[PATCH 1/4] main_cp.c: Unused value (CID #1255344)
Gedare Bloom
gedare at rtems.org
Tue Apr 6 16:01:43 UTC 2021
On Mon, Apr 5, 2021 at 1:38 PM Ryan Long <ryan.long at oarcorp.com> wrote:
>
> CID 1255344: Unused value in copy().
>
> Closes #4339
> ---
> cpukit/libmisc/shell/main_cp.c | 64 ++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 64 insertions(+)
>
> diff --git a/cpukit/libmisc/shell/main_cp.c b/cpukit/libmisc/shell/main_cp.c
> index cddbc95..2984a50 100644
> --- a/cpukit/libmisc/shell/main_cp.c
> +++ b/cpukit/libmisc/shell/main_cp.c
> @@ -310,11 +310,31 @@ copy(rtems_shell_cp_globals* cp_globals,
> case FTS_ERR:
> warnx("%s: %s",
> curr->fts_path, strerror(curr->fts_errno));
> + #if __rtems__
> + /*
> + * Coverity spotted that badcp is set by each loop
> + * iteration so setting it right before continue
> + * results in the value being unused. See CID 1255344
> + *
> + * The current NetBSD source (v1.62) was checked and
> + * the same issue appears to apply although the
> + * variable names have changed since this was imported
> + * to RTEMS.
> + *
> + * This pattern exists in multiple places in this file.
> + */
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> continue;
> case FTS_DC: /* Warn, continue. */
> warnx("%s: directory causes a cycle", curr->fts_path);
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> continue;
> default:
> ;
> @@ -366,7 +386,11 @@ copy(rtems_shell_cp_globals* cp_globals,
> if (target_mid - to.p_path + nlen >= PATH_MAX) {
> warnx("%s%s: name too long (not copied)",
> to.p_path, p);
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> continue;
> }
> (void)strncat(target_mid, p, nlen);
> @@ -418,7 +442,11 @@ copy(rtems_shell_cp_globals* cp_globals,
> to_stat.st_ino == curr->fts_statp->st_ino) {
> warnx("%s and %s are identical (not copied).",
> to.p_path, curr->fts_path);
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> if (S_ISDIR(curr->fts_statp->st_mode))
> (void)fts_set(ftsp, curr, FTS_SKIP);
> continue;
> @@ -428,7 +456,11 @@ copy(rtems_shell_cp_globals* cp_globals,
> warnx("cannot overwrite directory %s with "
> "non-directory %s",
> to.p_path, curr->fts_path);
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> continue;
> }
> dne = 0;
> @@ -441,10 +473,18 @@ copy(rtems_shell_cp_globals* cp_globals,
> ((fts_options & FTS_COMFOLLOW) &&
> curr->fts_level == 0)) {
> if (copy_file(cp_globals, curr, dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> } else {
> if (copy_link(cp_globals, curr, !dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> }
> break;
This case breaks, so it is not the same. badcp should be set here, in
case vflag is enabled then a print statement is expected. the same
goes for below.
> case S_IFDIR:
> @@ -452,7 +492,11 @@ copy(rtems_shell_cp_globals* cp_globals,
> warnx("%s is a directory (not copied).",
> curr->fts_path);
> (void)fts_set(ftsp, curr, FTS_SKIP);
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> break;
> }
> /*
> @@ -482,10 +526,18 @@ copy(rtems_shell_cp_globals* cp_globals,
> case S_IFCHR:
> if (Rflag) {
> if (copy_special(cp_globals, curr->fts_statp, !dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> } else {
> if (copy_file(cp_globals, curr, dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> }
> break;
> case S_IFSOCK:
> @@ -494,15 +546,27 @@ copy(rtems_shell_cp_globals* cp_globals,
> case S_IFIFO:
> if (Rflag) {
> if (copy_fifo(cp_globals, curr->fts_statp, !dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> } else {
> if (copy_file(cp_globals, curr, dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> }
> break;
> default:
> if (copy_file(cp_globals, curr, dne))
> + #if __rtems__
> + rval = 1;
> + #else
> badcp = rval = 1;
> + #endif
> break;
> }
> if (vflag && !badcp)
> --
> 1.8.3.1
>
> _______________________________________________
> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel
More information about the devel
mailing list