[Bug 2031683] Re: systemd FTBFS on ppc64el due to fuzzer failures

[Nick Rosbrook]

** Description changed:

+ [Impact]
+ 
+ systemd FTBFS on ppc64el due to failing fuzz tests.
+ 
+ [Test Plan]
+ 
+ The build-time tests and fuzzing should pass so that the build
+ ultimately passes on ppc64el.
+ 
+ [Where problems could occur]
+ 
+ The patch just removes the "pure" attribute from
+ bus_message_type_from_string, because that attribute was incorrect for
+ the behavior of the function. Hence, the regression potential should be
+ limited to building the package.
+ 
+ [Original Description]
+ 
  This PPA build[1] demonstrates the issue, the relevant part being:
  
  [ ... ]
  1061/1066 systemd:dist-check / check-help-networkctl                           OK                0.04s
  1062/1066 systemd:dist-check / check-version-networkctl                        OK                0.03s
  1063/1066 systemd:dist-check / check-help-kernel-install                       OK                0.03s
  1064/1066 systemd:dist-check / check-version-kernel-install                    OK                0.02s
  1065/1066 systemd:dist-check / check-directives                                OK                0.18s
  1066/1066 systemd:dist-check / parse-hwdb                                      OK                5.43s
  
  Summary of Failures:
  
-  661/1066 systemd:fuzzers / fuzz-bus-match_bugzilla1935084.input               FAIL              0.22s   killed by signal 11 SIGSEGV
-  667/1066 systemd:fuzzers / fuzz-bus-match_test.input                          FAIL              0.25s   killed by signal 11 SIGSEGV
+  661/1066 systemd:fuzzers / fuzz-bus-match_bugzilla1935084.input               FAIL              0.22s   killed by signal 11 SIGSEGV
+  667/1066 systemd:fuzzers / fuzz-bus-match_test.input                          FAIL              0.25s   killed by signal 11 SIGSEGV
  
  Ok:                 1032
- Expected Fail:      0   
- Fail:               2   
- Unexpected Pass:    0   
- Skipped:            32  
- Timeout:            0   
+ Expected Fail:      0
+ Fail:               2
+ Unexpected Pass:    0
+ Skipped:            32
+ Timeout:            0
  
  Full log written to /<<PKGBUILDDIR>>/obj-powerpc64le-linux-gnu/meson-logs/testlog.txt
  make[1]: *** [debian/rules:308: override_dh_auto_test] Error 2
  make[1]: Leaving directory '/<<PKGBUILDDIR>>'
  make: *** [debian/rules:312: binary-arch] Error 2
  dpkg-buildpackage: error: debian/rules binary-arch subprocess returned exit status 2
  
  This appears to be related to a toolchain change, because I get the same
  result when doing a no-change rebuild of systemd on ppc64el in Lunar
  (i.e. the FTBFS is not introduced by the changes I am testing).
  
  [1] https://launchpadlibrarian.net/682104329/buildlog_ubuntu-lunar-
  ppc64el.systemd_252.5-2ubuntu4~ppa11_BUILDING.txt.gz

** Changed in: systemd (Ubuntu Lunar)
       Status: Confirmed => Triaged

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/2031683

Title:
  systemd FTBFS on ppc64el due to fuzzer failures

Status in systemd package in Ubuntu:
  Fix Released
Status in systemd source package in Lunar:
  Triaged

Bug description:
  [Impact]

  systemd FTBFS on ppc64el due to failing fuzz tests.

  [Test Plan]

  The build-time tests and fuzzing should pass so that the build
  ultimately passes on ppc64el.

  [Where problems could occur]

  The patch just removes the "pure" attribute from
  bus_message_type_from_string, because that attribute was incorrect for
  the behavior of the function. Hence, the regression potential should
  be limited to building the package.

  [Original Description]

  This PPA build[1] demonstrates the issue, the relevant part being:

  [ ... ]
  1061/1066 systemd:dist-check / check-help-networkctl                           OK                0.04s
  1062/1066 systemd:dist-check / check-version-networkctl                        OK                0.03s
  1063/1066 systemd:dist-check / check-help-kernel-install                       OK                0.03s
  1064/1066 systemd:dist-check / check-version-kernel-install                    OK                0.02s
  1065/1066 systemd:dist-check / check-directives                                OK                0.18s
  1066/1066 systemd:dist-check / parse-hwdb                                      OK                5.43s

  Summary of Failures:

   661/1066 systemd:fuzzers / fuzz-bus-match_bugzilla1935084.input               FAIL              0.22s   killed by signal 11 SIGSEGV
   667/1066 systemd:fuzzers / fuzz-bus-match_test.input                          FAIL              0.25s   killed by signal 11 SIGSEGV

  Ok:                 1032
  Expected Fail:      0
  Fail:               2
  Unexpected Pass:    0
  Skipped:            32
  Timeout:            0

  Full log written to /<<PKGBUILDDIR>>/obj-powerpc64le-linux-gnu/meson-logs/testlog.txt
  make[1]: *** [debian/rules:308: override_dh_auto_test] Error 2
  make[1]: Leaving directory '/<<PKGBUILDDIR>>'
  make: *** [debian/rules:312: binary-arch] Error 2
  dpkg-buildpackage: error: debian/rules binary-arch subprocess returned exit status 2

  This appears to be related to a toolchain change, because I get the
  same result when doing a no-change rebuild of systemd on ppc64el in
  Lunar (i.e. the FTBFS is not introduced by the changes I am testing).

  [1] https://launchpadlibrarian.net/682104329/buildlog_ubuntu-lunar-
  ppc64el.systemd_252.5-2ubuntu4~ppa11_BUILDING.txt.gz

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/2031683/+subscriptions