[AArch64] Fix unordered comparisons to floating-point vcond.

[James Greenhalgh]

Hi,

When I added support for floating-point vcond on aarch64,
I didn't consider the various UNORDERED cases gcc required
support for.  This could in some circumstances lead to
an ICE.

This patch fixes that oversight, adding support for
the UNLT, UNLE, UNEQ, UNGE, UNGT, ORDERED and UNORDERED
cases to the AArch64 backend.

I've tested this patch with no regressions on aarch64-none-elf.

Is this OK to commit?

Regards,
James Greenhalgh

---

gcc/

2013-01-18  James Greenhalgh  <james.greenhalgh@arm.com>

	* config/aarch64/aarch64-simd.md
	(aarch64_vcond_internal<mode>): Handle unordered cases.
	* config/aarch64/iterators.md (v_cmp_result): New.

gcc/testsuite/

2013-01-18  James Greenhalgh  <james.greenhalgh@arm.com>

	* gcc.target/aarch64/vect-fcm-gt-f.c: Change expected output.
	* gcc.target/aarch64/vect-fcm-gt-d.c: Likewise.
	* gcc.target/aarch64/vect-fcm-ge-f.c: Likewise.
	* gcc.target/aarch64/vect-fcm-ge-d.c: Likewise.
	* gcc.target/aarch64/vect-fcm-eq-f.c: Likewise.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-AArch64-Fix-unordered-comparisons-to-floating-point-.patch
Type: text/x-patch
Size: 10171 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20130118/858f2fad/attachment.bin>