From d92982a335f3e5313546959a3b31b48f899cc7d3 Mon Sep 17 00:00:00 2001 From: Frank Mori Hess Date: Thu, 23 Aug 2001 16:27:25 +0000 Subject: [PATCH] added a comment --- comedi/drivers/amplc_pci230.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/comedi/drivers/amplc_pci230.c b/comedi/drivers/amplc_pci230.c index 7539198d..acddf4f6 100644 --- a/comedi/drivers/amplc_pci230.c +++ b/comedi/drivers/amplc_pci230.c @@ -918,8 +918,16 @@ static void pci230_interrupt(int irq, void *d, struct pt_regs *regs) * we can't even talk to the board (base addresses for native * IO regions aren't set...) */ + /* XXX doesn't make sense, since interrupt is allocated in the * _attach function, _after_ the IO regions are allocated */ + + /* XXX I (Frank Hess) do this in drivers to prevent a null + * dereference when the handler tries to use async. This + * probably won't help too much here as it is though, since + * it doesn't clear the interrupt before returning and so + * will lock up the (single cpu) computer. */ + if(dev->attached == 0) { comedi_error(dev, "premature interrupt"); return; -- 2.26.2