[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] QEMU extension

From: G Portokalidis
Subject: Re: [Qemu-devel] QEMU extension
Date: Wed, 1 Jun 2005 13:43:49 +0200

I think I'm gonna go ahead with Qemu. Bochs seems to be to slow for my purpose.
I was wondering if you know which windows XP versions have been know to work?
Also do you have any performance metrics on popular applications
runing under qemu(ex. apache)? I guess it's in the usenix paper, but i
can't download it.


On 5/26/05, Paul Brook <address@hidden> wrote:
> On Thursday 26 May 2005 17:12, G Portokalidis wrote:
> > Hello,
> >
> > I'm writing concerning a possible use for qemu in a project related to
> > network security.
> >
> > I'm looking for an emulator where I could load an entire (recent) OS,
> > like Linux 2.6 or Windows XP and run multiple, potentially CPU
> > intensive, services (IIS, Apache, MySQL, etc).
> >
> > For the needs of the project I need to be able to know every instruction
> > executed by the guest OS, and run custom code whenever an instruction of
> > particular interest appears (doesn't really matter whether it's C or
> > x86, but preferably the first).
> >
> > So my first question is whether we could run Linux 2.6 and most
> > importantly Windows XP on qemu without stability issues.
> Linux works fine. For windows XP it seems to depend which windows version
> you're using. Some versions work ok, others don't.
> > Second, does
> > the current design of qemu allows me to implement the functionality
> > described in the above paragraph.
> You may be better using bochs. That has instrumentation hooks that should
> allow you do do what you want. boch is significantly slower that qemu, but if
> you're instrumenting a significant number of instructions it's going to be
> dog slow anyway.
> Qemu already has infrastructure for a gdb ICE connection. You could probably
> hack that to do what you want.
> > Finally, what's the performance of qemu compared with a PC (how many
> > times slower)?
> It's generally 10-15x slower than the host.
> Paul

reply via email to

[Prev in Thread] Current Thread [Next in Thread]