1; RUN: llc < %s -mcpu=cortex-a15 -verify-machineinstrs -arm-atomic-cfg-tidy=0 | FileCheck %s 2 3; Check a spill right after a function call with large struct byval is correctly 4; generated. 5; PR16393 6 7; We expect 4-byte spill and reload to be generated. 8 9; CHECK: set_stored_macroblock_parameters 10; CHECK: str r{{.*}}, [sp, {{#[0-9]+}}] @ 4-byte Spill 11; CHECK: ldr r{{.*}}, [lr, {{#[0-9]+}}] @ 4-byte Reload 12 13target triple = "armv7l-unknown-linux-gnueabihf" 14 15%structA = type { double, [16 x [16 x i16]], [16 x [16 x i16]], [16 x [16 x i16]], i32****, i32***, i32, i16, [4 x i32], [4 x i32], i8**, [16 x i8], [16 x i8], i32, i64, i32, i16******, i16******, [2 x [4 x [4 x i8]]], i32, i32, i32, i32, i32, i32, i32, i32, i32 } 16%structB = type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8**, i8**, i32, i32***, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [9 x [16 x [16 x i16]]], [5 x [16 x [16 x i16]]], [9 x [8 x [8 x i16]]], [2 x [4 x [16 x [16 x i16]]]], [16 x [16 x i16]], [16 x [16 x i32]], i32****, i32***, i32***, i32***, i32****, i32****, %structC*, %structD*, %structK*, i32*, i32*, i32, i32, i32, i32, [4 x [4 x i32]], i32, i32, i32, i32, i32, double, i32, i32, i32, i32, i16******, i16******, i16******, i16******, [15 x i16], i32, i32, i32, i32, i32, i32, i32, i32, [6 x [32 x i32]], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [1 x i32], i32, i32, [2 x i32], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %structL*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, double**, double***, i32***, double**, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [3 x [2 x i32]], [2 x i32], i32, i32, i16, i32, i32, i32, i32, i32 } 17%structC = type { i32, i32, [100 x %structD*], i32, float, float, float } 18%structD = type { i32, i32, i32, i32, i32, i32, %structE*, %structH*, %structJ*, i32, i32*, i32*, i32*, i32, i32*, i32*, i32*, i32 (i32)*, [3 x [2 x i32]] } 19%structE = type { %structF*, %structG, %structG } 20%structF = type { i32, i32, i8, i32, i32, i8, i8, i32, i32, i8*, i32 } 21%structG = type { i32, i32, i32, i32, i32, i8*, i32*, i32, i32 } 22%structH = type { [3 x [11 x %structI]], [2 x [9 x %structI]], [2 x [10 x %structI]], [2 x [6 x %structI]], [4 x %structI], [4 x %structI], [3 x %structI] } 23%structI = type { i16, i8, i32 } 24%structJ = type { [2 x %structI], [4 x %structI], [3 x [4 x %structI]], [10 x [4 x %structI]], [10 x [15 x %structI]], [10 x [15 x %structI]], [10 x [5 x %structI]], [10 x [5 x %structI]], [10 x [15 x %structI]], [10 x [15 x %structI]] } 25%structK = type { i32, i32, i32, [2 x i32], i32, [8 x i32], %structK*, %structK*, i32, [2 x [4 x [4 x [2 x i32]]]], [16 x i8], [16 x i8], i32, i64, [4 x i32], [4 x i32], i64, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i16, double, i32, i32, i32, i32, i32, i32, i32, i32, i32 } 26%structL = type { i32, i32, i32, i32, i32, %structL* } 27%structM = type { i32, i32, i32, i32, i32, i32, [6 x [33 x i64]], [6 x [33 x i64]], [6 x [33 x i64]], [6 x [33 x i64]], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i16**, i16****, i16****, i16*****, i16***, i8*, i8***, i64***, i64***, i16****, i8**, i8**, %structM*, %structM*, %structM*, i32, i32, i32, i32, i32, i32, i32 } 28%structN = type { i32, [16 x [16 x i32]], [16 x [16 x i32]], [16 x [16 x i32]], [3 x [16 x [16 x i32]]], [4 x i16], [4 x i8], [4 x i8], [4 x i8], [16 x [16 x i16]], [16 x [16 x i16]], [16 x [16 x i32]] } 29 30@cofAC = external global i32****, align 4 31@cofDC = external global i32***, align 4 32@rdopt = external global %structA*, align 4 33@img = external global %structB* 34@enc_picture = external global %structM* 35@si_frame_indicator = external global i32, align 4 36@sp2_frame_indicator = external global i32, align 4 37@lrec = external global i32**, align 4 38@tr8x8 = external global %structN, align 4 39@best_mode = external global i16, align 2 40@best_c_imode = external global i32, align 4 41@best_i16offset = external global i32, align 4 42@bi_pred_me = external global i16, align 2 43@b8mode = external global [4 x i32], align 4 44@b8pdir = external global [4 x i32], align 4 45@b4_intra_pred_modes = external global [16 x i8], align 1 46@b8_intra_pred_modes8x8 = external global [16 x i8], align 1 47@b4_ipredmode = external global [16 x i8], align 1 48@b8_ipredmode8x8 = external global [4 x [4 x i8]], align 1 49@rec_mbY = external global [16 x [16 x i16]], align 2 50@lrec_rec = external global [16 x [16 x i32]], align 4 51@rec_mbU = external global [16 x [16 x i16]], align 2 52@rec_mbV = external global [16 x [16 x i16]], align 2 53@lrec_rec_U = external global [16 x [16 x i32]], align 4 54@lrec_uv = external global i32***, align 4 55@lrec_rec_V = external global [16 x [16 x i32]], align 4 56@cbp = external global i32, align 4 57@cbp_blk = external global i64, align 8 58@luma_transform_size_8x8_flag = external global i32, align 4 59@frefframe = external global [4 x [4 x i8]], align 1 60@brefframe = external global [4 x [4 x i8]], align 1 61 62; Function Attrs: nounwind 63declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i1) #0 64 65; Function Attrs: nounwind 66declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i1) #0 67 68; Function Attrs: nounwind 69declare void @SetMotionVectorsMB(%structK* nocapture, i32) #1 70 71; Function Attrs: nounwind 72define void @set_stored_macroblock_parameters() #1 { 73entry: 74 %0 = load %structB*, %structB** @img, align 4 75 %1 = load i32, i32* undef, align 4 76 %mb_data = getelementptr inbounds %structB, %structB* %0, i32 0, i32 61 77 %2 = load %structK*, %structK** %mb_data, align 4 78 br label %for.body 79 80for.body: ; preds = %for.body, %entry 81 br i1 undef, label %for.end, label %for.body 82 83for.end: ; preds = %for.body 84 br i1 undef, label %for.body20, label %if.end 85 86for.body20: ; preds = %for.end 87 unreachable 88 89if.end: ; preds = %for.end 90 br i1 undef, label %if.end40, label %for.cond31.preheader 91 92for.cond31.preheader: ; preds = %if.end 93 unreachable 94 95if.end40: ; preds = %if.end 96 br i1 undef, label %if.end43, label %if.then42 97 98if.then42: ; preds = %if.end40 99 br label %if.end43 100 101if.end43: ; preds = %if.then42, %if.end40 102 br i1 undef, label %if.end164, label %for.cond47.preheader 103 104for.cond47.preheader: ; preds = %if.end43 105 br i1 undef, label %for.body119, label %if.end164 106 107for.body119: ; preds = %for.body119, %for.cond47.preheader 108 br i1 undef, label %for.body119, label %if.end164 109 110if.end164: ; preds = %for.body119, %for.cond47.preheader, %if.end43 111 store i32*** null, i32**** @cofDC, align 4 112 %mb_type = getelementptr inbounds %structK, %structK* %2, i32 %1, i32 8 113 br i1 undef, label %if.end230, label %if.then169 114 115if.then169: ; preds = %if.end164 116 br i1 undef, label %for.cond185.preheader, label %for.cond210.preheader 117 118for.cond185.preheader: ; preds = %if.then169 119 unreachable 120 121for.cond210.preheader: ; preds = %if.then169 122 unreachable 123 124if.end230: ; preds = %if.end164 125 tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 undef, i8* align 4 bitcast ([4 x i32]* @b8mode to i8*), i32 16, i1 false) 126 %b8pdir = getelementptr inbounds %structK, %structK* %2, i32 %1, i32 15 127 %3 = bitcast [4 x i32]* %b8pdir to i8* 128 tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %3, i8* align 4 bitcast ([4 x i32]* @b8pdir to i8*), i32 16, i1 false) 129 br i1 undef, label %if.end236, label %if.then233 130 131if.then233: ; preds = %if.end230 132 unreachable 133 134if.end236: ; preds = %if.end230 135 %cmp242 = icmp ne i16 undef, 8 136 %4 = load i32, i32* @luma_transform_size_8x8_flag, align 4 137 %tobool245 = icmp ne i32 %4, 0 138 %or.cond812 = or i1 %cmp242, %tobool245 139 br i1 %or.cond812, label %if.end249, label %land.lhs.true246 140 141land.lhs.true246: ; preds = %if.end236 142 br i1 undef, label %if.end249, label %if.then248 143 144if.then248: ; preds = %land.lhs.true246 145 tail call void asm sideeffect "", "~{r1},~{r2},~{r3},~{r4},~{r5},~{r6},~{r7},~{r8},~{r9},~{r10},~{r11}"() nounwind 146 tail call void @RestoreMVBlock8x8(i32 1, i32 0, %structN* byval @tr8x8, i32 0) #0 147 tail call void @RestoreMVBlock8x8(i32 1, i32 2, %structN* byval @tr8x8, i32 0) #0 148 tail call void @RestoreMVBlock8x8(i32 1, i32 3, %structN* byval @tr8x8, i32 0) #0 149 br label %if.end249 150 151if.end249: ; preds = %if.then248, %land.lhs.true246, %if.end236 152 %5 = load i32, i32* @luma_transform_size_8x8_flag, align 4 153 %6 = load %structA*, %structA** @rdopt, align 4 154 %luma_transform_size_8x8_flag264 = getelementptr inbounds %structA, %structA* %6, i32 0, i32 21 155 store i32 %5, i32* %luma_transform_size_8x8_flag264, align 4 156 %7 = load i32, i32* undef, align 4 157 %add281 = add nsw i32 %7, 0 158 br label %for.body285 159 160for.body285: ; preds = %for.inc503, %if.end249 161 %8 = phi %structB* [ undef, %if.end249 ], [ %.pre1155, %for.inc503 ] 162 %i.21103 = phi i32 [ 0, %if.end249 ], [ %inc504, %for.inc503 ] 163 %block_x286 = getelementptr inbounds %structB, %structB* %8, i32 0, i32 37 164 %9 = load i32, i32* %block_x286, align 4 165 %add287 = add nsw i32 %9, %i.21103 166 %shr289 = ashr i32 %i.21103, 1 167 %add290 = add nsw i32 %shr289, 0 168 %arrayidx292 = getelementptr inbounds %structK, %structK* %2, i32 %1, i32 15, i32 %add290 169 %10 = load %structM*, %structM** @enc_picture, align 4 170 %ref_idx = getelementptr inbounds %structM, %structM* %10, i32 0, i32 35 171 %11 = load i8***, i8**** %ref_idx, align 4 172 %12 = load i8**, i8*** %11, align 4 173 %arrayidx313 = getelementptr inbounds i8*, i8** %12, i32 %add281 174 %13 = load i8*, i8** %arrayidx313, align 4 175 %arrayidx314 = getelementptr inbounds i8, i8* %13, i32 %add287 176 store i8 -1, i8* %arrayidx314, align 1 177 %14 = load %structB*, %structB** @img, align 4 178 %MbaffFrameFlag327 = getelementptr inbounds %structB, %structB* %14, i32 0, i32 100 179 %15 = load i32, i32* %MbaffFrameFlag327, align 4 180 %tobool328 = icmp eq i32 %15, 0 181 br i1 %tobool328, label %if.end454, label %if.then329 182 183if.then329: ; preds = %for.body285 184 %16 = load %structA*, %structA** @rdopt, align 4 185 br label %if.end454 186 187if.end454: ; preds = %if.then329, %for.body285 188 %17 = load i32, i32* %arrayidx292, align 4 189 %cmp457 = icmp eq i32 %17, 0 190 br i1 %cmp457, label %if.then475, label %lor.lhs.false459 191 192lor.lhs.false459: ; preds = %if.end454 193 %18 = load i32, i32* %mb_type, align 4 194 switch i32 %18, label %for.inc503 [ 195 i32 9, label %if.then475 196 i32 11, label %if.then475 197 i32 13, label %if.then475 198 i32 14, label %if.then475 199 ] 200 201if.then475: ; preds = %lor.lhs.false459, %lor.lhs.false459, %lor.lhs.false459, %lor.lhs.false459, %if.end454 202 store i16 0, i16* undef, align 2 203 br label %for.inc503 204 205for.inc503: ; preds = %if.then475, %lor.lhs.false459 206 %inc504 = add nsw i32 %i.21103, 1 207 %.pre1155 = load %structB*, %structB** @img, align 4 208 br label %for.body285 209} 210 211; Function Attrs: nounwind 212declare void @update_offset_params(i32, i32) #1 213 214; Function Attrs: nounwind 215declare void @RestoreMVBlock8x8(i32, i32, %structN* byval nocapture, i32) #1 216 217attributes #0 = { nounwind } 218attributes #1 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "unsafe-fp-math"="false" "use-soft-float"="false" } 219