2022-05-28 20:57:57

by Nam Cao

Subject: [PATCH] staging: vt6655: remove unnecessary type cast

Remove le64_to_cpu(), as the type is already u64, as reported by sparse:
drivers/staging/vt6655/card.c:758:16: warning: cast to restricted __le64

Signed-off-by: Nam Cao <[email protected]>
---
drivers/staging/vt6655/card.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/vt6655/card.c b/drivers/staging/vt6655/card.c
index 2cde0082fc03..852300c59980 100644
--- a/drivers/staging/vt6655/card.c
+++ b/drivers/staging/vt6655/card.c
@@ -755,7 +755,7 @@ u64 vt6655_get_current_tsf(struct vnt_private *priv)
return 0;
low = ioread32(iobase + MAC_REG_TSFCNTR);
high = ioread32(iobase + MAC_REG_TSFCNTR + 4);
- return le64_to_cpu(low + ((u64)high << 32));
+ return low + ((u64)high << 32);
}

/*
--
2.25.1

2022-05-30 13:54:38

by Nam Cao

Subject: Re: [PATCH] staging: vt6655: remove unnecessary type cast

Yes, this patch is a mistake. I clarified that but I think my email
got blocked because it's html.

Sorry for wasting your time.

Best regards,
Nam

2022-05-31 11:20:49

by Dan Carpenter

Subject: Re: [PATCH] staging: vt6655: remove unnecessary type cast

This is not the correct subject. The le64_to_cpu() is not a decorative
feature which can be added or removed without affecting functionality.
This patch either fixes a bug or it introduces a bug.

On Sat, May 28, 2022 at 10:28:31PM +0200, Nam Cao wrote:
> Remove le64_to_cpu(), as the type is already u64, as reported by sparse:
> drivers/staging/vt6655/card.c:758:16: warning: cast to restricted __le64
>
> Signed-off-by: Nam Cao <[email protected]>
> ---
> drivers/staging/vt6655/card.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/staging/vt6655/card.c b/drivers/staging/vt6655/card.c
> index 2cde0082fc03..852300c59980 100644
> --- a/drivers/staging/vt6655/card.c
> +++ b/drivers/staging/vt6655/card.c
> @@ -755,7 +755,7 @@ u64 vt6655_get_current_tsf(struct vnt_private *priv)
> return 0;
> low = ioread32(iobase + MAC_REG_TSFCNTR);
> high = ioread32(iobase + MAC_REG_TSFCNTR + 4);
> - return le64_to_cpu(low + ((u64)high << 32));
> + return low + ((u64)high << 32);

I fee like the original code would have been more readable if it were
written as:

return le64_to_cpu(((u64)high << 32) | low);

This seems like a correct way to combine two halves of a le64 value. So
the original code looks correct and the patch introduces a bug.

regards,
dan carpenter