Re: [PATCH] xen/many: xfree() can tolerate NULL pointers

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 19/01/15 10:54, Ian Campbell wrote:
> On Mon, 2015-01-19 at 10:42 +0000, Andrew Cooper wrote:
>> Replace instances of "if ( p ) xfree(p)" with just "xfree(p)"
>>
>> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
>> CC: Keir Fraser <keir@xen.org>
>> CC: Jan Beulich <JBeulich@suse.com>
>> CC: Tim Deegan <tim@xen.org>
>> CC: Ian Campbell <Ian.Campbell@citrix.com>
>> CC: Ian Jackson <Ian.Jackson@eu.citrix.com>
>>
>> ---
>>
>> This was from some experimentation with semantic patches.  'spatch' can't
>> currently parse some of our macros (e.g. XEN_GUEST_HANDLE()), which cases it
>> to skip large numbers of functions in the codebase
> How annoying! (and surprising)

It also can't parse the "case $X ... $Y:" syntax if the spaces around
the ellipsis is missing.  This is a little more understandable as this
gccism does introduce an ambiguity into the grammar, insofar that 1...3
is not a valid floating point constant.

We have a number of uses with #defined numbers uses without spaces. 
This turns out to be safe only because the preprocessor replacement puts
spaces around replaced tokens.

>
> Anyway, could you include the spatch in the commit log, for completeness
> and for future cargo culting ;-) (unless it's huge, I guess)

It was tiny, but sadly scummed to a `git clean`.  I can't even recall
which of the many kfree() variants I ended up modifying (s/k/x/), but
there is a huge library of generic and Linux specific semantic patches
at https://github.com/coccinelle/coccinellery

~Andrew