Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
In-Reply-To: <7bc075fc-7039-62d8-c708-1c36233b7126@wdc.com>
References: <1534377377-70108-1-git-send-email-atish.patra@wdc.com>
 <1534377377-70108-2-git-send-email-atish.patra@wdc.com> <CAAhSdy1Legi9jzYnkN2am3QKLw7Y8tpbUfmJVA0gBRB4nPnXjA@mail.gmail.com>
 <7bc075fc-7039-62d8-c708-1c36233b7126@wdc.com>
From:   Anup Patel <anup@brainfault.org>
Date:   Thu, 16 Aug 2018 11:09:48 +0530
Message-ID: <CAAhSdy0wb_dw0UypXQXLtRiQXRc5owxRdbCq=JRXw5gnYq9+JA@mail.gmail.com>
Subject: Re: [RFC PATCH 1/5] RISC-V: Add logical CPU indexing for RISC-V
To:     Atish Patra <atish.patra@wdc.com>
Cc:     "palmer@sifive.com" <palmer@sifive.com>,
        "linux-riscv@lists.infradead.org" <linux-riscv@lists.infradead.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Christoph Hellwig <hch@infradead.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        "linux-kernel@vger.kernel.org List" <linux-kernel@vger.kernel.org>,
        Damien Le Moal <Damien.LeMoal@wdc.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Aug 16, 2018 at 10:47 AM, Atish Patra <atish.patra@wdc.com> wrote:
> On 8/15/18 9:06 PM, Anup Patel wrote:
>>
>> On Thu, Aug 16, 2018 at 5:26 AM, Atish Patra <atish.patra@wdc.com> wrote:
>>>
>>> Currently, both linux cpu id and hardware cpu id are same.
>>> This is not recommended as it will lead to discontinuous cpu
>>> indexing in Linux. Moreover, kdump kernel will run from CPU0
>>> which would be absent if we follow existing scheme.
>>>
>>> Implement a logical mapping between Linux cpu id and hardware
>>> cpuid to decouple these two. Always mark the boot processor as
>>> cpu0 and all other cpus get the logical cpu id based on their
>>> booting order.
>>>
>>> Signed-off-by: Atish Patra <atish.patra@wdc.com>
>>> ---
>>>   arch/riscv/include/asm/smp.h | 17 ++++++++++++++++-
>>>   arch/riscv/kernel/setup.c    |  2 ++
>>>   arch/riscv/kernel/smp.c      | 19 +++++++++++++++++++
>>>   3 files changed, 37 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/riscv/include/asm/smp.h b/arch/riscv/include/asm/smp.h
>>> index 36016845..0763337b 100644
>>> --- a/arch/riscv/include/asm/smp.h
>>> +++ b/arch/riscv/include/asm/smp.h
>>> @@ -22,6 +22,12 @@
>>>   #include <linux/cpumask.h>
>>>   #include <linux/irqreturn.h>
>>>
>>> +/*
>>> + * Mapping between linux logical cpu index and hartid.
>>> + */
>>> +extern u64 __cpu_logical_map[NR_CPUS];
>>> +#define cpu_logical_map(cpu)    __cpu_logical_map[cpu]
>>> +
>>>   #ifdef CONFIG_SMP
>>>
>>>   /* SMP initialization hook for setup_arch */
>>> @@ -33,6 +39,8 @@ void arch_send_call_function_ipi_mask(struct cpumask
>>> *mask);
>>>   /* Hook for the generic smp_call_function_single() routine. */
>>>   void arch_send_call_function_single_ipi(int cpu);
>>>
>>> +int riscv_hartid_to_cpuid(int hartid);
>>> +void cpuid_to_hartid_mask(const struct cpumask *in, struct cpumask
>>> *out);
>>>   /*
>>>    * This is particularly ugly: it appears we can't actually get the
>>> definition
>>>    * of task_struct here, but we need access to the CPU this task is
>>> running on.
>>> @@ -41,6 +49,13 @@ void arch_send_call_function_single_ipi(int cpu);
>>>    */
>>>   #define raw_smp_processor_id() (*((int*)((char*)get_current() +
>>> TASK_TI_CPU)))
>>>
>>> -#endif /* CONFIG_SMP */
>>> +#else
>>> +
>>> +static inline int riscv_hartid_to_cpuid(int hartid) { return 0 ; }
>>> +static inline void cpuid_to_hartid_mask(const struct cpumask *in,
>>> +                                      struct cpumask *out) {
>>> +       cpumask_set_cpu(cpu_logical_map(0), out);
>>> +}
>>>
>>> +#endif /* CONFIG_SMP */
>>>   #endif /* _ASM_RISCV_SMP_H */
>>> diff --git a/arch/riscv/kernel/setup.c b/arch/riscv/kernel/setup.c
>>> index db20dc63..e21ed481 100644
>>> --- a/arch/riscv/kernel/setup.c
>>> +++ b/arch/riscv/kernel/setup.c
>>> @@ -82,6 +82,8 @@ EXPORT_SYMBOL(empty_zero_page);
>>>   /* The lucky hart to first increment this variable will boot the other
>>> cores */
>>>   atomic_t hart_lottery;
>>>
>>> +u64 __cpu_logical_map[NR_CPUS];
>>
>>
>> If hardware IDs are always machine word size then its better to use
>> "unsigned long" in-place of u64.
>>
>
> Good point. Thanks.
>
>> The __cpu_logical_map[] should be zero initially because zero is a
>> valid hardware ID. Better set all entries to -1 by assigning { -1 } to the
>> array.
>>
>
> ok. I will introduce INVALID_HART_ID (=-1) and initialize with that.
>
>> Also, I feel  __cpu_logical_map[] should be part of smp.c instead of
>> setup.c. Any particular reason for having it in setup.c?
>>
>
> currently smp.c is only compiled in if CONFIG_SMP. That's why I kept it in
> setup.c

Ahh, got it. For now let it be here till we figure better place for it.

Regards,
Anup