[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v1 21/27] xen/riscv: introduce (de)initialization helpers for vINTC





On 4/2/26 4:58 PM, Jan Beulich wrote:
On 10.03.2026 18:08, Oleksii Kurochko wrote:
@@ -96,4 +97,7 @@ void intc_handle_external_irqs(struct cpu_user_regs *regs);
int intc_make_domu_dt_node(const struct kernel_info *kinfo); +int domain_vintc_init(struct domain *d);
+void domain_vintc_deinit(struct domain *d);

The former has a caller added, but the later hasn't.

As arch_domain_destroy() is stub at the moment, domain_vintc_deinit() isn't called.


@@ -87,3 +88,42 @@ int map_device_irqs_to_domain(struct domain *d, struct 
dt_device_node *dev,
      return d->arch.vintc->ops->map_device_irqs_to_domain(d, dev, need_mapping,
                                                           irq_ranges);
  }
+
+int __init domain_vintc_init(struct domain *d)

The __init here and below is questionable again.

This function is called during construction domain only, what happen when __init section exists. So it is okay to have __init here.

Any use cases where it isn't true?

It could be wrong for domain_vintc_deinit() as it could be called when __init section was freed.

~ Oleksii



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.