Re: [PATCH V8 02/10] USB: f81232: implement RX bulk-in EP

From: Johan Hovold <johan@kernel.org>
To: Peter Hung <hpeter@gmail.com>
Cc: Johan Hovold <johan@kernel.org>,
	gregkh@linuxfoundation.org, linux-usb@vger.kernel.org,
	linux-kernel@vger.kernel.org, tom_tsai@fintek.com.tw,
	peter_hong@fintek.com.tw,
	Peter Hung <hpeter+linux_kernel@gmail.com>
Subject: Re: [PATCH V8 02/10] USB: f81232: implement RX bulk-in EP
Date: Mon, 16 Mar 2015 09:52:08 +0100	[thread overview]
Message-ID: <20150316085208.GA14918@localhost> (raw)
In-Reply-To: <550645B2.1090202@gmail.com>

On Mon, Mar 16, 2015 at 10:53:38AM +0800, Peter Hung wrote:
> Hello,
> 
> Johan Hovold 於 2015/3/14 下午 07:48 寫道:
> > On Thu, Feb 26, 2015 at 06:02:08PM +0800, Peter Hung wrote:
> >> -	if (!urb->actual_length)
> >> +	if ((urb->actual_length < 2) || (urb->actual_length % 2))
> >>   		return;
> >
> > Not parsing short data (e.g. not divisible by 2) is OK I guess. You
> > could also just discard the last odd byte, but that's up to you.
> >
> > Either way, I think you should add a dev_warn here rather than just
> > silently drop it.
> 
> With F81232, when it first submit with interrupt URB, it will response 
> once with 1 bytes data. The data is hw current LSR, but it useless on 
> open. It's should necessary with receiving data. When the device is 
> working good, it's should acked with even size data.

Ok, thanks for clarifying that.

> To avoid confusing to user, I'll keep it without warning message.

Yes, skip the warning, but could you a short comment about this (e.g.
the 1-byte packet on open) before you do the size check?

Thanks,
Johan