I don’t think the fuzz target has ever found a real issue. The closest being #25869
It is also, by far, the slowest fuzz target. For example, looking at https://cirrus-ci.com/task/5533338067271680?logs=ci#L3974, it takes more than one hour:
0Run wallet_notifications with args ['/ci_container_base/ci/scratch/build-x86_64-pc-linux-gnu/src/test/fuzz/fuzz', '-runs=1', PosixPath('/ci_container_base/ci/scratch/qa-assets/fuzz_corpora/wallet_notifications')]INFO: Running with entropic power schedule (0xFF, 100).
1INFO: Seed: 1096115652
2INFO: Loaded 1 modules (625824 inline 8-bit counters): 625824 [0x5628396d9138, 0x562839771dd8),
3INFO: Loaded 1 PC tables (625824 PCs): 625824 [0x562839771dd8,0x56283a0fe7d8),
4INFO: 1287 files found in /ci_container_base/ci/scratch/qa-assets/fuzz_corpora/wallet_notifications
5INFO: -max_len is not provided; libFuzzer will not generate inputs larger than 1047827 bytes
6INFO: seed corpus: files: 1287 min: 1b max: 1047827b total: 11616898b rss: 172Mb
7[#16](/bitcoin-bitcoin/16/) pulse cov: 14328 ft: 25341 corp: 14/239b exec/s: 5 rss: 204Mb
8[#64](/bitcoin-bitcoin/64/) pulse cov: 19179 ft: 58412 corp: 61/3587b exec/s: 5 rss: 320Mb
9[#128](/bitcoin-bitcoin/128/) pulse cov: 19692 ft: 85738 corp: 125/16Kb exec/s: 3 rss: 544Mb
10[#256](/bitcoin-bitcoin/256/) pulse cov: 19923 ft: 107490 corp: 253/72Kb exec/s: 2 rss: 556Mb
11[#512](/bitcoin-bitcoin/512/) pulse cov: 20107 ft: 124704 corp: 509/330Kb exec/s: 2 rss: 590Mb
12Slowest unit: 10 s:
13artifact_prefix='./'; Test unit written to ./slow-unit-9fa5f7d7e4afa1626622ef1b3c70a7563eecf11d
14[#1024](/bitcoin-bitcoin/1024/) pulse cov: 20360 ft: 136324 corp: 1009/2488Kb exec/s: 0 rss: 726Mb
15Slowest unit: 23 s:
16artifact_prefix='./'; Test unit written to ./slow-unit-5d99a20de2c2b6bedb0cbaf0ba3743ae3ba13c7c
17Slowest unit: 26 s:
18artifact_prefix='./'; Test unit written to ./slow-unit-8889ecb61bdc0650355e0d0d27c012f3239d07a4
19Slowest unit: 42 s:
20artifact_prefix='./'; Test unit written to ./slow-unit-d16c084282ac1a85fcdc43c48e49836b08446686
21[#1289](/bitcoin-bitcoin/1289/) INITED cov: 20409 ft: 138281 corp: 1245/10323Kb exec/s: 0 rss: 880Mb
22[#1289](/bitcoin-bitcoin/1289/) DONE cov: 20409 ft: 138281 corp: 1245/10323Kb lim: 1047827 exec/s: 0 rss: 880Mb
23Done 1289 runs in 3813 second(s)
Looking at the flame graphs, it looks like the slow runs spend most of their time in the Knapsack solver. This seems reasonable, because it may run 1000 inner Knapsack iterations 200 times. So reduce the fuzz iterations from 200 to 20 to avoid fuzz timeouts and wasted resources.
