https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70595
Bug ID: 70595
Summary: Cilk Plus testsuite needs massive cleanup
Product: gcc
Version: 6.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: testsuite
Assignee: unassigned at gcc dot gnu.org
Reporter: ro at gcc dot gnu.org
Target Milestone: ---
Prompted by the tentative patch to support SPARC in libcilkrts (PR
target/68945),
I looked into the current Cilk Plus testsuite and found tons of issues, way
beyond
what it takes to support a new target:
* Every single execution test explicitly lists the targets to run on, often
even
twice (in the dg-do target selector and then again when adding -lcilkrts via
dg-options). This is completely unmaintainable and should be replaced by a
target selector. I'm using the current check_libcilkrts_available, renamed
to
cilkplus_runtime, for thta purpose. There's no need to add -lcilkrts at all;
-fcilkplus already does this when linking.
* Two tests (c-c++-common/cilk-plus/CK/pr63307.c and
c-c++-common/cilk-plus/SE/ef_error3.c)
are pure compile tests and don't need a target selector at all.
* This only leaves us with c-c++-common/cilk-plus/SE/ef_error2.c, where the
expected
warning is x86-specific, thus the target selector needs to stay.
With those changes and the cleaned-up sparc port, I could successfully run the
cilk-plus testsuite on Solaris/SPARC, too.
But there's more cleanup possible/necessary:
* Two tests (c-c++-common/cilk-plus/AN/decl-ptr-colon.c and
c-c++-common/cilk-plus/AN/vla.c)
live in a common directory, but are restricted to target c. Unless someone
can educate me otherwise, they simply should be moved to gcc.dg/cilk-plus,
losing the target selector.
* Every since testcase is compiled/linked with -fcilkplus, most (all?) of them
specifying this explicitly via ad-options. This is plain silly and should be
moved into the framework.
* Many tests are run at several optimization levels. This cries for using
the torture framework instead of doing it explicitly.
There may be more, but that's what I've found so far. I'm attaching my current
patch for reference, but this is clearly gcc-7 material.
Rainer
