[PATCH] FAIL: gcc.dg/torture/ftrapv-1.c -O0 (test for excess errors) on bare-metal targets

[PATCH] FAIL: gcc.dg/torture/ftrapv-1.c -O0 (test for excess errors) on bare-metal targets

[Kyrill Tkachov]

[-- Attachment #1: Type: text/plain, Size: 1305 bytes --]

Hi all,

The recently added -ftrapv test FAILS on bare metal targets because it 
requires the usage of fork.
It does have a dg-require-fork guard but apparently that function takes 
an unused argument and if you don't give it that, DejaGNU will happily 
ignore the directive and not bother checking for the presence of fork :S

The other places where I see dg-require-fork being used is in the 
libstdc++ testsuite. Perhaps it's worth just removing the argument from 
the definition of dg-require-fork in target-supports-dg.exp, but there 
is a whole host of functions that take an argument and then don't use it 
(dg-require-mkfifo is one, for example) so it seems like it would be an 
out-of-place change unless we do a big sweep across the testsuite 
library to clean that stuff up. This patch just fixes the 
dg-requires-fork invocation in gcc.dg/torture/ftrapv-1.c.

Richard (and testsuite maintainers), is this ok? or would you prefer to 
remove the unused arguments for the dg-requires-* functions in 
testsuite/lib/ ?

With this patch the test now appears as UNSUPPORTED on bare metal 
aarch64 and arm targets (that I've tested).

Thanks,
Kyrill

2014-08-05  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>

     * gcc.dg/torture/ftrapv-1.c: Correct usage of dg-require-fork.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: dg-require-fork.patch --]
[-- Type: text/x-patch; name=dg-require-fork.patch, Size: 636 bytes --]

commit f7e9661501f3ceb2221c78569ce6bf6e1140f040
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Mon Aug 4 15:38:40 2014 +0100

    [test] Fix dg-require-fork usage

diff --git a/gcc/testsuite/gcc.dg/torture/ftrapv-1.c b/gcc/testsuite/gcc.dg/torture/ftrapv-1.c
index 4fdccd8..c74535f 100644
--- a/gcc/testsuite/gcc.dg/torture/ftrapv-1.c
+++ b/gcc/testsuite/gcc.dg/torture/ftrapv-1.c
@@ -1,7 +1,7 @@
 /* { dg-do run } */
 /* { dg-additional-options "-ftrapv" } */
 /* { dg-require-effective-target trapping } */
-/* { dg-require-fork } */
+/* { dg-require-fork "" } */
 
 #include <stdlib.h>
 #include <unistd.h>

Re: [PATCH] FAIL: gcc.dg/torture/ftrapv-1.c -O0 (test for excess errors) on bare-metal targets

[Richard Biener]

On Tue, Aug 5, 2014 at 11:47 AM, Kyrill Tkachov <kyrylo.tkachov@arm.com> wrote:
> Hi all,
>
> The recently added -ftrapv test FAILS on bare metal targets because it
> requires the usage of fork.
> It does have a dg-require-fork guard but apparently that function takes an
> unused argument and if you don't give it that, DejaGNU will happily ignore
> the directive and not bother checking for the presence of fork :S
>
> The other places where I see dg-require-fork being used is in the libstdc++
> testsuite. Perhaps it's worth just removing the argument from the definition
> of dg-require-fork in target-supports-dg.exp, but there is a whole host of
> functions that take an argument and then don't use it (dg-require-mkfifo is
> one, for example) so it seems like it would be an out-of-place change unless
> we do a big sweep across the testsuite library to clean that stuff up. This
> patch just fixes the dg-requires-fork invocation in
> gcc.dg/torture/ftrapv-1.c.
>
> Richard (and testsuite maintainers), is this ok? or would you prefer to
> remove the unused arguments for the dg-requires-* functions in
> testsuite/lib/ ?
>
> With this patch the test now appears as UNSUPPORTED on bare metal aarch64
> and arm targets (that I've tested).

Eh, I thought I fixed that already... (must have failed to commit).

OK.

Thanks,
Richard.

> Thanks,
> Kyrill
>
> 2014-08-05  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
>
>     * gcc.dg/torture/ftrapv-1.c: Correct usage of dg-require-fork.