Fix a register-class comparison bug in PPCCTRLoops

Thanks to Jakob for isolating the underlying problem from the
test case in r177423. The original commit had introduced
asymmetric copy operations, but these turned out to be a work-around
to the real problem (the use of == instead of hasSubClassEq in PPCCTRLoops).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177679 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Hal Finkel
2013-03-21 23:23:34 +00:00
parent bb4e619cd9
commit 3ea1b064a0
4 changed files with 3 additions and 20 deletions

View File

@@ -2,7 +2,8 @@
target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
target triple = "powerpc64-unknown-linux-gnu"
; This test triggers the use of the asymmetric OR8_32 copy pattern.
; This tests that the GPRC/GPRC_NOR0 intersection subclass relationship with
; GPRC is handled correctly. When it was not, this test would assert.
@gen_random.last = external unnamed_addr global i64, align 8
@.str = external unnamed_addr constant [4 x i8], align 1