Hi Siddhesh (and others),
I'd like to propose that we start using this script for the GDB project [1], so
I tried
to run it on the binutils-gdb repository. It works great until it hits commit:
4934a27c8c1d5c8623366f5dbafae8af60b96bc0
[binutils][arm] arm support for ARMv8.m Custom Datapath Extension
When we run the script with -d, we see that it is processing the huge
`coprocessor_opcodes`
array in opcodes/arm-dis.c. The script progresses, but slower and slower as it
goes. It is
taking so much time that it would not be practical to use it (I did not run it
until completion,
so I don't know how much time it would take) This seems to show some O(n^2) or
worse behavior.
I'm wondering if that can be improved.
I haven't looked at the implementation of the script to try to figure out why
yet, I thought
I would start by reporting the problem.
Thanks!
Simon
[1] If you want to jump in the discussion, here are the threads:
https://www.sourceware.org/ml/binutils/2020-02/msg00231.html
https://sourceware.org/ml/gdb-patches/2020-02/msg00478.html
