Message ID | 53E93D3D.2050109@mentor.com |
---|---|
State | New |
Headers | show |
On Aug 11, 2014, at 3:01 PM, Janis Johnson <janis_johnson@mentor.com> wrote: > The check for effective target arm_v8_neon_ok passes even if __ARM_ARCH > is not 8 or greater, but then some tests fail because intrinsic functions > used in the test have not been declared. This patch requires that > __ARM_ARCH be 8 or greater. Tested for arm-none-linux-gnu for mainline > and 4.9 with a variety of multilib flags. > > OK for mainline and the 4.9 branch? Ok.
On Mon, Aug 11, 2014 at 11:01 PM, Janis Johnson <janis_johnson@mentor.com> wrote: > The check for effective target arm_v8_neon_ok passes even if __ARM_ARCH > is not 8 or greater, but then some tests fail because intrinsic functions > used in the test have not been declared. This patch requires that > __ARM_ARCH be 8 or greater. Tested for arm-none-linux-gnu for mainline > and 4.9 with a variety of multilib flags. Out of curiosity - A number of tests are gated off by the target triplet arm*-*-*eabi* , I wonder how many of them actually run if you test with arm-none-linux-gnu . Also given this is the pre-EABI linux triplet, so I'd prefer not to conflate this with the use of the triplet arm-none-linux-gnueabi(hf). regards Ramana > > OK for mainline and the 4.9 branch? > > Janis
On 08/14/2014 01:06 AM, Ramana Radhakrishnan wrote: > On Mon, Aug 11, 2014 at 11:01 PM, Janis Johnson > <janis_johnson@mentor.com> wrote: >> The check for effective target arm_v8_neon_ok passes even if __ARM_ARCH >> is not 8 or greater, but then some tests fail because intrinsic functions >> used in the test have not been declared. This patch requires that >> __ARM_ARCH be 8 or greater. Tested for arm-none-linux-gnu for mainline >> and 4.9 with a variety of multilib flags. > > > Out of curiosity - A number of tests are gated off by the target > triplet arm*-*-*eabi* , I wonder how many of them actually run if you > test with arm-none-linux-gnu . Also given this is the pre-EABI linux > triplet, so I'd prefer not to conflate this with the use of the > triplet arm-none-linux-gnueabi(hf). > I hadn't noticed those but yes, they are run for arm-none-linux-gnueabi. The effective-target arm_eabi checks for __ARM_EABI__, which is also true for arm-none-linux-gnueabi. I was wrong about my upstream testing in this case; I used arm-none-eabi, and testing with our sources used both arm-none-eabi and arm-none-linux-gnueabi. Janis
Index: gcc/testsuite/lib/target-supports.exp =================================================================== --- gcc/testsuite/lib/target-supports.exp (revision 213831) +++ gcc/testsuite/lib/target-supports.exp (working copy) @@ -2578,6 +2578,9 @@ if { [check_effective_target_arm32] } { foreach flags {"" "-mfloat-abi=softfp" "-mfpu=neon-fp-armv8" "-mfpu=neon-fp-armv8 -mfloat-abi=softfp"} { if { [check_no_compiler_messages_nocache arm_v8_neon_ok object { + #if __ARM_ARCH < 8 + #error not armv8 or later + #endif #include "arm_neon.h" void foo ()