[http] Simplify H1 codec by removing the reserve/commit pattern #9825
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
Signed-off-by: Asra Ali <[email protected]>
|
Looks like it needs a format fix and then will take a look, thank you! /wait |
Signed-off-by: Asra Ali <[email protected]>
mattklein123
approved these changes
Jan 27, 2020
There was a problem hiding this comment.
LGTM. Will defer to @alyssawilk for final approval. Thank you!
alyssawilk
approved these changes
Jan 28, 2020
This was referenced Nov 27, 2021
alyssawilk
pushed a commit
that referenced
this pull request
Jan 12, 2022
…ders (#19115) Commit Message: minor perf: reduce fine grained buffer access when encoding HTTP1 headers Additional Description: The HTTP1 codec performs a large number of buffer APIs calls during the encoding of HTTP headers, which introduces an additional CPU overhead (3-4%). This PR implements a buffer helper as a cache to reduce direct buffer writes to improving the overall performance. Check this #19103 (comment) for some more related info. This PR reverts #9825. At the same time, I also try my best to make the code simpler and easier to maintain. Risk Level: High. Testing: Waiting. Docs Changes: N/A. Release Notes: N/A. Signed-off-by: wbpcode <[email protected]>
joshperry
pushed a commit
to joshperry/envoy
that referenced
this pull request
Feb 13, 2022
…ders (envoyproxy#19115) Commit Message: minor perf: reduce fine grained buffer access when encoding HTTP1 headers Additional Description: The HTTP1 codec performs a large number of buffer APIs calls during the encoding of HTTP headers, which introduces an additional CPU overhead (3-4%). This PR implements a buffer helper as a cache to reduce direct buffer writes to improving the overall performance. Check this envoyproxy#19103 (comment) for some more related info. This PR reverts envoyproxy#9825. At the same time, I also try my best to make the code simpler and easier to maintain. Risk Level: High. Testing: Waiting. Docs Changes: N/A. Release Notes: N/A. Signed-off-by: wbpcode <[email protected]> Signed-off-by: Josh Perry <[email protected]>
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.
This simplifies the H/1 codec by removing the pre-buffer reserved_iovec_ in the HTTP1 codec.
This change reduces effective memory usage by removing the need to round up when reserving memory blocks in OwnedImpl::reserve. The
output_buffer_in the codec is a watermark buffer that on adds willcheckHighWatermark. Performance may suffer slightly due to an increase in number of calls tocheckHighWatermarkwhile serializing headers.(Before the change, watermarks were checked at most twice in ConnectionImpl::reserveBuffer (commit if needed and reserve), and not in the following calls to add/copy to the buffer. Now, add/copyToBuffer directly call the watermark buffer, so check watermarks are increased.)
Risk level: medium/high
Testing: existing tests pass (suggestions welcome for more if I can peak at the connection's buffer)
Signed-off-by: Asra Ali [email protected]