Recent Posts

Pages: [1] 2 3 ... 10
1
ATMELTM AT91SAM7X and AVR32 / Re: __irq_en
« Last post by mark on January 24, 2020, 08:17:38 PM »
Hi

Since the involved code is assembler code I wouldn't expect a compiler version to be able to have any effect on that code.

The endless loop is a default exception vector, which probably means that the access caused an interrupt (maybe a fault interrupt?) to be triggered.

The instruction mrs is described here:
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0489c/CIHGJHHH.html
but I don't see any restrictions that could cause its use to fail.

Check that the code is located on a 32 bit boundary in memory - it is preceded by

    .code 32

but maybe the newer version ignores this?

Regards

Mark
2
ATMELTM AT91SAM7X and AVR32 / __irq_en
« Last post by timadria on January 20, 2020, 04:03:46 PM »
I've just upgraded to Crossworks v4 and after uploading the utasker to the SAM7S processor this will halt on the __irq_en macro.
When executing
Code: [Select]
mrs r12, CPSR  the CPU halts and loops on
Code: [Select]
  b .    /* endless loop */
Any idee how this can be solved?
3
µTasker general / Re: CAN TX Timeout
« Last post by AlexS on January 17, 2020, 06:13:57 AM »
Thanks, Mark! Looking forward to it as it's a bit of a blocker at the moment. :) I did a bit of my own digging and it seems like there's no code path to handle a flush request on a CAN interface.  As you know I'm on the old uTasker version (Jan 2019), so not sure if you've added anything in between.

Regards,
Alex
4
µTasker general / Re: CAN TX Timeout
« Last post by mark on January 16, 2020, 03:46:18 PM »
Hi Alex

I will look into a method to identify and clear blocked tx buffers - it has been quite some time since working at the CAN driver level so I'll need a little time to refresh on the details.

Regards

Mark
5
µTasker general / Re: CAN TX Timeout
« Last post by AlexS on January 15, 2020, 02:27:05 PM »
Very happy to monitor with the application, but not sure how I can abort and flush the buffers. It is indeed a passive error, where the other device on the bus (and the termination resistor as a consequence) are not active initially after my device starts.

Thanks!
6
µTasker general / Re: CAN TX Timeout
« Last post by mark on January 15, 2020, 02:07:28 PM »
Alex

What is the state of the bus that doesn't allow the frame to be sent?
Do you see that there is an error interrupt that occurs that could be used to clear the buffer?

If it is a passive error (never can send and continuously waits until it can) it may be necessary to monitor with an application timer and subsequently abort/flush buffers.

Regard

Mark
7
µTasker general / CAN TX Timeout
« Last post by AlexS on January 14, 2020, 01:08:08 PM »
Hi,

Using uTasker (with excellent results) on an automotive project. On the CAN-side of things I'm seeing a peculiar behaviour in a specific corner case. The corner case is one where my device is the only device switched on the CAN bus with another, that should receive my messages, being turned on a few seconds later. What happens is that the TX buffers quickly get filled up, but, oddly, they don't ever timeout, contrary to what's described in the documentation. The documentation suggests that the buffers will be freed after a certain number of retries.

After looking through the code, I noticed that the CAN_TX_BUF_ACTIVE flag is only cleared in case there's no ACK coming, not if the message can't be sent. Page 10 of the uTasker CAN documentation suggests that the expected behaviour is for the buffer to be returned to an inactive state by the driver, informing the application that the message couldn't be delivered.

        if ((ulError & CAN_ACK_ERR) != 0) {                              // a transmission received no ack
            (...)
            while (i-- != 0) {
                if ((ptrCanQue->ucMode & CAN_TX_BUF_ACTIVE) != 0) {      // this buffer is presently transmitting a message
                    if (++(ptrCanQue->ucErrors) >= MAX_TX_CAN_TRIES) {   // we allow a few attempts before quitting (it also filters counting normal active buffers)
                        ptrMessageBuffer->ulCode_Len_TimeStamp = ((ptrMessageBuffer->ulCode_Len_TimeStamp & ~CAN_CODE_FIELD) | MB_TX_INACTIVE); // stop transmission
                        can_error_int_message[MSG_DESTINATION_TASK] = ptrCanQue->TaskToWake;
                        if ((ptrCanQue->ucMode & CAN_TX_BUF_REMOTE) != 0) {
                            can_error_int_message[MSG_INTERRUPT_EVENT] = CAN_TX_REMOTE_ERROR;
                            ptrCanQue->ucMode = (CAN_TX_BUF | CAN_RX_BUF_FULL | CAN_TX_BUF_REMOTE); // mark that it is an errored transmission buffer
                        }
                        else {
                            can_error_int_message[MSG_INTERRUPT_EVENT] = CAN_TX_ERROR;
                            ptrCanQue->ucMode = (CAN_TX_BUF | CAN_RX_BUF_FULL); // mark that it is an errored transmission buffer
                        }


8
µTasker general / Re: UART Tx/Rx in different pins for KL03
« Last post by mark on January 11, 2020, 03:16:24 PM »
Raffaele

If
#define SERIAL_INTERFACE                                                 // enable serial interface driver
is defined and the same code is used I would expect identical behavior.

Regards

Mark
9
µTasker general / Re: UART Tx/Rx in different pins for KL03
« Last post by Raffaele on January 11, 2020, 12:23:10 AM »
Hey Mark,
 
I see how the APPLICATION TASK uses the QUEUE_HANDLE fnSetNewSerialMode(unsigned char ucDriverMode) to set up the UART and if I change the task to wake in this function to TASK_MY_UART I can use my task upon receiving UART data.

However, all this works only if I keep HELLO_WORLD active.
I tried to have the same configs of hello_world and replicate a similar TTYTABLE with tInterfaceParameters.Task_to_wake = TASK_MY_UART  in my task, commented the HELLO_WORLD, but it doesn't work.

Any ideas why?
10
µTasker general / Re: UART Tx/Rx in different pins for KL03
« Last post by Raffaele on January 10, 2020, 06:34:56 PM »
Hi Mark,

I was able to make it work, my code works now as far as I keep the HELLO_WORLD activated
Pages: [1] 2 3 ... 10