FAILED: patch "[PATCH] tee: add overflow check in register_shm_helper()" failed to apply to 5.4-stable tree

The patch below does not apply to the 5.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 573ae4f13f630d6660008f1974c0a8a29c30e18a Mon Sep 17 00:00:00 2001
From: Jens Wiklander <[email protected]>
Date: Thu, 18 Aug 2022 13:08:59 +0200
Subject: [PATCH] tee: add overflow check in register_shm_helper()

With special lengths supplied by user space, register_shm_helper() has an integer overflow when calculating the number of pages covered by a supplied user space memory region.

This causes internal_get_user_pages_fast() a helper function of pin_user_pages_fast() to do a NULL pointer dereference:

  Unable to handle kernel NULL pointer dereference at virtual address 0000000000000010
  Modules linked in:
  CPU: 1 PID: 173 Comm: optee_example_a Not tainted 5.19.0 #11
  Hardware name: QEMU QEMU Virtual Machine, BIOS 0.0.0 02/06/2015
  pc : internal_get_user_pages_fast+0x474/0xa80
  Call trace:
   internal_get_user_pages_fast+0x474/0xa80
   pin_user_pages_fast+0x24/0x4c
   register_shm_helper+0x194/0x330
   tee_shm_register_user_buf+0x78/0x120
   tee_ioctl+0xd0/0x11a0
   __arm64_sys_ioctl+0xa8/0xec
   invoke_syscall+0x48/0x114

Fix this by adding an an explicit call to access_ok() in tee_shm_register_user_buf() to catch an invalid user space address early.

index f2b1bcefcadd..1175f3a46859 100644
@@ -326,6 +326,9 @@ struct tee_shm *tee_shm_register_user_buf(struct tee_context *ctx,
 	void *ret;
 	int id;
 
+	if (!access_ok((void __user *)addr, length))
+		return ERR_PTR(-EFAULT);
+
 	mutex_lock(&teedev->mutex);
 	id = idr_alloc(&teedev->idr, NULL, 1, 0, GFP_KERNEL);
 	mutex_unlock(&teedev->mutex);

Yeah, there's some major re-org made by commit 53e16519c2ec ("tee:
replace tee_shm_register()") and related in this area in v5.18.

I think you need to just add that

        if (!access_ok((void __user *)data.addr, data.length))
                return -EFAULT;

to tee_ioctl_shm_register() just before the call to tee_shm_register().

It's where it checks "data.flags" too:

        /* Currently no input flags are supported */
        if (data.flags)
                return -EINVAL;

so it lines up with that whole "check ioctl arguments in the memory block we just copied".

But Jens should probably double-check that.

Linus

That should work, but data.addr is a u64 so to avoid a warning like:
drivers/tee/tee_core.c:185:17: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
185 | if (!access_ok((void __user *)data.addr, data.length))

We should first cast it to an unsigned long or such first.

I'll send a backported patch to take care of the warning I mentioned above.

Thanks,
Jens