Page MenuHomeFreeBSD

tests/truss: xfail accepted ABIs test if linux(4) not implemented
ClosedPublic

Authored by siva on Fri, Sep 11, 5:02 PM.
Tags
None
Referenced Files
F171868189: D59594.id186473.diff
Mon, Sep 14, 5:23 AM
F171855773: D59594.id186473.diff
Mon, Sep 14, 2:48 AM
F171851952: D59594.id186473.diff
Mon, Sep 14, 2:05 AM
F171818468: D59594.diff
Sun, Sep 13, 6:58 PM
Unknown Object (File)
Sun, Sep 13, 12:42 PM
Unknown Object (File)
Sun, Sep 13, 3:49 AM
Unknown Object (File)
Sun, Sep 13, 2:43 AM
Unknown Object (File)
Sat, Sep 12, 4:25 PM
Subscribers

Details

Summary

Split accepted ABIs test into separate case and conditionally
xfail it on archs that don't have linux(4) support.

This fixes a CI test failure[0].

[0] https://ci.freebsd.org/job/FreeBSD-main-riscv64-test/16706/testReport/usr.bin.truss/truss_test/unknown_syscall/

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Skipped
Unit
Tests Skipped
Build Status
Buildable 76884
Build 73767: arc lint + arc unit

Event Timeline

siva requested review of this revision.Fri, Sep 11, 5:02 PM
dteske requested changes to this revision.Sun, Sep 13, 3:16 AM

Thanks for splitting this out of unknown_syscall, and for using uname -p. One suggested edit on the Linux names.

usr.bin/truss/tests/truss_test.sh
238

warning is correct but skip like require_truss

This revision now requires changes to proceed.Sun, Sep 13, 3:16 AM
usr.bin/truss/tests/truss_test.sh
237–239

This is what I meant to suggest

usr.bin/truss/tests/truss_test.sh
237–239

I like the rest of the suggestion, but I'd argue that atf_expect_fail is more correct here. Since "sysdecode has no Linux ABI names on $arch", I would "expect this testcase to fail". In the future when sysdecode gets support for this, the test should pass again rather than continue to be skipped. We should take action in the future to correct this conditional skip once the support is ready.

This revision is now accepted and ready to land.Mon, Sep 14, 3:38 PM