Fix segfault in HiveDataSource::next() when preloading an empty split #8850
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
✅ Deploy Preview for meta-velox canceled.
|
29de8a1 to
ccc431f
Compare
|
I am unable to add a reproducer. Even though the It is unclear how the |
be1bcc1 to
0498c29
Compare
|
https://prestodb.slack.com/archives/C03DYLFBN06/p1709772605962749 |
|
@Yuhta has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
xiaoxmeng
reviewed
Mar 24, 2024
| @@ -201,8 +201,9 @@ std::optional<RowVectorPtr> HiveDataSource::next( | |||
| uint64_t size, | |||
| velox::ContinueFuture& /*future*/) { | |||
| VELOX_CHECK(split_ != nullptr, "No split to process. Call addSplit first."); | |||
| VELOX_CHECK(splitReader_ != nullptr); | |||
There was a problem hiding this comment.
nit: VELOX_CHECK_NOT_NULL
| @@ -327,11 +328,8 @@ void HiveDataSource::setFromDataSource( | |||
| VELOX_CHECK(source, "Bad DataSource type"); | |||
There was a problem hiding this comment.
NYC: VELOX_CHECK_NOT_NULL
0498c29 to
b0be73f
Compare
xiaoxmeng
approved these changes
Mar 24, 2024
|
@Yuhta has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Joe-Abraham
pushed a commit
to Joe-Abraham/velox
that referenced
this pull request
Jun 7, 2024
…facebookincubator#8850) Summary: We are seeing the segfault below in our benchmark cluster. The segfault is due to the `splitReader_` being accessed as a nullptr; The `splitReader_` is not initialized for an empty split in the preload path. The previous `splitReader_` is used and this is fixed. We confirmed that this change fixes the segfault. However, I do not have a test where the previous `splitReader_` is a nullptr. The addSplit() path, where `splitReader_` is always created first, seems to be always taken first before the preload path. It is unclear how the `splitReader_` can become a nullptr in practice. *** SIGSEGV (0x0) received by PID 55 (TID 0x7ef6a27fc700) from PID 0; stack trace: *** @ 0x7f58305b776e google::(anonymous namespace)::FailureSignalHandler() @ 0x7f582cc1bd20 (unknown) @ 0x6b0646 facebook::velox::connector::hive::HiveDataSource::next() @ 0x3a6ebbc facebook::velox::exec::TableScan::getOutput() @ 0x3953f3b facebook::velox::exec::Driver::runInternal() @ 0x3956e2d facebook::velox::exec::Driver::run() @ 0x3956fef _ZN5folly6detail8function14FunctionTraitsIFvvEE9callSmallIZN8facebook5velox4exec6Driver7enqueueESt10shared_ptrIS9_EEUlvE_EEvRNS1_4DataE @ 0x7f582e83d4f6 folly::ThreadPoolExecutor::runTask() @ 0x4087575 folly::CPUThreadPoolExecutor::threadRun() @ 0x7f582e83f007 folly::detail::function::FunctionTraits<>::callSmall<>() @ 0x7f582b875b23 (unknown) @ 0x7f582cc111ca start_thread @ 0x7f582ae7c8d3 __GI___clone @ 0x0 (unknown) Pull Request resolved: facebookincubator#8850 Reviewed By: xiaoxmeng Differential Revision: D55255299 Pulled By: Yuhta fbshipit-source-id: ba09971faf517195d8a8ded692a9d31a758bdd1c
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We are seeing the segfault below in our benchmark cluster.
The segfault is due to the
splitReader_being accessed as a nullptr;The
splitReader_is not initialized for an empty split in the preload path.The previous
splitReader_is used and this is fixed.We confirmed that this change fixes the segfault.
However, I do not have a test where the previous
splitReader_is a nullptr.The addSplit() path, where
splitReader_is always created first, seems to be always taken first beforethe preload path.
It is unclear how the
splitReader_can become a nullptr in practice.*** SIGSEGV (@0x0) received by PID 55 (TID 0x7ef6a27fc700) from PID 0; stack trace: ***
@ 0x7f58305b776e google::(anonymous namespace)::FailureSignalHandler()
@ 0x7f582cc1bd20 (unknown)
@ 0x6b0646 facebook::velox::connector::hive::HiveDataSource::next()
@ 0x3a6ebbc facebook::velox::exec::TableScan::getOutput()
@ 0x3953f3b facebook::velox::exec::Driver::runInternal()
@ 0x3956e2d facebook::velox::exec::Driver::run()
@ 0x3956fef _ZN5folly6detail8function14FunctionTraitsIFvvEE9callSmallIZN8facebook5velox4exec6Driver7enqueueESt10shared_ptrIS9_EEUlvE_EEvRNS1_4DataE
@ 0x7f582e83d4f6 folly::ThreadPoolExecutor::runTask()
@ 0x4087575 folly::CPUThreadPoolExecutor::threadRun()
@ 0x7f582e83f007 folly::detail::function::FunctionTraits<>::callSmall<>()
@ 0x7f582b875b23 (unknown)
@ 0x7f582cc111ca start_thread
@ 0x7f582ae7c8d3 __GI___clone
@ 0x0 (unknown)