Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1459733 > unrolled thread

[PATCH v1 07/10] sched/fair: Remove useless 64-bit to 32-bit variable conversion

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH v1 07/10] sched/fair: Remove useless 64-bit to 32-bit variable conversion Yuyang Du <yuyang.du@intel.com> - 2016-08-10 22:00 +0200
    Re: [PATCH v1 07/10] sched/fair: Remove useless 64-bit to 32-bit  variable conversion Vincent Guittot <vincent.guittot@linaro.org> - 2016-08-25 09:20 +0200

#1459733 — [PATCH v1 07/10] sched/fair: Remove useless 64-bit to 32-bit variable conversion

In __decay_sum(), the 64-bit to 32-bit variable conversion makes no
performance nor correctness use.

Minor cleanup and no functionality change.

Signed-off-by: Yuyang Du <yuyang.du@intel.com>
---
 kernel/sched/fair.c |   13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 283e2c2..eea3349 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2616,16 +2616,11 @@ static const u32 __accumulated_sum_N32[] = {
  */
 static __always_inline u64 __decay_sum(u64 val, u64 n)
 {
-	unsigned int local_n;
-
 	if (!n)
 		return val;
 	else if (unlikely(n > SCHED_AVG_HALFLIFE * 63))
 		return 0;
 
-	/* after bounds checking we can collapse to 32-bit */
-	local_n = n;
-
 	/*
 	 * As y^HALFLIFE = 1/2, we can combine
 	 *    y^n = 1/2^(n/HALFLIFE) * y^(n%HALFLIFE)
@@ -2633,12 +2628,12 @@ static __always_inline u64 __decay_sum(u64 val, u64 n)
 	 *
 	 * To achieve constant time __decay_load.
 	 */
-	if (unlikely(local_n >= SCHED_AVG_HALFLIFE)) {
-		val >>= local_n / SCHED_AVG_HALFLIFE;
-		local_n %= SCHED_AVG_HALFLIFE;
+	if (unlikely(n >= SCHED_AVG_HALFLIFE)) {
+		val >>= n / SCHED_AVG_HALFLIFE;
+		n %= SCHED_AVG_HALFLIFE;
 	}
 
-	val = mul_u64_u32_shr(val, __decay_inv_multiply_N[local_n], 32);
+	val = mul_u64_u32_shr(val, __decay_inv_multiply_N[n], 32);
 	return val;
 }
 
-- 
1.7.9.5

[toc] | [next] | [standalone]

#1469887 — Re: [PATCH v1 07/10] sched/fair: Remove useless 64-bit to 32-bit variable conversion

On 10 August 2016 at 02:14, Yuyang Du <yuyang.du@intel.com> wrote:
> In __decay_sum(), the 64-bit to 32-bit variable conversion makes no
> performance nor correctness use.

Are you sure that there is no impact on 32bits system ?
Even this will be mainly shift manipulation


>
> Minor cleanup and no functionality change.
>
> Signed-off-by: Yuyang Du <yuyang.du@intel.com>
> ---
>  kernel/sched/fair.c |   13 ++++---------
>  1 file changed, 4 insertions(+), 9 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 283e2c2..eea3349 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -2616,16 +2616,11 @@ static const u32 __accumulated_sum_N32[] = {
>   */
>  static __always_inline u64 __decay_sum(u64 val, u64 n)
>  {
> -       unsigned int local_n;
> -
>         if (!n)
>                 return val;
>         else if (unlikely(n > SCHED_AVG_HALFLIFE * 63))
>                 return 0;
>
> -       /* after bounds checking we can collapse to 32-bit */
> -       local_n = n;
> -
>         /*
>          * As y^HALFLIFE = 1/2, we can combine
>          *    y^n = 1/2^(n/HALFLIFE) * y^(n%HALFLIFE)
> @@ -2633,12 +2628,12 @@ static __always_inline u64 __decay_sum(u64 val, u64 n)
>          *
>          * To achieve constant time __decay_load.
>          */
> -       if (unlikely(local_n >= SCHED_AVG_HALFLIFE)) {
> -               val >>= local_n / SCHED_AVG_HALFLIFE;
> -               local_n %= SCHED_AVG_HALFLIFE;
> +       if (unlikely(n >= SCHED_AVG_HALFLIFE)) {
> +               val >>= n / SCHED_AVG_HALFLIFE;
> +               n %= SCHED_AVG_HALFLIFE;
>         }
>
> -       val = mul_u64_u32_shr(val, __decay_inv_multiply_N[local_n], 32);
> +       val = mul_u64_u32_shr(val, __decay_inv_multiply_N[n], 32);
>         return val;
>  }
>
> --
> 1.7.9.5
>

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web