CI is broken (yet again)

[kmk3]

It looks like it was broken by #4229. Workflow run:

• https://github.com/netblue30/firejail/runs/2534847161

Log excerpt

gcc-10 -g -O2 -ggdb -W -Wall -Werror -O2 -DVERSION='"0.9.65"'  -DPREFIX='"/usr"' -DSYSCONFDIR='"/etc/firejail"' -DLIBDIR='"/usr/lib"' -DBINDIR='"/usr/bin"'  -DHAVE_OUTPUT -DHAVE_X11 -DHAVE_PRIVATE_HOME -DHAVE_APPARMOR  -DHAVE_USERTMPFS -DHAVE_DBUSPROXY -DHAVE_FIRETUNNEL -DHAVE_GLOBALCFG -DHAVE_CHROOT -DHAVE_NETWORK -DHAVE_USERNS -DHAVE_FILE_TRANSFER -DHAVE_WHITELIST -DHAVE_SELINUX  -fstack-protector-all -D_FORTIFY_SOURCE=2 -fPIE -Wformat -Wformat-security -fstack-clash-protection -fstack-protector-strong -fanalyzer   -c fs_whitelist.c -o fs_whitelist.o
In function ‘tmpfs_topdirs’:
fs_whitelist.c:398:1: error: leak of ‘topdirs’ [CWE-401] [-Werror=analyzer-malloc-leak]
  398 | }
      | ^
  ‘fs_whitelist’: events 1-4
    |
    |  523 | void fs_whitelist(void) {
    |      |      ^~~~~~~~~~~~
    |      |      |
    |      |      (1) entry to ‘fs_whitelist’
    |  524 |  ProfileEntry *entry = cfg.profile;
    |  525 |  if (!entry)
    |      |     ~ 
    |      |     |
    |      |     (2) following ‘false’ branch (when ‘entry’ is non-NULL)...
    |......
    |  528 |  if (asprintf(&runuser, "/run/user/%u", getuid()) == -1)
    |      |  ~~ ~ 
    |      |  |  |
    |      |  |  (4) following ‘false’ branch...
    |      |  (3) ...to here
    |
  ‘fs_whitelist’: event 5
    |
    |../include/common.h:39:164:
    |   39 | #define errExit(msg)    do { char msgout[500]; snprintf(msgout, 500, "Error %s: %s:%d %s", msg, __FILE__, __LINE__, __FUNCTION__); perror(msgout); exit(1);} while (0)
    |      |                                                                                                                                                                    ^
    |      |                                                                                                                                                                    |
    |      |                                                                                                                                                                    (5) ...to here
fs_whitelist.c:529:3: note: in expansion of macro ‘errExit’
    |  529 |   errExit("asprintf");
    |      |   ^~~~~~~
    |
  ‘fs_whitelist’: event 6
    |
    |  536 |  if (nowhitelist == NULL)
    |      |     ^
    |      |     |
    |      |     (6) following ‘false’ branch (when ‘nowhitelist’ is non-NULL)...
    |
  ‘fs_whitelist’: event 7
    |
    |../include/common.h:39:164:
    |   39 | #define errExit(msg)    do { char msgout[500]; snprintf(msgout, 500, "Error %s: %s:%d %s", msg, __FILE__, __LINE__, __FUNCTION__); perror(msgout); exit(1);} while (0)
    |      |                                                                                                                                                                    ^
    |      |                                                                                                                                                                    |
    |      |                                                                                                                                                                    (7) ...to here

Link to the previous issue: #4256

Cc: @glitsj16 @netblue30 @reinerh @rusty-snake @smitsohu (as participants
of #4229/#4256)

[reinerh]

I just checked this error and can confirm that it is a false positive by GCC 10.
When rebuilding the same file (fs_whitelist.c) with GCC 11, it no longer detects a memory leak and builds it successfully.

Unfortunately there are new warnings with GCC 11 (see #4274).

[kmk3]

@reinerh @smitsohu Thanks for the quick fixes!