6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-15 20:29:48 +00:00
Code Issues Projects Releases Wiki Activity
998ae69abe
llvm-6502/test/Transforms/InstCombine/constant-fold-address-space-pointer.ll
Chandler Carruth 9b2d091a9c [InstCombine] Do an about-face on how LLVM canonicalizes (cast (load 
...)) and (load (cast ...)): canonicalize toward the former.

Historically, we've tried to load using the type of the *pointer*, and
tried to match that type as closely as possible removing as many pointer
casts as we could and trading them for bitcasts of the loaded value.
This is deeply and fundamentally wrong.

Repeat after me: memory does not have a type! This was a hard lesson for
me to learn working on SROA.

There is only one thing that should actually drive the type used for
a pointer, and that is the type which we need to use to load from that
pointer. Matching up pointer types to the loaded value types is very
useful because it minimizes the physical size of the IR required for
no-op casts. Similarly, the only thing that should drive the type used
for a loaded value is *how that value is used*! Again, this minimizes
casts. And in fact, the *only* thing motivating types in any part of
LLVM's IR are the types used by the operations in the IR. We should
match them as closely as possible.

I've ended up removing some tests here as they were testing bugs or
behavior that is no longer present. Mostly though, this is just cleanup
to let the tests continue to function as intended.

The only fallout I've found so far from this change was SROA and I have
fixed it to not be impeded by the different type of load. If you find
more places where this change causes optimizations not to fire, those
too are likely bugs where we are assuming that the type of pointers is
"significant" for optimization purposes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220138 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-18 06:36:22 +00:00

242 lines
9.5 KiB
LLVM
Raw Blame History

; RUN: opt -S -instcombine %s -o - | FileCheck %s
target datalayout = "e-p:32:32:32-p1:64:64:64-p2:8:8:8-p3:16:16:16-p4:16:16:16-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:32"
@g = addrspace(3) global i32 89
@const_zero_i8_as1 = addrspace(1) constant i8 0
@const_zero_i32_as1 = addrspace(1) constant i32 0
@const_zero_i8_as2 = addrspace(2) constant i8 0
@const_zero_i32_as2 = addrspace(2) constant i32 0
@const_zero_i8_as3 = addrspace(3) constant i8 0
@const_zero_i32_as3 = addrspace(3) constant i32 0
; Test constant folding of inttoptr (ptrtoint constantexpr)
; The intermediate integer size is the same as the pointer size
define i32 addrspace(3)* @test_constant_fold_inttoptr_as_pointer_same_size() {
; CHECK-LABEL: @test_constant_fold_inttoptr_as_pointer_same_size(
; CHECK-NEXT: ret i32 addrspace(3)* @const_zero_i32_as3
%x = ptrtoint i32 addrspace(3)* @const_zero_i32_as3 to i32
%y = inttoptr i32 %x to i32 addrspace(3)*
ret i32 addrspace(3)* %y
}
; The intermediate integer size is larger than the pointer size
define i32 addrspace(2)* @test_constant_fold_inttoptr_as_pointer_smaller() {
; CHECK-LABEL: @test_constant_fold_inttoptr_as_pointer_smaller(
; CHECK-NEXT: ret i32 addrspace(2)* @const_zero_i32_as2
%x = ptrtoint i32 addrspace(2)* @const_zero_i32_as2 to i16
%y = inttoptr i16 %x to i32 addrspace(2)*
ret i32 addrspace(2)* %y
}
; Different address spaces that are the same size, but they are
; different so nothing should happen
define i32 addrspace(4)* @test_constant_fold_inttoptr_as_pointer_smaller_different_as() {
; CHECK-LABEL: @test_constant_fold_inttoptr_as_pointer_smaller_different_as(
; CHECK-NEXT: ret i32 addrspace(4)* inttoptr (i16 ptrtoint (i32 addrspace(3)* @const_zero_i32_as3 to i16) to i32 addrspace(4)*)
%x = ptrtoint i32 addrspace(3)* @const_zero_i32_as3 to i16
%y = inttoptr i16 %x to i32 addrspace(4)*
ret i32 addrspace(4)* %y
}
; Make sure we don't introduce a bitcast between different sized
; address spaces when folding this
define i32 addrspace(2)* @test_constant_fold_inttoptr_as_pointer_smaller_different_size_as() {
; CHECK-LABEL: @test_constant_fold_inttoptr_as_pointer_smaller_different_size_as(
; CHECK-NEXT: ret i32 addrspace(2)* inttoptr (i32 ptrtoint (i32 addrspace(3)* @const_zero_i32_as3 to i32) to i32 addrspace(2)*)
%x = ptrtoint i32 addrspace(3)* @const_zero_i32_as3 to i32
%y = inttoptr i32 %x to i32 addrspace(2)*
ret i32 addrspace(2)* %y
}
; The intermediate integer size is too small, nothing should happen
define i32 addrspace(3)* @test_constant_fold_inttoptr_as_pointer_larger() {
; CHECK-LABEL: @test_constant_fold_inttoptr_as_pointer_larger(
; CHECK-NEXT: ret i32 addrspace(3)* inttoptr (i8 ptrtoint (i32 addrspace(3)* @const_zero_i32_as3 to i8) to i32 addrspace(3)*)
%x = ptrtoint i32 addrspace(3)* @const_zero_i32_as3 to i8
%y = inttoptr i8 %x to i32 addrspace(3)*
ret i32 addrspace(3)* %y
}
define i8 @const_fold_ptrtoint() {
; CHECK-LABEL: @const_fold_ptrtoint(
; CHECK-NEXT: ret i8 4
ret i8 ptrtoint (i32 addrspace(2)* inttoptr (i4 4 to i32 addrspace(2)*) to i8)
}
; Test that mask happens when the destination pointer is smaller than
; the original
define i8 @const_fold_ptrtoint_mask() {
; CHECK-LABEL: @const_fold_ptrtoint_mask(
; CHECK-NEXT: ret i8 1
ret i8 ptrtoint (i32 addrspace(3)* inttoptr (i32 257 to i32 addrspace(3)*) to i8)
}
; Address space 0 is too small for the correct mask, should mask with
; 64-bits instead of 32
define i64 @const_fold_ptrtoint_mask_small_as0() {
; CHECK-LABEL: @const_fold_ptrtoint_mask_small_as0(
; CHECK: ret i64 -1
ret i64 ptrtoint (i32 addrspace(1)* inttoptr (i128 -1 to i32 addrspace(1)*) to i64)
}
define i32 addrspace(3)* @const_inttoptr() {
; CHECK-LABEL: @const_inttoptr(
; CHECK-NEXT: ret i32 addrspace(3)* inttoptr (i16 4 to i32 addrspace(3)*)
%p = inttoptr i16 4 to i32 addrspace(3)*
ret i32 addrspace(3)* %p
}
define i16 @const_ptrtoint() {
; CHECK-LABEL: @const_ptrtoint(
; CHECK-NEXT: ret i16 ptrtoint (i32 addrspace(3)* @g to i16)
%i = ptrtoint i32 addrspace(3)* @g to i16
ret i16 %i
}
define i16 @const_inttoptr_ptrtoint() {
; CHECK-LABEL: @const_inttoptr_ptrtoint(
; CHECK-NEXT: ret i16 9
ret i16 ptrtoint (i32 addrspace(3)* inttoptr (i16 9 to i32 addrspace(3)*) to i16)
}
define i1 @constant_fold_cmp_constantexpr_inttoptr() {
; CHECK-LABEL: @constant_fold_cmp_constantexpr_inttoptr(
; CHECK-NEXT: ret i1 true
%x = icmp eq i32 addrspace(3)* inttoptr (i16 0 to i32 addrspace(3)*), null
ret i1 %x
}
define i1 @constant_fold_inttoptr_null(i16 %i) {
; CHECK-LABEL: @constant_fold_inttoptr_null(
; CHECK-NEXT: ret i1 false
%x = icmp eq i32 addrspace(3)* inttoptr (i16 99 to i32 addrspace(3)*), inttoptr (i16 0 to i32 addrspace(3)*)
ret i1 %x
}
define i1 @constant_fold_ptrtoint_null() {
; CHECK-LABEL: @constant_fold_ptrtoint_null(
; CHECK-NEXT: ret i1 false
%x = icmp eq i16 ptrtoint (i32 addrspace(3)* @g to i16), ptrtoint (i32 addrspace(3)* null to i16)
ret i1 %x
}
define i1 @constant_fold_ptrtoint_null_2() {
; CHECK-LABEL: @constant_fold_ptrtoint_null_2(
; CHECK-NEXT: ret i1 false
%x = icmp eq i16 ptrtoint (i32 addrspace(3)* null to i16), ptrtoint (i32 addrspace(3)* @g to i16)
ret i1 %x
}
define i1 @constant_fold_ptrtoint() {
; CHECK-LABEL: @constant_fold_ptrtoint(
; CHECK-NEXT: ret i1 true
%x = icmp eq i16 ptrtoint (i32 addrspace(3)* @g to i16), ptrtoint (i32 addrspace(3)* @g to i16)
ret i1 %x
}
define i1 @constant_fold_inttoptr() {
; CHECK-LABEL: @constant_fold_inttoptr(
; CHECK-NEXT: ret i1 false
%x = icmp eq i32 addrspace(3)* inttoptr (i16 99 to i32 addrspace(3)*), inttoptr (i16 27 to i32 addrspace(3)*)
ret i1 %x
}
@g_float_as3 = addrspace(3) global float zeroinitializer
@g_v4f_as3 = addrspace(3) global <4 x float> zeroinitializer
define float @constant_fold_bitcast_ftoi_load() {
; CHECK-LABEL: @constant_fold_bitcast_ftoi_load(
; CHECK: load float addrspace(3)* bitcast (i32 addrspace(3)* @g to float addrspace(3)*), align 4
%a = load float addrspace(3)* bitcast (i32 addrspace(3)* @g to float addrspace(3)*), align 4
ret float %a
}
define i32 @constant_fold_bitcast_itof_load() {
; CHECK-LABEL: @constant_fold_bitcast_itof_load(
; CHECK: load i32 addrspace(3)* bitcast (float addrspace(3)* @g_float_as3 to i32 addrspace(3)*), align 4
%a = load i32 addrspace(3)* bitcast (float addrspace(3)* @g_float_as3 to i32 addrspace(3)*), align 4
ret i32 %a
}
define <4 x float> @constant_fold_bitcast_vector_as() {
; CHECK-LABEL: @constant_fold_bitcast_vector_as(
; CHECK: load <4 x float> addrspace(3)* @g_v4f_as3, align 16
%a = load <4 x float> addrspace(3)* bitcast (<4 x i32> addrspace(3)* bitcast (<4 x float> addrspace(3)* @g_v4f_as3 to <4 x i32> addrspace(3)*) to <4 x float> addrspace(3)*), align 4
ret <4 x float> %a
}
@i32_array_as3 = addrspace(3) global [10 x i32] zeroinitializer
define i32 @test_cast_gep_small_indices_as() {
; CHECK-LABEL: @test_cast_gep_small_indices_as(
; CHECK: load i32 addrspace(3)* getelementptr inbounds ([10 x i32] addrspace(3)* @i32_array_as3, i16 0, i16 0), align 16
%p = getelementptr [10 x i32] addrspace(3)* @i32_array_as3, i7 0, i7 0
%x = load i32 addrspace(3)* %p, align 4
ret i32 %x
}
%struct.foo = type { float, float, [4 x i32], i32 addrspace(3)* }
@constant_fold_global_ptr = addrspace(3) global %struct.foo {
float 0.0,
float 0.0,
[4 x i32] zeroinitializer,
i32 addrspace(3)* getelementptr ([10 x i32] addrspace(3)* @i32_array_as3, i64 0, i64 0)
}
define i32 @test_cast_gep_large_indices_as() {
; CHECK-LABEL: @test_cast_gep_large_indices_as(
; CHECK: load i32 addrspace(3)* getelementptr inbounds ([10 x i32] addrspace(3)* @i32_array_as3, i16 0, i16 0), align 16
%p = getelementptr [10 x i32] addrspace(3)* @i32_array_as3, i64 0, i64 0
%x = load i32 addrspace(3)* %p, align 4
ret i32 %x
}
define i32 @test_constant_cast_gep_struct_indices_as() {
; CHECK-LABEL: @test_constant_cast_gep_struct_indices_as(
; CHECK: load i32 addrspace(3)* getelementptr inbounds (%struct.foo addrspace(3)* @constant_fold_global_ptr, i16 0, i32 2, i16 2), align 8
%x = getelementptr %struct.foo addrspace(3)* @constant_fold_global_ptr, i18 0, i32 2, i12 2
%y = load i32 addrspace(3)* %x, align 4
ret i32 %y
}
@constant_data_as3 = addrspace(3) constant [5 x i32] [i32 1, i32 2, i32 3, i32 4, i32 5]
define i32 @test_read_data_from_global_as3() {
; CHECK-LABEL: @test_read_data_from_global_as3(
; CHECK-NEXT: ret i32 2
%x = getelementptr [5 x i32] addrspace(3)* @constant_data_as3, i32 0, i32 1
%y = load i32 addrspace(3)* %x, align 4
ret i32 %y
}
@a = addrspace(1) constant i32 9
@b = addrspace(1) constant i32 23
@c = addrspace(1) constant i32 34
@d = addrspace(1) constant i32 99
@ptr_array = addrspace(2) constant [4 x i32 addrspace(1)*] [ i32 addrspace(1)* @a, i32 addrspace(1)* @b, i32 addrspace(1)* @c, i32 addrspace(1)* @d]
@indirect = addrspace(0) constant i32 addrspace(1)* addrspace(2)* getelementptr inbounds ([4 x i32 addrspace(1)*] addrspace(2)* @ptr_array, i1 0, i32 2)
define i32 @constant_through_array_as_ptrs() {
; CHECK-LABEL: @constant_through_array_as_ptrs(
; CHECK-NEXT: ret i32 34
%p = load i32 addrspace(1)* addrspace(2)* addrspace(0)* @indirect, align 4
%a = load i32 addrspace(1)* addrspace(2)* %p, align 4
%b = load i32 addrspace(1)* %a, align 4
ret i32 %b
}
@shared_mem = external addrspace(3) global [0 x i8]
define float @canonicalize_addrspacecast(i32 %i) {
; CHECK-LABEL: @canonicalize_addrspacecast
; CHECK-NEXT: getelementptr inbounds float* addrspacecast (float addrspace(3)* bitcast ([0 x i8] addrspace(3)* @shared_mem to float addrspace(3)*) to float*), i32 %i
%p = getelementptr inbounds float* addrspacecast ([0 x i8] addrspace(3)* @shared_mem to float*), i32 %i
%v = load float* %p
ret float %v
}
Reference in New Issue View Git Blame Copy Permalink