[PATCH] xfsprogs: fix potential memory free issues in quota_group_type()

[PATCH] xfsprogs: fix potential memory free issues in quota_group_type()

[Li Zhong]

This patch tries to fix CID 997009 reported by Coverity scan, as 
suggested by sekharan

If getgroups() fails, gids might be malloced, and overwritten with &gid.
So the free() at the end of the function is trying to free a variable on
the stack, while the pointer to the allocated memory is lost. 

Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com>
---
 quota/quota.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/quota/quota.c b/quota/quota.c
index 7e52ad2..ae9b156 100644
--- a/quota/quota.c
+++ b/quota/quota.c
@@ -292,7 +292,8 @@ quota_group_type(
 	} else if ( ((ngroups = sysconf(_SC_NGROUPS_MAX)) < 0) ||
 		    ((gids = malloc(ngroups * sizeof(gid_t))) == NULL) ||
 		    ((ngroups = getgroups(ngroups, gids)) < 0)) {
-		dofree = (gids != NULL);
+		if (gids)
+			free(gids);
 		gid = getgid();
 		gids = &gid;
 		ngroups = 1;
-- 
1.8.1.4



_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH] xfsprogs: fix potential memory free issues in quota_group_type()

[Mark Tinguely]

On 09/22/13 01:45, Li Zhong wrote:
> This patch tries to fix CID 997009 reported by Coverity scan, as
> suggested by sekharan
>
> If getgroups() fails, gids might be malloced, and overwritten with &gid.
> So the free() at the end of the function is trying to free a variable on
> the stack, while the pointer to the allocated memory is lost.
>
> Signed-off-by: Li Zhong<zhong@linux.vnet.ibm.com>
> ---

Looks good to me.

And it seems that dofree would always be true in the following else 
statement.

Reviewed-by: Mark Tinguely <tinguely@sgi.com>

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs