Hi,
Alessio Margan @ IIT wrote:
> Hi,
>
> I've got a doubt about s526 driver on reading analog input.
>
> I read analog data in this way
>
> sampl_t data;
> a4l_insn_t insn_tab = {
> .type = A4L_INSN_READ,
> .idx_subd = idx_subd,
> .chan_desc = CHAN(idx_chan),
> .data_size = sizeof(data),
> .data = &data};
>
> /* Sends the read instruction to the Analogy layer */
> ret = a4l_snd_insn(&dsc, &insn_tab);
>
> on driver side data_size is the number of samples not the nr of byte to
> read,
> the driver expect data_size to be the number of samples and data to be a
> vector of samples
You are right. There is a bug in the driver. The field data_size is not
the acquisition count but the size of all the acquisitions in bytes.
So we should see:
for (n = 0; n < insn->data_size / sizeof(uint16_t); n++) {
I will fix it tonight.
>
> uint16_t *data = (uint16_t *)insn->data;
>
> /* convert n samples */
> for (n = 0; n < insn->data_size; n++) {
> /* trigger conversion */
> outw(value, ADDR_REG(REG_ADC));
> a4l_info(dev, "s526_ai_rinsn: Wrote 0x%04x to ADC\n", value);
> a4l_info(dev, "s526_ai_rinsn: ADC reg=0x%04x\n", inw(ADDR_REG(REG_ADC)));
>
> #define TIMEOUT 100
>
> /* wait for conversion to end */
> for (i = 0; i < TIMEOUT; i++) {
> status = inw(ADDR_REG(REG_ISR));
> if (status & ISR_ADC_DONE) {
> outw(ISR_ADC_DONE, ADDR_REG(REG_ISR));
> break;
> }
> }
> if (i == TIMEOUT) {
> a4l_warn(dev, "s526_ai_rinsn: ADC(0x%04x) timeout\n",
> inw(ADDR_REG(REG_ISR)));
> return -ETIMEDOUT;
> }
>
> /* read data */
> d = inw(ADDR_REG(REG_ADD));
> a4l_info(dev, "s526_ai_rinsn: AI[%d]=0x%04x\n", n, (unsigned short)(d &
> 0xFFFF));
>
> /* munge data */
> data[n] = d ^ 0x8000;
> }
>
> TIA
>
> Alessio
>
Alexis.
_______________________________________________
Xenomai-core mailing list
Xenomai-core@gna.org
https://mail.gna.org/listinfo/xenomai-core
