compression: honor read buffer offset in resize_buffer

[saddamr3e]

AddressSanitizer, appending a chunk after a decompression member is consumed:

==ERROR: AddressSanitizer: heap-buffer-overflow  WRITE of size 200
0x6110000...fd0 is located 0 bytes after 208-byte region
    #1 test_resize_buffer_read_offset compression.c
    #2 flb_decompression_context_resize_buffer flb_compression.c:113

Once a member is decoded read_buffer sits at a non-zero offset inside input_buffer. flb_decompression_context_get_available_space() and flb_decompression_context_get_append_buffer() both account for that offset, but resize_buffer() reallocated to new_size only. The packed-forward path in in_forward resizes to input_buffer_length + len and then copies len bytes at get_append_buffer(), so with a non-zero read offset that copy runs read_buffer_offset bytes past the reallocation. Reserve the offset in the reallocation and keep input_buffer_size in sync with it.

Regression test lives in tests/internal/compression.c; it faults under AddressSanitizer before the change and passes after.

---

Enter [N/A] in the box, if an item is not applicable to your change.

Testing
Before we can approve your change; please submit the following in a comment:

• [N/A] Example configuration file for the change
• [N/A] Debug log output from testing the change
• Attached Valgrind output that shows no leaks or memory corruption was found

Used AddressSanitizer instead of Valgrind: the new tests/internal/compression.c unit test built with -DSANITIZE_ADDRESS=On reports a heap-buffer-overflow write past the resize_buffer() allocation on the unpatched tree and is clean with the fix, and flb-it-gzip / flb-it-zstd stay green.

If this is a change to packaging of containers or native binaries then please confirm it works for all targets.

• [N/A] Run local packaging test showing all targets (including any new ones) build.
• [N/A] Set ok-package-test label to test for all targets (requires maintainer to do).

Documentation

• [N/A] Documentation required for this feature

Backporting

• Backport to latest stable release.

---

Fluent Bit is licensed under Apache 2.0, by submitting this pull request I understand that this code will be released under the terms of that license.

Summary by CodeRabbit

• Bug Fixes

  • Fixed gzip decompression buffer resizing when unprocessed data remains in the buffer.
  • Prevented potential memory corruption when appending data after decompression.

• Tests

  • Added coverage for decompression buffer resizing with pending trailing data.

[coderabbitai]

📝 Walkthrough

Walkthrough

The decompression buffer resize now accounts for unconsumed input offsets. A gzip regression test exercises resizing and appending after decompression, and CMake includes the test in internal unit tests.

Changes

Compression buffer fix

Layer / File(s)	Summary
Offset-aware buffer resizing src/flb_compression.c	Reallocation and input_buffer_size now include the current read_buffer offset.
Resize regression coverage tests/internal/compression.c, tests/internal/CMakeLists.txt	Adds a gzip decompression test for offset-aware resizing and registers it with the internal unit test build.

Estimated code review effort: 2 (Simple) | ~10 minutes

Suggested reviewers: edsiper, cosmo0920, koleini

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly matches the main fix: resize_buffer now honors the read buffer offset during compression-related decompression resizing.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

🧹 Nitpick comments (2)

src/flb_compression.c (2)

109-110: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Reuse the existing flb_decompression_context_get_read_buffer_offset helper.

The offset calculation at lines 109–110 duplicates the logic already encapsulated in flb_decompression_context_get_read_buffer_offset (lines 27–40), which also provides a NULL guard on context. Using the helper reduces duplication and adds the safety check.

♻️ Proposed refactor

-    read_buffer_offset = (uintptr_t) context->read_buffer -
-                         (uintptr_t) context->input_buffer;
+    read_buffer_offset = flb_decompression_context_get_read_buffer_offset(context);

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/flb_compression.c` around lines 109 - 110, Replace the duplicated pointer
subtraction in the decompression flow with the existing
flb_decompression_context_get_read_buffer_offset helper, passing the current
context and preserving assignment to read_buffer_offset. Do not retain the
manual calculation.

---

112-114: 🔒 Security & Privacy | 🔵 Trivial | ⚡ Quick win

Guard against integer overflow in read_buffer_offset + new_size.

Both read_buffer_offset and new_size are size_t. Their sum is used as the flb_realloc size (line 114) and stored in input_buffer_size (line 128). If the addition wraps, flb_realloc allocates a tiny buffer while the caller subsequently writes append_len bytes at get_append_buffer() — the same heap-buffer-overflow class this PR fixes. A simple overflow check before the realloc closes the gap.

🛡️ Proposed overflow guard

     if (new_size > context->input_buffer_length) {
+        if (read_buffer_offset > SIZE_MAX - new_size) {
+            return FLB_DECOMPRESSOR_FAILURE;
+        }
         new_buffer_address = flb_realloc(context->input_buffer,
                                          read_buffer_offset + new_size);

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/flb_compression.c` around lines 112 - 114, In the buffer-growth logic
around the realloc call, validate that read_buffer_offset + new_size cannot
exceed SIZE_MAX before performing the addition or passing it to flb_realloc.
Handle the overflow using the function’s existing allocation-failure path, and
ensure the same validated total is used when storing input_buffer_size.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@src/flb_compression.c`:
- Around line 109-110: Replace the duplicated pointer subtraction in the
decompression flow with the existing
flb_decompression_context_get_read_buffer_offset helper, passing the current
context and preserving assignment to read_buffer_offset. Do not retain the
manual calculation.
- Around line 112-114: In the buffer-growth logic around the realloc call,
validate that read_buffer_offset + new_size cannot exceed SIZE_MAX before
performing the addition or passing it to flb_realloc. Handle the overflow using
the function’s existing allocation-failure path, and ensure the same validated
total is used when storing input_buffer_size.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 37fe48c2-d329-4480-a68f-3b24beefb519

📥 Commits

Reviewing files that changed from the base of the PR and between e6db1af and 2dfd13f.

📒 Files selected for processing (3)

• src/flb_compression.c
• tests/internal/CMakeLists.txt
• tests/internal/compression.c