Re: [PATCH 1/3] staging: r8188eu: remove _io_ops structure

[Pavel Skripkin]

On 9/6/21 10:49 AM, Dan Carpenter wrote:
> On Sat, Sep 04, 2021 at 05:04:45PM +0200, Fabio M. De Francesco wrote:
> > -static int usb_writeN(struct intf_hdl *pintfhdl, u32 addr, u32 length, u8 *pdata)
> > +int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *pdata)
> >  {
> > +	struct io_priv *pio_priv = &adapter->iopriv;
> > +	struct intf_hdl *pintfhdl = &pio_priv->intf;
> >  	u16 wvalue = (u16)(addr & 0x0000ffff);
> >  	u8 buf[VENDOR_CMD_MAX_DATA_LEN] = {0};
> > +	int ret;
> >  
> >  	if (length > VENDOR_CMD_MAX_DATA_LEN)
> >  		return -EINVAL;
>
> The caller treats this negative return as success.

Oh, good catch, thank you so much for pointing it out... This driver 
uses unusual approach for error handling with private _SUCCESS and _FAIL 
macros and I forgot to call RTW_ERROR_CODE :(

> >  
> >  	memcpy(buf, pdata, length);
> > +	ret = usbctrl_vendorreq(pintfhdl, wvalue, buf, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
> >  
> > -	return usbctrl_vendorreq(pintfhdl, wvalue, buf, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
> > +	return RTW_STATUS_CODE(ret);
> >  }
>
> [ snip ]
>
>
> > -void usb_read_port_cancel(struct intf_hdl *pintfhdl)
> > +void rtw_read_port_cancel(struct adapter *adapter)
> >  {
> >  	int i;
> > -	struct recv_buf *precvbuf;
> > -	struct adapter	*padapter = pintfhdl->padapter;
> > -	precvbuf = (struct recv_buf *)padapter->recvpriv.precv_buf;
> > +	struct recv_buf *precvbuf = (struct recv_buf *) adapter->recvpriv.precv_buf;
> >  
> >  	DBG_88E("%s\n", __func__);
> >  
> > -	padapter->bReadPortCancel = true;
> > +	adapter->bReadPortCancel = true;
>
> In these functions it would be better to rename "padapter" as "adapter"
> in a follow on function.  Keep it the same for now just to make review
> easier.

Ok, will bring old name back in v4. Thank you for review :)




With regards,
Pavel Skripkin