2009-06-26 16:49:42     MAX3100 driver problem/fix

Document created by Aaronwu Employee on Aug 15, 2013
Version 1Show Document
  • View in full screen mode

2009-06-26 16:49:42     MAX3100 driver problem/fix

Todor Mollov (UNITED STATES)

Message: 76411   

 

Hi,

 

I've recently been trying to get the max3100 spi to uart bridge working on a custom bf534 board, and I found a problem in the driver.  The driver was doing a cpu to big endian conversion before passing the transfer data over to the spi bus driver, which was also doing a cpu to big endian conversion with the end result of little endian data being put on the lines.

 

 

 

The following is the max3100_sr() function I'm currently using that correctly controls the device.

 

drivers/serial/max3100.c@line 193

 

static int max3100_sr(struct max3100_port *s, u16 tx, u16 *rx)

{

    struct spi_message message;

    u16 etx, erx;

    int status;

    struct spi_transfer tran = {

        .tx_buf = &etx,

        .rx_buf = &erx,

        .len = 2,

        .bits_per_word = 16,

    };

 

    //etx = cpu_to_be16(tx);

    etx = tx;

    spi_message_init(&message);

    spi_message_add_tail(&tran, &message);

    status = spi_sync(s->spi, &message);

    if (status) {

        dev_warn(&s->spi->dev, "error while calling spi_sync\n");

        return -EIO;

    }

    //*rx = be16_to_cpu(erx);

    *rx = erx;

    s->tx_empty = (*rx & MAX3100_T) > 0;

    dev_dbg(&s->spi->dev, "%04x - %04x\n", tx, *rx);

    return 0;

}

 

 

 

It needs to be cleaned up for a patch, but I have deadlines to meet right now =/.  Hope this helps someone.

 

(I'm working off of the r8353 uclinux-distro snapshot so I'm not sure of the exact kernel revision, but it's a pretty fresh snapshot.)

QuoteReplyEditDelete

 

 

2009-06-26 17:02:22     Re: MAX3100 driver problem/fix

Mike Frysinger (UNITED STATES)

Message: 76412   

 

Christian Pellegrin <chripell@evolware.org> is the guy who maintains that driver ... maybe try sending him an e-mail

Attachments

    Outcomes