gdb: potential bug due to double address translation

[ludfjig]

In add_sw_breakpoint and remove_sw_breakpoint, the address is translated from GVA to GPA using translate_gva, but then the resulting GPA is passed to read_addrs/write_addrs which internally call translate_gva again, treating the already-translated GPA as a GVA.

I would expect this to cause incorrect memory access since the physical address gets translated a second time?

[IrvingMg]

Hi, I'd like to work on this issue.

[jsturtevant]

@IrvingMg go for it, I am not sure we've fully verified this is an issue. If you have any questions or trouble you can ask questions in slack or here on the issue