Hi everyone,
I am working with the BeagleBone Black trying to make the ADC and the
PRU-ICSS work with each other for sampling an analogue signal. The issue
that brings me here is a small-big problem I have with the external
interrupts of the PRU-ICSS, basically it does not receive these signals.
Although the internal interrupts work fine (PRU to PRU) and sending pulses
to external hardware such as the host.
In some few words the test program I am now running in the system
configures the ADC to start sampling data continuously and generate a
"End_of_sequence" interrupt to the PRU0.
I have checked that this interrupt is being triggered correctly by polling
the IRQSTATUS_RAW register of the ADC from the PRU directly, and checking
the sampled data in the FIFO. (In fact this is also another way of
bypassing the problem, but is not the most efficient nor elegant way to
work with this device I think...)
I also have this problem with the EDMA3 module, there is no interrupt from
this device to the PRU-ICSS. But this is another issue, and I think is the
same problem as the previous one.
Here is the code (the important parts at least) I am running in the
PRU-ICSS:
Definitions:
#define gmoADC 0x44E0D000 //Global memory Offset of the ADC
First the ADC configuration:
//CONFIGURATION ADC
//
==========================================================================
MOV MemPointer, gmoADC
//Control register
MOV r1, 0x0005
LBBO r1, MemPointer, 0x40, 2
//TS_CHARGE_DELAY
//Set default value (1h)018C3040
//STEPCONFIG1.
//AIN1 (Channel 2) (+input)
//Single ended
//VREFP (+ref) = 1.8 V
//VREFN (-ref & -input) = GND
MOV r1,0x00000001 // 0x018C3002 //After test last
bit will be changed
SBBO r1, MemPointer, 0x64, 4
//STEP enable register
LDI r1, 0x02 //STEP1
SBBO r1, MemPointer, 0x54, 4
//STEPDELAY1 Set as default
//Interrupt parameters
//IRQENABLE: End of sequence
MOV r1, 0x00000002
SBBO r1, MemPointer, 0x2C, 4
Configuration of the INTC (PRU):
//Global interrupt enable PRU
MOV r1.w0, 0x0001
SBCO r1, INTCbase, 0x10, 2
//Enable host 0 int
MOV r1, (0x00000000|0) //(0x00000000|Host_num)
SBCO r1, INTCbase, 0x34, 4
//Map channel 0 to host 0
LDI r2.w0, 0x0800
ADD r2.w0, r2.w0, 0 //Add host_num to host map registers offset
0x800
MOV r1.b0, 0 //Channel number
SBCO r1.b0, INTCbase, r2.w0, 1
//Map system evento to channel 0
LDI r2.w0, 0x0400
ADD r2.w0, r2.w0, SysEvent //Add system evento to channel map
register offset 0x400
LDI r1.b0, 0 //Channel number
SBCO r1.b0, INTCbase, r2.w0, 1
//Clear system event interrupt
MOV r1, (0x00000000|SysEvent)
SBCO r1, INTCbase, 0x24, 4
//Enable system event interrupt
SBCO r1, INTCbase, 0x28, 4
Now, if I keep polling the register IRQSTATUS_RAW of the ADC (and clearing
the interrupt flag), the program works fine. But if I do the same but
polling instead bit 30 of R31.... well that is why I am writing this....
I verify this by getting the data from the PRU's RAM using the host process.
I am running an Angstrom v2012.12 in a revision A5C BBB. Kernel version:
3.8.13.
Someone has the same problem, knows some possible solution (appart from the
polling)?...
Thank you very much!
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
