[PATCH] powerpc/sysfs: Remove redundant wait time clamps

[PATCH] powerpc/sysfs: Remove redundant wait time clamps

[Thorsten Blum]

The wait times are u64 and cannot be negative. Remove the redundant
clamps.

Signed-off-by: Thorsten Blum <blum@kernel.org>
---
 arch/powerpc/kernel/sysfs.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
index 329c1690b5ed..2100c5e2c821 100644
--- a/arch/powerpc/kernel/sysfs.c
+++ b/arch/powerpc/kernel/sysfs.c
@@ -338,7 +338,7 @@ static ssize_t show_pw20_wait_time(struct device *dev,
 		time = pw20_wt;
 	}
 
-	return sysfs_emit(buf, "%llu\n", time > 0 ? time : 0);
+	return sysfs_emit(buf, "%llu\n", time);
 }
 
 static void set_pw20_wait_entry_bit(void *val)
@@ -460,7 +460,7 @@ static ssize_t show_altivec_idle_wait_time(struct device *dev,
 		time = altivec_idle_wt;
 	}
 
-	return sysfs_emit(buf, "%llu\n", time > 0 ? time : 0);
+	return sysfs_emit(buf, "%llu\n", time);
 }
 
 static void set_altivec_idle_wait_entry_bit(void *val)

Re: [PATCH] powerpc/sysfs: Remove redundant wait time clamps

[Amit Machhiwal]

On 2026/09/16 05:12 PM, Thorsten Blum wrote:
> The wait times are u64 and cannot be negative. Remove the redundant
> clamps.
> 
> Signed-off-by: Thorsten Blum <blum@kernel.org>
> ---
>  arch/powerpc/kernel/sysfs.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
> index 329c1690b5ed..2100c5e2c821 100644
> --- a/arch/powerpc/kernel/sysfs.c
> +++ b/arch/powerpc/kernel/sysfs.c
> @@ -338,7 +338,7 @@ static ssize_t show_pw20_wait_time(struct device *dev,
>  		time = pw20_wt;
>  	}
>  
> -	return sysfs_emit(buf, "%llu\n", time > 0 ? time : 0);

The time variables in both show_pw20_wait_time() and show_altivec_idle_wait_time()
are u64, so the expression time > 0 ? time : 0 is indeed a no-op for all
possible values — when time == 0 the ternary also yields 0, and negative values
are impossible by type.  The simplification is correct and removes a subtly
misleading pattern that implies a signed type.

Nit: the commit message could also mention that when time == 0, the ternary
still yields 0, making the conditional identity-preserving for all possible
inputs — not just the non-negative ones.  Strengthens the rationale a little,
but not a blocker.

Reviewed-by: Amit Machhiwal <amachhiw@linux.ibm.com>

Thanks,
Amit

> +	return sysfs_emit(buf, "%llu\n", time);
>  }
>  
>  static void set_pw20_wait_entry_bit(void *val)
> @@ -460,7 +460,7 @@ static ssize_t show_altivec_idle_wait_time(struct device *dev,
>  		time = altivec_idle_wt;
>  	}
>  
> -	return sysfs_emit(buf, "%llu\n", time > 0 ? time : 0);
> +	return sysfs_emit(buf, "%llu\n", time);
>  }
>  
>  static void set_altivec_idle_wait_entry_bit(void *val)