Re: [PATCH v2 07/10] block: use sectors_to_npage() and PAGE_SECTORS to clean up code

From: Leizhen (ThunderTown)
Date: Fri May 15 2020 - 02:52:39 EST




On 2020/5/15 12:19, Matthew Wilcox wrote:
> On Thu, May 07, 2020 at 03:50:57PM +0800, Zhen Lei wrote:
>> +++ b/block/blk-settings.c
>> @@ -150,7 +150,7 @@ void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_secto
>> unsigned int max_sectors;
>>
>> if ((max_hw_sectors << 9) < PAGE_SIZE) {
>> - max_hw_sectors = 1 << (PAGE_SHIFT - 9);
>> + max_hw_sectors = PAGE_SECTORS;
>
> Surely this should be:
>
> if (max_hw_sectors < PAGE_SECTORS) {
> max_hw_sectors = PAGE_SECTORS;
>
> ... no?

I've noticed this place before. "(max_hw_sectors << 9) < PAGE_SIZE" can also make sure
that max_hw_sectors is not too large, that means (max_hw_sectors << 9) may overflow.

>
>> - page = read_mapping_page(mapping,
>> - (pgoff_t)(n >> (PAGE_SHIFT - 9)), NULL);
>> + page = read_mapping_page(mapping, (pgoff_t)sectors_to_npage(n), NULL);
>
> ... again, get the type right, and you won't need the cast.
OK, I'll consider it.

>
>
> .
>