this effectively revert "ARM: KVM: Enable in-kernel timers with user space gic" at https://github.com/qemu/qemu/commit/5d721b785fbe80170fc756e9444cf751d3a47568

qemu-kvm-arm-userirq-hacks.patch

diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index b09bad0..9606ec6 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -2523,7 +2523,7 @@ int kvm_has_intx_set_mask(void)
 
 bool kvm_arm_supports_user_irq(void)
 {
-    return kvm_check_extension(kvm_state, KVM_CAP_ARM_USER_IRQ);
+    return true; //kvm_check_extension(kvm_state, KVM_CAP_ARM_USER_IRQ);
 }
 
 #ifdef KVM_CAP_SET_GUEST_DEBUG
diff --git a/target/arm/kvm.c b/target/arm/kvm.c
index b2eaa50..e8fcaa4 100644
--- a/target/arm/kvm.c
+++ b/target/arm/kvm.c
@@ -191,7 +191,7 @@ int kvm_arch_init(MachineState *ms, KVMState *s)
      * PSCI wakes up secondary cores, so we always need to
      * have vCPUs waiting in kernel space
      */
-    kvm_halt_in_kernel_allowed = true;
+    //kvm_halt_in_kernel_allowed = true; //userirq hacks
 
     cap_has_mp_state = kvm_check_extension(s, KVM_CAP_MP_STATE);
 
@@ -618,7 +618,7 @@ MemTxAttrs kvm_arch_post_run(CPUState *cs, struct kvm_run *run)
     ARMCPU *cpu;
     uint32_t switched_level;
 
-    if (kvm_irqchip_in_kernel()) {
+    if (1) { //kvm_irqchip_in_kernel()) {
         /*
          * We only need to sync timer states with user-space interrupt
          * controllers, so return early and save cycles if we don't.