Change register allocation order for ARM VFP and NEON registers to put the

callee-saved registers at the end of the lists.  Also prefer to avoid using
the low registers that are in register subclasses required by certain
instructions, so that those registers will more likely be available when needed.
This change makes a huge improvement in spilling in some cases.  Thanks to
Jakob for helping me realize the problem.

Most of this patch is fixing the testsuite.  There are quite a few places
where we're checking for specific registers.  I changed those to wildcards
in places where that doesn't weaken the tests.  The spill-q.ll and
thumb2-spill-q.ll tests stopped spilling with this change, so I added a bunch
of live values to force spills on those tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116055 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 2 insertions, 2 deletions

diff --git a/test/CodeGen/ARM/fpconsts.ll b/test/CodeGen/ARM/fpconsts.ll
index f1d6a16f3e..9e7a8ae28b 100644
--- a/test/CodeGen/ARM/fpconsts.ll
+++ b/test/CodeGen/ARM/fpconsts.ll
@@ -11,7 +11,7 @@ entry:
 define double @t2(double %x) nounwind readnone optsize {
 entry:
 ; CHECK: t2:
-; CHECK: vmov.f64 d1, #3.000000e+00
+; CHECK: vmov.f64 d{{.*}}, #3.000000e+00
   %0 = fadd double %x, 3.000000e+00
   ret double %0
 }
@@ -19,7 +19,7 @@ entry:
 define double @t3(double %x) nounwind readnone optsize {
 entry:
 ; CHECK: t3:
-; CHECK: vmov.f64 d1, #-1.300000e+01
+; CHECK: vmov.f64 d{{.*}}, #-1.300000e+01
   %0 = fmul double %x, -1.300000e+01
   ret double %0
 }