This bisect locally to 52cc5fe0d1b6c346e9de7829afd7c8c409869f5f ("Reland [Memory] Speculative fix for sanitizer flakiness.").
Bill, do you see a reason why this would cause OOM issues?

If not, please assign to Marja to check whether this OOM is expected / acceptable. Note that the test case just has 79kB in size.

Oh, and it only fails after several runs.
Reproduce with:
out/tmp/v8_json_parser_fuzzer -timeout=25 -rss_limit_mb=2048 -runs=100 ~/Downloads/clusterfuzz-testcase-minimized-6080783256387584

Where the gn args for out/tmp are:
enable_nacl = false
ffmpeg_branding = "ChromeOS"
is_debug = false
is_msan = true
msan_track_origins = 2
optimize_for_fuzzing = true
pdf_enable_xfa = true
proprietary_codecs = true
use_libfuzzer = true
use_prebuilt_instrumented_libraries = true

Assigning to marja. My CL does two things - first, on POSIX systems, restricts random mmap addresses to a 1 Terabyte region in high memory, which is guaranteed not to interfere with sanitizer memory. Second, it adds a mutex to protect the RNG which is not thread safe. TSAN found a race (which would scramble some random numbers, making them extra-random I suppose).

Here is how addresses are isolated on sanitizer builds (raw_addr is generated by the RNG):

  raw_addr &= 0x007fffff0000ULL;
  raw_addr += 0x7e8000000000ULL;

I can't think of a reason why restricting the address space range (to a still enormous 1 Terabyte sub-range) would lead to OOM in these tests.

I don't know anything about this particular fuzzer, so assigning back to actual fuzzer folks :)

MSan OOM on a huge input -- doesn't not make a lot of sense, but I'll add enforcement of the input length to avoid such crashes.

Uploaded: https://chromium-review.googlesource.com/c/v8/v8/+/1138550

ClusterFuzz has detected this issue as fixed in range 575765:575774.

Detailed report: https://clusterfuzz.com/testcase?key=6080783256387584

Fuzzer: libFuzzer_v8_json_parser_fuzzer
Job Type: libfuzzer_chrome_msan
Platform Id: linux

Crash Type: Out-of-memory (exceeds 2048 MB)
Crash Address: 
Crash State:
  v8_json_parser_fuzzer
  
Sanitizer: memory (MSAN)

Regressed: https://clusterfuzz.com/revisions?job=libfuzzer_chrome_msan&range=525420:525446
Fixed: https://clusterfuzz.com/revisions?job=libfuzzer_chrome_msan&range=575765:575774

Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=6080783256387584

See https://chromium.googlesource.com/chromium/src/+/master/testing/libfuzzer/reference.md for more information.

If you suspect that the result above is incorrect, try re-doing that job on the test case report page.

ClusterFuzz testcase 6080783256387584 is verified as fixed, so closing issue as verified.

If this is incorrect, please add ClusterFuzz-Wrong label and re-open the issue.