Skip to content

Unpoison stack frame ptrs from libunwind for msan#49316

Merged
tavplubix merged 4 commits intomasterfrom
rs/48385
May 4, 2023
Merged

Unpoison stack frame ptrs from libunwind for msan#49316
tavplubix merged 4 commits intomasterfrom
rs/48385

Conversation

@rschu1ze
Copy link
Copy Markdown
Member

@rschu1ze rschu1ze commented Apr 28, 2023 •
edited
Loading

Turned out that in msan builds, stack frame pointers returned by unw_backtrace() in Common/StackTrace.cpp are unpoisened, but when returned from the same function under the define STD_EXCEPTION_HAS_STACK_TRACE in std::exception (see (*)) they were not. This was not really obvious since std::exception is in contrib/llvm-project/libcxx. Unposining directly in header "exception" didn't work, therefore this PR unpoisons everywhere where get_stack_trace_frames() is called.

This fixes #48385 for me, i.e. msan complained about reads of uninitialized data when the query cache squashes and compresses chunks produced from SELECT last_error_trace FROM system.errors SETTINGS allow_experimental_query_cache=1, use_query_cache=1. Actually, the error happened also locally without query cache (then with a different crash stack) - I am not sure why it never occurred in CI.

Fixes: #48385

(*) https://github.com/ClickHouse/llvm-project/pull/7/files#diff-8c81d74ef72967be2fa27a31275000aa18532a583490f9673c62d7b981a5b39c

Changelog category (leave one):

  • Not for changelog (changelog entry is not required)

@robot-clickhouse robot-clickhouse added the pr-not-for-changelog This PR should not be mentioned in the changelog label Apr 28, 2023
@rschu1ze
Unpoison stack frame ptrs from libunwind for msan
7609832 
Turned out that in msan builds, stack frame pointers explicitly returned
by unw_backtrace() in Common/StackTrace.cpp were unpoisened, but stack
frame pointers returned from the same function under the
STD_EXCEPTION_HAS_STACK_TRACE define in std::exception (see (*)) were
not. I tried unpoisoning directly in libcxx. Since this didn't work,
this PR unpoisons where get_stack_trace_frames() is called.

I don't know why this issue never occurred in CI but it fixes #48385 for
me (where msan complains about reads of uninitialized data when query
cache squashes and compresses chunks produced from SELECT last_error_trace
FROM system.errors).

Fixes: #48385

(*) https://github.com/ClickHouse/llvm-project/pull/7/files#diff-8c81d74ef72967be2fa27a31275000aa18532a583490f9673c62d7b981a5b39c
@robot-clickhouse-ci-1
Copy link
Copy Markdown
Contributor

robot-clickhouse-ci-1 commented May 1, 2023 •
edited by robot-ch-test-poll1
Loading

This is an automated comment for commit cf884d2 with description of existing statuses. It's updated for the latest CI running
The full report is available here
The overall status of the commit is 🔴 failure

Check nameDescriptionStatus
AST fuzzerRuns randomly generated queries to catch program errors. The build type is optionally given in parenthesis. If it fails, ask a maintainer for help.🟢 success
CI runningA meta-check that indicates the running CI. Normally, it's in success or pending state. The failed status indicates some problems with the PR🟢 success
ClickHouse build checkBuilds ClickHouse in various configurations for use in further steps. You have to fix the builds that fail. Build logs often has enough information to fix the error, but you might have to reproduce the failure locally. The cmake options can be found in the build log, grepping for cmake. Use these options and follow the general build process.🟢 success
Compatibility checkChecks that clickhouse binary runs on distributions with old libc versions. If it fails, ask a maintainer for help.🟢 success
Docker image for serversThe check to build and optionally push the mentioned image to docker hub🟢 success
Fast testNormally this is the first check that is ran for a PR. It builds ClickHouse and runs most of stateless functional tests, omitting some. If it fails, further checks are not started until it is fixed. Look at the report to see which tests fail, then reproduce the failure locally as described here.🟢 success
Flaky testsChecks if new added or modified tests are flaky by running them repeatedly, in parallel, with more randomization. Functional tests are run 100 times with address sanitizer, and additional randomization of thread scheduling. Integrational tests are run up to 10 times. If at least once a new test has failed, or was too long, this check will be red. We don't allow flaky tests, read https://clickhouse.com/blog/decorating-a-christmas-tree-with-the-help-of-flaky-tests/🟢 success
Install packagesChecks that the built packages are installable in a clear environment🟢 success
Integration testsThe integration tests report. In parenthesis the package type is given, and in square brackets are the optional part/total tests🟢 success
Mergeable CheckChecks if all other necessary checks are successful🟢 success
Performance ComparisonMeasure changes in query performance. The performance test report is described in detail here. In square brackets are the optional part/total tests🟢 success
Push to DockerhubThe check for building and pushing the CI related docker images to docker hub🟢 success
SQLancerFuzzing tests that detect logical bugs with SQLancer tool.🟢 success
SqllogicRun clickhouse on the sqllogic test set against sqlite and checks that all statements are passed.🟢 success
Stateful testsRuns stateful functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc.🟢 success
Stateless testsRuns stateless functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc.🔴 failure
Stress testRuns stateless functional tests concurrently from several clients to detect concurrency-related errors.🟢 success
Style CheckRuns a set of checks to keep the code style clean. If some of tests failed, see the related log from the report.🟢 success
Unit testsRuns the unit tests for different release types🟢 success
Upgrade checkRuns stress tests on server version from last release and then tries to upgrade it to the version from the PR. It checks if the new server can successfully startup without any errors, crashes or sanitizer asserts.🔴 failure

@serxa serxa mentioned this pull request May 2, 2023
Merged
@tavplubix tavplubix self-assigned this May 4, 2023
@tavplubix
Copy link
Copy Markdown
Member

tavplubix commented May 4, 2023

Stateless tests (release, analyzer) - this check is broken in master
Upgrade check (asan) - 404 Client Error: Not Found for url: https://objects.githubusercontent.com/github-production-release-asset-2e65be/60246359/571ba11a-a337-40ba-a3b7-cc3a906723f1?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAIWNJYAX4CSVEH53A%2F20230502%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20230502T164734Z&X-Amz-Expires=300&X-Amz-Signature=af2a071c0372cdc41486591af6e746d5d4869c55edc141cd1fb18dadb99c2897&X-Amz-SignedHeaders=host&actor_id=0&key_id=0&repo_id=60246359&response-content-disposition=attachment%3B%20filename%3Dclickhouse-common-static-dbg_23.4.2.11_amd64.deb&response-content-type=application%2Foctet-stream
Upgrade check (tsan) — OOM in dmesg - we don't have diagnostics for OOM in prev release

@tavplubix tavplubix merged commit b5a57da into master May 4, 2023
@tavplubix tavplubix deleted the rs/48385 branch May 4, 2023 13:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@tavplubix tavplubix tavplubix approved these changes

Assignees

@tavplubix tavplubix

Labels

pr-not-for-changelog This PR should not be mentioned in the changelog

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

use-of-uninitialized-value in QueryCache::Writer

4 participants

@rschu1ze @robot-clickhouse-ci-1 @tavplubix @robot-clickhouse