[Ocfs2-devel] [PATCH] ocfs2: fix possible double free in ocfs2_write_begin_nolock

[Ocfs2-devel] [PATCH] ocfs2: fix possible double free in ocfs2_write_begin_nolock

[Xue jiufei]

When ocfs2_write_cluster_by_desc() failed in ocfs2_write_begin_nolock()
because of ENOSPC, it goes to out_quota, freeing data_ac(meta_ac). Then
it calls ocfs2_try_to_free_truncate_log() to free space. If enough
space freed, it will try to write again. Unfortunately, some error
happenes before ocfs2_lock_allocators(), it goes to out and free 
data_ac(meta_ac) again.

Signed-off-by: joyce <xuejiufei@huawei.com>
---
 fs/ocfs2/aops.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index f37d3c0..8ad0a41 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -1897,10 +1897,14 @@ out_commit:
 out:
 	ocfs2_free_write_ctxt(wc);
 
-	if (data_ac)
+	if (data_ac) {
 		ocfs2_free_alloc_context(data_ac);
-	if (meta_ac)
+		data_ac = NULL;
+	}
+	if (meta_ac) {
 		ocfs2_free_alloc_context(meta_ac);
+		meta_ac = NULL;
+	}
 
 	if (ret == -ENOSPC && try_free) {
 		/*
-- 
1.7.9.7

[Ocfs2-devel] [PATCH] ocfs2: fix possible double free in ocfs2_write_begin_nolock

[Jeff Liu]

On 10/26/2013 05:09 PM, Xue jiufei wrote:

> When ocfs2_write_cluster_by_desc() failed in ocfs2_write_begin_nolock()
> because of ENOSPC, it goes to out_quota, freeing data_ac(meta_ac). Then
> it calls ocfs2_try_to_free_truncate_log() to free space. If enough
> space freed, it will try to write again. Unfortunately, some error
> happenes before ocfs2_lock_allocators(), it goes to out and free 
> data_ac(meta_ac) again.

Looks good to me, thanks.

Reviewed-by: Jie Liu <jeff.liu@oracle.com>

> 
> Signed-off-by: joyce <xuejiufei@huawei.com>
> ---
>  fs/ocfs2/aops.c |    8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
> index f37d3c0..8ad0a41 100644
> --- a/fs/ocfs2/aops.c
> +++ b/fs/ocfs2/aops.c
> @@ -1897,10 +1897,14 @@ out_commit:
>  out:
>  	ocfs2_free_write_ctxt(wc);
>  
> -	if (data_ac)
> +	if (data_ac) {
>  		ocfs2_free_alloc_context(data_ac);
> -	if (meta_ac)
> +		data_ac = NULL;
> +	}
> +	if (meta_ac) {
>  		ocfs2_free_alloc_context(meta_ac);
> +		meta_ac = NULL;
> +	}
>  
>  	if (ret == -ENOSPC && try_free) {
>  		/*

[Ocfs2-devel] [PATCH] ocfs2: fix possible double free in ocfs2_write_begin_nolock

[Joel Becker]

On Sat, Oct 26, 2013 at 05:09:52PM +0800, Xue jiufei wrote:
> When ocfs2_write_cluster_by_desc() failed in ocfs2_write_begin_nolock()
> because of ENOSPC, it goes to out_quota, freeing data_ac(meta_ac). Then
> it calls ocfs2_try_to_free_truncate_log() to free space. If enough
> space freed, it will try to write again. Unfortunately, some error
> happenes before ocfs2_lock_allocators(), it goes to out and free 
> data_ac(meta_ac) again.
> 
> Signed-off-by: joyce <xuejiufei@huawei.com>

Good catch.  This function could use some refactoring for
understandability.

Acked-by: Joel Becker <jlbec@evilplan.org>

> ---
>  fs/ocfs2/aops.c |    8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
> index f37d3c0..8ad0a41 100644
> --- a/fs/ocfs2/aops.c
> +++ b/fs/ocfs2/aops.c
> @@ -1897,10 +1897,14 @@ out_commit:
>  out:
>  	ocfs2_free_write_ctxt(wc);
>  
> -	if (data_ac)
> +	if (data_ac) {
>  		ocfs2_free_alloc_context(data_ac);
> -	if (meta_ac)
> +		data_ac = NULL;
> +	}
> +	if (meta_ac) {
>  		ocfs2_free_alloc_context(meta_ac);
> +		meta_ac = NULL;
> +	}
>  
>  	if (ret == -ENOSPC && try_free) {
>  		/*
> -- 
> 1.7.9.7
> 
> 
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> https://oss.oracle.com/mailman/listinfo/ocfs2-devel

-- 

Life's Little Instruction Book #464

	"Don't miss the magic of the moment by focusing on what's
	 to come."

			http://www.jlbec.org/
			jlbec at evilplan.org