2022-07-13 02:54:14

by Souptick Joarder

Subject: [PATCH] mm/khugepaged: Avoid possible memory leak in failure path

From: "Souptick Joarder (HPE)" <[email protected]>

smatch warnings:
mm/khugepaged.c:2409 madvise_collapse() warn: possible memory
leak of 'cc'

Avoiding possible memory leak.

Reported-by: kernel test robot <[email protected]>
Reported-by: Dan Carpenter <[email protected]>
Signed-off-by: Souptick Joarder (HPE) <[email protected]>
---
mm/khugepaged.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 0207fc0a5b2a..8da9e7467543 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -2397,12 +2397,6 @@ int madvise_collapse(struct vm_area_struct *vma, struct vm_area_struct **prev,
BUG_ON(vma->vm_start > start);
BUG_ON(vma->vm_end < end);

- cc = kmalloc(sizeof(*cc), GFP_KERNEL);
- if (!cc)
- return -ENOMEM;
- cc->is_khugepaged = false;
- cc->last_target_node = NUMA_NO_NODE;
-
*prev = vma;

/* TODO: Support file/shmem */
@@ -2415,6 +2409,12 @@ int madvise_collapse(struct vm_area_struct *vma, struct vm_area_struct **prev,
if (!hugepage_vma_check(vma, vma->vm_flags, false, false, false))
return -EINVAL;

+ cc = kmalloc(sizeof(*cc), GFP_KERNEL);
+ if (!cc)
+ return -ENOMEM;
+ cc->is_khugepaged = false;
+ cc->last_target_node = NUMA_NO_NODE;
+
mmgrab(mm);
lru_add_drain_all();

--
2.25.1

2022-07-13 16:42:08

by Zach O'Keefe

Subject: Re: [PATCH] mm/khugepaged: Avoid possible memory leak in failure path

On Jul 13 08:11, Souptick Joarder wrote:
> From: "Souptick Joarder (HPE)" <[email protected]>
>
> smatch warnings:
> mm/khugepaged.c:2409 madvise_collapse() warn: possible memory
> leak of 'cc'
>
> Avoiding possible memory leak.
>
> Reported-by: kernel test robot <[email protected]>
> Reported-by: Dan Carpenter <[email protected]>
> Signed-off-by: Souptick Joarder (HPE) <[email protected]>
> ---
> mm/khugepaged.c | 12 ++++++------
> 1 file changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index 0207fc0a5b2a..8da9e7467543 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -2397,12 +2397,6 @@ int madvise_collapse(struct vm_area_struct *vma, struct vm_area_struct **prev,
> BUG_ON(vma->vm_start > start);
> BUG_ON(vma->vm_end < end);
>
> - cc = kmalloc(sizeof(*cc), GFP_KERNEL);
> - if (!cc)
> - return -ENOMEM;
> - cc->is_khugepaged = false;
> - cc->last_target_node = NUMA_NO_NODE;
> -
> *prev = vma;
>
> /* TODO: Support file/shmem */
> @@ -2415,6 +2409,12 @@ int madvise_collapse(struct vm_area_struct *vma, struct vm_area_struct **prev,
> if (!hugepage_vma_check(vma, vma->vm_flags, false, false, false))
> return -EINVAL;
>
> + cc = kmalloc(sizeof(*cc), GFP_KERNEL);
> + if (!cc)
> + return -ENOMEM;
> + cc->is_khugepaged = false;
> + cc->last_target_node = NUMA_NO_NODE;
> +
> mmgrab(mm);
> lru_add_drain_all();
>
> --
> 2.25.1
>
>

Thanks for the fix!

I've also sent out another patch on top which addresses the actual missing
kfree()