Re: [PATCH] tools/misc: fix xenwatchdogd invocation

[Jan Beulich <jbeulich@xxxxxxxx>]

On 26.03.2024 06:15, zithro / Cyril Rébert wrote:
> --- a/tools/misc/xenwatchdogd.c
> +++ b/tools/misc/xenwatchdogd.c
> @@ -8,26 +8,34 @@
>  #include <stdlib.h>
>  #include <unistd.h>
>  #include <signal.h>
> +#include <string.h>
>  #include <stdio.h>
>  
> +#define _GNU_SOURCE

This wants defining first thing in a file (or even on the compiler command
line), to (properly) affect all headers.

> +#include <getopt.h>
> +#if !defined(__GNUC__) && !defined(__GNUG__)

Why __GNUG__ in a C file? And anyway, why is this construct needed? You
don't ...

> +#define __attribute__(arg) /* empty */

... use any attributes down from here.

You mention xentop as where you've taken them from. I view those as
questionable. If in fact you had used any other utility's source for
reference, you wouldn't have encountered such.

> +#endif
> +
>  xc_interface *h;
>  int id = 0;
>  
>  void daemonize(void)
>  {
>      switch (fork()) {
> -    case -1:
> -     err(1, "fork");
> -    case 0:
> -     break;
> -    default:
> -     exit(0);
> +        case -1:
> +            err(1, "fork");
> +        case 0:
> +            break;
> +        default:
> +            exit(0);
>      }

The case labels were properly indented, weren't they? And all other re-
indentation you do wants splitting from the functional change.

> @@ -52,39 +60,68 @@ void catch_usr1(int sig)
>  
>  int main(int argc, char **argv)
>  {
> +
>      int t, s;

What would this new blank line be good for?

>      int ret;
> +    

I'd even question this one.

> +    char *usage = "usage: %s <timeout> <sleep>";

static const char[]

> +    int opt, optind = 0;
> +
> +    struct option lopts[] = {

static const

> +        { "help",          no_argument,       NULL, 'h' },
> +    };
> +    const char *sopts = "h";
> +
> +    while ((opt = getopt_long(argc, argv, sopts, lopts, &optind)) != -1) {
> +        switch (opt) {
> +        case '?':
> +        case 'h':
> +            errx(1, usage, argv[0]);

This isn't an error and hence wants to produce an exit code of 0.

> +            break;
> +        default:
> +            errx(1, usage, argv[0]);
> +        }

Please be consistent with break statements: Either you omit them
uniformly when a noreturn function is call, or you put them there
in all cases.

> +    }
>  
>      if (argc < 2)
> -     errx(1, "usage: %s <timeout> <sleep>", argv[0]);
> +        errx(1, usage, argv[0]);
>  
>      daemonize();
>  
>      h = xc_interface_open(NULL, NULL, 0);
>      if (h == NULL)
> -     err(1, "xc_interface_open");
> +        err(1, "xc_interface_open");
> +
> +    t = strtoul(argv[1], &argv[1], 0);
> +    
> +    // argv1 NaN

NaN is a term normally used for floating point values only.

> +    if ( *argv[1] != '\0' )
> +        errx(1, "Error: timeout must be a number, got '%s'", argv[1]);

This still doesn't guarantee a numeric: An empty string as argument
would still yield a value of 0 if I'm not mistaken. As would a
string consisting of just blanks.

> -    t = strtoul(argv[1], NULL, 0);
>      if (t == ULONG_MAX)
> -     err(1, "strtoul");
> +        err(1, "strtoul");
>  
>      s = t / 2;
>      if (argc == 3) {
> -     s = strtoul(argv[2], NULL, 0);
> -     if (s == ULONG_MAX)
> -         err(1, "strtoul");
> +        s = strtoul(argv[2], &argv[2], 0);
> +        // argv2 NaN
> +        if ( *argv[2] != '\0' ){
> +            errx(1, "Error: sleep must be a number, got '%s'", argv[2]);
> +        }

Like above you don't really need braces here. If, however, you add
them, the opening one wants to be separated by a blank from the
closing parenthesis. And following other style in this file, there
do not want to be blanks immediately inside the parentheses.

Jan