Subject: Re: [PATCH 1/2] kdump: vmcore: move copy_to() from vmcore.c to
 uaccess.h
To:     Andrew Morton <akpm@linux-foundation.org>
References: <1639143361-17773-1-git-send-email-yangtiezhu@loongson.cn>
 <1639143361-17773-2-git-send-email-yangtiezhu@loongson.cn>
 <20211210085903.e7820815e738d7dc6da06050@linux-foundation.org>
Cc:     Dave Young <dyoung@redhat.com>, Baoquan He <bhe@redhat.com>,
        Vivek Goyal <vgoyal@redhat.com>,
        linux-arm-kernel@lists.infradead.org, linux-ia64@vger.kernel.org,
        linux-mips@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
        linux-riscv@lists.infradead.org, linux-sh@vger.kernel.org,
        x86@kernel.org, linux-fsdevel@vger.kernel.org,
        kexec@lists.infradead.org, linux-kernel@vger.kernel.org,
        Xuefeng Li <lixuefeng@loongson.cn>
From:   Tiezhu Yang <yangtiezhu@loongson.cn>
Message-ID: <249b17ea-171a-49f7-b438-488c03fa1f9b@loongson.cn>
Date:   Sat, 11 Dec 2021 07:50:26 +0800
User-Agent: Mozilla/5.0 (X11; Linux mips64; rv:45.0) Gecko/20100101
 Thunderbird/45.4.0
MIME-Version: 1.0
In-Reply-To: <20211210085903.e7820815e738d7dc6da06050@linux-foundation.org>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 12/11/2021 12:59 AM, Andrew Morton wrote:
> On Fri, 10 Dec 2021 21:36:00 +0800 Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
>
>> In arch/*/kernel/crash_dump*.c, there exist similar code about
>> copy_oldmem_page(), move copy_to() from vmcore.c to uaccess.h,
>> and then we can use copy_to() to simplify the related code.
>>
>> ...
>>
>> --- a/fs/proc/vmcore.c
>> +++ b/fs/proc/vmcore.c
>> @@ -238,20 +238,6 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
>>  	return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
>>  }
>>
>> -/*
>> - * Copy to either kernel or user space
>> - */
>> -static int copy_to(void *target, void *src, size_t size, int userbuf)
>> -{
>> -	if (userbuf) {
>> -		if (copy_to_user((char __user *) target, src, size))
>> -			return -EFAULT;
>> -	} else {
>> -		memcpy(target, src, size);
>> -	}
>> -	return 0;
>> -}
>> -
>>  #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
>>  static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
>>  {
>> diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
>> index ac03940..4a6c3e4 100644
>> --- a/include/linux/uaccess.h
>> +++ b/include/linux/uaccess.h
>> @@ -201,6 +201,20 @@ copy_to_user(void __user *to, const void *from, unsigned long n)
>>  	return n;
>>  }
>>
>> +/*
>> + * Copy to either kernel or user space
>> + */
>> +static inline int copy_to(void *target, void *src, size_t size, int userbuf)
>> +{
>> +	if (userbuf) {
>> +		if (copy_to_user((char __user *) target, src, size))
>> +			return -EFAULT;
>> +	} else {
>> +		memcpy(target, src, size);
>> +	}
>> +	return 0;
>> +}
>> +
>
> Ordinarily I'd say "this is too large to be inlined".  But the function
> has only a single callsite per architecture so inlining it won't cause
> bloat at present.
>
> But hopefully copy_to() will get additional callers in the future, in
> which case it shouldn't be inlined.  So I'm thinking it would be best
> to start out with this as a regular non-inlined function, in
> lib/usercopy.c.
>
> Also, copy_to() is a very poor name for a globally-visible helper
> function.  Better would be copy_to_user_or_kernel(), although that's
> perhaps a bit long.
>
> And the `userbuf' arg should have type bool, yes?
>

Hi Andrew,

Thank you very much for your reply and suggestion, I agree with you,
I will send v2 later.

Thanks,
Tiezhu