2010-03-24 14:02:39

by jing zhang

Subject: [PATCH] ext4: memory leakage in ext4_mb_free_blocks()

From: Jing Zhang <[email protected]>

Date: Wed Mar 24 21:54:22 2010

There is memory leakage in ext4_mb_free_blocks(), if error, double
free of blocks, occurs in ext4_mb_free_metadata().

Cc: Theodore Ts'o <[email protected]>
Cc: Andreas Dilger <[email protected]>
Cc: Dave Kleikamp <[email protected]>
Signed-off-by: Jing Zhang <[email protected]>

---

--- linux-2.6.32/fs/ext4/mballoc.c 2009-12-03 11:51:22.000000000 +0800
+++ ext4_mm_leak/mballoc8.c 2010-03-24 21:46:50.000000000 +0800
@@ -4419,7 +4419,8 @@ ext4_mb_free_metadata(handle_t *handle,
spin_lock(&sbi->s_md_lock);
list_add(&new_entry->list, &handle->h_transaction->t_private_list);
spin_unlock(&sbi->s_md_lock);
- return 0;
+ /* info the caller new_entry is used */
+ return 1;
}

/*
@@ -4547,7 +4548,8 @@ do_more:

ext4_lock_group(sb, block_group);
mb_clear_bits(bitmap_bh->b_data, bit, count);
- ext4_mb_free_metadata(handle, &e4b, new_entry);
+ if (0 == ext4_mb_free_metadata(handle, &e4b, new_entry))
+ kmem_cache_free(ext4_free_ext_cachep, new_entry);
} else {
/* need to update group_info->bb_free and bitmap
* with group lock held. generate_buddy look at

2010-03-24 17:02:02

by Aneesh Kumar K.V

Subject: Re: [PATCH] ext4: memory leakage in ext4_mb_free_blocks()

On Wed, 24 Mar 2010 21:55:55 +0800, jing zhang <[email protected]> wrote:
> From: Jing Zhang <[email protected]>
>
> Date: Wed Mar 24 21:54:22 2010
>
> There is memory leakage in ext4_mb_free_blocks(), if error, double
> free of blocks, occurs in ext4_mb_free_metadata().
>
> Cc: Theodore Ts'o <[email protected]>
> Cc: Andreas Dilger <[email protected]>
> Cc: Dave Kleikamp <[email protected]>
> Signed-off-by: Jing Zhang <[email protected]>
>
> ---
>
> --- linux-2.6.32/fs/ext4/mballoc.c 2009-12-03 11:51:22.000000000 +0800
> +++ ext4_mm_leak/mballoc8.c 2010-03-24 21:46:50.000000000 +0800
> @@ -4419,7 +4419,8 @@ ext4_mb_free_metadata(handle_t *handle,
> spin_lock(&sbi->s_md_lock);
> list_add(&new_entry->list, &handle->h_transaction->t_private_list);
> spin_unlock(&sbi->s_md_lock);
> - return 0;
> + /* info the caller new_entry is used */
> + return 1;
> }
>

return 0 on success and -1 on failure


> /*
> @@ -4547,7 +4548,8 @@ do_more:
>
> ext4_lock_group(sb, block_group);
> mb_clear_bits(bitmap_bh->b_data, bit, count);
> - ext4_mb_free_metadata(handle, &e4b, new_entry);
> + if (0 == ext4_mb_free_metadata(handle, &e4b, new_entry))
> + kmem_cache_free(ext4_free_ext_cachep,
> new_entry);

I guess it would be better to be

if (ext4_mb_free_metadat() == -1)
kmem_cache_free(..)



> } else {
> /* need to update group_info->bb_free and bitmap
> * with group lock held. generate_buddy look at


-aneesh

2010-03-25 14:09:39

by jing zhang

Subject: Re: [PATCH] ext4: memory leakage in ext4_mb_free_blocks()

2010/3/25, Aneesh Kumar K. V <[email protected]>:
> On Wed, 24 Mar 2010 21:55:55 +0800, jing zhang <[email protected]> wrote:
>> From: Jing Zhang <[email protected]>
>>
>> Date: Wed Mar 24 21:54:22 2010
>>
>> There is memory leakage in ext4_mb_free_blocks(), if error, double
>> free of blocks, occurs in ext4_mb_free_metadata().
>>
>> Cc: Theodore Ts'o <[email protected]>
>> Cc: Andreas Dilger <[email protected]>
>> Cc: Dave Kleikamp <[email protected]>
>> Signed-off-by: Jing Zhang <[email protected]>
>>
>> ---
>>
>> --- linux-2.6.32/fs/ext4/mballoc.c 2009-12-03 11:51:22.000000000 +0800
>> +++ ext4_mm_leak/mballoc8.c 2010-03-24 21:46:50.000000000 +0800
>> @@ -4419,7 +4419,8 @@ ext4_mb_free_metadata(handle_t *handle,
>> spin_lock(&sbi->s_md_lock);
>> list_add(&new_entry->list, &handle->h_transaction->t_private_list);
>> spin_unlock(&sbi->s_md_lock);
>> - return 0;
>> + /* info the caller new_entry is used */
>> + return 1;
>> }
>>
>
> return 0 on success and -1 on failure

Agree

>
>
>> /*
>> @@ -4547,7 +4548,8 @@ do_more:
>>
>> ext4_lock_group(sb, block_group);
>> mb_clear_bits(bitmap_bh->b_data, bit, count);
>> - ext4_mb_free_metadata(handle, &e4b, new_entry);
>> + if (0 == ext4_mb_free_metadata(handle, &e4b, new_entry))
>> + kmem_cache_free(ext4_free_ext_cachep,
>> new_entry);
>
> I guess it would be better to be
>
> if (ext4_mb_free_metadat() == -1)
> kmem_cache_free(..)

Fine
- zj

>
>
>
>> } else {
>> /* need to update group_info->bb_free and bitmap
>> * with group lock held. generate_buddy look at
>
>
> -aneesh
>