On 2020/5/15 12:06, Matthew Wilcox wrote:
> On Thu, May 07, 2020 at 03:50:56PM +0800, Zhen Lei wrote:
>> @@ -266,7 +266,7 @@ int swap_writepage(struct page *page, struct
>> writeback_control *wbc)
>>
>> static sector_t swap_page_sector(struct page *page)
>> {
>> - return (sector_t)__page_file_index(page) << (PAGE_SHIFT - 9);
>> + return npage_to_sectors((sector_t)__page_file_index(page));
>
> If you make npage_to_sectors() a proper function instead of a macro,
> you can do the casting inside the function instead of in the callers
> (which is prone to bugs).
Oh, yes. __page_file_index(page) maybe called many times in marco, althouth
currently
it is not. So that, not all are suitable for page_to_sector(). And for this
example,
still need to use "<< PAGE_SECTORS_SHIFT".
>
> Also, this is a great example of why page_to_sector() was a better name
> than npage_to_sectors(). This function doesn't return a count of sectors,
> it returns a sector number within the swap device.
OK, so I will change to page_to_sector()/sector_to_page().
>
>
> .
>
