Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / ee4e2e718d00a44c9ac4e663534a0391d72a3549 / . / llvm / utils / update_test_checks.py
blob: a37f2b4c043e21992cd065920aab2e78cbaa6016 [file] [log] [blame]
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]1#!/usr/bin/env python2.7
2
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]3"""A script to generate FileCheck statements for 'opt' regression tests.
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]4
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]5This script is a utility to update LLVM opt test cases with new
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]6FileCheck patterns. It can either update all of the tests in the file or
7a single test function.
Sanjay Patel40641582016-04-05 18:00:47 +0000[diff] [blame]8
9Example usage:
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]10$ update_test_checks.py --opt=../bin/opt test/foo.ll
Sanjay Patel40641582016-04-05 18:00:47 +0000[diff] [blame]11
12Workflow:
131. Make a compiler patch that requires updating some number of FileCheck lines
14 in regression test files.
152. Save the patch and revert it from your local work area.
163. Update the RUN-lines in the affected regression tests to look canonical.
17 Example: "; RUN: opt < %s -instcombine -S | FileCheck %s"
184. Refresh the FileCheck lines for either the entire file or select functions by
19 running this script.
205. Commit the fresh baseline of checks.
216. Apply your patch from step 1 and rebuild your local binaries.
227. Re-run this script on affected regression tests.
238. Check the diffs to ensure the script has done something reasonable.
249. Submit a patch including the regression test diffs for review.
25
26A common pattern is to have the script insert complete checking of every
27instruction. Then, edit it down to only check the relevant instructions.
28The script is designed to make adding checks to a test case fast, it is *not*
29designed to be authoratitive about what constitutes a good test!
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]30"""
31
32import argparse
33import itertools
34import os # Used to advertise this file's name ("autogenerated_note").
35import string
36import subprocess
37import sys
38import tempfile
39import re
40
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]41from UpdateTestChecks import common
42
Sanjay Patel16be4df92016-04-05 19:50:21 +0000[diff] [blame]43ADVERT = '; NOTE: Assertions have been autogenerated by '
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]44
45# RegEx: this is where the magic happens.
46
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]47SCRUB_IR_COMMENT_RE = re.compile(r'\s*;.*')
48
Sanjay Patele54e6f52016-03-25 17:00:12 +0000[diff] [blame]49IR_FUNCTION_RE = re.compile('^\s*define\s+(?:internal\s+)?[^@]*@([\w-]+)\s*\(')
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]50OPT_FUNCTION_RE = re.compile(
Sanjay Patele54e6f52016-03-25 17:00:12 +0000[diff] [blame]51 r'^\s*define\s+(?:internal\s+)?[^@]*@(?P<func>[\w-]+?)\s*\('
Daniel Berlin101db5f2017-01-09 19:24:19 +0000[diff] [blame]52 r'(\s+)?[^)]*[^{]*\{\n(?P<body>.*?)^\}$',
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]53 flags=(re.M | re.S))
Daniel Berlinfe4e7d02017-01-07 19:04:59 +0000[diff] [blame]54# Match things that look at identifiers, but only if they are followed by
55# spaces, commas, paren, or end of the string
Daniel Berlinb18135f2017-01-15 07:40:46 +0000[diff] [blame]56IR_VALUE_RE = re.compile(r'(\s+)%([\w\.]+?)([,\s\(\)]|\Z)')
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]57
58
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]59
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]60def scrub_body(body, opt_basename):
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]61 # Scrub runs of whitespace out of the assembly, but leave the leading
62 # whitespace in place.
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]63 body = common.SCRUB_WHITESPACE_RE.sub(r' ', body)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]64 # Expand the tabs used for indentation.
65 body = string.expandtabs(body, 2)
66 # Strip trailing whitespace.
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]67 body = common.SCRUB_TRAILING_WHITESPACE_RE.sub(r'', body)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]68 return body
69
70
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]71
72# Create a FileCheck variable name based on an IR name.
73def get_value_name(var):
74 if var.isdigit():
75 var = 'TMP' + var
76 var = var.replace('.', '_')
77 return var.upper()
78
79
80# Create a FileCheck variable from regex.
81def get_value_definition(var):
82 return '[[' + get_value_name(var) + ':%.*]]'
83
84
85# Use a FileCheck variable.
86def get_value_use(var):
87 return '[[' + get_value_name(var) + ']]'
88
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]89# Replace IR value defs and uses with FileCheck variables.
90def genericize_check_lines(lines):
Daniel Berlinfe4e7d02017-01-07 19:04:59 +0000[diff] [blame]91 # This gets called for each match that occurs in
92 # a line. We transform variables we haven't seen
93 # into defs, and variables we have seen into uses.
94 def transform_line_vars(match):
95 var = match.group(2)
96 if var in vars_seen:
97 rv = get_value_use(var)
98 else:
99 vars_seen.add(var)
100 rv = get_value_definition(var)
101 # re.sub replaces the entire regex match
102 # with whatever you return, so we have
103 # to make sure to hand it back everything
104 # including the commas and spaces.
105 return match.group(1) + rv + match.group(3)
106
107 vars_seen = set()
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]108 lines_with_def = []
Daniel Berlinfe4e7d02017-01-07 19:04:59 +0000[diff] [blame]109
110 for i, line in enumerate(lines):
Sanjay Patel17681172016-03-27 20:44:35 +0000[diff] [blame]111 # An IR variable named '%.' matches the FileCheck regex string.
112 line = line.replace('%.', '%dot')
Daniel Berlinfe4e7d02017-01-07 19:04:59 +0000[diff] [blame]113 # Ignore any comments, since the check lines will too.
114 scrubbed_line = SCRUB_IR_COMMENT_RE.sub(r'', line)
115 lines[i] = IR_VALUE_RE.sub(transform_line_vars, scrubbed_line)
116 return lines
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]117
118
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]119def add_checks(output_lines, prefix_list, func_dict, func_name, opt_basename):
120 # Label format is based on IR string.
121 check_label_format = "; %s-LABEL: @%s("
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]122
123 printed_prefixes = []
124 for checkprefixes, _ in prefix_list:
125 for checkprefix in checkprefixes:
126 if checkprefix in printed_prefixes:
127 break
128 if not func_dict[checkprefix][func_name]:
129 continue
130 # Add some space between different check prefixes, but not after the last
131 # check line (before the test code).
132 #if len(printed_prefixes) != 0:
133 # output_lines.append(';')
134 printed_prefixes.append(checkprefix)
135 output_lines.append(check_label_format % (checkprefix, func_name))
136 func_body = func_dict[checkprefix][func_name].splitlines()
137
138 # For IR output, change all defs to FileCheck variables, so we're immune
139 # to variable naming fashions.
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]140 func_body = genericize_check_lines(func_body)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]141
Sanjay Patel96241e72016-04-05 16:49:07 +0000[diff] [blame]142 # This could be selectively enabled with an optional invocation argument.
143 # Disabled for now: better to check everything. Be safe rather than sorry.
144
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]145 # Handle the first line of the function body as a special case because
146 # it's often just noise (a useless asm comment or entry label).
Sanjay Patel96241e72016-04-05 16:49:07 +0000[diff] [blame]147 #if func_body[0].startswith("#") or func_body[0].startswith("entry:"):
148 # is_blank_line = True
149 #else:
150 # output_lines.append('; %s: %s' % (checkprefix, func_body[0]))
151 # is_blank_line = False
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]152
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]153 is_blank_line = False
Sanjay Patel96241e72016-04-05 16:49:07 +0000[diff] [blame]154
155 for func_line in func_body:
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]156 if func_line.strip() == '':
157 is_blank_line = True
158 continue
159 # Do not waste time checking IR comments.
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]160 func_line = SCRUB_IR_COMMENT_RE.sub(r'', func_line)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]161
162 # Skip blank lines instead of checking them.
163 if is_blank_line == True:
164 output_lines.append('; %s: %s' % (checkprefix, func_line))
165 else:
166 output_lines.append('; %s-NEXT: %s' % (checkprefix, func_line))
167 is_blank_line = False
168
169 # Add space between different check prefixes and also before the first
170 # line of code in the test function.
171 output_lines.append(';')
172 break
173 return output_lines
174
175
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]176def main():
Sanjay Patel40641582016-04-05 18:00:47 +0000[diff] [blame]177 from argparse import RawTextHelpFormatter
178 parser = argparse.ArgumentParser(description=__doc__, formatter_class=RawTextHelpFormatter)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]179 parser.add_argument('-v', '--verbose', action='store_true',
180 help='Show verbose output')
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]181 parser.add_argument('--opt-binary', default='opt',
182 help='The opt binary used to generate the test case')
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]183 parser.add_argument(
184 '--function', help='The function in the test file to update')
185 parser.add_argument('tests', nargs='+')
186 args = parser.parse_args()
187
Sanjay Patel16be4df92016-04-05 19:50:21 +0000[diff] [blame]188 autogenerated_note = (ADVERT + 'utils/' + os.path.basename(__file__))
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]189
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]190 opt_basename = os.path.basename(args.opt_binary)
191 if (opt_basename != "opt"):
192 print >>sys.stderr, 'ERROR: Unexpected opt name: ' + opt_basename
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]193 sys.exit(1)
194
195 for test in args.tests:
196 if args.verbose:
197 print >>sys.stderr, 'Scanning for RUN lines in test file: %s' % (test,)
198 with open(test) as f:
199 input_lines = [l.rstrip() for l in f]
200
Bryant Wong291264b2016-12-29 19:32:34 +0000[diff] [blame]201 raw_lines = [m.group(1)
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]202 for m in [common.RUN_LINE_RE.match(l) for l in input_lines] if m]
Bryant Wong291264b2016-12-29 19:32:34 +0000[diff] [blame]203 run_lines = [raw_lines[0]] if len(raw_lines) > 0 else []
204 for l in raw_lines[1:]:
Bryant Wong507256b2016-12-29 20:05:51 +0000[diff] [blame]205 if run_lines[-1].endswith("\\"):
206 run_lines[-1] = run_lines[-1].rstrip("\\") + " " + l
207 else:
208 run_lines.append(l)
Bryant Wong291264b2016-12-29 19:32:34 +0000[diff] [blame]209
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]210 if args.verbose:
211 print >>sys.stderr, 'Found %d RUN lines:' % (len(run_lines),)
212 for l in run_lines:
213 print >>sys.stderr, ' RUN: ' + l
214
215 prefix_list = []
216 for l in run_lines:
217 (tool_cmd, filecheck_cmd) = tuple([cmd.strip() for cmd in l.split('|', 1)])
218
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]219 if not tool_cmd.startswith(opt_basename + ' '):
220 print >>sys.stderr, 'WARNING: Skipping non-%s RUN line: %s' % (opt_basename, l)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]221 continue
222
223 if not filecheck_cmd.startswith('FileCheck '):
224 print >>sys.stderr, 'WARNING: Skipping non-FileChecked RUN line: ' + l
225 continue
226
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]227 tool_cmd_args = tool_cmd[len(opt_basename):].strip()
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]228 tool_cmd_args = tool_cmd_args.replace('< %s', '').replace('%s', '').strip()
229
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]230 check_prefixes = [item for m in common.CHECK_PREFIX_RE.finditer(filecheck_cmd)
Nikolai Bozhenov33ee40e2017-01-14 09:39:35 +0000[diff] [blame]231 for item in m.group(1).split(',')]
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]232 if not check_prefixes:
233 check_prefixes = ['CHECK']
234
235 # FIXME: We should use multiple check prefixes to common check lines. For
236 # now, we just ignore all but the last.
237 prefix_list.append((check_prefixes, tool_cmd_args))
238
239 func_dict = {}
240 for prefixes, _ in prefix_list:
241 for prefix in prefixes:
242 func_dict.update({prefix: dict()})
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]243 for prefixes, opt_args in prefix_list:
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]244 if args.verbose:
Sanjay Patelcae64a02017-06-12 17:44:30 +0000[diff] [blame]245 print >>sys.stderr, 'Extracted opt cmd: ' + opt_basename + ' ' + opt_args
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]246 print >>sys.stderr, 'Extracted FileCheck prefixes: ' + str(prefixes)
247
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]248 raw_tool_output = common.invoke_tool(args.opt_binary, opt_args, test)
249 common.build_function_body_dictionary(
250 OPT_FUNCTION_RE, scrub_body, [opt_basename], raw_tool_output,
251 prefixes, func_dict, args.verbose)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]252
253 is_in_function = False
254 is_in_function_start = False
255 prefix_set = set([prefix for prefixes, _ in prefix_list for prefix in prefixes])
256 if args.verbose:
257 print >>sys.stderr, 'Rewriting FileCheck prefixes: %s' % (prefix_set,)
258 output_lines = []
259 output_lines.append(autogenerated_note)
260
261 for input_line in input_lines:
262 if is_in_function_start:
263 if input_line == '':
264 continue
265 if input_line.lstrip().startswith(';'):
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]266 m = common.CHECK_RE.match(input_line)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]267 if not m or m.group(1) not in prefix_set:
268 output_lines.append(input_line)
269 continue
270
271 # Print out the various check lines here.
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]272 output_lines = add_checks(output_lines, prefix_list, func_dict, func_name, opt_basename)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]273 is_in_function_start = False
274
275 if is_in_function:
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]276 if common.should_add_line_to_output(input_line, prefix_set):
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]277 # This input line of the function body will go as-is into the output.
Sanjay Pateld8592712016-03-27 20:43:02 +0000[diff] [blame]278 # Except make leading whitespace uniform: 2 spaces.
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]279 input_line = common.SCRUB_LEADING_WHITESPACE_RE.sub(r' ', input_line)
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]280 output_lines.append(input_line)
281 else:
282 continue
283 if input_line.strip() == '}':
284 is_in_function = False
285 continue
286
Sanjay Patel16be4df92016-04-05 19:50:21 +0000[diff] [blame]287 # Discard any previous script advertising.
288 if input_line.startswith(ADVERT):
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]289 continue
290
291 # If it's outside a function, it just gets copied to the output.
292 output_lines.append(input_line)
293
294 m = IR_FUNCTION_RE.match(input_line)
295 if not m:
296 continue
Fangrui Songee4e2e72018-01-30 00:40:05 +0000[diff] [blame^]297 func_name = m.group(1)
298 if args.function is not None and func_name != args.function:
Sanjay Patelfff7a3d2016-03-24 23:19:26 +0000[diff] [blame]299 # When filtering on a specific function, skip all others.
300 continue
301 is_in_function = is_in_function_start = True
302
303 if args.verbose:
304 print>>sys.stderr, 'Writing %d lines to %s...' % (len(output_lines), test)
305
306 with open(test, 'wb') as f:
307 f.writelines([l + '\n' for l in output_lines])
308
309
310if __name__ == '__main__':
311 main()