run_on_cpu() doesn't seem to work reliably until the CPU has been fully
created if the single-threaded TCG main loop is already running.

Therefore, let's use run_on_cpu() for KVM only - KVM requires it due to
the initial CPU reset ioctl.

Signed-off-by: David Hildenbrand <david@redhat.com>
---
 target/s390x/cpu.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
index 40d6980229..7536dd0f03 100644
--- a/target/s390x/cpu.c
+++ b/target/s390x/cpu.c
@@ -218,11 +218,12 @@ static void s390_cpu_realizefn(DeviceState *dev, Error **errp)
 #endif
     s390_cpu_gdb_init(cs);
     qemu_init_vcpu(cs);
-#if !defined(CONFIG_USER_ONLY)
-    run_on_cpu(cs, s390_do_cpu_full_reset, RUN_ON_CPU_NULL);
-#else
-    cpu_reset(cs);
-#endif
+
+    if (kvm_enabled()) {
+        run_on_cpu(cs, s390_do_cpu_full_reset, RUN_ON_CPU_NULL);
+    } else {
+        cpu_reset(cs);
+    }
 
     scc->parent_realize(dev, &err);
 out:
-- 
2.17.1

On 20.06.2018 12:08, David Hildenbrand wrote:
> run_on_cpu() doesn't seem to work reliably until the CPU has been fully
> created if the single-threaded TCG main loop is already running.
> 
> Therefore, let's use run_on_cpu() for KVM only - KVM requires it due to
> the initial CPU reset ioctl.
> 
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
>  target/s390x/cpu.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
> index 40d6980229..7536dd0f03 100644
> --- a/target/s390x/cpu.c
> +++ b/target/s390x/cpu.c
> @@ -218,11 +218,12 @@ static void s390_cpu_realizefn(DeviceState *dev, Error **errp)
>  #endif
>      s390_cpu_gdb_init(cs);
>      qemu_init_vcpu(cs);
> -#if !defined(CONFIG_USER_ONLY)
> -    run_on_cpu(cs, s390_do_cpu_full_reset, RUN_ON_CPU_NULL);
> -#else
> -    cpu_reset(cs);
> -#endif
> +
> +    if (kvm_enabled()) {
> +        run_on_cpu(cs, s390_do_cpu_full_reset, RUN_ON_CPU_NULL);
> +    } else {
> +        cpu_reset(cs);
> +    }

Could you please add a comment with the information from the patch
description before that code? Otherwise, I guess everybody who only
looks at the code (and not at the git history) will be quite confused
about this, too.

 Thomas