2021-06-26 03:55:07

by Kemeng Shi

Subject: [PATCH v2] libnvdimm, badrange: replace div_u64_rem with DIV_ROUND_UP

__add_badblock_range use div_u64_rem to round up end_sector and it
will introduces unnecessary rem define and costly '%' operation.
So clean it with DIV_ROUND_UP.

Signed-off-by: Kemeng Shi <[email protected]>
---
V1->V2:
-- fix that end_sector is assigned twice, sorry for that.

drivers/nvdimm/badrange.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/nvdimm/badrange.c b/drivers/nvdimm/badrange.c
index aaf6e215a8c6..af622ae511aa 100644
--- a/drivers/nvdimm/badrange.c
+++ b/drivers/nvdimm/badrange.c
@@ -187,12 +187,9 @@ static void __add_badblock_range(struct badblocks *bb, u64 ns_offset, u64 len)
const unsigned int sector_size = 512;
sector_t start_sector, end_sector;
u64 num_sectors;
- u32 rem;

start_sector = div_u64(ns_offset, sector_size);
- end_sector = div_u64_rem(ns_offset + len, sector_size, &rem);
- if (rem)
- end_sector++;
+ end_sector = DIV_ROUND_UP(ns_offset + len, sector_size);
num_sectors = end_sector - start_sector;

if (unlikely(num_sectors > (u64)INT_MAX)) {
--
2.23.0

--
Best wishes
Kemeng Shi

2021-07-08 13:13:38

by Kemeng Shi

Subject: Re: [PATCH v2] libnvdimm, badrange: replace div_u64_rem with DIV_ROUND_UP

A gentle ping

on 2021/6/26 11:53, Kemeng Shi wrote:
> __add_badblock_range use div_u64_rem to round up end_sector and it
> will introduces unnecessary rem define and costly '%' operation.
> So clean it with DIV_ROUND_UP.
>
> Signed-off-by: Kemeng Shi <[email protected]>
> ---
> V1->V2:
> -- fix that end_sector is assigned twice, sorry for that.
>
> drivers/nvdimm/badrange.c | 5 +----
> 1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/drivers/nvdimm/badrange.c b/drivers/nvdimm/badrange.c
> index aaf6e215a8c6..af622ae511aa 100644
> --- a/drivers/nvdimm/badrange.c
> +++ b/drivers/nvdimm/badrange.c
> @@ -187,12 +187,9 @@ static void __add_badblock_range(struct badblocks *bb, u64 ns_offset, u64 len)
> const unsigned int sector_size = 512;
> sector_t start_sector, end_sector;
> u64 num_sectors;
> - u32 rem;
>
> start_sector = div_u64(ns_offset, sector_size);
> - end_sector = div_u64_rem(ns_offset + len, sector_size, &rem);
> - if (rem)
> - end_sector++;
> + end_sector = DIV_ROUND_UP(ns_offset + len, sector_size);
> num_sectors = end_sector - start_sector;
>
> if (unlikely(num_sectors > (u64)INT_MAX)) {
>

--
Best wishes
Kemeng Shi