diff options
author | Michael Ensing <michael.ensing@leviathansecurity.com> | 2020-05-01 16:00:17 -0700 |
---|---|---|
committer | Michael Ensing <michael.ensing@leviathansecurity.com> | 2020-05-13 22:30:13 -0700 |
commit | f0b8122ab9b1696a80988e2b51d3da3d715f46a8 (patch) | |
tree | 496eeceafd826f217a091d9cd0d48d7254526651 /system/tools/scripts/dump_metrics_ascii.py | |
parent | c5ca7d3a766f84d58119d18d8df738975204c7a9 (diff) |
Add fuzzers for 10 libosi components
Added fuzzers, most of which may be run either on host or on device.
Note that coverage will differ based on arch. Additionally, alarm
will not successfully run on host due to the reliance on hooking
system functions, and compat may not be useful on-device, as libosi's
compat.h function definitions, and therefore the fuzzer calls, are
wrapped in an '#if __GLIBC__'.
Test: Tested on a Pixel 3a with no starting corpus and a short run,
resulting in the following statistics:
- alarm 15% cov ~280 exec/s
- allocation_tracker: 61% cov ~2.5k exec/s
- allocator: 34% cov ~24k exec/s
- array: 22% cov ~16k exec/s
- buffer: 22% cov ~26k exec/s
- compat: N/A N/A
- fixed_queue: 45% cov ~8k exec/s
- future 22% cov ~83k exec/s
- list 46% cov ~21k exec/s
- ringbuffer 38% cov ~20k exec/s
Signed-off-by: Michael Ensing <michael.ensing@leviathansecurity.com>
Change-Id: I9203c804941beaa470940452f39bf352d725eb02
Diffstat (limited to 'system/tools/scripts/dump_metrics_ascii.py')
0 files changed, 0 insertions, 0 deletions