Fix interrupt handling

Logic for interrupts:

- interrupt.S raises spl (thus IF cleared)
- interrupt.S EOI
- interrupt.S calls the handler
   - for pure in-kernel handlers, they do whatever they want with IF
     cleared.
   - when a userland handler is registers, queue_intr masks the irq.
- interrupt.S lowers spl with splx_cli, thus IF still cleared
- iret, that sets IF

- later on, userland acks the IRQ, that unmasks the irq

The key to this change is that all interrupts, including IPIs, are
treated the same way. Eg. the spl level is now raised before an IPI and
restored after. Also, EOI is not needed inside irq_acknowledge.

With this change and the experimental change not to dispatch threads
direct to idle processors in the scheduler, I no longer observe kernel
faults, but an occasional hang does occur.

Message-Id: <20231002033906.124427-1-damien@zamaudio.com>

1 files changed, 14 insertions, 2 deletions

diff --git a/device/intr.c b/device/intr.c
index 15029440..9035c036 100644
--- a/device/intr.c
+++ b/device/intr.c
@@ -50,6 +50,20 @@ search_intr (struct irqdev *dev, ipc_port_t dst_port)
   return NULL;
 }
 
+
+/*
+ * Interrupt handling logic:
+ *
+ * interrupt.S raises spl (thus IF cleared)
+ * interrupt.S EOI
+ * interrupt.S calls the handler
+ *   - for pure in-kernel handlers, they do whatever they want with IF cleared.
+ *   - when a userland handler is registered, queue_intr masks the irq.
+ * interrupt.S lowers spl with splx_cli, thus IF still cleared
+ * iret, that also sets IF
+ *
+ * later on, (irq_acknowledge), userland acks the IRQ, that unmasks the irq
+ */
 kern_return_t
 irq_acknowledge (ipc_port_t receive_port)
 {
@@ -76,8 +90,6 @@ irq_acknowledge (ipc_port_t receive_port)
   if (ret)
     return ret;
 
-  (*(irqtab.irqdev_ack)) (&irqtab, e->id);
-
   __enable_irq (irqtab.irq[e->id]);
 
   return D_SUCCESS;