1// RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx1100 < %s | FileCheck --check-prefix=ASM %s 2// RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx1100 -filetype=obj < %s > %t 3// RUN: llvm-objdump -s -j .rodata %t | FileCheck --check-prefix=OBJDUMP %s 4 5// When going from asm -> asm, the expressions should remain the same (i.e., symbolic). 6// When going from asm -> obj, the expressions should get resolved (through fixups), 7 8// OBJDUMP: Contents of section .rodata 9// expr_defined_later 10// OBJDUMP-NEXT: 0000 2b000000 2c000000 00000000 00000000 11// OBJDUMP-NEXT: 0010 00000000 00000000 00000000 00000000 12// OBJDUMP-NEXT: 0020 00000000 00000000 00000000 00000000 13// OBJDUMP-NEXT: 0030 05f0afe4 811f007f 000c0000 00000000 14// expr_defined 15// OBJDUMP-NEXT: 0040 2a000000 2b000000 00000000 00000000 16// OBJDUMP-NEXT: 0050 00000000 00000000 00000000 00000000 17// OBJDUMP-NEXT: 0060 00000000 00000000 00000000 00000000 18// OBJDUMP-NEXT: 0070 05f0afe4 811f007f 000c0000 00000000 19 20.text 21// ASM: .text 22 23.amdhsa_code_object_version 4 24// ASM: .amdhsa_code_object_version 4 25 26.p2align 8 27.type expr_defined_later,@function 28expr_defined_later: 29 s_endpgm 30 31.p2align 8 32.type expr_defined,@function 33expr_defined: 34 s_endpgm 35 36.rodata 37// ASM: .rodata 38 39.p2align 6 40.amdhsa_kernel expr_defined_later 41 .amdhsa_group_segment_fixed_size defined_value+2 42 .amdhsa_private_segment_fixed_size defined_value+3 43 .amdhsa_system_vgpr_workitem_id defined_2_bits 44 .amdhsa_float_round_mode_32 defined_2_bits 45 .amdhsa_float_round_mode_16_64 defined_2_bits 46 .amdhsa_float_denorm_mode_32 defined_2_bits 47 .amdhsa_float_denorm_mode_16_64 defined_2_bits 48 .amdhsa_system_sgpr_workgroup_id_x defined_boolean 49 .amdhsa_system_sgpr_workgroup_id_y defined_boolean 50 .amdhsa_system_sgpr_workgroup_id_z defined_boolean 51 .amdhsa_system_sgpr_workgroup_info defined_boolean 52 .amdhsa_fp16_overflow defined_boolean 53 .amdhsa_workgroup_processor_mode defined_boolean 54 .amdhsa_memory_ordered defined_boolean 55 .amdhsa_forward_progress defined_boolean 56 .amdhsa_exception_fp_ieee_invalid_op defined_boolean 57 .amdhsa_exception_fp_denorm_src defined_boolean 58 .amdhsa_exception_fp_ieee_div_zero defined_boolean 59 .amdhsa_exception_fp_ieee_overflow defined_boolean 60 .amdhsa_exception_fp_ieee_underflow defined_boolean 61 .amdhsa_exception_fp_ieee_inexact defined_boolean 62 .amdhsa_exception_int_div_zero defined_boolean 63 .amdhsa_enable_private_segment defined_boolean 64 .amdhsa_uses_dynamic_stack defined_boolean 65 .amdhsa_next_free_vgpr defined_value+4 66 .amdhsa_next_free_sgpr defined_value+5 67 .amdhsa_reserve_vcc defined_boolean 68.end_amdhsa_kernel 69 70.set defined_value, 41 71.set defined_2_bits, 3 72.set defined_boolean, 1 73 74.p2align 6 75.amdhsa_kernel expr_defined 76 .amdhsa_group_segment_fixed_size defined_value+1 77 .amdhsa_private_segment_fixed_size defined_value+2 78 .amdhsa_system_vgpr_workitem_id defined_2_bits 79 .amdhsa_float_round_mode_32 defined_2_bits 80 .amdhsa_float_round_mode_16_64 defined_2_bits 81 .amdhsa_float_denorm_mode_32 defined_2_bits 82 .amdhsa_float_denorm_mode_16_64 defined_2_bits 83 .amdhsa_system_sgpr_workgroup_id_x defined_boolean 84 .amdhsa_system_sgpr_workgroup_id_y defined_boolean 85 .amdhsa_system_sgpr_workgroup_id_z defined_boolean 86 .amdhsa_system_sgpr_workgroup_info defined_boolean 87 .amdhsa_fp16_overflow defined_boolean 88 .amdhsa_workgroup_processor_mode defined_boolean 89 .amdhsa_memory_ordered defined_boolean 90 .amdhsa_forward_progress defined_boolean 91 .amdhsa_exception_fp_ieee_invalid_op defined_boolean 92 .amdhsa_exception_fp_denorm_src defined_boolean 93 .amdhsa_exception_fp_ieee_div_zero defined_boolean 94 .amdhsa_exception_fp_ieee_overflow defined_boolean 95 .amdhsa_exception_fp_ieee_underflow defined_boolean 96 .amdhsa_exception_fp_ieee_inexact defined_boolean 97 .amdhsa_exception_int_div_zero defined_boolean 98 .amdhsa_enable_private_segment defined_boolean 99 .amdhsa_uses_dynamic_stack defined_boolean 100 .amdhsa_next_free_vgpr defined_value+3 101 .amdhsa_next_free_sgpr defined_value+4 102 .amdhsa_reserve_vcc defined_boolean 103.end_amdhsa_kernel 104 105// ASM: .amdhsa_kernel expr_defined_later 106// ASM-NEXT: .amdhsa_group_segment_fixed_size defined_value+2 107// ASM-NEXT: .amdhsa_private_segment_fixed_size defined_value+3 108// ASM-NEXT: .amdhsa_kernarg_size 0 109// ASM-NEXT: .amdhsa_user_sgpr_count 0 110// ASM-NEXT: .amdhsa_user_sgpr_dispatch_ptr 0 111// ASM-NEXT: .amdhsa_user_sgpr_queue_ptr 0 112// ASM-NEXT: .amdhsa_user_sgpr_kernarg_segment_ptr 0 113// ASM-NEXT: .amdhsa_user_sgpr_dispatch_id 0 114// ASM-NEXT: .amdhsa_user_sgpr_private_segment_size 0 115// ASM-NEXT: .amdhsa_wavefront_size32 1 116// ASM-NEXT: .amdhsa_enable_private_segment ((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&1 117// ASM-NEXT: .amdhsa_system_sgpr_workgroup_id_x (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&128)>>7 118// ASM-NEXT: .amdhsa_system_sgpr_workgroup_id_y (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&256)>>8 119// ASM-NEXT: .amdhsa_system_sgpr_workgroup_id_z (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&512)>>9 120// ASM-NEXT: .amdhsa_system_sgpr_workgroup_info (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&1024)>>10 121// ASM-NEXT: .amdhsa_system_vgpr_workitem_id (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&6144)>>11 122// ASM-NEXT: .amdhsa_next_free_vgpr defined_value+4 123// ASM-NEXT: .amdhsa_next_free_sgpr defined_value+5 124// ASM-NEXT: .amdhsa_reserve_vcc defined_boolean 125// ASM-NEXT: .amdhsa_float_round_mode_32 (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&12288)>>12 126// ASM-NEXT: .amdhsa_float_round_mode_16_64 (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&49152)>>14 127// ASM-NEXT: .amdhsa_float_denorm_mode_32 (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&196608)>>16 128// ASM-NEXT: .amdhsa_float_denorm_mode_16_64 (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&786432)>>18 129// ASM-NEXT: .amdhsa_dx10_clamp (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&2097152)>>21 130// ASM-NEXT: .amdhsa_ieee_mode (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&8388608)>>23 131// ASM-NEXT: .amdhsa_fp16_overflow (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&67108864)>>26 132// ASM-NEXT: .amdhsa_workgroup_processor_mode (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&536870912)>>29 133// ASM-NEXT: .amdhsa_memory_ordered (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&1073741824)>>30 134// ASM-NEXT: .amdhsa_forward_progress (((((((((((((((((((1621884928|(defined_2_bits<<12))&(~49152))|(defined_2_bits<<14))&(~196608))|(defined_2_bits<<16))&(~786432))|(defined_2_bits<<18))&(~67108864))|(defined_boolean<<26))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~2147483648))|(defined_boolean<<31))&(~63))|(((alignto(max(defined_value+4, 1), 8))/8)-1))&(~960))&2147483648)>>31 135// ASM-NEXT: .amdhsa_shared_vgpr_count 0 136// ASM-NEXT: .amdhsa_exception_fp_ieee_invalid_op (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&16777216)>>24 137// ASM-NEXT: .amdhsa_exception_fp_denorm_src (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&33554432)>>25 138// ASM-NEXT: .amdhsa_exception_fp_ieee_div_zero (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&67108864)>>26 139// ASM-NEXT: .amdhsa_exception_fp_ieee_overflow (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&134217728)>>27 140// ASM-NEXT: .amdhsa_exception_fp_ieee_underflow (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&268435456)>>28 141// ASM-NEXT: .amdhsa_exception_fp_ieee_inexact (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&536870912)>>29 142// ASM-NEXT: .amdhsa_exception_int_div_zero (((((((((((((((((((((((((((128|(defined_2_bits<<11))&(~128))|(defined_boolean<<7))&(~256))|(defined_boolean<<8))&(~512))|(defined_boolean<<9))&(~1024))|(defined_boolean<<10))&(~16777216))|(defined_boolean<<24))&(~33554432))|(defined_boolean<<25))&(~67108864))|(defined_boolean<<26))&(~134217728))|(defined_boolean<<27))&(~268435456))|(defined_boolean<<28))&(~536870912))|(defined_boolean<<29))&(~1073741824))|(defined_boolean<<30))&(~1))|defined_boolean)&(~62))&1073741824)>>30 143// ASM-NEXT: .end_amdhsa_kernel 144 145// ASM: .set defined_value, 41 146// ASM-NEXT: .no_dead_strip defined_value 147// ASM-NEXT: .set defined_2_bits, 3 148// ASM-NEXT: .no_dead_strip defined_2_bits 149// ASM-NEXT: .set defined_boolean, 1 150// ASM-NEXT: .no_dead_strip defined_boolean 151 152// ASM: .amdhsa_kernel expr_defined 153// ASM-NEXT: .amdhsa_group_segment_fixed_size 42 154// ASM-NEXT: .amdhsa_private_segment_fixed_size 43 155// ASM-NEXT: .amdhsa_kernarg_size 0 156// ASM-NEXT: .amdhsa_user_sgpr_count 0 157// ASM-NEXT: .amdhsa_user_sgpr_dispatch_ptr 0 158// ASM-NEXT: .amdhsa_user_sgpr_queue_ptr 0 159// ASM-NEXT: .amdhsa_user_sgpr_kernarg_segment_ptr 0 160// ASM-NEXT: .amdhsa_user_sgpr_dispatch_id 0 161// ASM-NEXT: .amdhsa_user_sgpr_private_segment_size 0 162// ASM-NEXT: .amdhsa_wavefront_size32 1 163// ASM-NEXT: .amdhsa_enable_private_segment 1 164// ASM-NEXT: .amdhsa_system_sgpr_workgroup_id_x 1 165// ASM-NEXT: .amdhsa_system_sgpr_workgroup_id_y 1 166// ASM-NEXT: .amdhsa_system_sgpr_workgroup_id_z 1 167// ASM-NEXT: .amdhsa_system_sgpr_workgroup_info 1 168// ASM-NEXT: .amdhsa_system_vgpr_workitem_id 3 169// ASM-NEXT: .amdhsa_next_free_vgpr 44 170// ASM-NEXT: .amdhsa_next_free_sgpr 45 171// ASM-NEXT: .amdhsa_reserve_vcc 1 172// ASM-NEXT: .amdhsa_float_round_mode_32 3 173// ASM-NEXT: .amdhsa_float_round_mode_16_64 3 174// ASM-NEXT: .amdhsa_float_denorm_mode_32 3 175// ASM-NEXT: .amdhsa_float_denorm_mode_16_64 3 176// ASM-NEXT: .amdhsa_dx10_clamp 1 177// ASM-NEXT: .amdhsa_ieee_mode 1 178// ASM-NEXT: .amdhsa_fp16_overflow 1 179// ASM-NEXT: .amdhsa_workgroup_processor_mode 1 180// ASM-NEXT: .amdhsa_memory_ordered 1 181// ASM-NEXT: .amdhsa_forward_progress 1 182// ASM-NEXT: .amdhsa_shared_vgpr_count 0 183// ASM-NEXT: .amdhsa_exception_fp_ieee_invalid_op 1 184// ASM-NEXT: .amdhsa_exception_fp_denorm_src 1 185// ASM-NEXT: .amdhsa_exception_fp_ieee_div_zero 1 186// ASM-NEXT: .amdhsa_exception_fp_ieee_overflow 1 187// ASM-NEXT: .amdhsa_exception_fp_ieee_underflow 1 188// ASM-NEXT: .amdhsa_exception_fp_ieee_inexact 1 189// ASM-NEXT: .amdhsa_exception_int_div_zero 1 190// ASM-NEXT: .end_amdhsa_kernel 191