Re: [PATCH 18/25] [PATCH] turn priviled operations into macros in entry.S
From: Steven Rostedt <rostedt@goodmis.org>
Date: 2007-08-08 14:01:45
-- On Wed, 8 Aug 2007, Andi Kleen wrote:
quoted
When I said "this part of the code I don't fully understand" I was not talking about entry.S. I understand entry.S very well, but the comment was originally on the paranoid_restore code. Which I thought had to deal with NMIs and such that I didn't worry about that I simply did the default.The paranoid path is used for more than just NMIs; it's also used for MCEs, stack faults, double faults or debug exceptions. Anything that might happen with a invalid stack or unknown GS state or system in other unknown state. If you can guarantee your hypervisor never injects any of those it could be ignored; but at least losing debug exceptions would be probably not nice.
For now it does ignore them. But that's something to work on for later versions of lguest64. I want lguest out in the public and working for the general case. I don't plan on ignoring the paranoid section forever. But it's more on an anomaly for now. But you are right, I want the debug stuff working. But I also need to walk before I can run ;-)
quoted
quoted
quoted
paranoid_restore\trace: RESTORE_ALL 8 - iretq + INTERRUPT_RETURNI suspect Xen will need much more changes anyways because of its ring 3 guest. Are these changes sufficient for lguest?This was really a general comment not especially applying to the paranoid path.
Ah, OK, I assumed you were talking about just the previous code. But rereading your post, I see it was more general. Sorry, for the miscommunication. -- Steve