Re: [PATCH] checkpatch: Test for kmalloc/memset(0) pairs

From: Julia Lawall
Date: Sun Mar 20 2011 - 08:00:39 EST

On Sun, 20 Mar 2011, Nicolas Palix wrote:

> Hi,
> On Sun, Mar 20, 2011 at 12:06 PM, Pekka Enberg <penberg@xxxxxxxxxx> wrote:
> > Hi,
> >
> > On Sun, Mar 20, 2011 at 12:54 PM, Ingo Molnar <mingo@xxxxxxx> wrote:
> >>> Indeed that one seems to be quite out of date.  You can get the most
> >>> recent version here:
> >>
> >> With tools/coccinelle/ you would never run into such problems of distributing
> >> the latest stable version to your fellow kernel developers: it would always be
> >> available in tools/coccinelle/.
> >>
> >> Integration, synergy, availability, distribution and half a dozen other
> >> buzzwords come to mind as to why it's a good idea to have kernel-focused
> >> tools hosted in the kernel repo :-)
> Our usage is mainly kernel-focused but not the tool. It is C-program focused
> and we have used it on other programs like Wine, OpenSSL, VLC. Others
> use it on other projects like Davecot or close-source projets.
> So, IMHO Coccinelle should no be part of Linux.
> Integrating kernel-focused SmPL scripts is on the other hand a great idea
> to check the kernel and to ease kernel developer life.
> It is what have been done so far. It is certainly possible to improve that,
> at least by adding more and more scripts.
> >>
> >> IMO it's an option to consider.
> >
> > That's my thinking too. Yes, 80 KLOC of OCaml in the kernel tree
> > sounds crazy but I think the practical advantages might be enough to
> > justify it. Btw, would git-submodule be something to consider here?
> >
> At every RC, we push the Coccinelle code on github. Using git-submodule
> seems the way to go thus. Moreover, it will ease the maintenance of
> scripts as we may assume users have one of the latest versions.

The latter is indeed the more serious problem. But even if the Coccinelle
code is available, the person will still have to have ocaml installed to
be able to compile it.