On 24/04/14 19:11, Hartley Sweeten wrote:
On Thursday, April 24, 2014 4:50 AM, Ian Abbott wrote:
On 2014-04-24 00:07, H Hartley Sweeten wrote:
The helper functions that handle reading the analog input samples for
the interrupt function both call cfc_handle_events() and clear the
interrupt request at various times.
Move this to the main interrupt handler to make sure the events are
posted and the interrupt request is cleared.
Signed-off-by: H Hartley Sweeten <hsweeten@xxxxxxxxxxxxxxxxxxx>
Cc: Ian Abbott <abbotti@xxxxxxxxx>
Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/staging/comedi/drivers/adv_pci1710.c | 22 ++++------------------
1 file changed, 4 insertions(+), 18 deletions(-)
[snip]
@@ -945,6 +927,10 @@ static irqreturn_t interrupt_service_pci1710(int irq, void *d)
else
pci1710_handle_fifo(dev, s);
+ cfc_handle_events(dev, s);
+
+ outb(0, dev->iobase + PCI171x_CLRINT);
+
return IRQ_HANDLED;
}
It would be better to clear the interrupt _before_ calling
pci1710_handle_every_sample() or pci1710_handle_fifo() in case new data
arrives (and reasserts the interrupt) just after you've read the old data.
I tend to agree but the original code flow has the interrupt getting cleared after
the samples are read.
That's half true. I.e. it's true for pci1710_handle_fifo(), but not for
pci1710_handle_every_sample(). (Actually, both those functions read
from the FIFO, so the function name is a bit unfortunate.)
Maybe moving the point where the interrupt is cleared
should be a separate patch?
Depends. It may be easier for Greg to just take the first 11, in which
case things can be cleaned up a bit when the remainder are reposted as a
second series.
--
-=( Ian Abbott @ MEV Ltd. E-mail: <abbotti@xxxxxxxxx> )=-
-=( Tel: +44 (0)161 477 1898 FAX: +44 (0)161 718 3587 )=-
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
