diff --git a/hw/core/machine.c b/hw/core/machine.c
index b03d9192bafb..bb0dc8f6a93d 100644
--- a/hw/core/machine.c
+++ b/hw/core/machine.c
@@ -41,7 +41,9 @@
 #include "hw/virtio/virtio-pci.h"
 #include "qom/object_interfaces.h"
 
-GlobalProperty hw_compat_7_0[] = {};
+GlobalProperty hw_compat_7_0[] = {
+    { "arm-gicv3-common", "force-8-bit-prio", "on" },
+};
 const size_t hw_compat_7_0_len = G_N_ELEMENTS(hw_compat_7_0);
 
 GlobalProperty hw_compat_6_2[] = {
diff --git a/hw/intc/arm_gicv3_common.c b/hw/intc/arm_gicv3_common.c
index 5634c6fc7882..351843db4aac 100644
--- a/hw/intc/arm_gicv3_common.c
+++ b/hw/intc/arm_gicv3_common.c
@@ -563,6 +563,11 @@ static Property arm_gicv3_common_properties[] = {
     DEFINE_PROP_UINT32("revision", GICv3State, revision, 3),
     DEFINE_PROP_BOOL("has-lpi", GICv3State, lpi_enable, 0),
     DEFINE_PROP_BOOL("has-security-extensions", GICv3State, security_extn, 0),
+    /*
+     * Compatibility property: force 8 bits of physical priority, even
+     * if the CPU being emulated should have fewer.
+     */
+    DEFINE_PROP_BOOL("force-8-bit-prio", GICv3State, force_8bit_prio, 0),
     DEFINE_PROP_ARRAY("redist-region-count", GICv3State, nb_redist_regions,
                       redist_region_count, qdev_prop_uint32, uint32_t),
     DEFINE_PROP_LINK("sysmem", GICv3State, dma, TYPE_MEMORY_REGION,
diff --git a/hw/intc/arm_gicv3_cpuif.c b/hw/intc/arm_gicv3_cpuif.c
index 69a15f7a4442..66e06b787c72 100644
--- a/hw/intc/arm_gicv3_cpuif.c
+++ b/hw/intc/arm_gicv3_cpuif.c
@@ -2798,6 +2798,7 @@ void gicv3_init_cpuif(GICv3State *s)
          *  cpu->gic_num_lrs
          *  cpu->gic_vpribits
          *  cpu->gic_vprebits
+         *  cpu->gic_pribits
          */
 
         /* Note that we can't just use the GICv3CPUState as an opaque pointer
@@ -2810,11 +2811,17 @@ void gicv3_init_cpuif(GICv3State *s)
         define_arm_cp_regs(cpu, gicv3_cpuif_reginfo);
 
         /*
-         * For the moment, retain the existing behaviour of 8 priority bits;
-         * in a following commit we will take this from the CPU state,
-         * as we do for the virtual priority bits.
+         * The CPU implementation specifies the number of supported
+         * bits of physical priority. For backwards compatibility
+         * of migration, we have a compat property that forces use
+         * of 8 priority bits regardless of what the CPU really has.
          */
-        cs->pribits = 8;
+        if (s->force_8bit_prio) {
+            cs->pribits = 8;
+        } else {
+            cs->pribits = cpu->gic_pribits ?: 5;
+        }
+
         /*
          * The GICv3 has separate ID register fields for virtual priority
          * and preemption bit values, but only a single ID register field
diff --git a/include/hw/intc/arm_gicv3_common.h b/include/hw/intc/arm_gicv3_common.h
index 46677ec345c5..ab5182a28a2c 100644
--- a/include/hw/intc/arm_gicv3_common.h
+++ b/include/hw/intc/arm_gicv3_common.h
@@ -248,6 +248,7 @@ struct GICv3State {
     uint32_t revision;
     bool lpi_enable;
     bool security_extn;
+    bool force_8bit_prio;
     bool irq_reset_nonsecure;
     bool gicd_no_migration_shift_bug;
 
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index a99b430e54e7..a42464eb57a8 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -1002,6 +1002,7 @@ struct ArchCPU {
     int gic_num_lrs; /* number of list registers */
     int gic_vpribits; /* number of virtual priority bits */
     int gic_vprebits; /* number of virtual preemption bits */
+    int gic_pribits; /* number of physical priority bits */
 
     /* Whether the cfgend input is high (i.e. this CPU should reset into
      * big-endian mode).  This setting isn't used directly: instead it modifies
diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index 804a54922cb8..7628f4fa39db 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -87,6 +87,7 @@ static void aarch64_a57_initfn(Object *obj)
     cpu->gic_num_lrs = 4;
     cpu->gic_vpribits = 5;
     cpu->gic_vprebits = 5;
+    cpu->gic_pribits = 5;
     define_cortex_a72_a57_a53_cp_reginfo(cpu);
 }
 
@@ -140,6 +141,7 @@ static void aarch64_a53_initfn(Object *obj)
     cpu->gic_num_lrs = 4;
     cpu->gic_vpribits = 5;
     cpu->gic_vprebits = 5;
+    cpu->gic_pribits = 5;
     define_cortex_a72_a57_a53_cp_reginfo(cpu);
 }
 
@@ -191,6 +193,7 @@ static void aarch64_a72_initfn(Object *obj)
     cpu->gic_num_lrs = 4;
     cpu->gic_vpribits = 5;
     cpu->gic_vprebits = 5;
+    cpu->gic_pribits = 5;
     define_cortex_a72_a57_a53_cp_reginfo(cpu);
 }
 
@@ -252,6 +255,7 @@ static void aarch64_a76_initfn(Object *obj)
     cpu->gic_num_lrs = 4;
     cpu->gic_vpribits = 5;
     cpu->gic_vprebits = 5;
+    cpu->gic_pribits = 5;
 
     /* From B5.1 AdvSIMD AArch64 register summary */
     cpu->isar.mvfr0 = 0x10110222;
@@ -317,6 +321,7 @@ static void aarch64_neoverse_n1_initfn(Object *obj)
     cpu->gic_num_lrs = 4;
     cpu->gic_vpribits = 5;
     cpu->gic_vprebits = 5;
+    cpu->gic_pribits = 5;
 
     /* From B5.1 AdvSIMD AArch64 register summary */
     cpu->isar.mvfr0 = 0x10110222;
@@ -1008,6 +1013,7 @@ static void aarch64_a64fx_initfn(Object *obj)
     cpu->gic_num_lrs = 4;
     cpu->gic_vpribits = 5;
     cpu->gic_vprebits = 5;
+    cpu->gic_pribits = 5;
 
     /* Suppport of A64FX's vector length are 128,256 and 512bit only */
     aarch64_add_sve_properties(obj);