KVM: s390: Properly lock mm context allow_gmap_hpage_1m setting

We have to do down_write on the mm semaphore to set a bitfield in the
mm context.

Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Fixes: a449938 ("KVM: s390: Add huge page enablement control")
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

arch/s390/include/asm/mmu.h

@@ -16,7 +16,13 @@ typedef struct {
 	unsigned long asce;
 	unsigned long asce_limit;
 	unsigned long vdso_base;
-	/* The mmu context allocates 4K page tables. */
+	/*
+	 * The following bitfields need a down_write on the mm
+	 * semaphore when they are written to. As they are only
+	 * written once, they can be read without a lock.
+	 *
+	 * The mmu context allocates 4K page tables.
+	 */
 	unsigned int alloc_pgste:1;
 	/* The mmu context uses extended page tables. */
 	unsigned int has_pgste:1;

arch/s390/kvm/kvm-s390.c

@@ -695,7 +695,9 @@ static int kvm_vm_ioctl_enable_cap(struct kvm *kvm, struct kvm_enable_cap *cap)
 			r = -EINVAL;
 		else {
 			r = 0;
+			down_write(&kvm->mm->mmap_sem);
 			kvm->mm->context.allow_gmap_hpage_1m = 1;
+			up_write(&kvm->mm->mmap_sem);
 			/*
 			 * We might have to create fake 4k page
 			 * tables. To avoid that the hardware works on