I'm trying to port the TI hardware abstraction layer (HAL) code in the
CC2480 application to a different processor.
The confusion I'm having is with the timers. This code initializes the
TimerA by measuring the counts for 1 second, in the halTimerInit routine:
static void halTimerInit(void)
{
uint8 loop = 128; // Count 8mHz by 65536 timer overflow.
if (halEventFlags & HAL_EVT_NO_8MHz)
{
loop = 16; // Count 1mHz by 65536 timer overflow.
}
BCSCTL1 |= DIVA_3; // ACLK/8
BCSCTL3 |= LFXT1S_2; // ACLK = VLO
// Setup TimerB to calibrate TimerA.
TBCCR0 = 65535;
// Setup TimerA to be calibrated.
TACCR0 = 65535;
TACTL = TASSEL_1 + MC_1; // ACLK, up count mode.
// Calibrate TimerA.
TBCTL = TBSSEL_2 + MC_1; // SMCLK, up count mode.
do {
while ((TBCTL & TBIFG) == 0);
TBCTL &= ~TBIFG;
} while (--loop);
TBCTL = TBCLR; // Clear TimerB configuration.
TACTL &= ~MC_1; // Stop TimerA.
TACCR0_INIT = TAR;
TACTL |= MC_1; // Resume TimerA.
// Configure TimerA for 1-msec ISR for driving the HAL S/W timers, timing
key press+hold, etc.
TACCR0 = TACCR0_INIT; // 1 second by 12kHz / 8 -> 1.5kHz
// TACTL = TASSEL_1 + MC_1; // ACLK, up count mode.
TACCTL0 = CCIE; // TACCR0 interrupt enabled
}
My problem is that this says it's configuring the timer for a 1msec
interrupt, but as far as I can tell it's really doing a 1 second
interrupt.
This is further confused by the halDelay routine, which says it sleeps for a
number of msec, but has a very cryptic comment with a calculation
that almost makes sense, but then does things that make no sense at all to
me.
void halDelay(uint8 msecs, uint8 sleep)
{
// Expected scaling for msecs is 3/2, but depends on result of TimerA
calibration.
// TBD - not enough code space: uint16 stop = msecs * (TACCR0_INIT / 10) /
(1000 / 10);
uint16 stop = msecs * 3 / 2;
HAL_DISABLE_INTERRUPTS();
TACTL &= ~MC_1; // Stop the timer.
if (TAR != TACCR0_INIT)
{
stop += TAR;
if (stop == TACCR0_INIT)
{
stop++;
}
}
TACCR0 = stop;
TACTL |= MC_1; // Re-start the timer.
HAL_ENABLE_INTERRUPTS();
if (sleep)
{
do {
HAL_LOW_POWER_MODE();
} while (TACCR0 != TACCR0_INIT);
}
}
Is there a better source for hal code that has already been ported to other
parts, or am I just missing something that should be obvious here? I looked
for a 1000x clock speedup, but didn't find it anywhere.
Thanks,
Craig
--
R. Craig Campbell
Lead Engineer
Daedalus
note our new address: 4923 Centre Avenue, Pittsburgh, PA 15213
voice 412.687.7000 ext.16
fax 412.687.7500
website http://www.daed.com
directions http://www.daed.com/contact_maps.tt2
This message may contain private or proprietary information. Do not
forward this message without my permission. If it appears that I sent
you this message by mistake, please let me know and delete this
message.
