Gabe, I must confess the new concept how the SE translating proxy is supposed to work in the new version, is escaping me.
Here is what I am trying to do: I have a non-standard extension to the GDB RSP protocol which allows the gdb client to read/write the inferior's memory without sending M packets over the RSP socket, by SHMEM IPC with the backing store in gem5. For that, the GDB client needs to know which virtual addresses in the guest map to which backing store address. In the old SETranslatingProxy this was easy: the pTable contained exactly that. See https://github.com/shingarov/gem5/commit/b21074a9835ebe52e8f023e588bcad6a608d738e Now with the new design, I am scratching my head how to do something equivalent. It looks like the translation is buried very far behind the functionality of the MMU. Do you think there is still a way to get at it? Boris _______________________________________________ gem5-dev mailing list gem5-dev@gem5.org http://m5sim.org/mailman/listinfo/gem5-dev
