RDMA/umem: Fix potential addition overflow

Given a large enough memory allocation, it is possible to wrap the
pinned_vm counter.  Check for addition overflow to prevent such
eventualities.

Fixes: 40ddacf2dd ("RDMA/umem: Don't hold mmap_sem for too long")
Reported-by: Jason Gunthorpe <jgg@ziepe.ca>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
This commit is contained in:
Doug Ledford 2018-09-21 11:30:13 -04:00 committed by Jason Gunthorpe
parent 3312d1c6bd
commit c6ce580716

View File

@ -85,6 +85,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
struct page **page_list;
struct vm_area_struct **vma_list;
unsigned long lock_limit;
unsigned long new_pinned;
unsigned long cur_base;
struct mm_struct *mm;
unsigned long npages;
@ -160,12 +161,13 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
lock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
down_write(&mm->mmap_sem);
mm->pinned_vm += npages;
if ((mm->pinned_vm > lock_limit) && !capable(CAP_IPC_LOCK)) {
if (check_add_overflow(mm->pinned_vm, npages, &new_pinned) ||
(new_pinned > lock_limit && !capable(CAP_IPC_LOCK))) {
up_write(&mm->mmap_sem);
ret = -ENOMEM;
goto vma;
goto out;
}
mm->pinned_vm = new_pinned;
up_write(&mm->mmap_sem);
cur_base = addr & PAGE_MASK;