[SystemZ] Use MVC to spill loads and stores

Try to use MVC when spilling the destination of a simple load or the source
of a simple store.  As explained in the comment, this doesn't yet handle
the case where the load or store location is also a frame index, since
that could lead to two simultaneous scavenger spills, something the
backend can't handle yet.  spill-02.py tests that this restriction kicks in,
but unfortunately I've not yet found a case that would fail without it.
The volatile trick I used for other scavenger tests doesn't work here
because we can't use MVC for volatile accesses anyway.

I'm planning on relaxing the restriction later, hopefully with a test
that does trigger the problem...

Tests @f8 and @f9 also showed that L(G)RL and ST(G)RL were wrongly
classified as SimpleBDX{Load,Store}.  It wouldn't be easy to test for
that bug separately, which is why I didn't split out the fix as a
separate patch.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@185434 91177308-0d34-0410-b5e6-96231b3b80d8

3 files changed, 469 insertions, 0 deletions

diff --git a/test/CodeGen/SystemZ/Large/spill-01.py b/test/CodeGen/SystemZ/Large/spill-01.py
new file mode 100644
index 0000000..3c1d0b6
--- /dev/null
+++ b/test/CodeGen/SystemZ/Large/spill-01.py
@@ -0,0 +1,40 @@
+# Test cases where MVC is used for spill slots that end up being out of range.
+# RUN: python %s | llc -mtriple=s390x-linux-gnu | FileCheck %s
+
+# There are 8 usable call-saved GPRs, two of which are needed for the base
+# registers.  The first 160 bytes of the frame are needed for the ABI
+# call frame, and a further 8 bytes are needed for the emergency spill slot.
+# That means we will have at least one out-of-range slot if:
+#
+#    count == (4096 - 168) / 8 + 6 + 1 == 498
+#
+# Add in some extra room and check both %r15+4096 (the first out-of-range slot)
+# and %r15+4104.
+#
+# CHECK: f1:
+# CHECK: lay [[REG:%r[0-5]]], 4096(%r15)
+# CHECK: mvc 0(8,[[REG]]), {{[0-9]+}}({{%r[0-9]+}})
+# CHECK: brasl %r14, foo@PLT
+# CHECK: lay [[REG:%r[0-5]]], 4096(%r15)
+# CHECK: mvc {{[0-9]+}}(8,{{%r[0-9]+}}), 8([[REG]])
+# CHECK: br %r14
+count = 500
+
+print 'declare void @foo()'
+print ''
+print 'define void @f1(i64 *%base0, i64 *%base1) {'
+
+for i in range(count):
+    print '  %%ptr%d = getelementptr i64 *%%base%d, i64 %d' % (i, i % 2, i / 2)
+    print '  %%val%d = load i64 *%%ptr%d' % (i, i)
+    print ''
+
+print '  call void @foo()'
+print ''
+
+for i in range(count):
+    print '  store i64 %%val%d, i64 *%%ptr%d' % (i, i)
+
+print ''
+print '  ret void'
+print '}'
diff --git a/test/CodeGen/SystemZ/Large/spill-02.py b/test/CodeGen/SystemZ/Large/spill-02.py
new file mode 100644
index 0000000..0eba3ed
--- /dev/null
+++ b/test/CodeGen/SystemZ/Large/spill-02.py
@@ -0,0 +1,46 @@
+# Test cases where we spill from one frame index to another, both of which
+# would be out of range of MVC.  At present we don't use MVC in this case.
+# RUN: python %s | llc -mtriple=s390x-linux-gnu | FileCheck %s
+
+# There are 8 usable call-saved GPRs.  The first 160 bytes of the frame
+# are needed for the ABI call frame, and a further 8 bytes are needed
+# for the emergency spill slot.  That means we will have at least one
+# out-of-range slot if:
+#
+#    count == (4096 - 168) / 8 + 8 + 1 == 500
+#
+# Add in some extra just to be sure.
+#
+# CHECK: f1:
+# CHECK-NOT: mvc
+# CHECK: br %r14
+count = 510
+
+print 'declare void @foo(i64 *%base0, i64 *%base1)'
+print ''
+print 'define void @f1() {'
+
+for i in range(2):
+    print '  %%alloc%d = alloca [%d x i64]' % (i, count / 2)
+    print ('  %%base%d = getelementptr [%d x i64] * %%alloc%d, i64 0, i64 0'
+           % (i, count / 2, i))
+
+print '  call void @foo(i64 *%base0, i64 *%base1)'
+print ''
+
+for i in range(count):
+    print '  %%ptr%d = getelementptr i64 *%%base%d, i64 %d' % (i, i % 2, i / 2)
+    print '  %%val%d = load i64 *%%ptr%d' % (i, i)
+    print ''
+
+print '  call void @foo(i64 *%base0, i64 *%base1)'
+print ''
+
+for i in range (count):
+    print '  store i64 %%val%d, i64 *%%ptr%d' % (i, i)
+
+print ''
+print '  call void @foo(i64 *%base0, i64 *%base1)'
+print ''
+print '  ret void'
+print '}'
diff --git a/test/CodeGen/SystemZ/spill-01.ll b/test/CodeGen/SystemZ/spill-01.ll
new file mode 100644
index 0000000..a5e48ee
--- /dev/null
+++ b/test/CodeGen/SystemZ/spill-01.ll
@@ -0,0 +1,383 @@
+; Test spilling using MVC.
+;
+; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s
+
+declare void @foo()
+
+@g0 = global i32 0
+@g1 = global i32 1
+@g2 = global i32 2
+@g3 = global i32 3
+@g4 = global i32 4
+@g5 = global i32 5
+@g6 = global i32 6
+@g7 = global i32 7
+@g8 = global i32 8
+@g9 = global i32 9
+
+@h0 = global i64 0
+@h1 = global i64 1
+@h2 = global i64 2
+@h3 = global i64 3
+@h4 = global i64 4
+@h5 = global i64 5
+@h6 = global i64 6
+@h7 = global i64 7
+@h8 = global i64 8
+@h9 = global i64 9
+
+; This function shouldn't spill anything
+define void @f1(i32 *%ptr0) {
+; CHECK: f1:
+; CHECK: stmg
+; CHECK: aghi %r15, -160
+; CHECK-NOT: %r15
+; CHECK: brasl %r14, foo@PLT
+; CHECK-NOT: %r15
+; CHECK: lmg
+; CHECK: br %r14
+  %ptr1 = getelementptr i32 *%ptr0, i32 2
+  %ptr2 = getelementptr i32 *%ptr0, i32 4
+  %ptr3 = getelementptr i32 *%ptr0, i32 6
+  %ptr4 = getelementptr i32 *%ptr0, i32 8
+  %ptr5 = getelementptr i32 *%ptr0, i32 10
+  %ptr6 = getelementptr i32 *%ptr0, i32 12
+
+  %val0 = load i32 *%ptr0
+  %val1 = load i32 *%ptr1
+  %val2 = load i32 *%ptr2
+  %val3 = load i32 *%ptr3
+  %val4 = load i32 *%ptr4
+  %val5 = load i32 *%ptr5
+  %val6 = load i32 *%ptr6
+
+  call void @foo()
+
+  store i32 %val0, i32 *%ptr0
+  store i32 %val1, i32 *%ptr1
+  store i32 %val2, i32 *%ptr2
+  store i32 %val3, i32 *%ptr3
+  store i32 %val4, i32 *%ptr4
+  store i32 %val5, i32 *%ptr5
+  store i32 %val6, i32 *%ptr6
+
+  ret void
+}
+
+; Test a case where at least one i32 load and at least one i32 store
+; need spills.
+define void @f2(i32 *%ptr0) {
+; CHECK: f2:
+; CHECK: mvc [[OFFSET1:16[04]]](4,%r15), [[OFFSET2:[0-9]+]]({{%r[0-9]+}})
+; CHECK: brasl %r14, foo@PLT
+; CHECK: mvc [[OFFSET2]](4,{{%r[0-9]+}}), [[OFFSET1]](%r15)
+; CHECK: br %r14
+  %ptr1 = getelementptr i32 *%ptr0, i64 2
+  %ptr2 = getelementptr i32 *%ptr0, i64 4
+  %ptr3 = getelementptr i32 *%ptr0, i64 6
+  %ptr4 = getelementptr i32 *%ptr0, i64 8
+  %ptr5 = getelementptr i32 *%ptr0, i64 10
+  %ptr6 = getelementptr i32 *%ptr0, i64 12
+  %ptr7 = getelementptr i32 *%ptr0, i64 14
+  %ptr8 = getelementptr i32 *%ptr0, i64 16
+
+  %val0 = load i32 *%ptr0
+  %val1 = load i32 *%ptr1
+  %val2 = load i32 *%ptr2
+  %val3 = load i32 *%ptr3
+  %val4 = load i32 *%ptr4
+  %val5 = load i32 *%ptr5
+  %val6 = load i32 *%ptr6
+  %val7 = load i32 *%ptr7
+  %val8 = load i32 *%ptr8
+
+  call void @foo()
+
+  store i32 %val0, i32 *%ptr0
+  store i32 %val1, i32 *%ptr1
+  store i32 %val2, i32 *%ptr2
+  store i32 %val3, i32 *%ptr3
+  store i32 %val4, i32 *%ptr4
+  store i32 %val5, i32 *%ptr5
+  store i32 %val6, i32 *%ptr6
+  store i32 %val7, i32 *%ptr7
+  store i32 %val8, i32 *%ptr8
+
+  ret void
+}
+
+; Test a case where at least one i64 load and at least one i64 store
+; need spills.
+define void @f3(i64 *%ptr0) {
+; CHECK: f3:
+; CHECK: mvc 160(8,%r15), [[OFFSET:[0-9]+]]({{%r[0-9]+}})
+; CHECK: brasl %r14, foo@PLT
+; CHECK: mvc [[OFFSET]](8,{{%r[0-9]+}}), 160(%r15)
+; CHECK: br %r14
+  %ptr1 = getelementptr i64 *%ptr0, i64 2
+  %ptr2 = getelementptr i64 *%ptr0, i64 4
+  %ptr3 = getelementptr i64 *%ptr0, i64 6
+  %ptr4 = getelementptr i64 *%ptr0, i64 8
+  %ptr5 = getelementptr i64 *%ptr0, i64 10
+  %ptr6 = getelementptr i64 *%ptr0, i64 12
+  %ptr7 = getelementptr i64 *%ptr0, i64 14
+  %ptr8 = getelementptr i64 *%ptr0, i64 16
+
+  %val0 = load i64 *%ptr0
+  %val1 = load i64 *%ptr1
+  %val2 = load i64 *%ptr2
+  %val3 = load i64 *%ptr3
+  %val4 = load i64 *%ptr4
+  %val5 = load i64 *%ptr5
+  %val6 = load i64 *%ptr6
+  %val7 = load i64 *%ptr7
+  %val8 = load i64 *%ptr8
+
+  call void @foo()
+
+  store i64 %val0, i64 *%ptr0
+  store i64 %val1, i64 *%ptr1
+  store i64 %val2, i64 *%ptr2
+  store i64 %val3, i64 *%ptr3
+  store i64 %val4, i64 *%ptr4
+  store i64 %val5, i64 *%ptr5
+  store i64 %val6, i64 *%ptr6
+  store i64 %val7, i64 *%ptr7
+  store i64 %val8, i64 *%ptr8
+
+  ret void
+}
+
+
+; Test a case where at least at least one f32 load and at least one f32 store
+; need spills.  The 8 call-saved FPRs could be used for 8 of the %vals
+; (and are at the time of writing), but it would really be better to use
+; MVC for all 10.
+define void @f4(float *%ptr0) {
+; CHECK: f4:
+; CHECK: mvc [[OFFSET1:16[04]]](4,%r15), [[OFFSET2:[0-9]+]]({{%r[0-9]+}})
+; CHECK: brasl %r14, foo@PLT
+; CHECK: mvc [[OFFSET2]](4,{{%r[0-9]+}}), [[OFFSET1]](%r15)
+; CHECK: br %r14
+  %ptr1 = getelementptr float *%ptr0, i64 2
+  %ptr2 = getelementptr float *%ptr0, i64 4
+  %ptr3 = getelementptr float *%ptr0, i64 6
+  %ptr4 = getelementptr float *%ptr0, i64 8
+  %ptr5 = getelementptr float *%ptr0, i64 10
+  %ptr6 = getelementptr float *%ptr0, i64 12
+  %ptr7 = getelementptr float *%ptr0, i64 14
+  %ptr8 = getelementptr float *%ptr0, i64 16
+  %ptr9 = getelementptr float *%ptr0, i64 18
+
+  %val0 = load float *%ptr0
+  %val1 = load float *%ptr1
+  %val2 = load float *%ptr2
+  %val3 = load float *%ptr3
+  %val4 = load float *%ptr4
+  %val5 = load float *%ptr5
+  %val6 = load float *%ptr6
+  %val7 = load float *%ptr7
+  %val8 = load float *%ptr8
+  %val9 = load float *%ptr9
+
+  call void @foo()
+
+  store float %val0, float *%ptr0
+  store float %val1, float *%ptr1
+  store float %val2, float *%ptr2
+  store float %val3, float *%ptr3
+  store float %val4, float *%ptr4
+  store float %val5, float *%ptr5
+  store float %val6, float *%ptr6
+  store float %val7, float *%ptr7
+  store float %val8, float *%ptr8
+  store float %val9, float *%ptr9
+
+  ret void
+}
+
+; Similarly for f64.
+define void @f5(double *%ptr0) {
+; CHECK: f5:
+; CHECK: mvc 160(8,%r15), [[OFFSET:[0-9]+]]({{%r[0-9]+}})
+; CHECK: brasl %r14, foo@PLT
+; CHECK: mvc [[OFFSET]](8,{{%r[0-9]+}}), 160(%r15)
+; CHECK: br %r14
+  %ptr1 = getelementptr double *%ptr0, i64 2
+  %ptr2 = getelementptr double *%ptr0, i64 4
+  %ptr3 = getelementptr double *%ptr0, i64 6
+  %ptr4 = getelementptr double *%ptr0, i64 8
+  %ptr5 = getelementptr double *%ptr0, i64 10
+  %ptr6 = getelementptr double *%ptr0, i64 12
+  %ptr7 = getelementptr double *%ptr0, i64 14
+  %ptr8 = getelementptr double *%ptr0, i64 16
+  %ptr9 = getelementptr double *%ptr0, i64 18
+
+  %val0 = load double *%ptr0
+  %val1 = load double *%ptr1
+  %val2 = load double *%ptr2
+  %val3 = load double *%ptr3
+  %val4 = load double *%ptr4
+  %val5 = load double *%ptr5
+  %val6 = load double *%ptr6
+  %val7 = load double *%ptr7
+  %val8 = load double *%ptr8
+  %val9 = load double *%ptr9
+
+  call void @foo()
+
+  store double %val0, double *%ptr0
+  store double %val1, double *%ptr1
+  store double %val2, double *%ptr2
+  store double %val3, double *%ptr3
+  store double %val4, double *%ptr4
+  store double %val5, double *%ptr5
+  store double %val6, double *%ptr6
+  store double %val7, double *%ptr7
+  store double %val8, double *%ptr8
+  store double %val9, double *%ptr9
+
+  ret void
+}
+
+; Repeat f2 with atomic accesses.  We shouldn't use MVC here.
+define void @f6(i32 *%ptr0) {
+; CHECK: f6:
+; CHECK-NOT: mvc
+; CHECK: br %r14
+  %ptr1 = getelementptr i32 *%ptr0, i64 2
+  %ptr2 = getelementptr i32 *%ptr0, i64 4
+  %ptr3 = getelementptr i32 *%ptr0, i64 6
+  %ptr4 = getelementptr i32 *%ptr0, i64 8
+  %ptr5 = getelementptr i32 *%ptr0, i64 10
+  %ptr6 = getelementptr i32 *%ptr0, i64 12
+  %ptr7 = getelementptr i32 *%ptr0, i64 14
+  %ptr8 = getelementptr i32 *%ptr0, i64 16
+
+  %val0 = load atomic i32 *%ptr0 unordered, align 4
+  %val1 = load atomic i32 *%ptr1 unordered, align 4
+  %val2 = load atomic i32 *%ptr2 unordered, align 4
+  %val3 = load atomic i32 *%ptr3 unordered, align 4
+  %val4 = load atomic i32 *%ptr4 unordered, align 4
+  %val5 = load atomic i32 *%ptr5 unordered, align 4
+  %val6 = load atomic i32 *%ptr6 unordered, align 4
+  %val7 = load atomic i32 *%ptr7 unordered, align 4
+  %val8 = load atomic i32 *%ptr8 unordered, align 4
+
+  call void @foo()
+
+  store atomic i32 %val0, i32 *%ptr0 unordered, align 4
+  store atomic i32 %val1, i32 *%ptr1 unordered, align 4
+  store atomic i32 %val2, i32 *%ptr2 unordered, align 4
+  store atomic i32 %val3, i32 *%ptr3 unordered, align 4
+  store atomic i32 %val4, i32 *%ptr4 unordered, align 4
+  store atomic i32 %val5, i32 *%ptr5 unordered, align 4
+  store atomic i32 %val6, i32 *%ptr6 unordered, align 4
+  store atomic i32 %val7, i32 *%ptr7 unordered, align 4
+  store atomic i32 %val8, i32 *%ptr8 unordered, align 4
+
+  ret void
+}
+
+; ...likewise volatile accesses.
+define void @f7(i32 *%ptr0) {
+; CHECK: f7:
+; CHECK-NOT: mvc
+; CHECK: br %r14
+  %ptr1 = getelementptr i32 *%ptr0, i64 2
+  %ptr2 = getelementptr i32 *%ptr0, i64 4
+  %ptr3 = getelementptr i32 *%ptr0, i64 6
+  %ptr4 = getelementptr i32 *%ptr0, i64 8
+  %ptr5 = getelementptr i32 *%ptr0, i64 10
+  %ptr6 = getelementptr i32 *%ptr0, i64 12
+  %ptr7 = getelementptr i32 *%ptr0, i64 14
+  %ptr8 = getelementptr i32 *%ptr0, i64 16
+
+  %val0 = load volatile i32 *%ptr0
+  %val1 = load volatile i32 *%ptr1
+  %val2 = load volatile i32 *%ptr2
+  %val3 = load volatile i32 *%ptr3
+  %val4 = load volatile i32 *%ptr4
+  %val5 = load volatile i32 *%ptr5
+  %val6 = load volatile i32 *%ptr6
+  %val7 = load volatile i32 *%ptr7
+  %val8 = load volatile i32 *%ptr8
+
+  call void @foo()
+
+  store volatile i32 %val0, i32 *%ptr0
+  store volatile i32 %val1, i32 *%ptr1
+  store volatile i32 %val2, i32 *%ptr2
+  store volatile i32 %val3, i32 *%ptr3
+  store volatile i32 %val4, i32 *%ptr4
+  store volatile i32 %val5, i32 *%ptr5
+  store volatile i32 %val6, i32 *%ptr6
+  store volatile i32 %val7, i32 *%ptr7
+  store volatile i32 %val8, i32 *%ptr8
+
+  ret void
+}
+
+; Check that LRL and STRL are not converted.
+define void @f8() {
+; CHECK: f8:
+; CHECK-NOT: mvc
+; CHECK: br %r14
+  %val0 = load i32 *@g0
+  %val1 = load i32 *@g1
+  %val2 = load i32 *@g2
+  %val3 = load i32 *@g3
+  %val4 = load i32 *@g4
+  %val5 = load i32 *@g5
+  %val6 = load i32 *@g6
+  %val7 = load i32 *@g7
+  %val8 = load i32 *@g8
+  %val9 = load i32 *@g9
+
+  call void @foo()
+
+  store i32 %val0, i32 *@g0
+  store i32 %val1, i32 *@g1
+  store i32 %val2, i32 *@g2
+  store i32 %val3, i32 *@g3
+  store i32 %val4, i32 *@g4
+  store i32 %val5, i32 *@g5
+  store i32 %val6, i32 *@g6
+  store i32 %val7, i32 *@g7
+  store i32 %val8, i32 *@g8
+  store i32 %val9, i32 *@g9
+
+  ret void
+}
+
+; Likewise LGRL and STGRL.
+define void @f9() {
+; CHECK: f9:
+; CHECK-NOT: mvc
+; CHECK: br %r14
+  %val0 = load i64 *@h0
+  %val1 = load i64 *@h1
+  %val2 = load i64 *@h2
+  %val3 = load i64 *@h3
+  %val4 = load i64 *@h4
+  %val5 = load i64 *@h5
+  %val6 = load i64 *@h6
+  %val7 = load i64 *@h7
+  %val8 = load i64 *@h8
+  %val9 = load i64 *@h9
+
+  call void @foo()
+
+  store i64 %val0, i64 *@h0
+  store i64 %val1, i64 *@h1
+  store i64 %val2, i64 *@h2
+  store i64 %val3, i64 *@h3
+  store i64 %val4, i64 *@h4
+  store i64 %val5, i64 *@h5
+  store i64 %val6, i64 *@h6
+  store i64 %val7, i64 *@h7
+  store i64 %val8, i64 *@h8
+  store i64 %val9, i64 *@h9
+
+  ret void
+}