1FileCheck - Flexible pattern matching file verifier 2=================================================== 3 4 5SYNOPSIS 6-------- 7 8 9**FileCheck** *match-filename* [*--check-prefix=XXX*] [*--strict-whitespace*] 10 11 12DESCRIPTION 13----------- 14 15 16**FileCheck** reads two files (one from standard input, and one specified on the 17command line) and uses one to verify the other. This behavior is particularly 18useful for the testsuite, which wants to verify that the output of some tool 19(e.g. llc) contains the expected information (for example, a movsd from esp or 20whatever is interesting). This is similar to using grep, but it is optimized 21for matching multiple different inputs in one file in a specific order. 22 23The *match-filename* file specifies the file that contains the patterns to 24match. The file to verify is always read from standard input. 25 26 27OPTIONS 28------- 29 30 31 32**-help** 33 34 Print a summary of command line options. 35 36 37 38**--check-prefix** *prefix* 39 40 FileCheck searches the contents of *match-filename* for patterns to match. By 41 default, these patterns are prefixed with "CHECK:". If you'd like to use a 42 different prefix (e.g. because the same input file is checking multiple 43 different tool or options), the **--check-prefix** argument allows you to specify 44 a specific prefix to match. 45 46 47 48**--strict-whitespace** 49 50 By default, FileCheck canonicalizes input horizontal whitespace (spaces and 51 tabs) which causes it to ignore these differences (a space will match a tab). 52 The --strict-whitespace argument disables this behavior. 53 54 55 56**-version** 57 58 Show the version number of this program. 59 60 61 62 63EXIT STATUS 64----------- 65 66 67If **FileCheck** verifies that the file matches the expected contents, it exits 68with 0. Otherwise, if not, or if an error occurs, it will exit with a non-zero 69value. 70 71 72TUTORIAL 73-------- 74 75 76FileCheck is typically used from LLVM regression tests, being invoked on the RUN 77line of the test. A simple example of using FileCheck from a RUN line looks 78like this: 79 80 81.. code-block:: llvm 82 83 ; RUN: llvm-as < %s | llc -march=x86-64 | FileCheck %s 84 85 86This syntax says to pipe the current file ("%s") into llvm-as, pipe that into 87llc, then pipe the output of llc into FileCheck. This means that FileCheck will 88be verifying its standard input (the llc output) against the filename argument 89specified (the original .ll file specified by "%s"). To see how this works, 90let's look at the rest of the .ll file (after the RUN line): 91 92 93.. code-block:: llvm 94 95 define void @sub1(i32* %p, i32 %v) { 96 entry: 97 ; CHECK: sub1: 98 ; CHECK: subl 99 %0 = tail call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %p, i32 %v) 100 ret void 101 } 102 103 define void @inc4(i64* %p) { 104 entry: 105 ; CHECK: inc4: 106 ; CHECK: incq 107 %0 = tail call i64 @llvm.atomic.load.add.i64.p0i64(i64* %p, i64 1) 108 ret void 109 } 110 111 112Here you can see some "CHECK:" lines specified in comments. Now you can see 113how the file is piped into llvm-as, then llc, and the machine code output is 114what we are verifying. FileCheck checks the machine code output to verify that 115it matches what the "CHECK:" lines specify. 116 117The syntax of the CHECK: lines is very simple: they are fixed strings that 118must occur in order. FileCheck defaults to ignoring horizontal whitespace 119differences (e.g. a space is allowed to match a tab) but otherwise, the contents 120of the CHECK: line is required to match some thing in the test file exactly. 121 122One nice thing about FileCheck (compared to grep) is that it allows merging 123test cases together into logical groups. For example, because the test above 124is checking for the "sub1:" and "inc4:" labels, it will not match unless there 125is a "subl" in between those labels. If it existed somewhere else in the file, 126that would not count: "grep subl" matches if subl exists anywhere in the 127file. 128 129The FileCheck -check-prefix option 130~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 131 132 133The FileCheck -check-prefix option allows multiple test configurations to be 134driven from one .ll file. This is useful in many circumstances, for example, 135testing different architectural variants with llc. Here's a simple example: 136 137 138.. code-block:: llvm 139 140 ; RUN: llvm-as < %s | llc -mtriple=i686-apple-darwin9 -mattr=sse41 \ 141 ; RUN: | FileCheck %s -check-prefix=X32 142 ; RUN: llvm-as < %s | llc -mtriple=x86_64-apple-darwin9 -mattr=sse41 \ 143 ; RUN: | FileCheck %s -check-prefix=X64 144 145 define <4 x i32> @pinsrd_1(i32 %s, <4 x i32> %tmp) nounwind { 146 %tmp1 = insertelement <4 x i32>; %tmp, i32 %s, i32 1 147 ret <4 x i32> %tmp1 148 ; X32: pinsrd_1: 149 ; X32: pinsrd $1, 4(%esp), %xmm0 150 151 ; X64: pinsrd_1: 152 ; X64: pinsrd $1, %edi, %xmm0 153 } 154 155 156In this case, we're testing that we get the expected code generation with 157both 32-bit and 64-bit code generation. 158 159 160The "CHECK-NEXT:" directive 161~~~~~~~~~~~~~~~~~~~~~~~~~~~ 162 163 164Sometimes you want to match lines and would like to verify that matches 165happen on exactly consecutive lines with no other lines in between them. In 166this case, you can use CHECK: and CHECK-NEXT: directives to specify this. If 167you specified a custom check prefix, just use "<PREFIX>-NEXT:". For 168example, something like this works as you'd expect: 169 170 171.. code-block:: llvm 172 173 define void @t2(<2 x double>* %r, <2 x double>* %A, double %B) { 174 %tmp3 = load <2 x double>* %A, align 16 175 %tmp7 = insertelement <2 x double> undef, double %B, i32 0 176 %tmp9 = shufflevector <2 x double> %tmp3, 177 <2 x double> %tmp7, 178 <2 x i32> < i32 0, i32 2 > 179 store <2 x double> %tmp9, <2 x double>* %r, align 16 180 ret void 181 182 ; CHECK: t2: 183 ; CHECK: movl 8(%esp), %eax 184 ; CHECK-NEXT: movapd (%eax), %xmm0 185 ; CHECK-NEXT: movhpd 12(%esp), %xmm0 186 ; CHECK-NEXT: movl 4(%esp), %eax 187 ; CHECK-NEXT: movapd %xmm0, (%eax) 188 ; CHECK-NEXT: ret 189 } 190 191 192CHECK-NEXT: directives reject the input unless there is exactly one newline 193between it an the previous directive. A CHECK-NEXT cannot be the first 194directive in a file. 195 196 197The "CHECK-NOT:" directive 198~~~~~~~~~~~~~~~~~~~~~~~~~~ 199 200 201The CHECK-NOT: directive is used to verify that a string doesn't occur 202between two matches (or before the first match, or after the last match). For 203example, to verify that a load is removed by a transformation, a test like this 204can be used: 205 206 207.. code-block:: llvm 208 209 define i8 @coerce_offset0(i32 %V, i32* %P) { 210 store i32 %V, i32* %P 211 212 %P2 = bitcast i32* %P to i8* 213 %P3 = getelementptr i8* %P2, i32 2 214 215 %A = load i8* %P3 216 ret i8 %A 217 ; CHECK: @coerce_offset0 218 ; CHECK-NOT: load 219 ; CHECK: ret i8 220 } 221 222 223 224FileCheck Pattern Matching Syntax 225~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 226 227 228The CHECK: and CHECK-NOT: directives both take a pattern to match. For most 229uses of FileCheck, fixed string matching is perfectly sufficient. For some 230things, a more flexible form of matching is desired. To support this, FileCheck 231allows you to specify regular expressions in matching strings, surrounded by 232double braces: **{{yourregex}}**. Because we want to use fixed string 233matching for a majority of what we do, FileCheck has been designed to support 234mixing and matching fixed string matching with regular expressions. This allows 235you to write things like this: 236 237 238.. code-block:: llvm 239 240 ; CHECK: movhpd {{[0-9]+}}(%esp), {{%xmm[0-7]}} 241 242 243In this case, any offset from the ESP register will be allowed, and any xmm 244register will be allowed. 245 246Because regular expressions are enclosed with double braces, they are 247visually distinct, and you don't need to use escape characters within the double 248braces like you would in C. In the rare case that you want to match double 249braces explicitly from the input, you can use something ugly like 250**{{[{][{]}}** as your pattern. 251 252 253FileCheck Variables 254~~~~~~~~~~~~~~~~~~~ 255 256 257It is often useful to match a pattern and then verify that it occurs again 258later in the file. For codegen tests, this can be useful to allow any register, 259but verify that that register is used consistently later. To do this, FileCheck 260allows named variables to be defined and substituted into patterns. Here is a 261simple example: 262 263 264.. code-block:: llvm 265 266 ; CHECK: test5: 267 ; CHECK: notw [[REGISTER:%[a-z]+]] 268 ; CHECK: andw {{.*}}[[REGISTER]] 269 270 271The first check line matches a regex (**%[a-z]+**) and captures it into 272the variable "REGISTER". The second line verifies that whatever is in REGISTER 273occurs later in the file after an "andw". FileCheck variable references are 274always contained in **[[ ]]** pairs, are named, and their names can be 275name, then it is a definition of the variable, if not, it is a use. 276 277FileCheck variables can be defined multiple times, and uses always get the 278latest value. Note that variables are all read at the start of a "CHECK" line 279and are all defined at the end. This means that if you have something like 280"**CHECK: [[XYZ:.\\*]]x[[XYZ]]**", the check line will read the previous 281value of the XYZ variable and define a new one after the match is performed. If 282you need to do something like this you can probably take advantage of the fact 283that FileCheck is not actually line-oriented when it matches, this allows you to 284define two separate CHECK lines that match on the same line. 285