On Sun, 11 Sep 2016, Joe Perches wrote:
> On Sun, 2016-09-11 at 15:05 +0200, Julia Lawall wrote:
> > Constify local structures.
>
> Thanks Julia.
>
> A few suggestions & questions:
>
> Perhaps the script should go into scripts/coccinelle/
> so that future cases could be caught by the robot
> and commit message referenced by the patch instances.
OK.
> Can you please compile the files modified using the
> appropriate defconfig/allyesconfig and show the
I currently send patches for this issue only for files that compile using
the x86 allyesconfig.
> movement from data to const by using
> $ size <object>.new/old
> and include that in the changelogs (maybe next time)?
OK, thanks for the suggestion.
> Is it possible for a rule to trace the instances where
> an address of a struct or struct member is taken by
> locally defined and declared function call where the
> callee does not modify any dereferenced object?
>
> ie:
>
> struct foo {
> int bar;
> char *baz;
> };
>
> struct foo qux[] = {
> { 1, "description 1" },
> { 2, "dewcription 2" },
> [ n, "etc" ]...,
> };
>
> void message(struct foo *msg)
> {
> printk("%d %s\n", msg->bar, msg->baz);
> }
>
> where some code uses
>
> message(qux[index]);
>
> So could a coccinelle script change:
>
> struct foo qux[] = { to const struct foo quz[] = {
>
> and
>
> void message(struct foo *msg) to void message(const struct foo *msg)
Yes, this could be possible too.
Thanks for the feedback.
julia
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
