/riscv/mm/ptdump.c     | 2 +-
> >  4 files changed, 8 insertions(+), 8 deletions(-)
> >
> > diff --git a/arch/riscv/kernel/traps.c b/arch/riscv/kernel/traps.c
> > index 1357abf79570..07fdded10c21 100644
> > --- a/arch/riscv/kernel/traps.c
> > +++ b/arch/riscv/kernel/traps.c
> > @@ -197,6 +197,6 @@ int is_valid_bugaddr(unsigned long pc)
> >  #endif /* CONFIG_GENERIC_BUG */
> >
> >  /* stvec & scratch is already set from head.S */
> > -void trap_init(void)
> > +void __init trap_init(void)
> >  {
> >  }  
> 
> The trap_init() is unused currently so you can drop this change
> and remove trap_init() as a separate patch.

the kernel init/main.c expects a trap_init() implementation in architecture
code. Some architecture's implementation is NULL, similar as riscv, for example,
arm, powerpc and so on. However I think you are right, the trap_init() can be
removed, we need a trivial series to provide a __weak but NULL trap_init()
implementation in init/main.c then remove all NULL implementation from
all arch. I can take the task to do the clean up.

> 
> > diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
> > index 067583ab1bd7..76bf2de8aa59 100644
> > --- a/arch/riscv/mm/init.c
> > +++ b/arch/riscv/mm/init.c
> > @@ -57,7 +57,7 @@ static void __init zone_sizes_init(void)
> >         free_area_init(max_zone_pfns);
> >  }
> >
> > -static void setup_zero_page(void)
> > +static void __init setup_zero_page(void)
> >  {
> >         memset((void *)empty_zero_page, 0, PAGE_SIZE);

I think the zero page is already initialized as "0" because empty_zero_page
sits in .bss section. So this setup_zero_page() function can be removed. I
will send a newer version later.

thanks