Hey Joel;
Thanks. Yes /dev/mem is not the way to operate. We have been trying to add GPIO support to parport_pc driver so we can enable IRQ40 and set it into edge trigger mode. So far we have been attempting to use GPIO Port F, bit 2.
However, regardless of what we do, it acts as a level-sensitive interrupt. Being that we have made like zero headway in the past three weeks, we got some outside help this past weekend, and found that the 9302 chip does not even bring INT2 out to the external connections. Apparently they can see IRQ 40 in the VIC status registers, but their belief is that no GPIO is involved at all in IRQ7/40 or int IRQ6/33, only VIC. I have inserted all manner of signals onto the PC104 pins for these two interrupts and the result is the same. Nothing we set in GPIO affects the behavior at all, which would tend to support our expert's claims.
So we have set our sights on IRQ5, as it appears to be the only IRQ that supports edge trigger mode via the GPIO settings. However, the TS-7260 says that IRQ5 is really IRQ22 which is not available, according to request_irq().
And the TS-7260 schematic confuses things further, as it shows INT0 being taken by something called FPGA_IRQ, which if I am not mistaken is an NA for the TS-7260. ISA_IRQ5 is shown to be connected to FGPIO_2, which would imply in my very tiny understanding that this is Bit 2 of GPIO port F. However, experiment shows that it appears to be bit 3 of GPIO port F.
See why I am confused? At this point we'll take any of the three interrupts. Have any of you gotten TS-7260 external interrupts working in edge-triggered mode?
Thanks!
On Apr 8, 2011, at 5:51 PM, Joel Morgan wrote: On Fri, 2011-04-08 at 20:50 +0000, Clark wrote:
>
> Trying to get a 1PPS GPS interrupt going on IRQ7/40 for TS7260!
> We read all the notes in EP9301 User's Guide, in particular
> the bottom of page 525 seems important. We have been inspecting
> the driver's results using a user-space mapping of /dev/mem,
> but now we have simplified, we use the /dev/mem technique sans
> the driver and the interrupt keeps getting disabled!?
Userspace is not allowed to directly mess with interrupts. Where in the
kernel they are being disabled I cant say off the top of my head. What
driver are you using?
--
--
Joel R. Morgan
Morgan Millwright Services, Inc.
Linux User #504110 http://counter.li.org/
__._,_.___
![]()
__,_._,___
|