It looks like the comInstEventQueue will actually help with this quite a
bit. It's currently used to stop the simulation after a certain number of
instructions, but it can just as easily (I think) be used to trap into GDB
after a single instruction, or to get it to break into GDB initially at the
I'm working on adding remote GDB support for x86, and I've noticed a couple
of rough edges I'd like to smooth out somehow.
First, when GDB attaches, a pollEvent will be added which will give the GDB
stub control immediately. The context being interrupted might not be on an
instruction boundary.
Actually we have some patches that enable remote gdb for x86, including SE
mode. They were done by an intern we had this past summer. They need a
little cleaning up and merging with the head before being posted, and I
haven't had time for that, but if you'd like I can post them anyway.
I don't
Yes, please do. I'll keep working on it, and if I get something working
I'll post some reviews as well. We can pick the parts of each that will
work best. I like the sound of your idea, Ali, but I'll have to look into
probe points. Are those supported on all the CPU models? I doubt they're