2019-11-27 02:40:52 +08:00
|
|
|
#!/usr/bin/env python3
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
|
|
|
|
"""A test case update script.
|
|
|
|
|
|
2018-01-30 08:40:05 +08:00
|
|
|
This script is a utility to update LLVM 'llc' based test cases with new
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
FileCheck patterns. It can either update all of the tests in the file or
|
|
|
|
|
a single test function.
|
|
|
|
|
"""
|
|
|
|
|
|
2019-01-03 22:11:33 +08:00
|
|
|
from __future__ import print_function
|
|
|
|
|
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
import argparse
|
2020-07-08 17:59:50 +08:00
|
|
|
import os # Used to advertise this file's name ("autogenerated_note").
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
2022-03-01 17:54:24 +08:00
|
|
|
from UpdateTestChecks import common
|
2016-03-25 01:15:42 +08:00
|
|
|
|
2020-06-02 02:50:53 +08:00
|
|
|
# llc is the only llc-like in the LLVM tree but downstream forks can add
|
|
|
|
|
# additional ones here if they have them.
|
2020-07-08 17:59:50 +08:00
|
|
|
LLC_LIKE_TOOLS = ('llc',)
|
2016-03-25 01:15:42 +08:00
|
|
|
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
def main():
|
|
|
|
|
parser = argparse.ArgumentParser(description=__doc__)
|
2020-06-02 02:50:53 +08:00
|
|
|
parser.add_argument('--llc-binary', default=None,
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
help='The "llc" binary to use to generate the test case')
|
|
|
|
|
parser.add_argument(
|
|
|
|
|
'--function', help='The function in the test file to update')
|
2017-10-24 22:32:52 +08:00
|
|
|
parser.add_argument(
|
2018-06-01 21:37:01 +08:00
|
|
|
'--extra_scrub', action='store_true',
|
|
|
|
|
help='Always use additional regex to further reduce diffs between various subtargets')
|
2020-12-13 01:11:13 +08:00
|
|
|
parser.add_argument(
|
|
|
|
|
'--x86_scrub_sp', action='store_true', default=True,
|
|
|
|
|
help='Use regex for x86 sp matching to reduce diffs between various subtargets')
|
|
|
|
|
parser.add_argument(
|
|
|
|
|
'--no_x86_scrub_sp', action='store_false', dest='x86_scrub_sp')
|
2018-07-24 05:14:35 +08:00
|
|
|
parser.add_argument(
|
2021-05-22 10:26:15 +08:00
|
|
|
'--x86_scrub_rip', action='store_true', default=False,
|
2020-12-13 01:11:13 +08:00
|
|
|
help='Use more regex for x86 rip matching to reduce diffs between various subtargets')
|
2018-07-24 05:14:35 +08:00
|
|
|
parser.add_argument(
|
|
|
|
|
'--no_x86_scrub_rip', action='store_false', dest='x86_scrub_rip')
|
2020-02-20 22:22:56 +08:00
|
|
|
parser.add_argument(
|
|
|
|
|
'--no_x86_scrub_mem_shuffle', action='store_true', default=False,
|
|
|
|
|
help='Reduce scrubbing shuffles with memory operands')
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
parser.add_argument('tests', nargs='+')
|
2020-07-08 17:59:50 +08:00
|
|
|
initial_args = common.parse_commandline_args(parser)
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
2019-08-07 22:44:50 +08:00
|
|
|
script_name = os.path.basename(__file__)
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
2020-07-08 17:59:50 +08:00
|
|
|
for ti in common.itertests(initial_args.tests, parser,
|
|
|
|
|
script_name='utils/' + script_name):
|
2016-12-23 04:59:39 +08:00
|
|
|
triple_in_ir = None
|
2020-07-08 17:59:50 +08:00
|
|
|
for l in ti.input_lines:
|
2018-01-30 08:40:05 +08:00
|
|
|
m = common.TRIPLE_IR_RE.match(l)
|
2016-12-23 04:59:39 +08:00
|
|
|
if m:
|
|
|
|
|
triple_in_ir = m.groups()[0]
|
|
|
|
|
break
|
|
|
|
|
|
|
|
|
|
run_list = []
|
2020-07-08 17:59:50 +08:00
|
|
|
for l in ti.run_lines:
|
2019-08-07 22:44:50 +08:00
|
|
|
if '|' not in l:
|
|
|
|
|
common.warn('Skipping unparseable RUN line: ' + l)
|
|
|
|
|
continue
|
|
|
|
|
|
2021-04-21 19:19:08 +08:00
|
|
|
commands = [cmd.strip() for cmd in l.split('|')]
|
|
|
|
|
assert len(commands) >= 2
|
|
|
|
|
preprocess_cmd = None
|
|
|
|
|
if len(commands) > 2:
|
|
|
|
|
preprocess_cmd = " | ".join(commands[:-2])
|
|
|
|
|
llc_cmd = commands[-2]
|
|
|
|
|
filecheck_cmd = commands[-1]
|
2019-09-19 07:44:16 +08:00
|
|
|
llc_tool = llc_cmd.split(' ')[0]
|
2016-12-23 04:59:39 +08:00
|
|
|
|
|
|
|
|
triple_in_cmd = None
|
2018-01-30 08:40:05 +08:00
|
|
|
m = common.TRIPLE_ARG_RE.search(llc_cmd)
|
2016-12-23 04:59:39 +08:00
|
|
|
if m:
|
|
|
|
|
triple_in_cmd = m.groups()[0]
|
|
|
|
|
|
UpdateTestChecks: fix AMDGPU handling
Summary:
Was looking into supporting `(srl (shl x, c1), c2)` with c1 != c2 in dagcombiner,
this test changes, but makes `update_llc_test_checks.py` unhappy.
**Many** AMDGPU tests specify `-march`, not `-mtriple`, which results in `update_llc_test_checks.py`
defaulting to x86 asm function detection heuristics, which don't work here.
I propose to fix this by adding an infrastructure to map from `-march` to `-mtriple`,
in the UpdateTestChecks tooling.
Reviewers: RKSimon, MaskRay, arsenm
Reviewed By: arsenm
Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62099
llvm-svn: 361101
2019-05-18 21:00:03 +08:00
|
|
|
march_in_cmd = None
|
|
|
|
|
m = common.MARCH_ARG_RE.search(llc_cmd)
|
|
|
|
|
if m:
|
|
|
|
|
march_in_cmd = m.groups()[0]
|
|
|
|
|
|
2022-03-01 17:54:24 +08:00
|
|
|
m = common.DEBUG_ONLY_ARG_RE.search(llc_cmd)
|
|
|
|
|
if m and m.groups()[0] == 'isel':
|
|
|
|
|
from UpdateTestChecks import isel as output_type
|
|
|
|
|
else:
|
|
|
|
|
from UpdateTestChecks import asm as output_type
|
|
|
|
|
|
[UpdateTestChecks] Emit warning when invalid value for -check-prefix(es) option
Summary:
The script is silent for the following issue:
FileCheck %s -check-prefix=CHECK,POPCOUNT
FileCheck will catch it later, but I think we can warn here too.
Now it warns:
./update_llc_test_checks.py file.ll
WARNING: Supplied prefix 'CHECK,POPCOUNT' is invalid. Prefix must contain only alphanumeric characters, hyphens and underscores. Did you mean --check-prefixes=CHECK,POPCOUNT?
Reviewers: lebedev.ri, spatel, RKSimon, craig.topper, nikic, gbedwell
Reviewed By: RKSimon
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64589
llvm-svn: 367244
2019-07-30 01:41:00 +08:00
|
|
|
common.verify_filecheck_prefixes(filecheck_cmd)
|
2020-06-02 02:50:53 +08:00
|
|
|
if llc_tool not in LLC_LIKE_TOOLS:
|
2019-08-07 22:44:50 +08:00
|
|
|
common.warn('Skipping non-llc RUN line: ' + l)
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
if not filecheck_cmd.startswith('FileCheck '):
|
2019-08-07 22:44:50 +08:00
|
|
|
common.warn('Skipping non-FileChecked RUN line: ' + l)
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
continue
|
|
|
|
|
|
2019-09-19 07:44:16 +08:00
|
|
|
llc_cmd_args = llc_cmd[len(llc_tool):].strip()
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
llc_cmd_args = llc_cmd_args.replace('< %s', '').replace('%s', '').strip()
|
2020-07-08 17:59:50 +08:00
|
|
|
if ti.path.endswith('.mir'):
|
[utils] Add minimal support for MIR inputs to update_llc_test_checks.py
update_{llc,mir}_test_checks.py applicability is determined by the
output (assembly or MIR), not the input, which makes
update_llc_test_checks.py the right tool to generate tests that start at
MIR and stop at the final assembly.
This commit adds the minimal support for this path. Main limitation that
remains:
- MIR has to have LLVM IR section, and the CHECK lines will be inserted
into the LLVM IR functions that correspond to the MIR functions.
Running
../utils/update_llc_test_checks.py --llc-binary ./bin/llc
on a slightly modified ../test/CodeGen/X86/bad-tls-fold.mir
produces the following diff:
+# NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+# RUN: llc %s -o - | FileCheck %s
--- |
target triple = "x86_64-unknown-linux-gnu"
@@ -6,17 +7,31 @@
@i = external thread_local global i32
define i32 @or() {
+ ; CHECK-LABEL: or:
+ ; CHECK: # %bb.0: # %entry
+ ; CHECK-NEXT: movq {{.*}}(%rip), %rax
+ ; CHECK-NEXT: orq $7, %rax
+ ; CHECK-NEXT: movq i@{{.*}}(%rip), %rcx
+ ; CHECK-NEXT: orq %rax, %rcx
+ ; CHECK-NEXT: movl %fs:(%rcx), %eax
+ ; CHECK-NEXT: retq
entry:
ret i32 undef
}
-
define i32 @and() {
+ ; CHECK-LABEL: and:
+ ; CHECK: # %bb.0: # %entry
+ ; CHECK-NEXT: movq {{.*}}(%rip), %rax
+ ; CHECK-NEXT: orq $7, %rax
+ ; CHECK-NEXT: movq i@{{.*}}(%rip), %rcx
+ ; CHECK-NEXT: andq %rax, %rcx
+ ; CHECK-NEXT: movl %fs:(%rcx), %eax
+ ; CHECK-NEXT: retq
entry:
ret i32 undef
}
...
(not applied)
llvm-svn: 372277
2019-09-19 07:44:17 +08:00
|
|
|
llc_cmd_args += ' -x mir'
|
2018-01-30 08:40:05 +08:00
|
|
|
check_prefixes = [item for m in common.CHECK_PREFIX_RE.finditer(filecheck_cmd)
|
2017-01-14 17:39:35 +08:00
|
|
|
for item in m.group(1).split(',')]
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
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.
|
2021-04-21 19:19:08 +08:00
|
|
|
run_list.append((check_prefixes, llc_tool, llc_cmd_args, preprocess_cmd,
|
|
|
|
|
triple_in_cmd, march_in_cmd))
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
2020-07-08 17:59:50 +08:00
|
|
|
if ti.path.endswith('.mir'):
|
2019-12-02 18:50:23 +08:00
|
|
|
check_indent = ' '
|
|
|
|
|
else:
|
|
|
|
|
check_indent = ''
|
[utils] Add minimal support for MIR inputs to update_llc_test_checks.py
update_{llc,mir}_test_checks.py applicability is determined by the
output (assembly or MIR), not the input, which makes
update_llc_test_checks.py the right tool to generate tests that start at
MIR and stop at the final assembly.
This commit adds the minimal support for this path. Main limitation that
remains:
- MIR has to have LLVM IR section, and the CHECK lines will be inserted
into the LLVM IR functions that correspond to the MIR functions.
Running
../utils/update_llc_test_checks.py --llc-binary ./bin/llc
on a slightly modified ../test/CodeGen/X86/bad-tls-fold.mir
produces the following diff:
+# NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+# RUN: llc %s -o - | FileCheck %s
--- |
target triple = "x86_64-unknown-linux-gnu"
@@ -6,17 +7,31 @@
@i = external thread_local global i32
define i32 @or() {
+ ; CHECK-LABEL: or:
+ ; CHECK: # %bb.0: # %entry
+ ; CHECK-NEXT: movq {{.*}}(%rip), %rax
+ ; CHECK-NEXT: orq $7, %rax
+ ; CHECK-NEXT: movq i@{{.*}}(%rip), %rcx
+ ; CHECK-NEXT: orq %rax, %rcx
+ ; CHECK-NEXT: movl %fs:(%rcx), %eax
+ ; CHECK-NEXT: retq
entry:
ret i32 undef
}
-
define i32 @and() {
+ ; CHECK-LABEL: and:
+ ; CHECK: # %bb.0: # %entry
+ ; CHECK-NEXT: movq {{.*}}(%rip), %rax
+ ; CHECK-NEXT: orq $7, %rax
+ ; CHECK-NEXT: movq i@{{.*}}(%rip), %rcx
+ ; CHECK-NEXT: andq %rax, %rcx
+ ; CHECK-NEXT: movl %fs:(%rcx), %eax
+ ; CHECK-NEXT: retq
entry:
ret i32 undef
}
...
(not applied)
llvm-svn: 372277
2019-09-19 07:44:17 +08:00
|
|
|
|
2020-12-17 02:20:12 +08:00
|
|
|
builder = common.FunctionTestBuilder(
|
2021-05-06 02:46:02 +08:00
|
|
|
run_list=run_list,
|
2020-12-17 02:20:12 +08:00
|
|
|
flags=type('', (object,), {
|
|
|
|
|
'verbose': ti.args.verbose,
|
2022-01-31 23:06:08 +08:00
|
|
|
'filters': ti.args.filters,
|
2020-12-17 02:20:12 +08:00
|
|
|
'function_signature': False,
|
2021-02-20 12:22:14 +08:00
|
|
|
'check_attributes': False,
|
2021-05-06 02:46:02 +08:00
|
|
|
'replace_value_regex': []}),
|
2021-06-20 18:14:33 +08:00
|
|
|
scrubber_args=[ti.args],
|
|
|
|
|
path=ti.path)
|
2020-12-17 02:20:12 +08:00
|
|
|
|
2021-04-21 19:19:08 +08:00
|
|
|
for prefixes, llc_tool, llc_args, preprocess_cmd, triple_in_cmd, march_in_cmd in run_list:
|
2019-12-02 18:50:23 +08:00
|
|
|
common.debug('Extracted LLC cmd:', llc_tool, llc_args)
|
|
|
|
|
common.debug('Extracted FileCheck prefixes:', str(prefixes))
|
2016-03-25 01:15:42 +08:00
|
|
|
|
2020-12-17 02:20:12 +08:00
|
|
|
raw_tool_output = common.invoke_tool(ti.args.llc_binary or llc_tool,
|
2021-04-21 19:19:08 +08:00
|
|
|
llc_args, ti.path, preprocess_cmd,
|
|
|
|
|
verbose=ti.args.verbose)
|
UpdateTestChecks: fix AMDGPU handling
Summary:
Was looking into supporting `(srl (shl x, c1), c2)` with c1 != c2 in dagcombiner,
this test changes, but makes `update_llc_test_checks.py` unhappy.
**Many** AMDGPU tests specify `-march`, not `-mtriple`, which results in `update_llc_test_checks.py`
defaulting to x86 asm function detection heuristics, which don't work here.
I propose to fix this by adding an infrastructure to map from `-march` to `-mtriple`,
in the UpdateTestChecks tooling.
Reviewers: RKSimon, MaskRay, arsenm
Reviewed By: arsenm
Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62099
llvm-svn: 361101
2019-05-18 21:00:03 +08:00
|
|
|
triple = triple_in_cmd or triple_in_ir
|
|
|
|
|
if not triple:
|
2022-03-01 17:54:24 +08:00
|
|
|
triple = common.get_triple_from_march(march_in_cmd)
|
2016-12-23 04:59:39 +08:00
|
|
|
|
2022-03-01 17:54:24 +08:00
|
|
|
scrubber, function_re = output_type.get_run_handler(triple)
|
2021-07-29 18:55:34 +08:00
|
|
|
builder.process_run_line(function_re, scrubber, raw_tool_output, prefixes, True)
|
2020-12-17 02:20:12 +08:00
|
|
|
|
|
|
|
|
func_dict = builder.finish_and_get_func_dict()
|
2022-05-12 18:21:55 +08:00
|
|
|
global_vars_seen_dict = {}
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
|
|
|
|
is_in_function = False
|
|
|
|
|
is_in_function_start = False
|
2016-11-08 02:08:19 +08:00
|
|
|
func_name = None
|
2016-12-23 04:59:39 +08:00
|
|
|
prefix_set = set([prefix for p in run_list for prefix in p[0]])
|
2019-12-02 18:50:23 +08:00
|
|
|
common.debug('Rewriting FileCheck prefixes:', str(prefix_set))
|
2016-03-25 01:15:42 +08:00
|
|
|
output_lines = []
|
2020-01-14 02:16:35 +08:00
|
|
|
|
|
|
|
|
include_generated_funcs = common.find_arg_in_test(ti,
|
|
|
|
|
lambda args: ti.args.include_generated_funcs,
|
|
|
|
|
'--include-generated-funcs',
|
|
|
|
|
True)
|
|
|
|
|
|
[UpdateTestChecks] Auto-generate stub bodies for unused prefixes
This is scoped to autogenerated tests.
The goal is to support having each RUN line specify a list of
check-prefixes where one can specify potentially redundant prefixes. For example,
for X86, if one specified prefixes for both AVX1 and AVX2, and the codegen happened to
match today, one of the prefixes would be used and the onther one not.
If the unused prefix were dropped, and later, codegen differences were
introduced, one would have to go figure out where to add what prefix
(paraphrasing
https://lists.llvm.org/pipermail/llvm-dev/2021-February/148326.html)
To avoid getting errors due to unused prefixes, whole directories can be
opted out (as discussed on that thread), but that means that tests that
aren't autogenerated in such directories could have undetected unused
prefix bugs.
This patch proposes an alternative that both avoids the above, dir-level
optout, and supports the main autogen scenario discussed first. The autogen
tool appends at the end of the test file the list of unused prefixes,
together with a note explaining that is the case. Each prefix is set up
to always pass.
This way, unexpected unused prefixes are easily discoverable, and
expected cases "just work".
Differential Revision: https://reviews.llvm.org/D124306
2022-04-23 03:49:15 +08:00
|
|
|
generated_prefixes = []
|
2020-01-14 02:16:35 +08:00
|
|
|
if include_generated_funcs:
|
|
|
|
|
# Generate the appropriate checks for each function. We need to emit
|
|
|
|
|
# these in the order according to the generated output so that CHECK-LABEL
|
|
|
|
|
# works properly. func_order provides that.
|
|
|
|
|
|
|
|
|
|
# We can't predict where various passes might insert functions so we can't
|
|
|
|
|
# be sure the input function order is maintained. Therefore, first spit
|
|
|
|
|
# out all the source lines.
|
|
|
|
|
common.dump_input_lines(output_lines, ti, prefix_set, ';')
|
|
|
|
|
|
|
|
|
|
# Now generate all the checks.
|
[UpdateTestChecks] Auto-generate stub bodies for unused prefixes
This is scoped to autogenerated tests.
The goal is to support having each RUN line specify a list of
check-prefixes where one can specify potentially redundant prefixes. For example,
for X86, if one specified prefixes for both AVX1 and AVX2, and the codegen happened to
match today, one of the prefixes would be used and the onther one not.
If the unused prefix were dropped, and later, codegen differences were
introduced, one would have to go figure out where to add what prefix
(paraphrasing
https://lists.llvm.org/pipermail/llvm-dev/2021-February/148326.html)
To avoid getting errors due to unused prefixes, whole directories can be
opted out (as discussed on that thread), but that means that tests that
aren't autogenerated in such directories could have undetected unused
prefix bugs.
This patch proposes an alternative that both avoids the above, dir-level
optout, and supports the main autogen scenario discussed first. The autogen
tool appends at the end of the test file the list of unused prefixes,
together with a note explaining that is the case. Each prefix is set up
to always pass.
This way, unexpected unused prefixes are easily discoverable, and
expected cases "just work".
Differential Revision: https://reviews.llvm.org/D124306
2022-04-23 03:49:15 +08:00
|
|
|
generated_prefixes = common.add_checks_at_end(
|
|
|
|
|
output_lines, run_list, builder.func_order(),
|
|
|
|
|
check_indent + ';',
|
|
|
|
|
lambda my_output_lines, prefixes, func:
|
|
|
|
|
output_type.add_checks(my_output_lines,
|
|
|
|
|
check_indent + ';',
|
|
|
|
|
prefixes, func_dict, func,
|
|
|
|
|
global_vars_seen_dict,
|
|
|
|
|
is_filtered=builder.is_filtered()))
|
2020-01-14 02:16:35 +08:00
|
|
|
else:
|
|
|
|
|
for input_info in ti.iterlines(output_lines):
|
|
|
|
|
input_line = input_info.line
|
|
|
|
|
args = input_info.args
|
|
|
|
|
if is_in_function_start:
|
|
|
|
|
if input_line == '':
|
|
|
|
|
continue
|
|
|
|
|
if input_line.lstrip().startswith(';'):
|
|
|
|
|
m = common.CHECK_RE.match(input_line)
|
|
|
|
|
if not m or m.group(1) not in prefix_set:
|
|
|
|
|
output_lines.append(input_line)
|
|
|
|
|
continue
|
|
|
|
|
|
|
|
|
|
# Print out the various check lines here.
|
[UpdateTestChecks] Auto-generate stub bodies for unused prefixes
This is scoped to autogenerated tests.
The goal is to support having each RUN line specify a list of
check-prefixes where one can specify potentially redundant prefixes. For example,
for X86, if one specified prefixes for both AVX1 and AVX2, and the codegen happened to
match today, one of the prefixes would be used and the onther one not.
If the unused prefix were dropped, and later, codegen differences were
introduced, one would have to go figure out where to add what prefix
(paraphrasing
https://lists.llvm.org/pipermail/llvm-dev/2021-February/148326.html)
To avoid getting errors due to unused prefixes, whole directories can be
opted out (as discussed on that thread), but that means that tests that
aren't autogenerated in such directories could have undetected unused
prefix bugs.
This patch proposes an alternative that both avoids the above, dir-level
optout, and supports the main autogen scenario discussed first. The autogen
tool appends at the end of the test file the list of unused prefixes,
together with a note explaining that is the case. Each prefix is set up
to always pass.
This way, unexpected unused prefixes are easily discoverable, and
expected cases "just work".
Differential Revision: https://reviews.llvm.org/D124306
2022-04-23 03:49:15 +08:00
|
|
|
generated_prefixes.extend(
|
|
|
|
|
output_type.add_checks(output_lines, check_indent + ';', run_list,
|
|
|
|
|
func_dict, func_name, global_vars_seen_dict,
|
|
|
|
|
is_filtered=builder.is_filtered()))
|
2020-01-14 02:16:35 +08:00
|
|
|
is_in_function_start = False
|
|
|
|
|
|
|
|
|
|
if is_in_function:
|
|
|
|
|
if common.should_add_line_to_output(input_line, prefix_set):
|
|
|
|
|
# This input line of the function body will go as-is into the output.
|
2016-03-25 01:15:42 +08:00
|
|
|
output_lines.append(input_line)
|
2020-01-14 02:16:35 +08:00
|
|
|
else:
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
continue
|
2020-01-14 02:16:35 +08:00
|
|
|
if input_line.strip() == '}':
|
|
|
|
|
is_in_function = False
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
continue
|
|
|
|
|
|
2020-01-14 02:16:35 +08:00
|
|
|
# If it's outside a function, it just gets copied to the output.
|
|
|
|
|
output_lines.append(input_line)
|
2016-03-25 01:15:42 +08:00
|
|
|
|
2020-01-14 02:16:35 +08:00
|
|
|
m = common.IR_FUNCTION_RE.match(input_line)
|
|
|
|
|
if not m:
|
|
|
|
|
continue
|
|
|
|
|
func_name = m.group(1)
|
|
|
|
|
if args.function is not None and func_name != args.function:
|
|
|
|
|
# When filtering on a specific function, skip all others.
|
|
|
|
|
continue
|
|
|
|
|
is_in_function = is_in_function_start = True
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
[UpdateTestChecks] Auto-generate stub bodies for unused prefixes
This is scoped to autogenerated tests.
The goal is to support having each RUN line specify a list of
check-prefixes where one can specify potentially redundant prefixes. For example,
for X86, if one specified prefixes for both AVX1 and AVX2, and the codegen happened to
match today, one of the prefixes would be used and the onther one not.
If the unused prefix were dropped, and later, codegen differences were
introduced, one would have to go figure out where to add what prefix
(paraphrasing
https://lists.llvm.org/pipermail/llvm-dev/2021-February/148326.html)
To avoid getting errors due to unused prefixes, whole directories can be
opted out (as discussed on that thread), but that means that tests that
aren't autogenerated in such directories could have undetected unused
prefix bugs.
This patch proposes an alternative that both avoids the above, dir-level
optout, and supports the main autogen scenario discussed first. The autogen
tool appends at the end of the test file the list of unused prefixes,
together with a note explaining that is the case. Each prefix is set up
to always pass.
This way, unexpected unused prefixes are easily discoverable, and
expected cases "just work".
Differential Revision: https://reviews.llvm.org/D124306
2022-04-23 03:49:15 +08:00
|
|
|
if ti.args.gen_unused_prefix_body:
|
|
|
|
|
output_lines.extend(ti.get_checks_for_unused_prefixes(
|
|
|
|
|
run_list, generated_prefixes))
|
|
|
|
|
|
2020-07-08 17:59:50 +08:00
|
|
|
common.debug('Writing %d lines to %s...' % (len(output_lines), ti.path))
|
|
|
|
|
with open(ti.path, 'wb') as f:
|
2019-01-31 00:15:59 +08:00
|
|
|
f.writelines(['{}\n'.format(l).encode('utf-8') for l in output_lines])
|
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
llvm-svn: 225618
2015-01-12 12:43:18 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
|
|
main()
|