intel_iommu: reject broken EIM

Cluster x2APIC cannot work without KVM's x2apic API when the maximal
APIC ID is greater than 8 and only KVM's LAPIC can support x2APIC, so we
forbid other APICs and also the old KVM case with less than 9, to
simplify the code.

There is no point in enabling EIM in forbidden APICs, so we keep it
enabled only for the KVM APIC;  unconditionally, because making the
option depend on KVM version would be a maintanance burden.

Old QEMUs would enable eim whenever intremap was on, which would trick
guests into thinking that they can enable cluster x2APIC even if any
interrupt destination would get clamped to 8 bits.
Depending on your configuration, QEMU could notice that the destination
LAPIC is not present and report it with a very non-obvious:

  KVM: injection failed, MSI lost (Operation not permitted)

Or the guest could say something about unexpected interrupts, because
clamping leads to aliasing so interrupts were being delivered to
incorrect VCPUs.

KVM_X2APIC_API is the feature that allows us to enable EIM for KVM.

QEMU 2.7 allowed EIM whenever interrupt remapping was enabled.  In order
to keep backward compatibility, we again allow guests to misbehave in
non-obvious ways, and make it the default for old machine types.

A user can enable the buggy mode it with "x-buggy-eim=on".

Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

6 files changed, 39 insertions, 1 deletions

diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index a70aa8494e..1655a65bce 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -34,6 +34,7 @@
 #include "hw/pci-host/q35.h"
 #include "sysemu/kvm.h"
 #include "hw/i386/apic_internal.h"
+#include "kvm_i386.h"
 
 /*#define DEBUG_INTEL_IOMMU*/
 #ifdef DEBUG_INTEL_IOMMU
@@ -2015,6 +2016,7 @@ static Property vtd_properties[] = {
     DEFINE_PROP_UINT32("version", IntelIOMMUState, version, 0),
     DEFINE_PROP_ON_OFF_AUTO("eim", IntelIOMMUState, intr_eim,
                             ON_OFF_AUTO_AUTO),
+    DEFINE_PROP_BOOL("x-buggy-eim", IntelIOMMUState, buggy_eim, false),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -2484,9 +2486,21 @@ static bool vtd_decide_config(IntelIOMMUState *s, Error **errp)
     }
 
     if (s->intr_eim == ON_OFF_AUTO_AUTO) {
-        s->intr_eim = x86_iommu->intr_supported ?
+        s->intr_eim = (kvm_irqchip_in_kernel() || s->buggy_eim)
+                      && x86_iommu->intr_supported ?
                                               ON_OFF_AUTO_ON : ON_OFF_AUTO_OFF;
     }
+    if (s->intr_eim == ON_OFF_AUTO_ON && !s->buggy_eim) {
+        if (!kvm_irqchip_in_kernel()) {
+            error_setg(errp, "eim=on requires accel=kvm,kernel-irqchip=split");
+            return false;
+        }
+        if (!kvm_enable_x2apic()) {
+            error_setg(errp, "eim=on requires support on the KVM side"
+                             "(X2APIC_API, first shipped in v4.7)");
+            return false;
+        }
+    }
 
     return true;
 }
diff --git a/include/hw/compat.h b/include/hw/compat.h
index ef3fae3e1b..0f06e113be 100644
--- a/include/hw/compat.h
+++ b/include/hw/compat.h
@@ -14,6 +14,10 @@
         .driver   = "ioapic",\
         .property = "version",\
         .value    = "0x11",\
+    },{\
+        .driver   = "intel-iommu",\
+        .property = "x-buggy-eim",\
+        .value    = "true",\
     },
 
 #define HW_COMPAT_2_6 \
diff --git a/include/hw/i386/intel_iommu.h b/include/hw/i386/intel_iommu.h
index b5ac60927b..1989c1eec1 100644
--- a/include/hw/i386/intel_iommu.h
+++ b/include/hw/i386/intel_iommu.h
@@ -290,6 +290,7 @@ struct IntelIOMMUState {
     uint32_t intr_size;             /* Number of IR table entries */
     bool intr_eime;                 /* Extended interrupt mode enabled */
     OnOffAuto intr_eim;             /* Toggle for EIM cabability */
+    bool buggy_eim;                 /* Force buggy EIM unless eim=off */
 };
 
 /* Find the VTD Address space associated with the given bus pointer,
diff --git a/target-i386/kvm-stub.c b/target-i386/kvm-stub.c
index cdf1506109..bda4dc2f0c 100644
--- a/target-i386/kvm-stub.c
+++ b/target-i386/kvm-stub.c
@@ -25,6 +25,11 @@ bool kvm_has_smm(void)
     return 1;
 }
 
+bool kvm_enable_x2apic(void)
+{
+    return false;
+}
+
 /* This function is only called inside conditionals which we
  * rely on the compiler to optimize out when CONFIG_KVM is not
  * defined.
diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index ee1f53e569..0fd6646486 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -122,6 +122,19 @@ bool kvm_allows_irq0_override(void)
     return !kvm_irqchip_in_kernel() || kvm_has_gsi_routing();
 }
 
+static bool kvm_x2apic_api_set_flags(uint64_t flags)
+{
+    KVMState *s = KVM_STATE(current_machine->accelerator);
+
+    return !kvm_vm_enable_cap(s, KVM_CAP_X2APIC_API, 0, flags);
+}
+
+bool kvm_enable_x2apic(void)
+{
+    return kvm_x2apic_api_set_flags(KVM_X2APIC_API_USE_32BIT_IDS |
+                                    KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK);
+}
+
 static int kvm_get_tsc(CPUState *cs)
 {
     X86CPU *cpu = X86_CPU(cs);
diff --git a/target-i386/kvm_i386.h b/target-i386/kvm_i386.h
index 36407e0a5d..5c369b1c5b 100644
--- a/target-i386/kvm_i386.h
+++ b/target-i386/kvm_i386.h
@@ -43,4 +43,5 @@ int kvm_device_msix_deassign(KVMState *s, uint32_t dev_id);
 
 void kvm_put_apicbase(X86CPU *cpu, uint64_t value);
 
+bool kvm_enable_x2apic(void);
 #endif