Re: [PATCH] checkpatch: Add --strict test for kmalloc/kzalloc with multiply

[Joe Perches <joe@perches.com>]

On Thu, 2014-05-15 at 15:58 -0700, Andrew Morton wrote:
> On Tue, 13 May 2014 16:48:49 -0700 Joe Perches <joe@perches.com> wrote:
> > Protect against sizeof overflows by preferring
> > kmalloc_array and kcalloc to kmalloc/kzalloc
> > with a sizeof multiply.
[]
> > +# check for k[mz]alloc with multiplies that could be kmalloc_array/kcalloc
> > +		if ($^V && $^V ge 5.10.0 &&
> > +		    $line =~ /\b($Lval)\s*\=\s*(?:$balanced_parens)?\s*(k[mz]alloc)\s*\(\s*($FuncArg)\s*\*\s*($FuncArg)/) {
> > +			my $oldfunc = $3;
> > +			my $a1 = $4;
> > +			my $a2 = $10;
> > +			my $newfunc = "kmalloc_array";
> > +			$newfunc = "kcalloc" if ($oldfunc eq "kzalloc");
> > +			if ($a1 =~ /^sizeof\s*\S/ || $a2 =~ /^sizeof\s*\S/) {
> > +				CHK("ALLOC_WITH_MULTIPLY",
> > +				    "Prefer $newfunc over $oldfunc with multiply\n" . $herecurr);
> > +			}
> > +		}
> > +
> 
> Why hide this behind --strict?

Non-obvious CHK/--strict tests are less controversial.

The block above it
"prefer foo = alloc(sizeof(*foo)) over foo = alloc(sizeof(struct bar))"
used CHK so I copied it.

I've no objection to making it WARN instead,