Re: [PATCH 2/6] rtlwifi: Add usb driver

[Stanislaw Gruszka <sgruszka@redhat.com>]

On Sun, Feb 06, 2011 at 12:44:20PM -0600, Larry Finger wrote:
> +enum {
> +	VENDOR_WRITE = 0x00,
> +	VENDOR_READ = 0x01,
> +};
[snip]
> +static int _usbctrl_vendorreq_async(struct usb_device *udev, u8 request,
> +					  u16 value, u16 index, void *pdata,
> +					  u16 len, u8 requesttype)
> +{
It whould be better to leave old name and remove requesttype argument.
We never call  _usbctrl_vendorreq_async with VENDOR_READ, and 
 _usbctrl_vendorreq_sync with VENDOR_WRITE.

If this is not preparation for future changes, that need to
cleaned up, and VENDOR_WRITE and VENDOR_READ definition removed at all.

> +	wvalue = (u16)(addr&0x0000ffff);
wvalue = (u16) addr;

> +static void _rtl_usb_io_handler_release(struct ieee80211_hw *hw)
> +{
> +	struct rtl_priv *rtlpriv = rtl_priv(hw);
> +
> +	if (&rtlpriv->io)
How it can be NULL? :-) Perhaps this should be rtlpriv check?

> +	/*  TODO: think of race condition... */
[snip]
> +	/*  TODO:	statistics */
[snip]
> +
[snip]
> +		/* TODO: Shall I ignore error and keep issue other urbs? */
[snip]
> +		/* TODO: reinsert skb to list or free the skb? */
[snip]
> +	/* TODO: if usb suspend, we may queue the skb and wake device up.
> +	 * Then xmit.
> +	 * TODO: may change this statement and check all usb state */

Lots of TODO, is this code working? :-)

Is this driver for devices that are currently on the market,
or for future product? I have 3 Realtek dongles, but all works
only with staging driver ...

Stanislaw