Previous commits (in particular "Create CPUs once in MPCore parent") allowed creating the MPCore private container vCPUs within the container. We don't need to call qemu_get_cpu(), which is unsafe in hegerogeneous context. Directly access the CortexMPPrivState::cpu array.
Signed-off-by: Philippe Mathieu-Daudé <phi...@linaro.org> --- hw/cpu/a15mpcore.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/cpu/a15mpcore.c b/hw/cpu/a15mpcore.c index 5f28a97adb..8dccf15103 100644 --- a/hw/cpu/a15mpcore.c +++ b/hw/cpu/a15mpcore.c @@ -44,7 +44,7 @@ static void a15mp_priv_realize(DeviceState *dev, Error **errp) * appropriate GIC PPI inputs */ for (i = 0; i < c->num_cores; i++) { - DeviceState *cpudev = DEVICE(qemu_get_cpu(i)); + DeviceState *cpudev = DEVICE(c->cpu[i]); int ppibase = c->gic_spi_num - 32 + i * 32; int irq; /* Mapping from the output timer irq lines from the CPU to the -- 2.41.0