hugetlbfs: fix memory leak for resv_map
authorMike Kravetz <mike.kravetz@oracle.com>
Sat, 6 Apr 2019 01:39:06 +0000 (18:39 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 8 May 2019 05:21:51 +0000 (07:21 +0200)
commitb51fdcbe45d1f09b5ecea5d2b796d37b8e6cbb4c
tree1b8eb3dbfb9d7ac5ecea4e56dae5206e5ec2573e
parent6a62bbe82343f8dc44df03e0f8b3a153a2024f4c
hugetlbfs: fix memory leak for resv_map

[ Upstream commit 58b6e5e8f1addd44583d61b0a03c0f5519527e35 ]

When mknod is used to create a block special file in hugetlbfs, it will
allocate an inode and kmalloc a 'struct resv_map' via resv_map_alloc().
inode->i_mapping->private_data will point the newly allocated resv_map.
However, when the device special file is opened bd_acquire() will set
inode->i_mapping to bd_inode->i_mapping.  Thus the pointer to the
allocated resv_map is lost and the structure is leaked.

Programs to reproduce:
        mount -t hugetlbfs nodev hugetlbfs
        mknod hugetlbfs/dev b 0 0
        exec 30<> hugetlbfs/dev
        umount hugetlbfs/

resv_map structures are only needed for inodes which can have associated
page allocations.  To fix the leak, only allocate resv_map for those
inodes which could possibly be associated with page allocations.

Link: http://lkml.kernel.org/r/20190401213101.16476-1-mike.kravetz@oracle.com
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Reported-by: Yufen Yu <yuyufen@huawei.com>
Suggested-by: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
fs/hugetlbfs/inode.c