]> git.hungrycats.org Git - linux/commitdiff
KVM: s390: Fix user triggerable bug in dead code
authorChristian Borntraeger <borntraeger@de.ibm.com>
Wed, 6 Aug 2014 14:17:58 +0000 (16:17 +0200)
committerZefan Li <lizefan@huawei.com>
Mon, 1 Dec 2014 10:02:20 +0000 (18:02 +0800)
commit 614a80e474b227cace52fd6e3c790554db8a396e upstream.

In the early days, we had some special handling for the
KVM_EXIT_S390_SIEIC exit, but this was gone in 2009 with commit
d7b0b5eb3000 (KVM: s390: Make psw available on all exits, not
just a subset).

Now this switch statement is just a sanity check for userspace
not messing with the kvm_run structure. Unfortunately, this
allows userspace to trigger a kernel BUG. Let's just remove
this switch statement.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
[lizf: Backported to 3.4:
 - adjust context
 - no KVM_EXIT_S390_TSCH and KVM_EXIT_DEBUG in 3.4]
Signed-off-by: Zefan Li <lizefan@huawei.com>
arch/s390/kvm/kvm-s390.c

index 53c973ebc0c8f1abdd403d1d9d9c34240cd341cb..0f250d12a18cbd6ca6d354c5c58c895515006141 100644 (file)
@@ -566,17 +566,6 @@ rerun_vcpu:
 
        BUG_ON(vcpu->kvm->arch.float_int.local_int[vcpu->vcpu_id] == NULL);
 
-       switch (kvm_run->exit_reason) {
-       case KVM_EXIT_S390_SIEIC:
-       case KVM_EXIT_UNKNOWN:
-       case KVM_EXIT_INTR:
-       case KVM_EXIT_S390_RESET:
-       case KVM_EXIT_S390_UCONTROL:
-               break;
-       default:
-               BUG();
-       }
-
        vcpu->arch.sie_block->gpsw.mask = kvm_run->psw_mask;
        vcpu->arch.sie_block->gpsw.addr = kvm_run->psw_addr;
        if (kvm_run->kvm_dirty_regs & KVM_SYNC_PREFIX) {