diff mbox

[testsuite,10/17] Add options for floatN when checking effective target for support

Message ID 1478878647-22547-11-git-send-email-james.greenhalgh@arm.com
State New
Headers show

Commit Message

James Greenhalgh Nov. 11, 2016, 3:37 p.m. UTC 
Hi,

As Joseph and I discussed in
https://gcc.gnu.org/ml/gcc-patches/2016-11/msg00239.html the
check_effective_target_float<N><x> tests should add any options which the
target requires to enable those float types (add_options_for_float<N><x>.

This patch ensures those options get added, which presently only changes
the testsuite behaviour for the (out-of-tree but submitted for review)
ARM support for _Float16.

Tested on arm-none-linux-gnueabihf with those patches applied, to show that
the _Float16 tests now run.

OK?

Thanks,
James

---

gcc/testsuite/

2016-11-07  James Greenhalgh  <james.greenhalgh@arm.com>

	* lib/target-supports.exp (check_effective_target_float16): Add
	options for _Float16.
	(check_effective_target_float32): Add options for _Float32.
	(check_effective_target_float64): Add options for _Float64.
	(check_effective_target_float128): Add options for _Float128.
	(check_effective_target_float32x): Add options for _Float32x.
	(check_effective_target_float64x): Add options for _Float64x.
	(check_effective_target_float128x): Add options for _Float128x.

Comments

Joseph Myers Nov. 11, 2016, 3:58 p.m. UTC | #1 
On Fri, 11 Nov 2016, James Greenhalgh wrote:

> 

> Hi,

> 

> As Joseph and I discussed in

> https://gcc.gnu.org/ml/gcc-patches/2016-11/msg00239.html the

> check_effective_target_float<N><x> tests should add any options which the

> target requires to enable those float types (add_options_for_float<N><x>.

> 

> This patch ensures those options get added, which presently only changes

> the testsuite behaviour for the (out-of-tree but submitted for review)

> ARM support for _Float16.

> 

> Tested on arm-none-linux-gnueabihf with those patches applied, to show that

> the _Float16 tests now run.

> 

> OK?


OK, but watch out for any issues reported by people testing on powerpc 
(where I expect this to have the effect of causing the _Float128 / 
_Float64x tests to start running).

-- 
Joseph S. Myers
joseph@codesourcery.com
diff mbox

Patch

diff --git a/gcc/testsuite/lib/target-supports.exp b/gcc/testsuite/lib/target-supports.exp
index b683c09..b917250 100644
--- a/gcc/testsuite/lib/target-supports.exp
+++ b/gcc/testsuite/lib/target-supports.exp
@@ -2483,43 +2483,43 @@  proc check_effective_target_has_q_floating_suffix { } {
 proc check_effective_target_float16 {} {
     return [check_no_compiler_messages_nocache float16 object {
         _Float16 x;
-    }]
+    } [add_options_for_float16 ""]]
 }
 
 proc check_effective_target_float32 {} {
     return [check_no_compiler_messages_nocache float32 object {
         _Float32 x;
-    }]
+    } [add_options_for_float32 ""]]
 }
 
 proc check_effective_target_float64 {} {
     return [check_no_compiler_messages_nocache float64 object {
         _Float64 x;
-    }]
+    } [add_options_for_float64 ""]]
 }
 
 proc check_effective_target_float128 {} {
     return [check_no_compiler_messages_nocache float128 object {
         _Float128 x;
-    }]
+    } [add_options_for_float128 ""]]
 }
 
 proc check_effective_target_float32x {} {
     return [check_no_compiler_messages_nocache float32x object {
         _Float32x x;
-    }]
+    } [add_options_for_float32x ""]]
 }
 
 proc check_effective_target_float64x {} {
     return [check_no_compiler_messages_nocache float64x object {
         _Float64x x;
-    }]
+    } [add_options_for_float64x ""]]
 }
 
 proc check_effective_target_float128x {} {
     return [check_no_compiler_messages_nocache float128x object {
         _Float128x x;
-    }]
+    } [add_options_for_float128x ""]]
 }
 
 # Likewise, but runtime support for any special options used as well