[Bug 1905044] Re: systemd 245.4-4ubuntu3.3 ADT test failure with linux-hwe-5.8

Tue Dec 15 13:28:36 UTC 2020

@rbalint, fyi since this is a testcase-only fix, I just pushed to the hirsute repo, assuming you wouldn't mind.
https://git.launchpad.net/~ubuntu-core-dev/ubuntu/+source/systemd/commit/?id=d55c6f8186cb2cb73f61d6ff5b2e0ded82ef265c

** Description changed:

  [impact]

- autopkgtest failure when running with 5.8 kernel
+ autopkgtest failure when running with 5.8 kernel, if systemd is built
+ with earlier kernel (e.g. 5.4)

  [test case]

  see autopkgtest results, e.g. links in original description below

  [regression potential]

  as this only fixes a test case, any regression would likely result in an
  incorrectly passing, or incorrectly failing, test.

  [scope]

  this is needed for b/f/g/h.

  this bug was introduced by upstream commit 23cc81e7c22 which first added
  testing for 'invalid' cap numbers, but incorrectly using
  capability_list_length() instead of cap_last_cap(). That commit was
  first included in v236, so this bug does not existing before Bionic.

  this is fixed upstream by commit
  ebc815cd1c647faa934a446ceea91ff4bc9dffa4, which was first included in
  v247, so this is needed for h and earlier.

  Also note that even though the 5.8 kernel is not planned to be released
  for Bionic, this test also runs at build time, and since the LP build
  farm builds inside chroots, if the build farm ever moved up to Focal
  with a 5.8 kernel, the build of systemd for Bionic would start failing,
  so this does need to be fixed in Bionic systemd as well.

  [other info]

  there is a non-test bug related to this in bug 1905245

  [original description]

  Testing failed on:
      amd64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/amd64/s/systemd/20201117_174614_4ece6@/log.gz
      arm64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/arm64/s/systemd/20201117_221555_48b91@/log.gz
      ppc64el: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/ppc64el/s/systemd/20201117_175806_e779f@/log.gz
      s390x: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/s390x/s/systemd/20201117_153051_90e7e@/log.gz

  The failing testcases are:

  - root-unittests

  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  FAIL: test-cap-list (code: 134)

  - upstream

  TEST-24-UNIT-TESTS:

  --- test-cap-list begin ---
  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  --- test-cap-list end ---

  Both seem to be failing with the same assertion.

  These tests are successful on Focal with linux 5.4, therefore they would
  regress when upgrading the kernel from 5.4 to 5.8.

** Description changed:

  [impact]

  autopkgtest failure when running with 5.8 kernel, if systemd is built
  with earlier kernel (e.g. 5.4)

  [test case]

  see autopkgtest results, e.g. links in original description below

  [regression potential]

  as this only fixes a test case, any regression would likely result in an
  incorrectly passing, or incorrectly failing, test.

  [scope]

  this is needed for b/f/g/h.

  this bug was introduced by upstream commit 23cc81e7c22 which first added
  testing for 'invalid' cap numbers, but incorrectly using
  capability_list_length() instead of cap_last_cap(). That commit was
- first included in v236, so this bug does not existing before Bionic.
+ first included in v236, so this bug does not exist before Bionic.

  this is fixed upstream by commit
  ebc815cd1c647faa934a446ceea91ff4bc9dffa4, which was first included in
  v247, so this is needed for h and earlier.

  Also note that even though the 5.8 kernel is not planned to be released
  for Bionic, this test also runs at build time, and since the LP build
  farm builds inside chroots, if the build farm ever moved up to Focal
  with a 5.8 kernel, the build of systemd for Bionic would start failing,
  so this does need to be fixed in Bionic systemd as well.

  [other info]

  there is a non-test bug related to this in bug 1905245

  [original description]

  Testing failed on:
      amd64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/amd64/s/systemd/20201117_174614_4ece6@/log.gz
      arm64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/arm64/s/systemd/20201117_221555_48b91@/log.gz
      ppc64el: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/ppc64el/s/systemd/20201117_175806_e779f@/log.gz
      s390x: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/s390x/s/systemd/20201117_153051_90e7e@/log.gz

  The failing testcases are:

  - root-unittests

  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  FAIL: test-cap-list (code: 134)

  - upstream

  TEST-24-UNIT-TESTS:

  --- test-cap-list begin ---
  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  --- test-cap-list end ---

  Both seem to be failing with the same assertion.

  These tests are successful on Focal with linux 5.4, therefore they would
  regress when upgrading the kernel from 5.4 to 5.8.

** Description changed:

  [impact]

  autopkgtest failure when running with 5.8 kernel, if systemd is built
  with earlier kernel (e.g. 5.4)

  [test case]

  see autopkgtest results, e.g. links in original description below

  [regression potential]

  as this only fixes a test case, any regression would likely result in an
  incorrectly passing, or incorrectly failing, test.

  [scope]

  this is needed for b/f/g/h.

  this bug was introduced by upstream commit 23cc81e7c22 which first added
  testing for 'invalid' cap numbers, but incorrectly using
  capability_list_length() instead of cap_last_cap(). That commit was
  first included in v236, so this bug does not exist before Bionic.

  this is fixed upstream by commit
  ebc815cd1c647faa934a446ceea91ff4bc9dffa4, which was first included in
  v247, so this is needed for h and earlier.

  Also note that even though the 5.8 kernel is not planned to be released
  for Bionic, this test also runs at build time, and since the LP build
  farm builds inside chroots, if the build farm ever moved up to Focal
- with a 5.8 kernel, the build of systemd for Bionic would start failing,
- so this does need to be fixed in Bionic systemd as well.
+ with a 5.8 kernel, the build of systemd for Bionic would start failing
+ (since it would still be building using the older kernel headers, inside
+ the chroot), so this does need to be fixed in Bionic systemd as well.

  [other info]

  there is a non-test bug related to this in bug 1905245

  [original description]

  Testing failed on:
      amd64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/amd64/s/systemd/20201117_174614_4ece6@/log.gz
      arm64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/arm64/s/systemd/20201117_221555_48b91@/log.gz
      ppc64el: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/ppc64el/s/systemd/20201117_175806_e779f@/log.gz
      s390x: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/s390x/s/systemd/20201117_153051_90e7e@/log.gz

  The failing testcases are:

  - root-unittests

  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  FAIL: test-cap-list (code: 134)

  - upstream

  TEST-24-UNIT-TESTS:

  --- test-cap-list begin ---
  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  --- test-cap-list end ---

  Both seem to be failing with the same assertion.

  These tests are successful on Focal with linux 5.4, therefore they would
  regress when upgrading the kernel from 5.4 to 5.8.

-- 
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/1905044

Title:
  systemd 245.4-4ubuntu3.3 ADT test failure with linux-hwe-5.8

Status in systemd package in Ubuntu:
  In Progress
Status in systemd source package in Bionic:
  In Progress
Status in systemd source package in Focal:
  In Progress
Status in systemd source package in Groovy:
  In Progress
Status in systemd source package in Hirsute:
  In Progress

Bug description:
  [impact]

  autopkgtest failure when running with 5.8 kernel, if systemd is built
  with earlier kernel (e.g. 5.4)

  [test case]

  see autopkgtest results, e.g. links in original description below

  [regression potential]

  as this only fixes a test case, any regression would likely result in
  an incorrectly passing, or incorrectly failing, test.

  [scope]

  this is needed for b/f/g/h.

  this bug was introduced by upstream commit 23cc81e7c22 which first
  added testing for 'invalid' cap numbers, but incorrectly using
  capability_list_length() instead of cap_last_cap(). That commit was
  first included in v236, so this bug does not exist before Bionic.

  this is fixed upstream by commit
  ebc815cd1c647faa934a446ceea91ff4bc9dffa4, which was first included in
  v247, so this is needed for h and earlier.

  Also note that even though the 5.8 kernel is not planned to be
  released for Bionic, this test also runs at build time, and since the
  LP build farm builds inside chroots, if the build farm ever moved up
  to Focal with a 5.8 kernel, the build of systemd for Bionic would
  start failing (since it would still be building using the older kernel
  headers, inside the chroot), so this does need to be fixed in Bionic
  systemd as well.

  [other info]

  there is a non-test bug related to this in bug 1905245

  [original description]

  Testing failed on:
      amd64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/amd64/s/systemd/20201117_174614_4ece6@/log.gz
      arm64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/arm64/s/systemd/20201117_221555_48b91@/log.gz
      ppc64el: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/ppc64el/s/systemd/20201117_175806_e779f@/log.gz
      s390x: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/s390x/s/systemd/20201117_153051_90e7e@/log.gz

  The failing testcases are:

  - root-unittests

  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  FAIL: test-cap-list (code: 134)

  - upstream

  TEST-24-UNIT-TESTS:

  --- test-cap-list begin ---
  Assertion 'capability_set_to_string_alloc(c, &t1) == 0' failed at src/test/test-cap-list.c:60, funct
  ion test_capability_set_one(). Aborting.
  --- test-cap-list end ---

  Both seem to be failing with the same assertion.

  These tests are successful on Focal with linux 5.4, therefore they
  would regress when upgrading the kernel from 5.4 to 5.8.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1905044/+subscriptions