Repository navigation
Opa gzip fix - #4357
Merged
Merged
Opa gzip fix#4357
Conversation
Decompress Content-Encoding: gzip request bodies in ExtractHttpBodyOptionally before passing them to OPA, preventing policy bypass where input.parsed_body was null for compressed bodies. Also fixes semaphore over-reservation by releasing unused bytes immediately after reading. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
…pressing Replace the gzip-decompression approach with an early rejection: when body parsing is enabled (opaAuthorizeRequestWithBody, opaServeResponseWithReqBody), any request carrying a Content-Encoding header is rejected with 415 Unsupported Media Type before body extraction. This is simpler and more conservative — OPA always receives a plain, parseable body or the request never reaches policy evaluation. The fix applies to all encodings, not only gzip, and covers both filter implementations that call ExtractHttpBodyOptionally with bodyParsing enabled. Requests without body parsing (opaAuthorizeRequest, opaServeResponse) are unaffected. The decompression logic is removed from ExtractHttpBodyOptionally. The semaphore over-reservation fix (release unused bytes after read) is kept as it is correct independent of encoding. Regression tests added for both opaauthorizerequest and opaserveresponse, verifying 415 for gzip/br/deflate encoded bodies with body parsing enabled, and pass-through for encoded bodies when body parsing is disabled. Fixes GHSA-j3r3-rjcc-vh4q. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Signed-off-by: nuwandi-wickramasinghe_zse <nuwandi.wickramasinghe@zalando.de> Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
Add a second layer of defence against GHSA-j3r3-rjcc-vh4q: after reading the request body, validate it with json.Valid before passing it to OPA. If the bytes are not valid JSON and the body is not truncated, reject with 400 Bad Request. This closes the bypass where an attacker omits Content-Encoding but sends a compressed (or otherwise non-JSON) payload declared as application/json. Without this check, OPA receives unparseable bytes, input.parsed_body is null, and deny-on-body-field policies fail open. Truncated bodies are excluded from the check because they carry a partial JSON fragment by design; OPA already receives x-envoy-auth-partial-body: true in that case. Regression tests added to both opaauthorizerequest and opaserveresponse covering: valid JSON accepted, gzip bytes without Content-Encoding rejected with 400, arbitrary non-JSON bytes rejected with 400. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Signed-off-by: nuwandi-wickramasinghe_zse <nuwandi.wickramasinghe@zalando.de> Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
Use %q instead of %s when formatting the untrusted Content-Encoding header value in the error message, so control characters and other special bytes from the client are safely escaped. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Signed-off-by: nuwandi-wickramasinghe_zse <nuwandi.wickramasinghe@zalando.de> Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
… helper tests - Add TestBodyExtractionPartialSemaphoreRelease to verify unused semaphore capacity is released immediately when actual body bytes < requested bytes - Update gzipCompress to accept *testing.T and fail fast on write/close errors Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Signed-off-by: nuwandi-wickramasinghe_zse <nuwandi.wickramasinghe@zalando.de> Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
Update gzipBytes to accept *testing.T and call t.Fatalf on write or close failure instead of silently discarding errors. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Signed-off-by: nuwandi-wickramasinghe_zse <nuwandi.wickramasinghe@zalando.de> Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
Collaborator
Author
|
👍 |
Signed-off-by: Mustafa Abdelrahman <mustafa.abdelrahman@zalando.de>
|
Docker image "registry-write.opensource.zalan.do/teapot/skipper:b5585b0f22a0332d703b75b35269ee49c09c1a3b" is not based on an approved base image. Any production deployment relying on this image will be blocked. To create a compliant Docker image of your application, you must reference an allowed Docker image as its base image in your Dockerfile. This base image must come from the Zalando Container Registry namespace |
szuecs
approved these changes
Oct 20, 2026
Collaborator
Author
|
👍 |
This was referenced Oct 20, 2026
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
to be done