• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 82 subscribers
  • Views 211 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS320F28069: CLA: global array values get shifted

Joost de Bruin
Prodigy 200 points
Part Number: TMS320F28069

I have a strange bug in CLA code. I initialize values of a global array in CLA memory as follows:

Fullscreen
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
interrupt void InitVars(void) {
// TODO: Calculate these during initialization from Fourier coefficients specified in MotorControl.h
float32 BackEmfProfInit[PROF_SIZE] =
{0.9778235f, 0.9403109f, 0.8992523f, 0.8542375f, 0.8049480f, 0.7511688f, 0.6927969f, 0.6298467f,
0.5624517f, 0.4908623f, 0.4154409f, 0.3366528f, 0.2550543f, 0.1712778f, 0.0860155f, 0.0000000f,
-0.0860155f, -0.1712778f, -0.2550543f, -0.3366528f, -0.4154409f, -0.4908623f, -0.5624517f, -0.6298467f,
-0.6927969f, -0.7511688f, -0.8049480f, -0.8542375f, -0.8992523f, -0.9403109f, -0.9778235f, -1.0122777f,
-1.0442214f, -1.0742445f, -1.1029587f, -1.1309776f, -1.1588958f, -1.1872691f, -1.2165963f, -1.2473018f,
-1.2797216f, -1.3140913f, -1.3505375f, -1.3890725f, -1.4295926f, -1.4718800f, -1.5156081f, -1.5603503f,
-1.6055917f, -1.6507441f, -1.6951625f, -1.7381641f, -1.7790479f, -1.8171158f, -1.8516927f, -1.8821464f,
-1.9079068f, -1.9284820f, -1.9434737f, -1.9525886f, -1.9556470f, -1.9525886f, -1.9434737f, -1.9284820f,
-1.9079068f, -1.8821464f, -1.8516927f, -1.8171158f, -1.7790479f, -1.7381641f, -1.6951625f, -1.6507441f,
-1.6055917f, -1.5603503f, -1.5156081f, -1.4718800f, -1.4295926f, -1.3890725f, -1.3505375f, -1.3140913f,
-1.2797216f, -1.2473018f, -1.2165963f, -1.1872691f, -1.1588958f, -1.1309776f, -1.1029587f, -1.0742445f,
-1.0442214f, -1.0122777f, -0.9778235f, -0.9403109f, -0.8992523f, -0.8542375f, -0.8049480f, -0.7511688f,
-0.6927969f, -0.6298467f, -0.5624517f, -0.4908623f, -0.4154409f, -0.3366528f, -0.2550543f, -0.1712778f,
-0.0860155f, 0.0000000f, 0.0860155f, 0.1712778f, 0.2550543f, 0.3366528f, 0.4154409f, 0.4908623f,
0.5624517f, 0.6298467f, 0.6927969f, 0.7511688f, 0.8049480f, 0.8542375f, 0.8992523f, 0.9403109f,
0.9778235f, 1.0122777f, 1.0442214f, 1.0742445f, 1.1029587f, 1.1309776f, 1.1588958f, 1.1872691f,
1.2165963f, 1.2473018f, 1.2797216f, 1.3140913f, 1.3505375f, 1.3890725f, 1.4295926f, 1.4718800f,
1.5156081f, 1.5603503f, 1.6055917f, 1.6507441f, 1.6951625f, 1.7381641f, 1.7790479f, 1.8171158f,
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

The values of BackEmfProfInit get shifted by three positions, i.e. 2->5, 3->6 etc. The first two values are rubbish. I am at a loss as to how this could happen. I configure CLA memory as follows:

Fullscreen
1
2
3
4
5
6
7
8
9
// Set memory available to CLA.
Cla1Regs.MMEMCFG.all = 0;
Cla1Regs.MMEMCFG.bit.PROGE = 1;
Cla1Regs.MMEMCFG.bit.RAM0E = 1;
Cla1Regs.MMEMCFG.bit.RAM1E = 1;
Cla1Regs.MMEMCFG.bit.RAM2E = 1;
Cla1Regs.MMEMCFG.bit.RAM0CPUE = 0;
Cla1Regs.MMEMCFG.bit.RAM1CPUE = 0;
Cla1Regs.MMEMCFG.bit.RAM2CPUE = 0;
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

And in the CMD linker file:

Fullscreen
1
2
3
4
5
CLARAM : origin = 0x008000, length = 0x001000 /* on-chip RAM block L0/L1/L2 */
...
.scratchpad : > CLARAM, PAGE = 1
.bss_cla : > CLARAM, PAGE = 1
.const_cla : > CLARAM, PAGE = 1
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

I checked, the memory is not double-booked. Anybody any idea?

  • 0
    Prodigy 200 points

    Ok, kind of making a habit here of answering my own questions, but separating the memory section according to physical blocks did the trick. Thus like this:

    Fullscreen
    1
    2
    3
    4
    5
    6
    7
    CLARAM2 : origin = 0x008000, length = 0x000800 /* on-chip RAM block L0 */
    CLARAM0 : origin = 0x008800, length = 0x000400 /* on-chip RAM block L1 */
    CLARAM1 : origin = 0x008C00, length = 0x000400 /* on-chip RAM block L2 */
    ...
    .scratchpad : > CLARAM0, PAGE = 1
    .bss_cla : > CLARAM2, PAGE = 1
    .const_cla : > CLARAM1, PAGE = 1
    XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

  • 0 in reply to Joost de Bruin
    TI__Expert 7695 points

    Thanks for the update.