It would have been possible for a rogue client-core to send in a symlink
target which is not NUL terminated. This returns EIO if the client-core
gives us corrupt data.
Leave debugfs and superblock code as is for now.
Other dcache.c and namei.c strncpy instances are safe because
ORANGEFS_NAME_MAX = NAME_MAX + 1; there is always enough space for a
name plus a NUL byte.
Signed-off-by: Martin Brandenburg <martin@xxxxxxxxxxxx>
---
fs/orangefs/orangefs-utils.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/fs/orangefs/orangefs-utils.c b/fs/orangefs/orangefs-utils.c
index 40f5163..f392a6a 100644
--- a/fs/orangefs/orangefs-utils.c
+++ b/fs/orangefs/orangefs-utils.c
@@ -315,9 +315,13 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size)
inode->i_size = (loff_t)strlen(new_op->
downcall.resp.getattr.link_target);
orangefs_inode->blksize = (1 << inode->i_blkbits);
- strlcpy(orangefs_inode->link_target,
+ ret = strscpy(orangefs_inode->link_target,
new_op->downcall.resp.getattr.link_target,
ORANGEFS_NAME_MAX);
+ if (ret == -E2BIG) {
+ ret = -EIO;
+ goto out;
+ }
inode->i_link = orangefs_inode->link_target;
}
break;
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
