Wolfgang Grandegger schrieb:
> Yegor Yefremov wrote:
>> I borrowed this code from LinCAN project. It helped me a lot by
>> investigating the problem with memory access. I don't know if it is worth
>> committing it to the repository, but it was nice to use it.
>>
>> Yegor
>>
>> diff --git a/drivers/net/can/sja1000/sja1000.c
>> b/drivers/net/can/sja1000/sja1000.c
>> index 145b1a7..86ef501 100644
>> --- a/drivers/net/can/sja1000/sja1000.c
>> +++ b/drivers/net/can/sja1000/sja1000.c
>> @@ -229,6 +229,8 @@ static int sja1000_get_berr_counter(const struct
>> net_device *dev,
>> static void chipset_init(struct net_device *dev)
>> {
>> struct sja1000_priv *priv = netdev_priv(dev);
>> + int i;
>> + unsigned char n, r;
>>
>> /* set clock divider and output control register */
>> priv->write_reg(priv, REG_CDR, priv->cdr | CDR_PELICAN);
>> @@ -245,6 +247,18 @@ static void chipset_init(struct net_device *dev)
>> priv->write_reg(priv, REG_ACCM3, 0xFF);
>>
>> priv->write_reg(priv, REG_OCR, priv->ocr | OCR_MODE_NORMAL);
>> +
>> + /* Simple check for chip presence */
>> + for (i=0, n=0x5a; i<8; i++, n+=0xf) {
>> + priv->write_reg(priv, REG_ACCC0 + i, n);
>> + }
>> + for (i=0, n=0x5a; i<8; i++, n+=0xf) {
>> + r = n ^ priv->read_reg(priv, REG_ACCC0 + i);
>> + if (r) {
>> + printk("sja1000p_chip_config: chip connection broken,"
>> + " readback differ 0x%02x\n", r);
>> + }
>> + }
>> }
>
> This would work on any memory location, right?
If you use 8bit access, then the function would say, it couldn't find the chip.
That is the stuff I needed.
> The plx_pci driver does
> have a proper chip identification out of the reset. But I'm not sure if
> it's worth adding it.
Now, that the problem was cleared, it is really not worth adding it.
Yegor
_______________________________________________
Socketcan-core mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-core
