Re: [PATCH] test: skip tst.symbols.c if no /proc/kallmodsyms

[Kris Van Hees <kris.van.hees@oracle.com>]

On Wed, Jul 31, 2024 at 03:36:49PM -0400, Eugene Loh wrote:
> Should we be adding a version of the test that's appropriate for
> /proc/kallsyms?

I don't think that would make sense because this test seems to validate the
kallmodsyms information against the information dtrace is able to determine.
But with kallsyms and modules.builtin.ranges, we don't really have a good
way to validate the module association because there is only one source for
it.

I guess we could implement a version of this test that merely checks that
the symbol-address association is accurate but that is possibly not even
really worth doing?  We obtain it using a straight read from the file and
adding the name and address to a list.  Validation would merely mean that
we validate that we read it correctly and stored it correctly.
> 
> On 7/31/24 15:02, Kris Van Hees wrote:
> > Signed-off-by: Kris Van Hees <kris.van.hees@oracle.com>
> > ---
> >   test/unittest/consumer/tst.symbols.x | 12 ++++++++++++
> >   1 file changed, 12 insertions(+)
> >   create mode 100755 test/unittest/consumer/tst.symbols.x
> > 
> > diff --git a/test/unittest/consumer/tst.symbols.x b/test/unittest/consumer/tst.symbols.x
> > new file mode 100755
> > index 00000000..e83d97f6
> > --- /dev/null
> > +++ b/test/unittest/consumer/tst.symbols.x
> > @@ -0,0 +1,12 @@
> > +#!/bin/bash
> > +#
> > +# Oracle Linux DTrace.
> > +# Copyright (c) 2024, Oracle and/or its affiliates. All rights reserved.
> > +# Licensed under the Universal Permissive License v 1.0 as shown at
> > +# http://oss.oracle.com/licenses/upl.
> > +
> > +# If /proc/kallmodsyms does not exist, there is nothing to test
> > +
> > +[[ -r /proc/kallmodsyms ]] || exit 2
> > +
> > +exit 0