[qubes-os] "Merging new testcases failed:"

[paraschetal]

I am unable to access the performance report and coverage statistics for the libqubes-rpc-filecopy fuzz target. Based on the logs, I think it might be because the fuzz targets are not being executed.

Command: /mnt/scratch0/clusterfuzz/slave-bot/builds/clusterfuzz-builds_qubes-os_dba2089d23e7b835b08c68a02d4a2dca7a3c102b/revisions/libqubes-rpc-filecopy -max_len=10240 -timeout=25 -rss_limit_mb=2048 -artifact_prefix=/ -max_total_time=2950 -print_final_stats=1 /new /qubes-os_libqubes-rpc-filecopy
Bot: oss-fuzz-linux-zone1-pre-qubes-os-6vdq
Time ran: 0.117007

INFO: Seed: 525719240
INFO: Loaded 1 modules (126 guards): [0x883530, 0x883728),
Loading corpus dir: /new
Loading corpus dir: /qubes-os_libqubes-rpc-filecopy
INFO: A corpus is not provided, starting from an empty corpus
#0  READ units: 1
EOF
Merging new testcases failed:

I have tested running the fuzz targets locally using the infra/helper.py script, and that does works fine.

user@work ~/p/oss-fuzz> python infra/helper.py run_fuzzer qubes-os libqubes-rpc-filecopy                                                              qubes-linux-utils
Running: docker run --rm -i --cap-add SYS_PTRACE -e FUZZING_ENGINE=libfuzzer -v /home/user/projects/oss-fuzz/build/out/qubes-os:/out -t gcr.io/oss-fuzz-base/base-runner run_fuzzer libqubes-rpc-filecopy
/out/libqubes-rpc-filecopy -rss_limit_mb=2048 -timeout=25 -max_len=10240
INFO: Seed: 662960318
INFO: Loaded 1 modules (126 guards): [0x8865f0, 0x8867e8), 
INFO: A corpus is not provided, starting from an empty corpus
#0  READ units: 1
#1  INITED cov: 29 ft: 27 corp: 1/1b exec/s: 0 rss: 30Mb
#10 NEW    cov: 29 ft: 28 corp: 2/126b exec/s: 0 rss: 30Mb L: 125 MS: 4 CMP-CMP-CopyPart-InsertRepeatedBytes- DE: "\x00\x00\x00\x00\x00\x00\x00 "-"handle_ill"-
#51 NEW    cov: 30 ft: 33 corp: 3/10366b exec/s: 51 rss: 31Mb L: 10240 MS: 5 CopyPart-CMP-CMP-ShuffleBytes-CrossOver- DE: "\x00\x00\x00\x00\x00\x00\x00\x07"-"rss_l"-
...

[oliverchang]

Will take a closer look today. In the meantime, please take a look https://github.com/google/oss-fuzz/blob/master/docs/fuzzer_environment.md in case there's something obvious here.

[oliverchang]

Sorry for the delay, but I finally got around to taking a look at this.

If I do:

./libqubes-rpc-filecopy < /dev/null

Then the fuzz target exits instantly. This is similar to how we run the target on our infrastructure.

Is target expecting input on stdin?

[paraschetal]

Hmm, weird...It shouldn't be expecting any input. I'll check what's wrong.

[paraschetal]

It should be fixed now. Still, let's just wait till the changes are reflected in the clusterfuzz logs.

[inferno-chromium]

@Dor1s - can you confirm if things look ok and then close this tmrw.

[Dor1s]

Confirmed, the most recent version works fine.