Re: [PATCH v4 17/18] staging: r8188eu: Shorten calls chain of rtw_read8/16/32()

["Fabio M. De Francesco" <fmdefrancesco@gmail.com>]

On Tuesday, September 14, 2021 11:32:58 AM CEST Dan Carpenter wrote:
> On Mon, Sep 13, 2021 at 08:10:01PM +0200, Fabio M. De Francesco wrote:
> > diff --git a/drivers/staging/r8188eu/hal/usb_ops_linux.c b/drivers/
staging/r8188eu/hal/usb_ops_linux.c
> > index 04402bab805e..75475b0083db 100644
> > --- a/drivers/staging/r8188eu/hal/usb_ops_linux.c
> > +++ b/drivers/staging/r8188eu/hal/usb_ops_linux.c
> > @@ -89,6 +89,56 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, 
u16 value, void *data, u1
> >  	return status;
> >  }
> >  
> > +static int usb_read(struct intf_hdl *intfhdl, u16 addr, void *data, u8 
size)
> > +{
> > +	int status;
> > +	u8 *io_buf; /* pointer to I/O buffer */
> > +	struct adapter *adapt = intfhdl->padapter;
> > +	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
> > +	struct usb_device *udev = dvobjpriv->pusbdev;
> > +
> > +	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx) 
{
> > +		status = -EPERM;
> > +		goto exit;
> > +	}
> 
> Just return directly.
> 
> 	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
> 		return -EPERM;

Dear Dan,

I agree with you, I'll return it directly in v5.

> > +
> > +	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
> > +
> > +	io_buf = dvobjpriv->usb_vendor_req_buf;
> > +
> > +	status = usb_control_msg_recv(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
> > +				      REALTEK_USB_VENQT_READ, 
addr,
> > +				      REALTEK_USB_VENQT_CMD_IDX, 
io_buf,
> > +				      size, 
RTW_USB_CONTROL_MSG_TIMEOUT,
> > +				      GFP_KERNEL);
> > +	if (!status) {
> > +		/*  Success this control transfer. */
> 
> It's always better to do error handling instead of success handling.

Yes, you're right again. I blindly followed the style that was in the parts 
of the code of usbctrl_vendorreq() that I'm reusing here.
 
> 	if (status) {
> 
> Remove the comment because now it's in the standard format.

OK, those comments are not necessary: I'll remove them.

> > +		rtw_reset_continual_urb_error(dvobjpriv);
> > +		memcpy(data, io_buf, size);
> > +		goto mutex_unlock;
> > +	}
> > +	/*  error cases */
> > +	if (status == (-ESHUTDOWN || -ENODEV || -ENOENT)) {
> 
> 	if (status == -ESHUTDOWN ||
> 	    status == -ENODEV ||
> 	    status == -ENOENT) {

This is a stupid mistake and Pavel soon noticed it. Yesterday I sent a 
message to ask reviewers for disregarding v4 and wait for v5 with the fix of 
this test. :(

However, I noticed that usb_control_msg_recv() might return in "status" some 
recoverable errors (like -ENOMEM and others); so I guess that the code must 
retry in a while loop (exactly as it did with usb_control_msg() in 
usbctrl_vendorreq()). 

Actually I'm not 100% sure that the "while" loop is needed. What I know is 
that I've been using this version (with _no_ loop) for about a week, not less 
than 12 hours a day and I didn't notice any problem...

Can you please tell if a loop for retries is *really* necessary?

> 
> > +		/*
> > +		 * device or controller has been disabled due to
> > +		 * some problem that could not be worked around,
> > +		 * device or bus doesn’t exist, endpoint does not
> > +		 * exist or is not enabled.
> > +		 */
> > +		adapt->bSurpriseRemoved = true;
> > +			goto mutex_unlock;
> 
> Indented wrong.

Yes, it must be fixed.

Regards,

Fabio

> 
> > +	}
> > +	GET_HAL_DATA(adapt)->srestpriv.wifi_error_status = 
USB_VEN_REQ_CMD_FAIL;
> > +	if (rtw_inc_and_chk_continual_urb_error(dvobjpriv)) {
> > +		adapt->bSurpriseRemoved = true;
> > +		goto mutex_unlock;
> > +	}
> > +mutex_unlock:
> > +	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
> > +exit:
> > +	return status;
> > +}
> 
> regards,
> dan carpenter
>