The ecspi controller is supposed to support burst length of 1 to 4096 bits,
meaning it is possible to configure it to use a value that isn't a multiple
of 8 (to my understanding). In that case, since tx_burst is always
decremented by 8, it will underflow. Sorry I didn't include an explanation.
Should I resubmit the patch?
>
> > DPRINTF("writing 0x%02x\n", (uint32_t)byte);
> > @@ -206,7 +206,7 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)
> > if (fifo32_is_full(&s->rx_fifo)) {
> > s->regs[ECSPI_STATREG] |= ECSPI_STATREG_RO;
> > } else {
> > - fifo32_push(&s->rx_fifo, (uint8_t)rx);
> > + fifo32_push(&s->rx_fifo, rx);
> > }
> >
> > if (s->burst_length <= 0) {
>
> This seems like a separate change to the first one;
> in general unrelated changes should be each in their
> own patch, rather than combined into a single patch.
Should I resubmit this as a patch?
>
> The fifo32_push() part of this change looks correct to me.
>
> thanks
> -- PMM
>
