Lots of unexpected passes(!), and 1 failure, in testsuite on Linux

On my work Linux installation, on a fresh checkout of 5b007ec5, I get

   9203 total tests, which gave rise to
   36791 test cases, of which
   27140 were skipped
       7 had missing libraries

    9419 expected passes
     201 expected failures

       0 caused framework failures
       0 caused framework warnings
      17 unexpected passes
       1 unexpected failures
       0 unexpected stat failures
      17 fragile tests

Unexpected passes:
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/plugins/frontend01.run                frontend01 [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/prog010/ghci.prog010.run         ghci.prog010 [unexpected] (ghci)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/scripts/ghci058.run              ghci058 [unexpected] (ghci)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/linking/ghcilink005.run          ghcilink005 [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/linking/dyn/load_short_name.run  load_short_name [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/prog001/prog001.run              prog001 [unexpected] (ghci)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/prog001/prog001.run              prog001 [unexpected] (ghci-ext)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/prog002/prog002.run              prog002 [unexpected] (ghci)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/rts/T13676.run                        T13676 [unexpected] (ghci)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/T13786/T13786.run                T13786 [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/ghci/linking/dyn/T1407.run            T1407 [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/package/T20010/T20010-ghci.run        T20010-ghci [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/rts/linker/T20494.run                 T20494 [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/rts/linker/T5435_v_asm_a.run          T5435_v_asm_a [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/rts/linker/T5435_v_asm_b.run          T5435_v_asm_b [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/rts/linker/T5435_v_gcc.run            T5435_v_gcc [unexpected] (normal)
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/rts/T7040_ghci.run                    T7040_ghci [unexpected] (ghci)

Unexpected failures:
   /tmp/ghctest-j30ss1rr/test   spaces/testsuite/tests/driver/j-space/jspace.run  jspace [bad exit code (2)] (normal)

Choosing somewhat at random, I looked at ghci058. I see this in its all.T:

test('ghci058', [extra_files(['../shell.hs']),
                 cmd_prefix('ghciWayFlags=' + config.ghci_way_flags),
                 when(opsys('linux') and not ghc_dynamic(), expect_broken(20706))],
     ghci_script, ['ghci058.script'])

So my guess is that I'm in the not ghc_dynamic() case. Should I be? Is there something strange about my configuration? Is there an important missing configuration in our CI matrix that this has gone uncaught?