|
221
|
1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
|
2 ; RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -amdgpu-spill-sgpr-to-vgpr=true < %s | FileCheck -check-prefix=SPILL-TO-VGPR %s
|
|
|
3 ; RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -amdgpu-spill-sgpr-to-vgpr=false < %s | FileCheck -check-prefix=NO-SPILL-TO-VGPR %s
|
|
|
4
|
|
|
5 ; Check frame setup where SGPR spills to VGPRs are disabled or enabled.
|
|
|
6
|
|
|
7 declare hidden void @external_void_func_void() #0
|
|
|
8
|
|
|
9 define void @callee_with_stack_and_call() #0 {
|
|
|
10 ; SPILL-TO-VGPR-LABEL: callee_with_stack_and_call:
|
|
|
11 ; SPILL-TO-VGPR: ; %bb.0:
|
|
|
12 ; SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
|
|
|
13 ; SPILL-TO-VGPR-NEXT: s_or_saveexec_b64 s[4:5], -1
|
|
|
14 ; SPILL-TO-VGPR-NEXT: buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
|
|
|
15 ; SPILL-TO-VGPR-NEXT: s_mov_b64 exec, s[4:5]
|
|
|
16 ; SPILL-TO-VGPR-NEXT: v_writelane_b32 v40, s33, 2
|
|
|
17 ; SPILL-TO-VGPR-NEXT: v_writelane_b32 v40, s30, 0
|
|
|
18 ; SPILL-TO-VGPR-NEXT: s_mov_b32 s33, s32
|
|
223
|
19 ; SPILL-TO-VGPR-NEXT: s_addk_i32 s32, 0x400
|
|
221
|
20 ; SPILL-TO-VGPR-NEXT: v_mov_b32_e32 v0, 0
|
|
|
21 ; SPILL-TO-VGPR-NEXT: s_getpc_b64 s[4:5]
|
|
|
22 ; SPILL-TO-VGPR-NEXT: s_add_u32 s4, s4, external_void_func_void@rel32@lo+4
|
|
|
23 ; SPILL-TO-VGPR-NEXT: s_addc_u32 s5, s5, external_void_func_void@rel32@hi+12
|
|
|
24 ; SPILL-TO-VGPR-NEXT: v_writelane_b32 v40, s31, 1
|
|
|
25 ; SPILL-TO-VGPR-NEXT: buffer_store_dword v0, off, s[0:3], s33
|
|
|
26 ; SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
27 ; SPILL-TO-VGPR-NEXT: s_swappc_b64 s[30:31], s[4:5]
|
|
|
28 ; SPILL-TO-VGPR-NEXT: v_readlane_b32 s4, v40, 0
|
|
|
29 ; SPILL-TO-VGPR-NEXT: v_readlane_b32 s5, v40, 1
|
|
223
|
30 ; SPILL-TO-VGPR-NEXT: s_addk_i32 s32, 0xfc00
|
|
221
|
31 ; SPILL-TO-VGPR-NEXT: v_readlane_b32 s33, v40, 2
|
|
|
32 ; SPILL-TO-VGPR-NEXT: s_or_saveexec_b64 s[6:7], -1
|
|
|
33 ; SPILL-TO-VGPR-NEXT: buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
|
|
|
34 ; SPILL-TO-VGPR-NEXT: s_mov_b64 exec, s[6:7]
|
|
|
35 ; SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
36 ; SPILL-TO-VGPR-NEXT: s_setpc_b64 s[4:5]
|
|
|
37 ;
|
|
|
38 ; NO-SPILL-TO-VGPR-LABEL: callee_with_stack_and_call:
|
|
|
39 ; NO-SPILL-TO-VGPR: ; %bb.0:
|
|
|
40 ; NO-SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
|
|
|
41 ; NO-SPILL-TO-VGPR-NEXT: v_mov_b32_e32 v0, s33
|
|
|
42 ; NO-SPILL-TO-VGPR-NEXT: buffer_store_dword v0, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
|
|
|
43 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b32 s33, s32
|
|
223
|
44 ; NO-SPILL-TO-VGPR-NEXT: s_addk_i32 s32, 0x800
|
|
221
|
45 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b64 s[6:7], exec
|
|
|
46 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b64 exec, 3
|
|
|
47 ; NO-SPILL-TO-VGPR-NEXT: buffer_store_dword v1, off, s[0:3], s33 offset:16
|
|
|
48 ; NO-SPILL-TO-VGPR-NEXT: v_writelane_b32 v1, s30, 0
|
|
|
49 ; NO-SPILL-TO-VGPR-NEXT: v_writelane_b32 v1, s31, 1
|
|
|
50 ; NO-SPILL-TO-VGPR-NEXT: buffer_store_dword v1, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
|
|
|
51 ; NO-SPILL-TO-VGPR-NEXT: buffer_load_dword v1, off, s[0:3], s33 offset:16
|
|
|
52 ; NO-SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
53 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b64 exec, s[6:7]
|
|
|
54 ; NO-SPILL-TO-VGPR-NEXT: v_mov_b32_e32 v0, 0
|
|
|
55 ; NO-SPILL-TO-VGPR-NEXT: s_getpc_b64 s[4:5]
|
|
|
56 ; NO-SPILL-TO-VGPR-NEXT: s_add_u32 s4, s4, external_void_func_void@rel32@lo+4
|
|
|
57 ; NO-SPILL-TO-VGPR-NEXT: s_addc_u32 s5, s5, external_void_func_void@rel32@hi+12
|
|
|
58 ; NO-SPILL-TO-VGPR-NEXT: buffer_store_dword v0, off, s[0:3], s33
|
|
|
59 ; NO-SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
60 ; NO-SPILL-TO-VGPR-NEXT: s_swappc_b64 s[30:31], s[4:5]
|
|
|
61 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b64 s[6:7], exec
|
|
|
62 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b64 exec, 3
|
|
|
63 ; NO-SPILL-TO-VGPR-NEXT: buffer_store_dword v1, off, s[0:3], s33 offset:16
|
|
|
64 ; NO-SPILL-TO-VGPR-NEXT: buffer_load_dword v1, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
|
|
|
65 ; NO-SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
66 ; NO-SPILL-TO-VGPR-NEXT: v_readlane_b32 s4, v1, 0
|
|
|
67 ; NO-SPILL-TO-VGPR-NEXT: v_readlane_b32 s5, v1, 1
|
|
|
68 ; NO-SPILL-TO-VGPR-NEXT: buffer_load_dword v1, off, s[0:3], s33 offset:16
|
|
|
69 ; NO-SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
70 ; NO-SPILL-TO-VGPR-NEXT: s_mov_b64 exec, s[6:7]
|
|
223
|
71 ; NO-SPILL-TO-VGPR-NEXT: s_addk_i32 s32, 0xf800
|
|
221
|
72 ; NO-SPILL-TO-VGPR-NEXT: buffer_load_dword v0, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
|
|
|
73 ; NO-SPILL-TO-VGPR-NEXT: s_waitcnt vmcnt(0)
|
|
|
74 ; NO-SPILL-TO-VGPR-NEXT: v_readfirstlane_b32 s33, v0
|
|
|
75 ; NO-SPILL-TO-VGPR-NEXT: s_setpc_b64 s[4:5]
|
|
|
76 %alloca = alloca i32, addrspace(5)
|
|
|
77 store volatile i32 0, i32 addrspace(5)* %alloca
|
|
|
78 call void @external_void_func_void()
|
|
|
79 ret void
|
|
|
80 }
|
|
|
81
|
|
|
82 attributes #0 = { nounwind }
|
