New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
tests: Add option --valgrind to run the functional tests under Valgrind #17633
Conversation
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers. ConflictsReviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do all tests currently pass with this new valgrind option?
61a48c1
to
5db506b
Compare
There are a few failing tests: most of them due to timing assumptions that fail in the presence of the Note that doing that has the added nice benefit of turning a few flaky tests into stable ones since said timing assumptions are intermittently violated under normal test execution due to high load, etc. |
ACK 5db506b |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Concept ACK, thank you for working on adding valgrind for the tests. Unfortunately, every test run with --valgrind
fails for me with AssertionError: [node 0] Error: no RPC connection
. Pointers welcome :)
$ valgrind --version
valgrind-3.14.0
"..", "..", "..", "contrib", "valgrind.supp") | ||
suppressions_file = os.getenv("VALGRIND_SUPPRESSIONS_FILE", | ||
default_suppressions_file) | ||
self.args = ["valgrind", "--suppressions={}".format(suppressions_file), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
suggestion if you have to retouch this:
- self.args = ["valgrind", "--suppressions={}".format(suppressions_file),
- "--gen-suppressions=all", "--exit-on-first-error=yes",
- "--error-exitcode=1", "--quiet"] + self.args
+ self.args = [
+ "valgrind",
+ "--suppressions={}".format(suppressions_file),
+ "--gen-suppressions=all",
+ "--exit-on-first-error=yes",
+ "--error-exitcode=1",
+ "--quiet",
+ ] + self.args
(like lines 89-97 before it)
@jonatack What does |
@MarcoFalke thanks! I don't see |
Try |
Also, could you try with the suppressions file from https://github.com/bitcoin/bitcoin/pull/17490/files#diff-8982317ece972c9b26bd8545350812f2R31 ? Maybe remove the line I linked to if it still doesn't work. |
ACK 5db506b Reviewed code, reproduced the PR description steps after modifying Thank you @MarcoFalke for your suggestions. |
I'm probably missing something, but isn't that what bitcoin/test/functional/combine_logs.py Lines 98 to 111 in 1189b6a
|
Ah right. Forgot that I already wrote that. And it indeed was also included in https://gist.github.com/jonatack/915df3494f9f7274a07e29ae092a2a47#file-pr17633-combine_log_output-txt-L263 |
…s under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR #17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that #17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…se of uninitialized memory 870f0cd build: Add MemorySanitizer (MSan) in Travis to detect use of uninitialized memory (practicalswift) Pull request description: Add MemorySanitizer (MSan) in Travis to detect use of uninitialized memory. First UBSan, then ASan followed by TSan... and now: yes, the wait is over -- **MSan is finally here!** :) Some historical context: * 2017: Continuous compilation with Clang Thread Safety analysis enabled (#10866, #10923) * 2018: Continuous testing with trapping on signed integer overflows (`-ftrapv`) (#12686) * 2018: Continuous testing of use of locale dependent functions (#13041) * 2018: Continuous testing of format strings (#13705) * 2018: Continuous compilation with MSVC `TreatWarningAsError` (#14151) * 2018: Continuous testing under UndefinedBehaviorSanitizer – UBSan (#14252, #14673, #17006) * 2018: Continuous testing under AddressSanitizer – ASan (#14794, #17205, #17674) * 2018: Continuous testing under ThreadSanitizer – TSan (#14829) * 2019: Continuous testing in an unsigned char environment (`-funsigned-char`) (#15134) * 2019: Continuous compile-time testing of assumptions we're making (#15391) * 2019: Continuous testing of fuzz test cases under Valgrind (#17633, #18159, #18166) * 2020: Finally... MemorySanitizer – MSAN! :) What is the next step? What tools should we add to CI to keep bugs from entering `master`? :) ACKs for top commit: MarcoFalke: ACK 870f0cd Tree-SHA512: 38327c8b75679d97d469fe42e704cacd1217447a5a603701dd8a58ee50b3be2c10248f8d68a479ed081c0c4b254589d3081c9183f991640b06ef689061f75578
…ests under Valgrind Summary: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` bitcoin/bitcoin@5db506b --- Depends on D7083 Backport of Core [[bitcoin/bitcoin#17633 | PR17633]] Test Plan: Have valgrind >3.14 installed ninja test_runner.py --valgrind notice that tests run 10x slower Reviewers: #bitcoin_abc, Fabien Reviewed By: #bitcoin_abc, Fabien Differential Revision: https://reviews.bitcoinabc.org/D7084
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795 # Conflicts: # test/functional/test_framework/test_framework.py # test/functional/test_framework/test_node.py
…etect use of uninitialized memory 870f0cd build: Add MemorySanitizer (MSan) in Travis to detect use of uninitialized memory (practicalswift) Pull request description: Add MemorySanitizer (MSan) in Travis to detect use of uninitialized memory. First UBSan, then ASan followed by TSan... and now: yes, the wait is over -- **MSan is finally here!** :) Some historical context: * 2017: Continuous compilation with Clang Thread Safety analysis enabled (bitcoin#10866, bitcoin#10923) * 2018: Continuous testing with trapping on signed integer overflows (`-ftrapv`) (bitcoin#12686) * 2018: Continuous testing of use of locale dependent functions (bitcoin#13041) * 2018: Continuous testing of format strings (bitcoin#13705) * 2018: Continuous compilation with MSVC `TreatWarningAsError` (bitcoin#14151) * 2018: Continuous testing under UndefinedBehaviorSanitizer – UBSan (bitcoin#14252, bitcoin#14673, bitcoin#17006) * 2018: Continuous testing under AddressSanitizer – ASan (bitcoin#14794, bitcoin#17205, bitcoin#17674) * 2018: Continuous testing under ThreadSanitizer – TSan (bitcoin#14829) * 2019: Continuous testing in an unsigned char environment (`-funsigned-char`) (bitcoin#15134) * 2019: Continuous compile-time testing of assumptions we're making (bitcoin#15391) * 2019: Continuous testing of fuzz test cases under Valgrind (bitcoin#17633, bitcoin#18159, bitcoin#18166) * 2020: Finally... MemorySanitizer – MSAN! :) What is the next step? What tools should we add to CI to keep bugs from entering `master`? :) ACKs for top commit: MarcoFalke: ACK 870f0cd Tree-SHA512: 38327c8b75679d97d469fe42e704cacd1217447a5a603701dd8a58ee50b3be2c10248f8d68a479ed081c0c4b254589d3081c9183f991640b06ef689061f75578
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
…al tests under Valgrind 5db506b tests: Add option --valgrind to run nodes under valgrind in the functional tests (practicalswift) Pull request description: What is better than fixing bugs? Fixing entire bug classes of course! :) Add option `--valgrind` to run the functional tests under Valgrind. Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically. Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :) My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :) Hopefully `test/functional/test_runner.py --valgrind` will become a natural part of the pre-release QA process. **Usage:** ``` $ test/functional/test_runner.py --help … --valgrind run nodes under the valgrind memory error detector: expect at least a ~10x slowdown, valgrind 3.14 or later required ``` **Live demo:** First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR bitcoin#17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have"). ``` $ git diff diff --git a/src/consensus/validation.h b/src/consensus/validation.h index 3401eb64c..940adea33 100644 --- a/src/consensus/validation.h +++ b/src/consensus/validation.h @@ -114,7 +114,7 @@ inline ValidationState::~ValidationState() {}; class TxValidationState : public ValidationState { private: - TxValidationResult m_result = TxValidationResult::TX_RESULT_UNSET; + TxValidationResult m_result; public: bool Invalid(TxValidationResult result, const std::string &reject_reason="", ``` Second, let's test as normal without Valgrind: ``` $ test/functional/p2p_segwit.py -l INFO 2019-11-28T09:30:42.810000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test__fc8q3qo … 2019-11-28T09:31:57.187000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) … 2019-11-28T09:32:08.265000Z TestFramework (INFO): Tests successful ``` Third, let's test with `--valgrind` and see if the test fail (as we expect) when the unitialized value is used: ``` $ test/functional/p2p_segwit.py -l INFO --valgrind 2019-11-28T09:32:33.018000Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_gtjecx2l … 2019-11-28T09:40:36.702000Z TestFramework (INFO): Subtest: test_non_standard_witness_blinding (Segwit active = True) 2019-11-28T09:40:37.813000Z TestFramework (ERROR): Assertion failed ConnectionRefusedError: [Errno 111] Connection refused ``` ACKs for top commit: MarcoFalke: ACK 5db506b jonatack: ACK 5db506b Tree-SHA512: 2eaecacf4da166febad88b2a8ee6d7ac2bcd38d4c1892ca39516b6343e8f8c8814edf5eaf14c90f11a069a0389d24f0713076112ac284de987e72fc5f6cc3795
155a11f doc: Added running functional tests in valgrind (Elichai Turkel) Pull request description: Technically the notes only show an "example" of how to run valgrind with the suppression file, but now that bitcoin#17633 is merged then maybe this can encourage more people to run also the functional tests in valgrind Top commit has no ACKs. Tree-SHA512: b8417249b720d0ed5e10b732648f2e07e8889bfc7aa7e94192d1c049b4b7837971678d30c535f273c227848f1290cf11e14369fd6c1924b734f2e47e2af41401
What is better than fixing bugs? Fixing entire bug classes of course! :)
Add option
--valgrind
to run the functional tests under Valgrind.Regular functional testing under Valgrind would have caught many of the uninitialized reads we've seen historically.
Let's kill this bug class once and for all: let's never use an uninitialized value ever again. Or at least not one that would be triggered by running the functional tests! :)
My hope is that this addition will make it super-easy to run the functional tests under Valgrind and thus increase the probability of people making use of it :)
Hopefully
test/functional/test_runner.py --valgrind
will become a natural part of the pre-release QA process.Usage:
Live demo:
First, let's re-introduce a memory bug by reverting the recent P2P uninitialized read bug fix from PR #17624 ("net: Fix an uninitialized read in ProcessMessage(…, "tx", …) when receiving a transaction we already have").
Second, let's test as normal without Valgrind:
Third, let's test with
--valgrind
and see if the test fail (as we expect) when the unitialized value is used: