Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <c5e1b113-f457-2cc7-808b-b13f83adbb2b@loongson.cn>
Date:   Fri, 2 Dec 2022 17:03:24 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.13.0
Subject: Re: [PATCH] LoongArch: export symbol with function
 smp_send_reschedule
Content-Language: en-US
To:     WANG Xuerui <kernel@xen0n.name>,
        Huacai Chen <chenhuacai@kernel.org>
Cc:     loongarch@lists.linux.dev, linux-kernel@vger.kernel.org
References: <20221202075854.2253116-1-maobibo@loongson.cn>
 <07e0d775-6bb7-051b-9ec9-71153efbdf21@xen0n.name>
From:   maobibo <maobibo@loongson.cn>
In-Reply-To: <07e0d775-6bb7-051b-9ec9-71153efbdf21@xen0n.name>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk


在 2022/12/2 16:25, WANG Xuerui 写道:
> On 2022/12/2 15:58, Bibo Mao wrote:
>> Function smp_send_reschedule is standard kernel ABI, which is
>> defined header file include/linux/smp.h, however on LoongArch
>> it is defined as inline function, so that kernel module can
>> not use this function.
>>
>> Now define smp_send_reschedule as general function, and add
>> EXPORT_SYMBOL_GPL on this function, so that kernel modules can
>> use it.
>>
>> Signed-off-by: Bibo Mao <maobibo@loongson.cn>
>> ---
>>   arch/loongarch/include/asm/smp.h | 10 ----------
>>   arch/loongarch/kernel/smp.c      | 11 +++++++++++
>>   2 files changed, 11 insertions(+), 10 deletions(-)
>>
>> diff --git a/arch/loongarch/include/asm/smp.h b/arch/loongarch/include/asm/smp.h
>> index 3dd172d9ffea..d82687390b4a 100644
>> --- a/arch/loongarch/include/asm/smp.h
>> +++ b/arch/loongarch/include/asm/smp.h
>> @@ -78,16 +78,6 @@ extern void calculate_cpu_foreign_map(void);
>>    */
>>   extern void show_ipi_list(struct seq_file *p, int prec);
>>   -/*
>> - * This function sends a 'reschedule' IPI to another CPU.
>> - * it goes straight through and wastes no time serializing
>> - * anything. Worst case is that we lose a reschedule ...
>> - */
>> -static inline void smp_send_reschedule(int cpu)
>> -{
>> -    loongson_send_ipi_single(cpu, SMP_RESCHEDULE);
>> -}
>> -
>>   static inline void arch_send_call_function_single_ipi(int cpu)
>>   {
>>       loongson_send_ipi_single(cpu, SMP_CALL_FUNCTION);
>> diff --git a/arch/loongarch/kernel/smp.c b/arch/loongarch/kernel/smp.c
>> index 6ed72f7ff278..51dd3c3f06cb 100644
>> --- a/arch/loongarch/kernel/smp.c
>> +++ b/arch/loongarch/kernel/smp.c
>> @@ -141,6 +141,17 @@ void loongson_send_ipi_single(int cpu, unsigned int action)
>>       ipi_write_action(cpu_logical_map(cpu), (u32)action);
>>   }
>>   +/*
>> + * This function sends a 'reschedule' IPI to another CPU.
>> + * it goes straight through and wastes no time serializing
>> + * anything. Worst case is that we lose a reschedule ...
>> + */
>> +void smp_send_reschedule(int cpu)
>> +{
>> +    loongson_send_ipi_single(cpu, SMP_RESCHEDULE);
>> +}
>> +EXPORT_SYMBOL_GPL(smp_send_reschedule);
>> +
> 
> While the change is in itself okay (one less case of mips legacy, getting in line with ia64, powerpc and riscv that all EXPORT_SYMBOL_GPL this), I'd suggest you batch this patch with the subsequent changes you plan to enable with this one, so reviewers would have more context and hopefully avoid churn. (I, by my familiarity with Loongson and LoongArch development, know you're probably aiming to use this with KVM, but others probably don't know, and again it's always better to have more context.)
> 

yes, kvm module depends on function smp_send_reschedule, only that it is not mature now. And this function is standard API, not arch specified API, it is normal for modules to use it :)

regards
bibo,mao
>>   void loongson_send_ipi_mask(const struct cpumask *mask, unsigned int action)
>>   {
>>       unsigned int i;
>