Re: Patch "USB: serial: mos7840: Supported MCS7810 device"

[Alan Cox]

> +static void mos7810_control_callback(struct urb *urb)
> +{
> +	if (!urb)
> +		return;

How can this ever occur. If its not theoretically possible then this
check is just going to hide bugs in future.

> +static void mos7810_led_off(unsigned long arg)
> +{
> +	struct moschip_port *mcs = (struct moschip_port *) arg;
> +
> +	if (!mcs)
> +		return;

Ditto

> +	/* Turn on MCS7810 LED */
> +	if (serial->num_ports == 1 && mos7840_port->mos7810_led_flag == 0) {
> +		mos7840_port->mos7810_led_flag = 1;
> +		mos7810_set_reg(mos7840_port, 0x0301,
> MODEM_CONTROL_REGISTER);
> +		mod_timer(&mos7840_port->mos7810_led_timer1, jiffies + 500);
> +	}

The assumption that serial->num_ports = 1 means "has LED" is going to
cause problems later if another deivce appears with 1 port and no LED. I
think it would be much clearer to have something like

	if (mos7840_port->has_led)

for these tests of the driver

> +static int mos7810_check(struct usb_serial *serial)
> +{
> +	int i, pass_count = 0;
> +	__u16 Data = 0, MCR_Data = 0;
> +	__u16 test_pattern = 0x55AA, write_pattern;

generally linux style is all lower case for egister names. Not a big deal

> +		/* If this a MCS7810 device, both test patterns must match
> */
> +		if (((test_pattern >> i) & 0x0001) != ((~(Data >> 1)) &
> 0x0001))

		if (((test_pattern >> i) ^ (~Data >> 1)) & 0x0001) ?

again trivia...


> +			mos7840_port->mos7810_led_timer1.expires =
> +							jiffies + 500;

This 500 value depends upon the HZ value of the kernel which won't always
be 1000 on all platforms. You probably want

			jiffies + msecs_to_jiffies(500);

(or indeed HZ/2)

> +				/* Turn off MCS7810 LED */
> +				usb_control_msg(serial->dev,
> +					usb_sndctrlpipe(serial->dev, 0),
> +					MCS_WRREQ, MCS_WR_RTYPE, 0x0300,
> +					MODEM_CONTROL_REGISTER,	NULL, 0,
> +					MOS_WDR_TIMEOUT);
> +

Might be clearer if these occurences were a separate small function - gcc
will inline them if it makes sense, and it will be clearer.


Alan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/