Re: [PATCH 1/2] exec: Add KUnit test for bprm_stack_limits()

[Guenter Roeck]

On 3/18/26 11:33, Kees Cook wrote:
> On Mon, Mar 16, 2026 at 12:16:58PM -0700, Guenter Roeck wrote:
> > Hi,
> >
> > On Sun, May 19, 2024 at 07:16:11PM -0700, Kees Cook wrote:
> > > Since bprm_stack_limits() operates with very limited side-effects, add
> > > it as the first exec.c KUnit test. Add to Kconfig and adjust MAINTAINERS
> > > file to include it.
> > >
> > > Tested on 64-bit UML:
> > > $ tools/testing/kunit/kunit.py run exec
> > >
> > > Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
> > > ---
> > > Cc: Eric Biederman <ebiederm@xxxxxxxxxxxx>
> > > Cc: Justin Stitt <justinstitt@xxxxxxxxxx>
> > > Cc: Alexander Viro <viro@xxxxxxxxxxxxxxxxxx>
> > > Cc: Christian Brauner <brauner@xxxxxxxxxx>
> > > Cc: Jan Kara <jack@xxxxxxx>
> > > Cc: linux-fsdevel@xxxxxxxxxxxxxxx
> > > Cc: linux-mm@xxxxxxxxx
> > > ---
> > ...
> > > +++ b/fs/exec.c
> > ...
> > > +	  .expected_argmin = ULONG_MAX - (_STK_LIM / 4 * 3) + sizeof(void *) },
> > > +	{ { .p = ULONG_MAX, .rlim_stack.rlim_cur = 4 * (_STK_LIM / 4 *  + sizeof(void *)),
> >                                                                     ^^^^^^
> >
> > Google's experimental AI review agent tagged the above change as syntactically wrong.
> > How does this even compile ?
>
> Wow. I have no idea how this ever built either. I will get it fixed.
I ended up writing a test program.

#define TESTVAL 10
...
long int x = TESTVAL / 4 * + sizeof(void *);

does build, as does

long int x = TESTVAL / 4 * - sizeof(void *);

or even

long int x = TESTVAL / 4 * 3 - - sizeof(void *);

long int x = TESTVAL / 4 * - - sizeof(void *);

It doesn't even have problems with

long int x = TESTVAL / 4 * - - + - + - sizeof(void *);

Weird. Looks like the compiler takes the "+" or "-" as sign value,
not as operator, and once can have as many of those as one desires.

Guenter