Re: [PATCH] stmmac: remove custom implementation of print_hex_dump()

[Joe Perches <joe@perches.com>]

On Mon, 2014-11-03 at 19:53 +0200, Andy Shevchenko wrote:
> On Mon, 2014-11-03 at 09:39 -0800, Joe Perches wrote:
> > On Mon, 2014-11-03 at 19:28 +0200, Andy Shevchenko wrote:
> > > There is a kernel helper to dump buffers in a hexdecimal format. This patch
> > > substitutes the open coded function by calling that helper.
> > []
> > > diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
> > []
> > > @@ -191,14 +191,8 @@ static void stmmac_clk_csr_set(struct stmmac_priv *priv)
> > >  
> > >  static void print_pkt(unsigned char *buf, int len)
> > >  {
> > > -	int j;
> > > -	pr_debug("len = %d byte, buf addr: 0x%p", len, buf);
> > > -	for (j = 0; j < len; j++) {
> > > -		if ((j % 16) == 0)
> > > -			pr_debug("\n %03x:", j);
> > > -		pr_debug(" %02x", buf[j]);
> > > -	}
> > > -	pr_debug("\n");
> > > +	pr_debug("len = %d byte, buf addr: 0x%p\n", len, buf);
> > > +	print_hex_dump(KERN_DEBUG, " ", DUMP_PREFIX_OFFSET, 16, 1, buf, len, 0);
> > 
> > Prefix should be ""
> 
> Oh, initially there is an indentation in one space.
> Maybe Giuseppe would comment on this.

I think that's not particularly important.

This changes the output anyway as the the printed offset
is now 8 chars wide not 3.

> > Another (better?) option would be to use:
> > 
> > 	print_hex_dump_bytes("", DUMP_PREFIX_OFFSET, buf, len);
> 
> > so that it can be dynamically controlled like
> > the pr_debug statements.
> 
> Ah, yes, but unfortunately it will print ASCII part always. Giuseppe,
> what do you think?

I'm not bothered by logging message output changes.
It's not a guaranteed stable output.

> P.S. [off topic] Joe, just would like to know if you have you seen my
> last version of the patch series against hexdump.c which adds
> seq_hex_dump() call [1]. If so, could you comment on it?
> [1] https://lkml.org/lkml/2014/9/4/309

seq_<foo> output however is guaranteed.

So any conversion to the seq_hex_dump function
would need to be exactly the same.

New code could use seq_hex_dump.

Other than that, it seems sensible enough.