On 2020/11/19 上午12:29, David Sterba wrote:
> On Fri, Nov 13, 2020 at 08:51:47PM +0800, Qu Wenruo wrote:
>> @@ -1781,8 +1781,9 @@ static int scrub_checksum_data(struct scrub_block *sblock)
>>  	struct scrub_ctx *sctx = sblock->sctx;
>>  	struct btrfs_fs_info *fs_info = sctx->fs_info;
>>  	SHASH_DESC_ON_STACK(shash, fs_info->csum_shash);
>> -	u8 csum[BTRFS_CSUM_SIZE];
>>  	struct scrub_page *spage;
>> +	u32 sectorsize = fs_info->sectorsize;
>> +	u8 csum[BTRFS_CSUM_SIZE];
>>  	char *kaddr;
>>  
>>  	BUG_ON(sblock->page_count < 1);
>> @@ -1794,11 +1795,15 @@ static int scrub_checksum_data(struct scrub_block *sblock)
>>  
>>  	shash->tfm = fs_info->csum_shash;
>>  	crypto_shash_init(shash);
>> -	crypto_shash_digest(shash, kaddr, PAGE_SIZE, csum);
>> +
>> +	/*
>> +	 * In scrub_pages() and scrub_pages_for_parity() we ensure
>> +	 * each spage only contains just one sector of data.
>> +	 */
>> +	crypto_shash_digest(shash, kaddr, sectorsize, csum);
> 
> Temporary variable is not needed for single use (sectorsize).
> 

Personally speaking, whether such temporary variable is needed should be
determined at compile time.

For reader, I didn't see anything wrong using such variable, especially
it can save some "fs_info->" typing and saves some new line.

Thanks,
Qu