optee: Fix multi page dynamic shm pool alloc
authorSumit Garg <sumit.garg@linaro.org>
Mon, 30 Dec 2019 13:22:40 +0000 (18:52 +0530)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 26 Jan 2020 09:01:09 +0000 (10:01 +0100)
[ Upstream commit 5a769f6ff439cedc547395a6dc78faa26108f741 ]

optee_shm_register() expected pages to be passed as an array of page
pointers rather than as an array of contiguous pages. So fix that via
correctly passing pages as per expectation.

Fixes: a249dd200d03 ("tee: optee: Fix dynamic shm pool allocations")
Reported-by: Vincent Cao <vincent.t.cao@intel.com>
Signed-off-by: Sumit Garg <sumit.garg@linaro.org>
Tested-by: Vincent Cao <vincent.t.cao@intel.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/tee/optee/shm_pool.c

index 0332a53..d767eeb 100644 (file)
@@ -28,9 +28,22 @@ static int pool_op_alloc(struct tee_shm_pool_mgr *poolm,
        shm->size = PAGE_SIZE << order;
 
        if (shm->flags & TEE_SHM_DMA_BUF) {
+               unsigned int nr_pages = 1 << order, i;
+               struct page **pages;
+
+               pages = kcalloc(nr_pages, sizeof(pages), GFP_KERNEL);
+               if (!pages)
+                       return -ENOMEM;
+
+               for (i = 0; i < nr_pages; i++) {
+                       pages[i] = page;
+                       page++;
+               }
+
                shm->flags |= TEE_SHM_REGISTER;
-               rc = optee_shm_register(shm->ctx, shm, &page, 1 << order,
+               rc = optee_shm_register(shm->ctx, shm, pages, nr_pages,
                                        (unsigned long)shm->kaddr);
+               kfree(pages);
        }
 
        return rc;