Since non-threaded IRQ handlers can't use IRQF_ONESHOT and we return directly from the IRQ handler after queueing a worker thread, the IRQ handler will be called continuously on level-triggered interrupts.
At least on the author's platform, this causes CPU stalls detected by rcu_preempt, since the IRQ handler will be called too many times.
This patch emulates IRQF_ONESHOT behaviour by disabling the IRQ in the IRQ handler and re-enabling it again at the end of the worker thread.
---
drivers/tty/serial/sc16is7xx.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c
index edb5305..8a6aa71 100644
--- a/drivers/tty/serial/sc16is7xx.c
+++ b/drivers/tty/serial/sc16is7xx.c
@@ -695,12 +695,15 @@ static void sc16is7xx_ist(struct kthread_work *ws)
for (i = 0; i < s->devtype->nr_uart; ++i)
sc16is7xx_port_irq(s, i);
+
+ enable_irq(s->p[0].port.irq);
}
static irqreturn_t sc16is7xx_irq(int irq, void *dev_id)
{
struct sc16is7xx_port *s = (struct sc16is7xx_port *)dev_id;
+ disable_irq_nosync(irq);
queue_kthread_work(&s->kworker, &s->irq_work);
return IRQ_HANDLED;
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-serial" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
