2018-10-10 17:27:21

by Khazhismel Kumykov

[permalink] [raw]
Subject: [PATCH] fs/fat: add cond_resched to fat_count_free_clusters

On non-preempt kernels this loop can take a long time (more than 50
ticks) processing through entries.

Signed-off-by: Khazhismel Kumykov <[email protected]>
---
fs/fat/fatent.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c
index defc2168de91..f58c0cacc531 100644
--- a/fs/fat/fatent.c
+++ b/fs/fat/fatent.c
@@ -682,6 +682,7 @@ int fat_count_free_clusters(struct super_block *sb)
if (ops->ent_get(&fatent) == FAT_ENT_FREE)
free++;
} while (fat_ent_next(sbi, &fatent));
+ cond_resched();
}
sbi->free_clusters = free;
sbi->free_clus_valid = 1;
--
2.19.0.605.g01d371f741-goog



2018-10-11 19:37:45

by OGAWA Hirofumi

[permalink] [raw]
Subject: Re: [PATCH] fs/fat: add cond_resched to fat_count_free_clusters

Khazhismel Kumykov <[email protected]> writes:

> On non-preempt kernels this loop can take a long time (more than 50
> ticks) processing through entries.
>
> Signed-off-by: Khazhismel Kumykov <[email protected]>
> ---
> fs/fat/fatent.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c
> index defc2168de91..f58c0cacc531 100644
> --- a/fs/fat/fatent.c
> +++ b/fs/fat/fatent.c
> @@ -682,6 +682,7 @@ int fat_count_free_clusters(struct super_block *sb)
> if (ops->ent_get(&fatent) == FAT_ENT_FREE)
> free++;
> } while (fat_ent_next(sbi, &fatent));
> + cond_resched();
> }
> sbi->free_clusters = free;
> sbi->free_clus_valid = 1;

Acked-by: OGAWA Hirofumi <[email protected]>

Thanks, looks good.
--
OGAWA Hirofumi <[email protected]>