RE: [PATCH v3 -next 4/4] net: emaclite: Fix restricted cast warning of sparse

[Radhey Shyam Pandey <radheys@xilinx.com>]

> -----Original Message-----
> From: Andrew Lunn <andrew@lunn.ch>
> Sent: Friday, January 31, 2020 7:19 PM
> To: Radhey Shyam Pandey <radheys@xilinx.com>
> Cc: davem@davemloft.net; netdev@vger.kernel.org; linux-
> kernel@vger.kernel.org; Anirudha Sarangi <anirudh@xilinx.com>; Michal Simek
> <michals@xilinx.com>; gregkh@linuxfoundation.org;
> mchehab+samsung@kernel.org; John Linn <linnj@xilinx.com>; linux-arm-
> kernel@lists.infradead.org
> Subject: Re: [PATCH v3 -next 4/4] net: emaclite: Fix restricted cast warning of
> sparse
> 
> On Fri, Jan 31, 2020 at 05:17:50PM +0530, Radhey Shyam Pandey wrote:
> > Explicitly cast xemaclite_readl return value when it's passed to ntohl.
> > Fixes below reported sparse warnings:
> >
> > xilinx_emaclite.c:411:24: sparse: sparse: cast to restricted __be32
> > xilinx_emaclite.c:420:36: sparse: sparse: cast to restricted __be32
> >
> > Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
> > Reported-by: kbuild test robot <lkp@intel.com>
> > ---
> >  drivers/net/ethernet/xilinx/xilinx_emaclite.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/xilinx/xilinx_emaclite.c
> b/drivers/net/ethernet/xilinx/xilinx_emaclite.c
> > index 96e9d21..3273d4f 100644
> > --- a/drivers/net/ethernet/xilinx/xilinx_emaclite.c
> > +++ b/drivers/net/ethernet/xilinx/xilinx_emaclite.c
> > @@ -408,7 +408,8 @@ static u16 xemaclite_recv_data(struct net_local
> *drvdata, u8 *data, int maxlen)
> >
> >  	/* Get the protocol type of the ethernet frame that arrived
> >  	 */
> > -	proto_type = ((ntohl(xemaclite_readl(addr + XEL_HEADER_OFFSET +
> > +	proto_type = ((ntohl((__force __be32)xemaclite_readl(addr +
> > +			XEL_HEADER_OFFSET +
> >  			XEL_RXBUFF_OFFSET)) >> XEL_HEADER_SHIFT) &
> >  			XEL_RPLR_LENGTH_MASK);
> >
> > @@ -417,7 +418,7 @@ static u16 xemaclite_recv_data(struct net_local
> *drvdata, u8 *data, int maxlen)
> >  	 */
> >  	if (proto_type > ETH_DATA_LEN) {
> >  		if (proto_type == ETH_P_IP) {
> > -			length = ((ntohl(xemaclite_readl(addr +
> > +			length = ((ntohl((__force __be32)xemaclite_readl(addr
> +
> >  					XEL_HEADER_IP_LENGTH_OFFSET +
> >  					XEL_RXBUFF_OFFSET)) >>
> >  					XEL_HEADER_SHIFT) &
> 
> If i understand this code correctly, you need the ntohl because you
> are poking around inside the packet. All the other uses of
> xemaclite_readl() are for descriptors etc.
> 
> It would be cleaner if you defined a xemaclite_readlbe32. If you use
> ioread32be() it will do the endinness swap for you, so you don't need
> the ntohl() and the horrible cast.
Thanks for the review. Yes, defining xemaclite_readlbe32 would eliminate the
cast need.  I will address it in the next version. 

> 
>     Andrew