Fix bug in TLS context - allow incoming write buffer to be relocated #5482
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
eddyashton
commented
Jul 28, 2023
eddyashton
added
auto-backport
|
ssl_debug@73621 aka 20230731.6 vs main ewma over 20 builds from 73098 to 73609 Click to see tablemain
ssl_debug
|
achamayou
approved these changes
Jul 31, 2023
ghost
mentioned this pull request
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Outdated suggestions cannot be applied.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.
Spotted some odd TLS errors during multi-threaded perf tests, where clients were being disconnected. Tracked it down to this. There were cases where we got a
WANTS_WRITE/WANTS_READresponse from an attempted write, which means a handshake is still in-progress or similar, and we need to retry the write later. The default semantics are very strict, and we must "retry the write" with exactly the same pointer and size. Our source for this is apending_writesbuffer, so if it got resized and reallocated, we'd pass a new pointer and this would blow up. There's a magic flag to make it less strict, which it seems we want. This is well explained in this StackOverflow answer, once you've convinced OpenSSL to log its errors correctly.