[PATCH] xfs: use kmalloc_array() instead of kmalloc() for map allocation

[PATCH] xfs: use kmalloc_array() instead of kmalloc() for map allocation

[Kriish Sharma]

Using kmalloc_array() better reflects the intent to allocate an array of
map entries, and improves consistency with similar allocations across the
kernel.

No functional change intended.

Signed-off-by: Kriish Sharma <kriish.sharma2006@gmail.com>
---
 fs/xfs/xfs_qm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index 23ba84ec919a..34ec61e455ff 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -1218,7 +1218,7 @@ xfs_qm_reset_dqcounts_buf(
 	if (qip->i_nblocks == 0)
 		return 0;
 
-	map = kmalloc(XFS_DQITER_MAP_SIZE * sizeof(*map),
+	map = kmalloc_array(XFS_DQITER_MAP_SIZE, sizeof(*map),
 			GFP_KERNEL | __GFP_NOFAIL);
 
 	lblkno = 0;
-- 
2.34.1

Re: [PATCH] xfs: use kmalloc_array() instead of kmalloc() for map allocation

[Carlos Maiolino]

Hello.

On Sat, Oct 18, 2025 at 07:45:28PM +0000, Kriish Sharma wrote:
> Using kmalloc_array() better reflects the intent to allocate an array of
> map entries, and improves consistency with similar allocations across the
> kernel.
> 
> No functional change intended.

Thanks for the patch.
Have you ran xfstests against this patch? All patches sent to xfs should
at least pass basic testing via xfstests. I.e. ensure there is no new
failure between an unpatched kernel and a kernel with your patch.

If you did, please send me the results summary, if you did not, please
run it, and send a V2 of this patch  including in the patch description
the summary of the results.

Any changes includes risk, and so every patch sent to xfs should be at
least basically tested against xfstests auto group.

Thanks,
Carlos

> 
> Signed-off-by: Kriish Sharma <kriish.sharma2006@gmail.com>
> ---
>  fs/xfs/xfs_qm.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
> index 23ba84ec919a..34ec61e455ff 100644
> --- a/fs/xfs/xfs_qm.c
> +++ b/fs/xfs/xfs_qm.c
> @@ -1218,7 +1218,7 @@ xfs_qm_reset_dqcounts_buf(
>  	if (qip->i_nblocks == 0)
>  		return 0;
> 
> -	map = kmalloc(XFS_DQITER_MAP_SIZE * sizeof(*map),
> +	map = kmalloc_array(XFS_DQITER_MAP_SIZE, sizeof(*map),
>  			GFP_KERNEL | __GFP_NOFAIL);
> 
>  	lblkno = 0;
> --
> 2.34.1
> 

Re: [PATCH] xfs: use kmalloc_array() instead of kmalloc() for map allocation

[Darrick J. Wong]

On Sat, Oct 18, 2025 at 07:45:28PM +0000, Kriish Sharma wrote:
> Using kmalloc_array() better reflects the intent to allocate an array of
> map entries, and improves consistency with similar allocations across the
> kernel.
> 
> No functional change intended.
> 
> Signed-off-by: Kriish Sharma <kriish.sharma2006@gmail.com>
> ---
>  fs/xfs/xfs_qm.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
> index 23ba84ec919a..34ec61e455ff 100644
> --- a/fs/xfs/xfs_qm.c
> +++ b/fs/xfs/xfs_qm.c
> @@ -1218,7 +1218,7 @@ xfs_qm_reset_dqcounts_buf(
>  	if (qip->i_nblocks == 0)
>  		return 0;
>  
> -	map = kmalloc(XFS_DQITER_MAP_SIZE * sizeof(*map),
> +	map = kmalloc_array(XFS_DQITER_MAP_SIZE, sizeof(*map),

Why would we replace a constant that can be computed at compile time
with a dynamic computation that is now fallible despite NOFAIL?

--D

>  			GFP_KERNEL | __GFP_NOFAIL);
>  
>  	lblkno = 0;
> -- 
> 2.34.1
> 
> 

Re: [PATCH] xfs: use kmalloc_array() instead of kmalloc() for map allocation

[Nirjhar Roy (IBM)]

On Sat, 2025-10-18 at 19:45 +0000, Kriish Sharma wrote:
> Using kmalloc_array() better reflects the intent to allocate an array of
> map entries, and improves consistency with similar allocations across the
> kernel.
> 
> No functional change intended.
> 
> Signed-off-by: Kriish Sharma <kriish.sharma2006@gmail.com>
> ---
>  fs/xfs/xfs_qm.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
> index 23ba84ec919a..34ec61e455ff 100644
> --- a/fs/xfs/xfs_qm.c
> +++ b/fs/xfs/xfs_qm.c
> @@ -1218,7 +1218,7 @@ xfs_qm_reset_dqcounts_buf(
>  	if (qip->i_nblocks == 0)
>  		return 0;
>  
> -	map = kmalloc(XFS_DQITER_MAP_SIZE * sizeof(*map),
> +	map = kmalloc_array(XFS_DQITER_MAP_SIZE, sizeof(*map),
>  			GFP_KERNEL | __GFP_NOFAIL);
I think kmalloc_array is more useful when the size of memory to be allocated is dynamic with
kmalloc_array doing some additional checks on the size that is being passed. In this case, both
XFS_QUITER_MAP_SIZE  and sizeof(*map) are constant i.e, we are allocating constant size memory, so
maybe this isn't quite necessary?

Definition of kmalloc_array()

static inline void *kmalloc_array(size_t n, size_t size, gfp_t flags)
{
	if (size != 0 && n > SIZE_MAX / size)
		return NULL;
	return kmalloc(n * size, flags);
}


--NR
>  
>  	lblkno = 0;