Merge 4.9.50 into android-4.9

Changes in 4.9.50
	mtd: nand: mxc: Fix mxc_v1 ooblayout
	mtd: nand: qcom: fix read failure without complete bootchain
	mtd: nand: qcom: fix config error for BCH
	nvme-fabrics: generate spec-compliant UUID NQNs
	btrfs: resume qgroup rescan on rw remount
	selftests/x86/fsgsbase: Test selectors 1, 2, and 3
	mm/memory.c: fix mem_cgroup_oom_disable() call missing
	locktorture: Fix potential memory leak with rw lock test
	ALSA: msnd: Optimize / harden DSP and MIDI loops
	Bluetooth: Properly check L2CAP config option output buffer length
	ARM64: dts: marvell: armada-37xx: Fix GIC maintenance interrupt
	ARM: 8692/1: mm: abort uaccess retries upon fatal signal
	NFS: Fix 2 use after free issues in the I/O code
	NFS: Sync the correct byte range during synchronous writes
	xfs: XFS_IS_REALTIME_INODE() should be false if no rt device present
	Linux 4.9.50

Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

mm/memory.c

@@ -3596,6 +3596,11 @@ int handle_mm_fault(struct vm_area_struct *vma, unsigned long address,
 	/* do counter updates before entering really critical section. */
 	check_sync_rss_stat(current);
 
+	if (!arch_vma_access_permitted(vma, flags & FAULT_FLAG_WRITE,
+					    flags & FAULT_FLAG_INSTRUCTION,
+					    flags & FAULT_FLAG_REMOTE))
+		return VM_FAULT_SIGSEGV;
+
 	/*
 	 * Enable the memcg OOM handling for faults triggered in user
 	 * space.  Kernel faults are handled more gracefully.
@@ -3603,11 +3608,6 @@ int handle_mm_fault(struct vm_area_struct *vma, unsigned long address,
 	if (flags & FAULT_FLAG_USER)
 		mem_cgroup_oom_enable();
 
-	if (!arch_vma_access_permitted(vma, flags & FAULT_FLAG_WRITE,
-					    flags & FAULT_FLAG_INSTRUCTION,
-					    flags & FAULT_FLAG_REMOTE))
-		return VM_FAULT_SIGSEGV;
-
 	if (unlikely(is_vm_hugetlb_page(vma)))
 		ret = hugetlb_fault(vma->vm_mm, vma, address, flags);
 	else