linux-kernel - Re: [PATCH 1/38] KVM: Create kvm-intel.ko module

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 30 Nov 2006 16:44:25 +0100
From:	Ingo Molnar <mingo@...e.hu>
To:	Christoph Hellwig <hch@...radead.org>,
	Avi Kivity <avi@...ranet.com>, kvm-devel@...ts.sourceforge.net,
	linux-kernel@...r.kernel.org, akpm@...l.org
Subject: Re: [PATCH 1/38] KVM: Create kvm-intel.ko module

* Christoph Hellwig <hch@...radead.org> wrote:

> > please move this from drivers/kvm/ to kernel/kvm/ [or even into a 
> > toplevel kvm/ directory] - KVM is not a "driver", KVM enhances the 
> > core Linux kernel with hypervisor functionality.
> 
> Actually it's exactly a driver.  It's a character driver that exposes 
> the virtualization features of modern x86 hardware. [...]

you are fundamentally wrong. In the end KVM is a fundamental and complex 
infrastructure that enables Linux to provide full hardware capabilities 
to another OS via the resources of this OS. This concept justifies a 
system call and a place in linux/kernel/. It's not fundamentally limited 
to x86 either. Full virtualization (and paravirtualization) makes sense 
on any platform. And there's no reason KVM be limited to full 
virtualization alone - both paravirtualization and accelerated guest 
drivers need a sane hypercall API.

> [...] Pretty similar to things like the msr or mtrr driver that expose 
> cpu features as character drivers aswell.

you can expose everything as character drivers and ioctls, but that 
doesnt make it the right solution. It might /start out/ as a driver, 
because that's an easy to hack model, but the moment something becomes 
important enough (and virtualization certainly is such a model) it 
demands a system call.

Just like inotify started out as an ioctl hack, but then was 
(rightfully) moved to the system-call space. [ Which btw. was on your 
request ;-) ]

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/