Date: Tue, 25 Sep 2012 19:11:08 -0700 From: Garrett Cooper <yanegomi@gmail.com> To: Ed Maste <emaste@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r240944 - in head/tools/regression/net: . if_tap Message-ID: <CAGH67wQ6yv0pAxB=kU8o6CBAf=6otijCvBxzThhE674fum4rmQ@mail.gmail.com> In-Reply-To: <CAPyFy2C5Yxr8a2nF09pCx6rmz898CYQ=PccwVi8qNketJ1c8KA@mail.gmail.com> References: <201209260025.q8Q0PAp0021186@svn.freebsd.org> <CAGH67wQDaSprjhCXGGLGtu8ggnM-0bk8x1fTezSd8ZFpghnAVQ@mail.gmail.com> <CAPyFy2C5Yxr8a2nF09pCx6rmz898CYQ=PccwVi8qNketJ1c8KA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Sep 25, 2012 at 6:40 PM, Ed Maste <emaste@freebsd.org> wrote: > On 25 September 2012 21:11, Garrett Cooper <yanegomi@gmail.com> wrote: >> On Tue, Sep 25, 2012 at 5:25 PM, Ed Maste <emaste@freebsd.org> wrote: >>> Author: emaste >>> Date: Wed Sep 26 00:25:09 2012 >>> New Revision: 240944 >>> URL: http://svn.freebsd.org/changeset/base/240944 > ... >> This test is technically only valid if sysctl >> kern.features.invariant_support == 1, right? > > It's only possible to prove that the bug described in kern/172075 is > fixed if kern.features.invariant_support == 1, yes. Without > invariants though the test still verifies that a tap(4) device can be > properly destroyed when open and not. > > I have another change coming that will cause cat(1) to terminate > immediately when the device is destroyed, and this script will become > a valid test of that behaviour with or without INVARIANTS. > > Ideally the script would report pass or fail for the individual tap(4) > create and destroy invocations, and separately pass or unsupported for > kern/172075 based on kern.features.invariant_support. I'd like to > figure out how to implement that once our test framework is in place. Here's a basic /bin/sh based example for how to use ATF: http://p4db.freebsd.org/fileViewer.cgi?FSPC=//depot/user/gcooper/atf-head/src/tests/examples/sh_test/sh_test.sh&REV=2 . Please note that using a function like that won't work though unless it's in a separate script as ATF forks a separate shell, and thus all definitions not inlined in the atf_check `test` are lost (it's a caveat that annoys me a small bit). If you want me to CC you on the thread where gnn@ and marcel@ are helping review the FreeBSD ATF porting work, please let me know! Thanks! -Garrett
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAGH67wQ6yv0pAxB=kU8o6CBAf=6otijCvBxzThhE674fum4rmQ>