Re: [PATCH v2 1/2] tools/nolibc: add support for program_invocation_{,short_}name

[Willy Tarreau]

Hi Thomas,

On Fri, Mar 13, 2026 at 09:26:28PM +0100, Thomas Weißschuh wrote:
> Add support for the GNU extensions 'program_invocation_name' and
> 'program_invocation_short_name'. These are useful to print error
> messages, which by convention include the program name.
> 
> As these are global variables which take up memory even if not used,
> similar to 'errno', gate them behind NOLIBC_IGNORE_ERRNO.
> 
> Signed-off-by: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
> ---
>  tools/include/nolibc/crt.h                   | 26 ++++++++++++++++++++++
>  tools/include/nolibc/errno.h                 |  2 ++
>  tools/testing/selftests/nolibc/nolibc-test.c | 33 ++++++++++++++++++++++++++++
>  3 files changed, 61 insertions(+)
> 
> diff --git a/tools/include/nolibc/crt.h b/tools/include/nolibc/crt.h
> index d9262998dae9..842f86e41f2f 100644
> --- a/tools/include/nolibc/crt.h
> +++ b/tools/include/nolibc/crt.h
> @@ -17,6 +17,7 @@ const unsigned long *_auxv __attribute__((weak));
>  void _start(void);
>  static void __stack_chk_init(void);
>  static void exit(int);
> +static char *strrchr(const char *s, int c);
>  
>  extern void (*const __preinit_array_start[])(int, char **, char**) __attribute__((weak));
>  extern void (*const __preinit_array_end[])(int, char **, char**) __attribute__((weak));
> @@ -27,6 +28,24 @@ extern void (*const __init_array_end[])(int, char **, char**) __attribute__((wea
>  extern void (*const __fini_array_start[])(void) __attribute__((weak));
>  extern void (*const __fini_array_end[])(void) __attribute__((weak));
>  
> +extern char *program_invocation_name __attribute__((weak));
> +extern char *program_invocation_short_name __attribute__((weak));
> +
> +static __inline__
> +char *__nolibc_program_invocation_short_name(char *long_name)
> +{
> +	char *short_name;
> +
> +	if (!long_name)
> +		return NULL;
> +
> +	short_name = strrchr(long_name, '/');
> +	if (!short_name || !short_name[0])
> +		return NULL;

Here it should return long_name, not NULL, since you want a valid name
to use later. I'm seeing it passed to strcmp() for example.

Also because of this, I'm wondering for the first test about !long_name.
Either we consider that it's not possible to have a NULL long_name and
we don't need to test for it, or we consider it is valid, and we should
return a non-null string (e.g. "") so that the rest of the program
survives it. Both approaches are fine to me, but IMHO testing for NULL
to return a NULL that will crash the program later instead of just now
is not very useful. Maybe returning the empty string is still slightly
better because it allows well-designed programs that check their argc
before using argv[0] to survive.

Cheers,
Willy