Updated unicorn_dumper_gdb.py to support new gef api and replaced deprecated functions . The functions that are not in the new gef api are read_memory(), and current_arch(). Also replaced some deprecated functions with the updated versions of them.
replaced read_memory() with GefMemoryManager.read() as read_memory(). read_memory() is in legacy-gef-api
replaced current_arch with gef.arch.registers
replaced get_process_maps() with gef.memory.maps (just depreacated)
replaced get_register() with gef.arch.register()
When the computer is suspended during a fuzzing session,
the time spent in suspended state is counted as a "run time"
on a statistics screen.
The time returned by `gettimeofday(2)` is affected by discontinuous
jumps in the system time. It is better using `clock_gettime(2)`.
The patch replace `gettimeofday` with `clock_gettime` [1].
`clock_gettime` uses a CLOCK_MONOTONIC_COARSE clock type,
it is faster than CLOCK_MONOTONIC, but still has resolution (~1ms)
that is adequate for our purposes. However, CLOCK_MONOTONIC_COARSE
is a Linux-specific clock variant, so on macOS it is replaced
with CLOCK_MONOTONIC, and with CLOCK_MONOTONIC_FAST on FreeBSD [2].
Closes#1241
1. https://pubs.opengroup.org/onlinepubs/9699919799/functions/clock_gettime.html
2. https://man.freebsd.org/cgi/man.cgi?query=clock_gettime
- Introduce the -l option to set min block deletion length using
powers of 2 (e.g., 1, 2, 4, 8, 16, ...).
- This enables a trade-off between minimization thoroughness and speed.
- Adjusting del_len_limit allows for faster processing, as doubling it
roughly halves the minimization time.
Fixes a bug where the new fork server would decrement dict_size until
zero then try to use it as the upper bound for the number of bytes to
pass to add_extra_func, causing it to never store any of the tokens.
