Re: [PATCH 13/24] RISC-V: ACPI: smpboot: Add ACPI support in smp_setup()

[Sunil V L <sunilvl@ventanamicro.com>]

On Wed, Feb 08, 2023 at 10:10:14PM +0000, Conor Dooley wrote:
> On Mon, Jan 30, 2023 at 11:52:14PM +0530, Sunil V L wrote:
> > Add function to parse the RINTC structure in
> > the MADT table and create the required initializations to
> > enable SMP boot on ACPI based platforms.
> > 
> > Signed-off-by: Sunil V L <sunilvl@ventanamicro.com>
> > ---
> >  arch/riscv/include/asm/acpi.h |  7 ++++
> >  arch/riscv/kernel/smpboot.c   | 73 ++++++++++++++++++++++++++++++++++-
> >  2 files changed, 79 insertions(+), 1 deletion(-)
> > 
> > diff --git a/arch/riscv/include/asm/acpi.h b/arch/riscv/include/asm/acpi.h
> > index c5cb9f96d404..d1f1e53ec657 100644
> > --- a/arch/riscv/include/asm/acpi.h
> > +++ b/arch/riscv/include/asm/acpi.h
> > @@ -58,6 +58,13 @@ static inline bool acpi_has_cpu_in_madt(void)
> >  }
> >  
> >  static inline void arch_fix_phys_package_id(int num, u32 slot) { }
> > +
> > +#ifdef CONFIG_ACPI_NUMA
> > +int acpi_numa_get_nid(unsigned int cpu);
> > +#else
> > +static inline int acpi_numa_get_nid(unsigned int cpu) { return NUMA_NO_NODE; }
> > +#endif /* CONFIG_ACPI_NUMA */
> > +
> >  #endif
> >  
> >  #endif /*_ASM_ACPI_H*/
> > diff --git a/arch/riscv/kernel/smpboot.c b/arch/riscv/kernel/smpboot.c
> > index 26214ddefaa4..e48cf88d0bc1 100644
> > --- a/arch/riscv/kernel/smpboot.c
> > +++ b/arch/riscv/kernel/smpboot.c
> > @@ -8,6 +8,7 @@
> >   * Copyright (C) 2017 SiFive
> >   */
> >  
> > +#include <linux/acpi.h>
> >  #include <linux/arch_topology.h>
> >  #include <linux/module.h>
> >  #include <linux/init.h>
> > @@ -70,6 +71,73 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
> >  	}
> >  }
> >  
> > +#ifdef CONFIG_ACPI
> > +static unsigned int cpu_count = 1;
> > +
> > +static int __init
> > +acpi_parse_rintc(union acpi_subtable_headers *header,
> > +			     const unsigned long end)
> 
> This all fits on one line. And also avoids the checkpatch complaint from
> what you have currently done...
> 
Okay.

> > +{
> > +	unsigned long hart;
> > +	bool found_boot_cpu = false;
> > +
> > +	struct acpi_madt_rintc *processor;
> > +
> > +	processor = (struct acpi_madt_rintc *)header;
> 
> Why not combine the above two lines?
> 
> > +	/* RINTC entry which has !ACPI_MADT_ENABLED is not enabled so skip */
> 
> This comment is a bit -ENOPARSE. Please reword it in a way that is
> understandable to mere mortals like myself.
>
Okay, let me try :-).
 
> > +	if (!(processor->flags & ACPI_MADT_ENABLED))
> > +		return 0;
> > +
> > +	hart = processor->hart_id;
> > +	if (hart < 0)
> > +		return 0;
> 
> Newline here please
> 
> > +	if (hart == cpuid_to_hartid_map(0)) {
> > +		BUG_ON(found_boot_cpu);
> > +		found_boot_cpu = 1;
> 
> This is a bool, why not assign a bool value to it so it looks more
> intentional? I know this is copied from the dt code, but that should
> really be on too IMO.
>
Okay.
 
> > +		early_map_cpu_to_node(0, acpi_numa_get_nid(cpu_count));
> > +		return 0;
> > +	}
> 
> And a newline here too...
>
Okay. 
> > +	if (cpu_count >= NR_CPUS) {
> > +		pr_warn("Invalid cpuid [%d] for hartid [%lu]\n",
> > +			cpu_count, hart);
> > +		return 0;
> > +	}
> > +
> > +	cpuid_to_hartid_map(cpu_count) = hart;
> > +	early_map_cpu_to_node(cpu_count, acpi_numa_get_nid(cpu_count));
> > +	cpu_count++;
> 
> ...and also here please!
>
Okay.

> > +	return 0;
> > +}
> > +
> > +static void __init acpi_parse_and_init_cpus(void)
> > +{
> > +	int cpuid;
> > +
> > +	cpu_set_ops(0);
> 
> While I'm at it suggesting newline additions, adding them before
> comments would be great too.
>
Okay.
 
> > +	/*
> > +	 * do a walk of MADT to determine how many CPUs
> > +	 * we have including disabled CPUs, and get information
> > +	 * we need for SMP init.
> > +	 */
> > +	acpi_table_parse_madt(ACPI_MADT_TYPE_RINTC,
> > +				      acpi_parse_rintc, 0);
> > +
> > +	/*
> > +	 * NUMA - TODO
> > +	 */
> 
> TODO before merging, or TODO at some indeterminate point in the future?
>
Will remove this comment. We need to add NUMA after basic support is
done.
 
> Anyways, this is all nits & this largely seem to resemble the dt code,
> so with the nits fixed (and an s/ACPI: // in $subject):
> Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
> 
Thanks!
Sunil