connection: Do not reset delayed closed timer if doWrite consumes 0 bytes from the output buffer. #11833
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
…ytes from the output buffer. Signed-off-by: Antonio Vicente <avd@google.com>
|
/cc @davinci26 Hey Sotiris, Can you double check that this change addresses your problem? |
antoniovicente
commented
Jul 1, 2020
| delayed_close_timer_->enableTimer(delayed_close_timeout_); | ||
| ASSERT(delayed_close_timer_ != nullptr && delayed_close_timer_->enabled()); | ||
| if (result.bytes_processed_ > 0) { | ||
| delayed_close_timer_->enableTimer(delayed_close_timeout_); |
There was a problem hiding this comment.
I don't think there's a way to distinguish between these two enableTimer calls in tests. Ideally we would move delayed_close_state_ to a WaitingForClose state or change the fd mask to:
file_event_->setEnabled(enable_half_close_ ? 0 : Event::FileReadyType::Closed);
to avoid getting any Write events past this point.
There was a problem hiding this comment.
nah, but I'll trust you made sure your test would fail without the code change :-)
|
/assign @davinci26 |
|
@davinci26 cannot be assigned to this issue. |
|
LGTM it address the issue. I have verified it locally in my devbox |
…_only_if_made_progress Signed-off-by: Antonio Vicente <avd@google.com>
alyssawilk
approved these changes
Jul 7, 2020
| delayed_close_timer_->enableTimer(delayed_close_timeout_); | ||
| ASSERT(delayed_close_timer_ != nullptr && delayed_close_timer_->enabled()); | ||
| if (result.bytes_processed_ > 0) { | ||
| delayed_close_timer_->enableTimer(delayed_close_timeout_); |
There was a problem hiding this comment.
nah, but I'll trust you made sure your test would fail without the code change :-)
Bumped up to medium. Thanks! |
jwendell
pushed a commit
to istio/envoy
that referenced
this pull request
Jul 21, 2020
* Revert "test: shard http2_integration_test (envoyproxy#11939)" This reverts commit 2026ec2. * Revert "fuzz: added fuzz test for listener filter original_dst (envoyproxy#11847)" This reverts commit 673cab8. * Revert "docs: updating release instructions (envoyproxy#11938)" This reverts commit 09b96a5. * Revert "test: remove superfluous test dependencies (envoyproxy#11912)" This reverts commit 5e9fb8a. * Revert "dynamic_forward_proxy: cleanup integration test (envoyproxy#11891)" This reverts commit 11a4667. * Revert "preliminary PR to Porting Envoy to C++17 (envoyproxy#11840)" This reverts commit 9ad964d. * Revert "connection: Do not reset delayed closed timer if doWrite consumes 0 bytes from the output buffer. (envoyproxy#11833)" This reverts commit 5960533. * Revert "threadlocal: avoiding a dynamic cast in opt builds (envoyproxy#11900)" This reverts commit 363b104. * Revert "release: kicking off 1.16.0 (envoyproxy#11930)" This reverts commit ef74d8f. Signed-off-by: Lizan Zhou <lizan@tetrate.io>
scheler
pushed a commit
to scheler/envoy
that referenced
this pull request
Aug 4, 2020
…ytes from the output buffer. (envoyproxy#11833) Commit Message: connection: Do not reset delayed closed timer if doWrite consumes 0 bytes from the output buffer. Additional Description: Only reset the delayed close timer if the write attempt made progress. This works around spurious fd Write events which are delivered to a connection even after it manages to fully drain the output buffer and should be waiting for client FIN or the delay close timer to expire. This is known to happen when listening for level events instead of edge-trigger fd events. Risk Level: medium, changes to timeout behavior. Testing: unit Docs Changes: n/a Release Notes: n/a Fixes envoyproxy#11829 Signed-off-by: Antonio Vicente <avd@google.com> Signed-off-by: scheler <santosh.cheler@appdynamics.com>
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.
Commit Message: connection: Do not reset delayed closed timer if doWrite consumes 0 bytes from the output buffer.
Additional Description: Only reset the delayed close timer if the write attempt made progress. This works around spurious fd Write events which are delivered to a connection even after it manages to fully drain the output buffer and should be waiting for client FIN or the delay close timer to expire. This is known to happen when listening for level events instead of edge-trigger fd events.
Risk Level: medium, changes to timeout behavior.
Testing: unit
Docs Changes: n/a
Release Notes: n/a
Fixes #11829