1 ; RUN: llc < %s -mtriple=thumbv7-apple-darwin9 -mcpu=cortex-a8 -relocation-model=pic -disable-fp-elim
3 %struct.FILE = type { i8*, i32, i32, i16, i16, %struct.__sbuf, i32, i8*, i32 (i8*)*, i32 (i8*, i8*, i32)*, i64 (i8*, i64, i32)*, i32 (i8*, i8*, i32)*, %struct.__sbuf, %struct.__sFILEX*, i32, [3 x i8], [1 x i8], %struct.__sbuf, i32, i64 }
4 %struct.JHUFF_TBL = type { [17 x i8], [256 x i8], i32 }
5 %struct.JQUANT_TBL = type { [64 x i16], i32 }
6 %struct.__sFILEX = type opaque
7 %struct.__sbuf = type { i8*, i32 }
8 %struct.anon = type { [8 x i32], [48 x i8] }
9 %struct.backing_store_info = type { void (%struct.jpeg_common_struct*, %struct.backing_store_info*, i8*, i32, i32)*, void (%struct.jpeg_common_struct*, %struct.backing_store_info*, i8*, i32, i32)*, void (%struct.jpeg_common_struct*, %struct.backing_store_info*)*, %struct.FILE*, [64 x i8] }
10 %struct.jpeg_color_deconverter = type { void (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*, i8***, i32, i8**, i32)* }
11 %struct.jpeg_color_quantizer = type { void (%struct.jpeg_decompress_struct*, i32)*, void (%struct.jpeg_decompress_struct*, i8**, i8**, i32)*, void (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*)* }
12 %struct.jpeg_common_struct = type { %struct.jpeg_error_mgr*, %struct.jpeg_memory_mgr*, %struct.jpeg_progress_mgr*, i32, i32 }
13 %struct.jpeg_component_info = type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %struct.JQUANT_TBL*, i8* }
14 %struct.jpeg_d_coef_controller = type { void (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*, i8***)*, %struct.jvirt_barray_control** }
15 %struct.jpeg_d_main_controller = type { void (%struct.jpeg_decompress_struct*, i32)*, void (%struct.jpeg_decompress_struct*, i8**, i32*, i32)* }
16 %struct.jpeg_d_post_controller = type { void (%struct.jpeg_decompress_struct*, i32)*, void (%struct.jpeg_decompress_struct*, i8***, i32*, i32, i8**, i32*, i32)* }
17 %struct.jpeg_decomp_master = type { void (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*)*, i32 }
18 %struct.jpeg_decompress_struct = type { %struct.jpeg_error_mgr*, %struct.jpeg_memory_mgr*, %struct.jpeg_progress_mgr*, i32, i32, %struct.jpeg_source_mgr*, i32, i32, i32, i32, i32, i32, i32, double, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8**, i32, i32, i32, i32, i32, [64 x i32]*, [4 x %struct.JQUANT_TBL*], [4 x %struct.JHUFF_TBL*], [4 x %struct.JHUFF_TBL*], i32, %struct.jpeg_component_info*, i32, i32, [16 x i8], [16 x i8], [16 x i8], i32, i32, i8, i16, i16, i32, i8, i32, i32, i32, i32, i32, i8*, i32, [4 x %struct.jpeg_component_info*], i32, i32, i32, [10 x i32], i32, i32, i32, i32, i32, %struct.jpeg_decomp_master*, %struct.jpeg_d_main_controller*, %struct.jpeg_d_coef_controller*, %struct.jpeg_d_post_controller*, %struct.jpeg_input_controller*, %struct.jpeg_marker_reader*, %struct.jpeg_entropy_decoder*, %struct.jpeg_inverse_dct*, %struct.jpeg_upsampler*, %struct.jpeg_color_deconverter*, %struct.jpeg_color_quantizer* }
19 %struct.jpeg_entropy_decoder = type { void (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*, [64 x i16]**)* }
20 %struct.jpeg_error_mgr = type { void (%struct.jpeg_common_struct*)*, void (%struct.jpeg_common_struct*, i32)*, void (%struct.jpeg_common_struct*)*, void (%struct.jpeg_common_struct*, i8*)*, void (%struct.jpeg_common_struct*)*, i32, %struct.anon, i32, i32, i8**, i32, i8**, i32, i32 }
21 %struct.jpeg_input_controller = type { i32 (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*)*, i32, i32 }
22 %struct.jpeg_inverse_dct = type { void (%struct.jpeg_decompress_struct*)*, [10 x void (%struct.jpeg_decompress_struct*, %struct.jpeg_component_info*, i16*, i8**, i32)*] }
23 %struct.jpeg_marker_reader = type { void (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*)*, [16 x i32 (%struct.jpeg_decompress_struct*)*], i32, i32, i32, i32 }
24 %struct.jpeg_memory_mgr = type { i8* (%struct.jpeg_common_struct*, i32, i32)*, i8* (%struct.jpeg_common_struct*, i32, i32)*, i8** (%struct.jpeg_common_struct*, i32, i32, i32)*, [64 x i16]** (%struct.jpeg_common_struct*, i32, i32, i32)*, %struct.jvirt_sarray_control* (%struct.jpeg_common_struct*, i32, i32, i32, i32, i32)*, %struct.jvirt_barray_control* (%struct.jpeg_common_struct*, i32, i32, i32, i32, i32)*, void (%struct.jpeg_common_struct*)*, i8** (%struct.jpeg_common_struct*, %struct.jvirt_sarray_control*, i32, i32, i32)*, [64 x i16]** (%struct.jpeg_common_struct*, %struct.jvirt_barray_control*, i32, i32, i32)*, void (%struct.jpeg_common_struct*, i32)*, void (%struct.jpeg_common_struct*)*, i32 }
25 %struct.jpeg_progress_mgr = type { void (%struct.jpeg_common_struct*)*, i32, i32, i32, i32 }
26 %struct.jpeg_source_mgr = type { i8*, i32, void (%struct.jpeg_decompress_struct*)*, i32 (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*, i32)*, i32 (%struct.jpeg_decompress_struct*, i32)*, void (%struct.jpeg_decompress_struct*)* }
27 %struct.jpeg_upsampler = type { void (%struct.jpeg_decompress_struct*)*, void (%struct.jpeg_decompress_struct*, i8***, i32*, i32, i8**, i32*, i32)*, i32 }
28 %struct.jvirt_barray_control = type { [64 x i16]**, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %struct.jvirt_barray_control*, %struct.backing_store_info }
29 %struct.jvirt_sarray_control = type { i8**, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %struct.jvirt_sarray_control*, %struct.backing_store_info }
31 define arm_apcscc void @jpeg_idct_float(%struct.jpeg_decompress_struct* nocapture %cinfo, %struct.jpeg_component_info* nocapture %compptr, i16* nocapture %coef_block, i8** nocapture %output_buf, i32 %output_col) nounwind {
33 %workspace = alloca [64 x float], align 4 ; <[64 x float]*> [#uses=11]
34 %0 = load i8** undef, align 4 ; <i8*> [#uses=5]
37 bb: ; preds = %bb, %entry
38 %indvar = phi i32 [ 0, %entry ], [ %indvar.next, %bb ] ; <i32> [#uses=11]
39 %tmp39 = add i32 %indvar, 8 ; <i32> [#uses=0]
40 %tmp41 = add i32 %indvar, 16 ; <i32> [#uses=2]
41 %scevgep42 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp41 ; <float*> [#uses=1]
42 %tmp43 = add i32 %indvar, 24 ; <i32> [#uses=1]
43 %scevgep44 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp43 ; <float*> [#uses=1]
44 %tmp45 = add i32 %indvar, 32 ; <i32> [#uses=1]
45 %scevgep46 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp45 ; <float*> [#uses=1]
46 %tmp47 = add i32 %indvar, 40 ; <i32> [#uses=1]
47 %scevgep48 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp47 ; <float*> [#uses=1]
48 %tmp49 = add i32 %indvar, 48 ; <i32> [#uses=1]
49 %scevgep50 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp49 ; <float*> [#uses=1]
50 %tmp51 = add i32 %indvar, 56 ; <i32> [#uses=1]
51 %scevgep52 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp51 ; <float*> [#uses=1]
52 %wsptr.119 = getelementptr [64 x float]* %workspace, i32 0, i32 %indvar ; <float*> [#uses=1]
53 %tmp54 = shl i32 %indvar, 2 ; <i32> [#uses=1]
54 %scevgep76 = getelementptr i8* undef, i32 %tmp54 ; <i8*> [#uses=1]
55 %quantptr.118 = bitcast i8* %scevgep76 to float* ; <float*> [#uses=1]
56 %scevgep79 = getelementptr i16* %coef_block, i32 %tmp41 ; <i16*> [#uses=0]
57 %inptr.117 = getelementptr i16* %coef_block, i32 %indvar ; <i16*> [#uses=1]
58 %1 = load i16* null, align 2 ; <i16> [#uses=1]
59 %2 = load i16* undef, align 2 ; <i16> [#uses=1]
60 %3 = load i16* %inptr.117, align 2 ; <i16> [#uses=1]
61 %4 = sitofp i16 %3 to float ; <float> [#uses=1]
62 %5 = load float* %quantptr.118, align 4 ; <float> [#uses=1]
63 %6 = fmul float %4, %5 ; <float> [#uses=1]
64 %7 = fsub float %6, undef ; <float> [#uses=2]
65 %8 = fmul float undef, 0x3FF6A09E60000000 ; <float> [#uses=1]
66 %9 = fsub float %8, 0.000000e+00 ; <float> [#uses=2]
67 %10 = fadd float undef, 0.000000e+00 ; <float> [#uses=2]
68 %11 = fadd float %7, %9 ; <float> [#uses=2]
69 %12 = fsub float %7, %9 ; <float> [#uses=2]
70 %13 = sitofp i16 %1 to float ; <float> [#uses=1]
71 %14 = fmul float %13, undef ; <float> [#uses=2]
72 %15 = sitofp i16 %2 to float ; <float> [#uses=1]
73 %16 = load float* undef, align 4 ; <float> [#uses=1]
74 %17 = fmul float %15, %16 ; <float> [#uses=1]
75 %18 = fadd float %14, undef ; <float> [#uses=2]
76 %19 = fsub float %14, undef ; <float> [#uses=2]
77 %20 = fadd float undef, %17 ; <float> [#uses=2]
78 %21 = fadd float %20, %18 ; <float> [#uses=3]
79 %22 = fsub float %20, %18 ; <float> [#uses=1]
80 %23 = fmul float %22, 0x3FF6A09E60000000 ; <float> [#uses=1]
81 %24 = fadd float %19, undef ; <float> [#uses=1]
82 %25 = fmul float %24, 0x3FFD906BC0000000 ; <float> [#uses=2]
83 %26 = fmul float undef, 0x3FF1517A80000000 ; <float> [#uses=1]
84 %27 = fsub float %26, %25 ; <float> [#uses=1]
85 %28 = fmul float %19, 0xC004E7AEA0000000 ; <float> [#uses=1]
86 %29 = fadd float %28, %25 ; <float> [#uses=1]
87 %30 = fsub float %29, %21 ; <float> [#uses=3]
88 %31 = fsub float %23, %30 ; <float> [#uses=3]
89 %32 = fadd float %27, %31 ; <float> [#uses=1]
90 %33 = fadd float %10, %21 ; <float> [#uses=1]
91 store float %33, float* %wsptr.119, align 4
92 %34 = fsub float %10, %21 ; <float> [#uses=1]
93 store float %34, float* %scevgep52, align 4
94 %35 = fadd float %11, %30 ; <float> [#uses=1]
95 store float %35, float* null, align 4
96 %36 = fsub float %11, %30 ; <float> [#uses=1]
97 store float %36, float* %scevgep50, align 4
98 %37 = fadd float %12, %31 ; <float> [#uses=1]
99 store float %37, float* %scevgep42, align 4
100 %38 = fsub float %12, %31 ; <float> [#uses=1]
101 store float %38, float* %scevgep48, align 4
102 %39 = fadd float undef, %32 ; <float> [#uses=1]
103 store float %39, float* %scevgep46, align 4
104 store float undef, float* %scevgep44, align 4
105 %indvar.next = add i32 %indvar, 1 ; <i32> [#uses=1]
106 br i1 undef, label %bb6, label %bb
109 %.sum10 = add i32 %output_col, 1 ; <i32> [#uses=1]
110 %.sum8 = add i32 %output_col, 6 ; <i32> [#uses=1]
111 %.sum6 = add i32 %output_col, 2 ; <i32> [#uses=1]
112 %.sum = add i32 %output_col, 3 ; <i32> [#uses=1]
115 bb8: ; preds = %bb8, %bb6
116 %ctr.116 = phi i32 [ 0, %bb6 ], [ %88, %bb8 ] ; <i32> [#uses=3]
117 %scevgep = getelementptr i8** %output_buf, i32 %ctr.116 ; <i8**> [#uses=1]
118 %tmp = shl i32 %ctr.116, 3 ; <i32> [#uses=5]
119 %tmp2392 = or i32 %tmp, 4 ; <i32> [#uses=1]
120 %scevgep24 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp2392 ; <float*> [#uses=1]
121 %tmp2591 = or i32 %tmp, 2 ; <i32> [#uses=1]
122 %scevgep26 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp2591 ; <float*> [#uses=1]
123 %tmp2790 = or i32 %tmp, 6 ; <i32> [#uses=1]
124 %scevgep28 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp2790 ; <float*> [#uses=1]
125 %tmp3586 = or i32 %tmp, 7 ; <i32> [#uses=0]
126 %wsptr.215 = getelementptr [64 x float]* %workspace, i32 0, i32 %tmp ; <float*> [#uses=1]
127 %40 = load i8** %scevgep, align 4 ; <i8*> [#uses=4]
128 %41 = load float* %wsptr.215, align 4 ; <float> [#uses=1]
129 %42 = load float* %scevgep24, align 4 ; <float> [#uses=1]
130 %43 = fadd float %41, %42 ; <float> [#uses=1]
131 %44 = load float* %scevgep26, align 4 ; <float> [#uses=1]
132 %45 = load float* %scevgep28, align 4 ; <float> [#uses=1]
133 %46 = fadd float %44, %45 ; <float> [#uses=1]
134 %47 = fsub float %43, %46 ; <float> [#uses=2]
135 %48 = fsub float undef, 0.000000e+00 ; <float> [#uses=1]
136 %49 = fadd float 0.000000e+00, undef ; <float> [#uses=1]
137 %50 = fptosi float %49 to i32 ; <i32> [#uses=1]
138 %51 = add i32 %50, 4 ; <i32> [#uses=1]
139 %52 = lshr i32 %51, 3 ; <i32> [#uses=1]
140 %53 = and i32 %52, 1023 ; <i32> [#uses=1]
141 %.sum14 = add i32 %53, 128 ; <i32> [#uses=1]
142 %54 = getelementptr i8* %0, i32 %.sum14 ; <i8*> [#uses=1]
143 %55 = load i8* %54, align 1 ; <i8> [#uses=1]
144 store i8 %55, i8* null, align 1
145 %56 = getelementptr i8* %40, i32 %.sum10 ; <i8*> [#uses=1]
146 store i8 0, i8* %56, align 1
147 %57 = load i8* null, align 1 ; <i8> [#uses=1]
148 %58 = getelementptr i8* %40, i32 %.sum8 ; <i8*> [#uses=1]
149 store i8 %57, i8* %58, align 1
150 %59 = fadd float undef, %48 ; <float> [#uses=1]
151 %60 = fptosi float %59 to i32 ; <i32> [#uses=1]
152 %61 = add i32 %60, 4 ; <i32> [#uses=1]
153 %62 = lshr i32 %61, 3 ; <i32> [#uses=1]
154 %63 = and i32 %62, 1023 ; <i32> [#uses=1]
155 %.sum7 = add i32 %63, 128 ; <i32> [#uses=1]
156 %64 = getelementptr i8* %0, i32 %.sum7 ; <i8*> [#uses=1]
157 %65 = load i8* %64, align 1 ; <i8> [#uses=1]
158 %66 = getelementptr i8* %40, i32 %.sum6 ; <i8*> [#uses=1]
159 store i8 %65, i8* %66, align 1
160 %67 = fptosi float undef to i32 ; <i32> [#uses=1]
161 %68 = add i32 %67, 4 ; <i32> [#uses=1]
162 %69 = lshr i32 %68, 3 ; <i32> [#uses=1]
163 %70 = and i32 %69, 1023 ; <i32> [#uses=1]
164 %.sum5 = add i32 %70, 128 ; <i32> [#uses=1]
165 %71 = getelementptr i8* %0, i32 %.sum5 ; <i8*> [#uses=1]
166 %72 = load i8* %71, align 1 ; <i8> [#uses=1]
167 store i8 %72, i8* undef, align 1
168 %73 = fadd float %47, undef ; <float> [#uses=1]
169 %74 = fptosi float %73 to i32 ; <i32> [#uses=1]
170 %75 = add i32 %74, 4 ; <i32> [#uses=1]
171 %76 = lshr i32 %75, 3 ; <i32> [#uses=1]
172 %77 = and i32 %76, 1023 ; <i32> [#uses=1]
173 %.sum3 = add i32 %77, 128 ; <i32> [#uses=1]
174 %78 = getelementptr i8* %0, i32 %.sum3 ; <i8*> [#uses=1]
175 %79 = load i8* %78, align 1 ; <i8> [#uses=1]
176 store i8 %79, i8* undef, align 1
177 %80 = fsub float %47, undef ; <float> [#uses=1]
178 %81 = fptosi float %80 to i32 ; <i32> [#uses=1]
179 %82 = add i32 %81, 4 ; <i32> [#uses=1]
180 %83 = lshr i32 %82, 3 ; <i32> [#uses=1]
181 %84 = and i32 %83, 1023 ; <i32> [#uses=1]
182 %.sum1 = add i32 %84, 128 ; <i32> [#uses=1]
183 %85 = getelementptr i8* %0, i32 %.sum1 ; <i8*> [#uses=1]
184 %86 = load i8* %85, align 1 ; <i8> [#uses=1]
185 %87 = getelementptr i8* %40, i32 %.sum ; <i8*> [#uses=1]
186 store i8 %86, i8* %87, align 1
187 %88 = add i32 %ctr.116, 1 ; <i32> [#uses=2]
188 %exitcond = icmp eq i32 %88, 8 ; <i1> [#uses=1]
189 br i1 %exitcond, label %return, label %bb8
191 return: ; preds = %bb8