2022-06-25 09:55:56

by Christophe JAILLET

[permalink] [raw]
Subject: [PATCH] powerpc/xive: Fix some incorrect memory allocation

'xibm->count' really looks like a number of bits (see how it is used in
__xive_irq_bitmap_alloc()), so use the bitmap API to allocate and free this
bitmap.

This improves semantic and potentially avoids some over memory allocation.

Fixes: eac1e731b59e ("powerpc/xive: guest exploitation of the XIVE interrupt controller")
Signed-off-by: Christophe JAILLET <[email protected]>
---
I don't cross compile, so this patch is NOT compile-tested.
---
arch/powerpc/sysdev/xive/spapr.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/sysdev/xive/spapr.c b/arch/powerpc/sysdev/xive/spapr.c
index 7d5128676e83..fdbebb14b7de 100644
--- a/arch/powerpc/sysdev/xive/spapr.c
+++ b/arch/powerpc/sysdev/xive/spapr.c
@@ -57,7 +57,7 @@ static int __init xive_irq_bitmap_add(int base, int count)
spin_lock_init(&xibm->lock);
xibm->base = base;
xibm->count = count;
- xibm->bitmap = kzalloc(xibm->count, GFP_KERNEL);
+ xibm->bitmap = bitmap_zalloc(xibm->count, GFP_KERNEL);
if (!xibm->bitmap) {
kfree(xibm);
return -ENOMEM;
@@ -75,7 +75,7 @@ static void xive_irq_bitmap_remove_all(void)

list_for_each_entry_safe(xibm, tmp, &xive_irq_bitmaps, list) {
list_del(&xibm->list);
- kfree(xibm->bitmap);
+ bitmap_free(xibm->bitmap);
kfree(xibm);
}
}
--
2.34.1


2022-06-25 11:56:30

by Cédric Le Goater

[permalink] [raw]
Subject: Re: [PATCH] powerpc/xive: Fix some incorrect memory allocation

Hello Christophe,

On 6/25/22 11:49, Christophe JAILLET wrote:
> 'xibm->count' really looks like a number of bits (see how it is used in
> __xive_irq_bitmap_alloc()), so use the bitmap API to allocate and free this
> bitmap.
>
> This improves semantic and potentially avoids some over memory allocation.

This is correct and

> Fixes: eac1e731b59e ("powerpc/xive: guest exploitation of the XIVE interrupt controller")
> Signed-off-by: Christophe JAILLET <[email protected]>
> ---
> I don't cross compile, so this patch is NOT compile-tested.


Nathan provided a fix for it already :

http://patchwork.ozlabs.org/project/linuxppc-dev/patch/[email protected]/

Thanks,

C.

> ---
> arch/powerpc/sysdev/xive/spapr.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/powerpc/sysdev/xive/spapr.c b/arch/powerpc/sysdev/xive/spapr.c
> index 7d5128676e83..fdbebb14b7de 100644
> --- a/arch/powerpc/sysdev/xive/spapr.c
> +++ b/arch/powerpc/sysdev/xive/spapr.c
> @@ -57,7 +57,7 @@ static int __init xive_irq_bitmap_add(int base, int count)
> spin_lock_init(&xibm->lock);
> xibm->base = base;
> xibm->count = count;
> - xibm->bitmap = kzalloc(xibm->count, GFP_KERNEL);
> + xibm->bitmap = bitmap_zalloc(xibm->count, GFP_KERNEL);
> if (!xibm->bitmap) {
> kfree(xibm);
> return -ENOMEM;
> @@ -75,7 +75,7 @@ static void xive_irq_bitmap_remove_all(void)
>
> list_for_each_entry_safe(xibm, tmp, &xive_irq_bitmaps, list) {
> list_del(&xibm->list);
> - kfree(xibm->bitmap);
> + bitmap_free(xibm->bitmap);
> kfree(xibm);
> }
> }