Add a new utility script that helps update very simple regression tests.
This script is currently specific to x86 and limited to use with very
small regression or feature tests using 'llc' and 'FileCheck' in
a reasonably canonical way. It is in no way general purpose or robust at
this point. However, it works quite well for simple examples. Here is
the intended workflow:
- Make a change that requires updating N test files and M functions'
assertions within those files.
- Stash the change.
- Update those N test files' RUN-lines to look "canonical"[1].
- Refresh the FileCheck lines for either the entire file or select
functions by running this script.
- The script will parse the RUN lines and run the 'llc' binary you
give it according to each line, collecting the asm.
- It will then annotate each function with the appropriate FileCheck
comments to check every instruction from the start of the first
basic block to the last return.
- There will be numerous cases where the script either fails to remove
the old lines, or inserts checks which need to be manually editted,
but the manual edits tend to be deletions or replacements of
registers with FileCheck variables which are fast manual edits.
- A common pattern is to have the script insert complete checking of
every instruction, and then edit it down to only check the relevant
ones.
- Be careful to do all of these cleanups though! The script is
designed to make transferring and formatting the asm output of llc
into a test case fast, it is *not* designed to be authoratitive
about what constitutes a good test!
- Commit the nice fresh baseline of checks.
- Unstash your change and rebuild llc.
- Re-run script to regenerate the FileCheck annotations
- Remember to re-cleanup these annotations!!!
- Check the diff to make sure this is sane, checking the things you
expected it to, and check that the newly updated tests actually pass.
- Profit!
Also, I'm *terrible* at writing Python, and frankly I didn't spend a lot
of time making this script beautiful or well engineered. But it's useful
to me and may be useful to others so I thought I'd send it out.
http://reviews.llvm.org/D5546
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225618 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-12 04:43:18 +00:00
|
|
|
#!/usr/bin/env python2.7
|
|
|
|
|
|
|
|
|
|
"""A test case update script.
|
|
|
|
|
|
|
|
|
|
This script is a utility to update LLVM X86 'llc' based test cases with new
|
|
|
|
|
FileCheck patterns. It can either update all of the tests in the file or
|
|
|
|
|
a single test function.
|
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
import argparse
|
|
|
|
|
import itertools
|
|
|
|
|
import string
|
|
|
|
|
import subprocess
|
|
|
|
|
import sys
|
|
|
|
|
import tempfile
|
|
|
|
|
import re
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def llc(args, cmd_args, ir):
|
|
|
|
|
with open(ir) as ir_file:
|
|
|
|
|
stdout = subprocess.check_output(args.llc_binary + ' ' + cmd_args,
|
|
|
|
|
shell=True, stdin=ir_file)
|
|
|
|
|
return stdout
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ASM_SCRUB_WHITESPACE_RE = re.compile(r'(?!^(| \w))[ \t]+', flags=re.M)
|
2015-02-04 10:46:48 +00:00
|
|
|
ASM_SCRUB_TRAILING_WHITESPACE_RE = re.compile(r'[ \t]+$', flags=re.M)
|
Add a new utility script that helps update very simple regression tests.
This script is currently specific to x86 and limited to use with very
small regression or feature tests using 'llc' and 'FileCheck' in
a reasonably canonical way. It is in no way general purpose or robust at
this point. However, it works quite well for simple examples. Here is
the intended workflow:
- Make a change that requires updating N test files and M functions'
assertions within those files.
- Stash the change.
- Update those N test files' RUN-lines to look "canonical"[1].
- Refresh the FileCheck lines for either the entire file or select
functions by running this script.
- The script will parse the RUN lines and run the 'llc' binary you
give it according to each line, collecting the asm.
- It will then annotate each function with the appropriate FileCheck
comments to check every instruction from the start of the first
basic block to the last return.
- There will be numerous cases where the script either fails to remove
the old lines, or inserts checks which need to be manually editted,
but the manual edits tend to be deletions or replacements of
registers with FileCheck variables which are fast manual edits.
- A common pattern is to have the script insert complete checking of
every instruction, and then edit it down to only check the relevant
ones.
- Be careful to do all of these cleanups though! The script is
designed to make transferring and formatting the asm output of llc
into a test case fast, it is *not* designed to be authoratitive
about what constitutes a good test!
- Commit the nice fresh baseline of checks.
- Unstash your change and rebuild llc.
- Re-run script to regenerate the FileCheck annotations
- Remember to re-cleanup these annotations!!!
- Check the diff to make sure this is sane, checking the things you
expected it to, and check that the newly updated tests actually pass.
- Profit!
Also, I'm *terrible* at writing Python, and frankly I didn't spend a lot
of time making this script beautiful or well engineered. But it's useful
to me and may be useful to others so I thought I'd send it out.
http://reviews.llvm.org/D5546
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225618 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-12 04:43:18 +00:00
|
|
|
ASM_SCRUB_SHUFFLES_RE = (
|
|
|
|
|
re.compile(
|
|
|
|
|
r'^(\s*\w+) [^#\n]+#+ ((?:[xyz]mm\d+|mem) = .*)$',
|
|
|
|
|
flags=re.M))
|
|
|
|
|
ASM_SCRUB_SP_RE = re.compile(r'\d+\(%(esp|rsp)\)')
|
|
|
|
|
ASM_SCRUB_RIP_RE = re.compile(r'[.\w]+\(%rip\)')
|
|
|
|
|
ASM_SCRUB_KILL_COMMENT_RE = re.compile(r'^ *#+ +kill:.*\n')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def scrub_asm(asm):
|
|
|
|
|
# Scrub runs of whitespace out of the assembly, but leave the leading
|
|
|
|
|
# whitespace in place.
|
|
|
|
|
asm = ASM_SCRUB_WHITESPACE_RE.sub(r' ', asm)
|
|
|
|
|
# Expand the tabs used for indentation.
|
|
|
|
|
asm = string.expandtabs(asm, 2)
|
|
|
|
|
# Detect shuffle asm comments and hide the operands in favor of the comments.
|
|
|
|
|
asm = ASM_SCRUB_SHUFFLES_RE.sub(r'\1 {{.*#+}} \2', asm)
|
|
|
|
|
# Generically match the stack offset of a memory operand.
|
|
|
|
|
asm = ASM_SCRUB_SP_RE.sub(r'{{[0-9]+}}(%\1)', asm)
|
|
|
|
|
# Generically match a RIP-relative memory operand.
|
|
|
|
|
asm = ASM_SCRUB_RIP_RE.sub(r'{{.*}}(%rip)', asm)
|
|
|
|
|
# Strip kill operands inserted into the asm.
|
|
|
|
|
asm = ASM_SCRUB_KILL_COMMENT_RE.sub('', asm)
|
2015-02-04 10:46:48 +00:00
|
|
|
# Strip trailing whitespace.
|
|
|
|
|
asm = ASM_SCRUB_TRAILING_WHITESPACE_RE.sub(r'', asm)
|
Add a new utility script that helps update very simple regression tests.
This script is currently specific to x86 and limited to use with very
small regression or feature tests using 'llc' and 'FileCheck' in
a reasonably canonical way. It is in no way general purpose or robust at
this point. However, it works quite well for simple examples. Here is
the intended workflow:
- Make a change that requires updating N test files and M functions'
assertions within those files.
- Stash the change.
- Update those N test files' RUN-lines to look "canonical"[1].
- Refresh the FileCheck lines for either the entire file or select
functions by running this script.
- The script will parse the RUN lines and run the 'llc' binary you
give it according to each line, collecting the asm.
- It will then annotate each function with the appropriate FileCheck
comments to check every instruction from the start of the first
basic block to the last return.
- There will be numerous cases where the script either fails to remove
the old lines, or inserts checks which need to be manually editted,
but the manual edits tend to be deletions or replacements of
registers with FileCheck variables which are fast manual edits.
- A common pattern is to have the script insert complete checking of
every instruction, and then edit it down to only check the relevant
ones.
- Be careful to do all of these cleanups though! The script is
designed to make transferring and formatting the asm output of llc
into a test case fast, it is *not* designed to be authoratitive
about what constitutes a good test!
- Commit the nice fresh baseline of checks.
- Unstash your change and rebuild llc.
- Re-run script to regenerate the FileCheck annotations
- Remember to re-cleanup these annotations!!!
- Check the diff to make sure this is sane, checking the things you
expected it to, and check that the newly updated tests actually pass.
- Profit!
Also, I'm *terrible* at writing Python, and frankly I didn't spend a lot
of time making this script beautiful or well engineered. But it's useful
to me and may be useful to others so I thought I'd send it out.
http://reviews.llvm.org/D5546
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225618 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-12 04:43:18 +00:00
|
|
|
return asm
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def main():
|
|
|
|
|
parser = argparse.ArgumentParser(description=__doc__)
|
|
|
|
|
parser.add_argument('-v', '--verbose', action='store_true',
|
|
|
|
|
help='Show verbose output')
|
|
|
|
|
parser.add_argument('--llc-binary', default='llc',
|
|
|
|
|
help='The "llc" binary to use to generate the test case')
|
|
|
|
|
parser.add_argument(
|
|
|
|
|
'--function', help='The function in the test file to update')
|
|
|
|
|
parser.add_argument('tests', nargs='+')
|
|
|
|
|
args = parser.parse_args()
|
|
|
|
|
|
|
|
|
|
run_line_re = re.compile('^\s*;\s*RUN:\s*(.*)$')
|
|
|
|
|
ir_function_re = re.compile('^\s*define\s+(?:internal\s+)?[^@]*@(\w+)\s*\(')
|
|
|
|
|
asm_function_re = re.compile(
|
|
|
|
|
r'^_?(?P<f>[^:]+):[ \t]*#+[ \t]*@(?P=f)\n[^:]*?'
|
|
|
|
|
r'(?P<body>^##?[ \t]+[^:]+:.*?)\s*'
|
2015-02-15 00:08:01 +00:00
|
|
|
r'^\s*(?:[^:\n]+?:\s*\n\s*\.size|\.cfi_endproc|\.globl|\.comm|\.(?:sub)?section)',
|
Add a new utility script that helps update very simple regression tests.
This script is currently specific to x86 and limited to use with very
small regression or feature tests using 'llc' and 'FileCheck' in
a reasonably canonical way. It is in no way general purpose or robust at
this point. However, it works quite well for simple examples. Here is
the intended workflow:
- Make a change that requires updating N test files and M functions'
assertions within those files.
- Stash the change.
- Update those N test files' RUN-lines to look "canonical"[1].
- Refresh the FileCheck lines for either the entire file or select
functions by running this script.
- The script will parse the RUN lines and run the 'llc' binary you
give it according to each line, collecting the asm.
- It will then annotate each function with the appropriate FileCheck
comments to check every instruction from the start of the first
basic block to the last return.
- There will be numerous cases where the script either fails to remove
the old lines, or inserts checks which need to be manually editted,
but the manual edits tend to be deletions or replacements of
registers with FileCheck variables which are fast manual edits.
- A common pattern is to have the script insert complete checking of
every instruction, and then edit it down to only check the relevant
ones.
- Be careful to do all of these cleanups though! The script is
designed to make transferring and formatting the asm output of llc
into a test case fast, it is *not* designed to be authoratitive
about what constitutes a good test!
- Commit the nice fresh baseline of checks.
- Unstash your change and rebuild llc.
- Re-run script to regenerate the FileCheck annotations
- Remember to re-cleanup these annotations!!!
- Check the diff to make sure this is sane, checking the things you
expected it to, and check that the newly updated tests actually pass.
- Profit!
Also, I'm *terrible* at writing Python, and frankly I didn't spend a lot
of time making this script beautiful or well engineered. But it's useful
to me and may be useful to others so I thought I'd send it out.
http://reviews.llvm.org/D5546
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225618 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-12 04:43:18 +00:00
|
|
|
flags=(re.M | re.S))
|
|
|
|
|
check_prefix_re = re.compile('--check-prefix=(\S+)')
|
|
|
|
|
check_re = re.compile(r'^\s*;\s*([^:]+?)(?:-NEXT|-NOT|-DAG|-LABEL)?:')
|
|
|
|
|
|
|
|
|
|
for test in args.tests:
|
|
|
|
|
if args.verbose:
|
|
|
|
|
print >>sys.stderr, 'Scanning for RUN lines in test file: %s' % (test,)
|
|
|
|
|
with open(test) as f:
|
|
|
|
|
test_lines = [l.rstrip() for l in f]
|
|
|
|
|
|
|
|
|
|
run_lines = [m.group(1)
|
|
|
|
|
for m in [run_line_re.match(l) for l in test_lines] if m]
|
|
|
|
|
if args.verbose:
|
|
|
|
|
print >>sys.stderr, 'Found %d RUN lines:' % (len(run_lines),)
|
|
|
|
|
for l in run_lines:
|
|
|
|
|
print >>sys.stderr, ' RUN: ' + l
|
|
|
|
|
|
|
|
|
|
checks = []
|
|
|
|
|
for l in run_lines:
|
|
|
|
|
(llc_cmd, filecheck_cmd) = tuple([cmd.strip() for cmd in l.split('|', 1)])
|
|
|
|
|
if not llc_cmd.startswith('llc '):
|
|
|
|
|
print >>sys.stderr, 'WARNING: Skipping non-llc RUN line: ' + l
|
|
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
if not filecheck_cmd.startswith('FileCheck '):
|
|
|
|
|
print >>sys.stderr, 'WARNING: Skipping non-FileChecked RUN line: ' + l
|
|
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
llc_cmd_args = llc_cmd[len('llc'):].strip()
|
|
|
|
|
llc_cmd_args = llc_cmd_args.replace('< %s', '').replace('%s', '').strip()
|
|
|
|
|
|
|
|
|
|
check_prefixes = [m.group(1)
|
|
|
|
|
for m in check_prefix_re.finditer(filecheck_cmd)]
|
|
|
|
|
if not check_prefixes:
|
|
|
|
|
check_prefixes = ['CHECK']
|
|
|
|
|
|
|
|
|
|
# FIXME: We should use multiple check prefixes to common check lines. For
|
|
|
|
|
# now, we just ignore all but the last.
|
|
|
|
|
checks.append((check_prefixes, llc_cmd_args))
|
|
|
|
|
|
|
|
|
|
asm = {}
|
|
|
|
|
for prefixes, _ in checks:
|
|
|
|
|
for prefix in prefixes:
|
|
|
|
|
asm.update({prefix: dict()})
|
|
|
|
|
for prefixes, llc_args in checks:
|
|
|
|
|
if args.verbose:
|
|
|
|
|
print >>sys.stderr, 'Extracted LLC cmd: llc ' + llc_args
|
|
|
|
|
print >>sys.stderr, 'Extracted FileCheck prefixes: ' + str(prefixes)
|
|
|
|
|
raw_asm = llc(args, llc_args, test)
|
|
|
|
|
# Build up a dictionary of all the function bodies.
|
|
|
|
|
for m in asm_function_re.finditer(raw_asm):
|
|
|
|
|
if not m:
|
|
|
|
|
continue
|
|
|
|
|
f = m.group('f')
|
|
|
|
|
f_asm = scrub_asm(m.group('body'))
|
2015-02-03 21:26:45 +00:00
|
|
|
if f.startswith('stress'):
|
|
|
|
|
# We only use the last line of the asm for stress tests.
|
|
|
|
|
f_asm = '\n'.join(f_asm.splitlines()[-1:])
|
Add a new utility script that helps update very simple regression tests.
This script is currently specific to x86 and limited to use with very
small regression or feature tests using 'llc' and 'FileCheck' in
a reasonably canonical way. It is in no way general purpose or robust at
this point. However, it works quite well for simple examples. Here is
the intended workflow:
- Make a change that requires updating N test files and M functions'
assertions within those files.
- Stash the change.
- Update those N test files' RUN-lines to look "canonical"[1].
- Refresh the FileCheck lines for either the entire file or select
functions by running this script.
- The script will parse the RUN lines and run the 'llc' binary you
give it according to each line, collecting the asm.
- It will then annotate each function with the appropriate FileCheck
comments to check every instruction from the start of the first
basic block to the last return.
- There will be numerous cases where the script either fails to remove
the old lines, or inserts checks which need to be manually editted,
but the manual edits tend to be deletions or replacements of
registers with FileCheck variables which are fast manual edits.
- A common pattern is to have the script insert complete checking of
every instruction, and then edit it down to only check the relevant
ones.
- Be careful to do all of these cleanups though! The script is
designed to make transferring and formatting the asm output of llc
into a test case fast, it is *not* designed to be authoratitive
about what constitutes a good test!
- Commit the nice fresh baseline of checks.
- Unstash your change and rebuild llc.
- Re-run script to regenerate the FileCheck annotations
- Remember to re-cleanup these annotations!!!
- Check the diff to make sure this is sane, checking the things you
expected it to, and check that the newly updated tests actually pass.
- Profit!
Also, I'm *terrible* at writing Python, and frankly I didn't spend a lot
of time making this script beautiful or well engineered. But it's useful
to me and may be useful to others so I thought I'd send it out.
http://reviews.llvm.org/D5546
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225618 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-12 04:43:18 +00:00
|
|
|
if args.verbose:
|
|
|
|
|
print >>sys.stderr, 'Processing asm for function: ' + f
|
|
|
|
|
for l in f_asm.splitlines():
|
|
|
|
|
print >>sys.stderr, ' ' + l
|
|
|
|
|
for prefix in prefixes:
|
|
|
|
|
if f in asm[prefix] and asm[prefix][f] != f_asm:
|
|
|
|
|
if prefix == prefixes[-1]:
|
|
|
|
|
print >>sys.stderr, ('WARNING: Found conflicting asm under the '
|
|
|
|
|
'same prefix!')
|
|
|
|
|
else:
|
|
|
|
|
asm[prefix][f] = None
|
|
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
asm[prefix][f] = f_asm
|
|
|
|
|
|
|
|
|
|
is_in_function = False
|
|
|
|
|
is_in_function_start = False
|
|
|
|
|
prefix_set = set([prefix for prefixes, _ in checks for prefix in prefixes])
|
|
|
|
|
if args.verbose:
|
|
|
|
|
print >>sys.stderr, 'Rewriting FileCheck prefixes: %s' % (prefix_set,)
|
|
|
|
|
fixed_lines = []
|
|
|
|
|
for l in test_lines:
|
|
|
|
|
if is_in_function_start:
|
|
|
|
|
if l.lstrip().startswith(';'):
|
|
|
|
|
m = check_re.match(l)
|
|
|
|
|
if not m or m.group(1) not in prefix_set:
|
|
|
|
|
fixed_lines.append(l)
|
|
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
# Print out the various check lines here
|
|
|
|
|
printed_prefixes = []
|
|
|
|
|
for prefixes, _ in checks:
|
|
|
|
|
for prefix in prefixes:
|
|
|
|
|
if prefix in printed_prefixes:
|
|
|
|
|
break
|
|
|
|
|
if not asm[prefix][name]:
|
|
|
|
|
continue
|
|
|
|
|
if len(printed_prefixes) != 0:
|
|
|
|
|
fixed_lines.append(';')
|
|
|
|
|
printed_prefixes.append(prefix)
|
|
|
|
|
fixed_lines.append('; %s-LABEL: %s:' % (prefix, name))
|
|
|
|
|
asm_lines = asm[prefix][name].splitlines()
|
|
|
|
|
fixed_lines.append('; %s: %s' % (prefix, asm_lines[0]))
|
|
|
|
|
for asm_line in asm_lines[1:]:
|
|
|
|
|
fixed_lines.append('; %s-NEXT: %s' % (prefix, asm_line))
|
|
|
|
|
break
|
|
|
|
|
is_in_function_start = False
|
|
|
|
|
|
|
|
|
|
if is_in_function:
|
|
|
|
|
# Skip any blank comment lines in the IR.
|
|
|
|
|
if l.strip() == ';':
|
|
|
|
|
continue
|
|
|
|
|
# And skip any CHECK lines. We'll build our own.
|
|
|
|
|
m = check_re.match(l)
|
|
|
|
|
if m and m.group(1) in prefix_set:
|
|
|
|
|
continue
|
|
|
|
|
# Collect the remaining lines in the function body and look for the end
|
|
|
|
|
# of the function.
|
|
|
|
|
fixed_lines.append(l)
|
|
|
|
|
if l.strip() == '}':
|
|
|
|
|
is_in_function = False
|
|
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
fixed_lines.append(l)
|
|
|
|
|
|
|
|
|
|
m = ir_function_re.match(l)
|
|
|
|
|
if not m:
|
|
|
|
|
continue
|
|
|
|
|
name = m.group(1)
|
|
|
|
|
if args.function is not None and name != args.function:
|
|
|
|
|
# When filtering on a specific function, skip all others.
|
|
|
|
|
continue
|
|
|
|
|
is_in_function = is_in_function_start = True
|
|
|
|
|
|
|
|
|
|
if args.verbose:
|
|
|
|
|
print>>sys.stderr, 'Writing %d fixed lines to %s...' % (
|
|
|
|
|
len(fixed_lines), test)
|
|
|
|
|
with open(test, 'w') as f:
|
|
|
|
|
f.writelines([l + '\n' for l in fixed_lines])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
|
|
main()
|