Re: [PATCH] vsprintf: avoid misleading "(null)" for %px

[Adam Borowski]

On Mon, Feb 05, 2018 at 09:03:05PM +1100, Tobin C. Harding wrote:
> On Mon, Feb 05, 2018 at 10:44:38AM +0100, Petr Mladek wrote:
> > On Sun 2018-02-04 18:45:21, Adam Borowski wrote:
> > > Like %pK already does, print "00000000" instead.
> > >
> > > This confused people -- the convention is that "(null)" means you tried to
> > > dereference a null pointer as opposed to printing the address.
> > 
> > By other words, this avoids regressions when people convert
> > %x to %px. Do I get it right, please?

It's a regression in the sense that it confuses people.  %px never could
dereference a pointer so the information provided doesn't change, merely its
presentation.

> > > diff --git a/lib/vsprintf.c b/lib/vsprintf.c
> > > index 77ee6ced11b1..d7a708f82559 100644
> > > --- a/lib/vsprintf.c
> > > +++ b/lib/vsprintf.c
> > > @@ -1849,7 +1849,7 @@ char *pointer(const char *fmt, char *buf, char *end, void *ptr,
> > >  {
> > >  	const int default_width = 2 * sizeof(void *);
> > >  
> > > -	if (!ptr && *fmt != 'K') {
> > > +	if (!ptr && *fmt != 'K' && *fmt != 'x') {
> 
> I don't know if it matters but with this it won't be immediately
> apparent that a null pointer was printed (since zero could hash to
> anything).

My change touches %px only, where your concern doesn't apply.

You're right, though, when it comes to %pK:
    printk("%%pK: %pK, %%px: %px\n", 0, 0);
says
    %pK: 00000000ba8bdc0a, %px: 0000000000000000

So what should we do?  Avoid hashing 0?  Print a special value?


Meow!
-- 
âââââââ The bill with 3 years prison for mentioning Polish concentration
âââââââ camps is back.  What about KL Warschau (operating until 1956)?
âââââââ Zgoda?  Åambinowice?  Most ex-German KLs?  If those were "soviet
âââââââ puppets", Bereza Kartuska?  Sikorski's camps in UK (thanks Brits!)?