All of lore.kernel.org
 help / color / mirror / Atom feed
From: Arnd Bergmann <arnd@arndb.de>
To: Avi Kivity <avi@qumranet.com>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 0/7] KVM: Kernel-based Virtual Machine
Date: Sat, 21 Oct 2006 18:16:27 +0200	[thread overview]
Message-ID: <200610211816.27964.arnd@arndb.de> (raw)
In-Reply-To: <4537BD27.7050509@qumranet.com>

On Thursday 19 October 2006 20:00, Avi Kivity wrote:
> Working code is fairly hairy, since it's emulating a PC.  That'll be on
> sourceforge once they approve my new project.
>
> In general one does
>
>   open("/dev/kvm")
>   ioctl(KVM_SET_MEMORY_REGION) for main memory
>   ioctl(KVM_SET_MEMORY_REGION) for the framebuffer
>   ioctl(KVM_CREATE_VCPU) for the obvious reason
>   if (debugger)
>     ioctl(KVM_DEBUG_GUEST) to singlestep or breakpoint the guest
>   while (1) {
>      ioctl(KVM_RUN)
>      switch (exit reason) {
>          handle mmio, I/O etc. might call
>             ioctl(KVM_INTERRUPT) to queue an external interrupt
>             ioctl(KVM_{GET,SET}_{REGS,SREGS}) to query/modify registers
>             ioctl(KVM_GET_DIRTY_LOG) to see which guest memory pages
> have changed
>      }
>
> I have some simple test code, I'll clean it up and post it.

This looks _a_lot_ like what we're doing for the SPUs in the cell processor,
except that we're using different calls into the kernel. Have you looked
into what we have implemented there? The code is in
arch/powerpc/platforms/cell/spufs. I think it would be a good abstraction
to use for you as well, maybe we could even move to a common infrastructure,
as I have heard from a few other projects that want to do similar things.

The main differences to your interface are:

- A file system is used instead of a character device
- Directories, not open file descriptors represent contexts
- Two new syscalls were introduced (spu_create/spu_run)
- instead of ioctls, files represent different bits of information,
  you can read/write, poll or mmap them.

Your example above could translate to something like:

   int kvm_fd = kvm_create("/kvm/my_vcpu")
   int mem_fd = openat(kvm_fd, "mem", O_RDWR);
   void *mem = mmap(mem_fd, ...); // main memory
   void *fbmem = mmap(mem_fd, ...); // frame buffer memory
   int regs_fd = openat(kvm_fd, "regs", O_RDWR);
   int irq_fd = openat(kvm_fd, "regs", O_WRONLY);

   if (debugger) {
     int fd = openat(fvm_fd, "debug", O_WRONLY);
     write(fd, "1", 1);
     close(fd);
   }
   while (1) {
      int exit_reason = kvm_run(kvm_fd, &kvm_descriptor);
      switch (exit reason) {
          handle mmio, I/O etc. might call
             write(irq_fd, &interrupt_packet, sizeof (interrupt_packet));
             pread(regs_fd, &rax, sizeof rax, KVM_REG_RAX);
   }

	Arnd <><

  parent reply	other threads:[~2006-10-21 16:16 UTC|newest]

Thread overview: 86+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-10-19 13:45 [PATCH 0/7] KVM: Kernel-based Virtual Machine Avi Kivity
2006-10-19 13:47 ` [PATCH 1/7] KVM: userspace interface Avi Kivity
2006-10-19 14:30   ` John Stoffel
2006-10-19 14:43     ` Avi Kivity
2006-10-19 23:26       ` Greg KH
2006-10-19 14:50     ` Alan Cox
2006-10-19 14:51       ` Avi Kivity
2006-10-19 15:25         ` John Stoffel
2006-10-19 18:49       ` Anthony Liguori
2006-10-19 19:10         ` Avi Kivity
2006-10-19 19:17           ` Anthony Liguori
2006-10-20  7:36             ` Avi Kivity
2006-10-20 15:33               ` Anthony Liguori
2006-10-22  8:10                 ` Avi Kivity
2006-10-19 20:36         ` Alan Cox
2006-10-19 18:46   ` Anthony Liguori
2006-10-19 19:04     ` Avi Kivity
2006-10-19 19:09       ` Anthony Liguori
2006-10-19 19:26         ` Avi Kivity
2006-10-19 19:31           ` Anthony Liguori
2006-10-19 22:15             ` Alan Cox
2006-10-20  7:42             ` Avi Kivity
2006-10-20 15:35               ` Anthony Liguori
2006-10-19 20:10       ` Andi Kleen
2006-10-19 20:14   ` Jan Engelhardt
2006-10-20  7:16     ` Avi Kivity
2006-10-21 15:50       ` Arnd Bergmann
2006-10-22  8:19         ` Avi Kivity
2006-10-21 13:37   ` Steven Rostedt
2006-10-22  8:14     ` Avi Kivity
2006-10-19 13:48 ` [PATCH 2/7] KVM: Intel virtual mode extensions definitions Avi Kivity
2006-10-19 20:19   ` Jan Engelhardt
2006-10-19 21:54     ` Alan Cox
2006-10-20  7:17     ` Avi Kivity
2006-10-21 13:48   ` Steven Rostedt
2006-10-22  8:17     ` Avi Kivity
2006-10-19 13:49 ` [PATCH 3/7] KVM: kvm data structures Avi Kivity
2006-10-19 13:53 ` [PATCH 5/7] KVM: mmu virtualization Avi Kivity
2006-10-19 20:26   ` Jan Engelhardt
2006-10-20  7:24     ` Avi Kivity
2006-10-19 13:54 ` [PATCH 6/7] KVM: x86 emulator Avi Kivity
2006-10-19 13:56 ` [PATCH 7/7] KVM: plumbing Avi Kivity
2006-10-19 13:58 ` [PATCH 0/7] KVM: Kernel-based Virtual Machine Avi Kivity
2006-10-19 16:05 ` Andi Kleen
2006-10-19 16:09   ` Avi Kivity
2006-10-19 19:02     ` Anthony Liguori
2006-10-19 19:14       ` Avi Kivity
2006-10-19 19:28         ` Anthony Liguori
2006-10-20  7:37           ` Avi Kivity
2006-10-19 17:31 ` Muli Ben-Yehuda
2006-10-19 18:00   ` Avi Kivity
2006-10-19 18:12     ` Randy Dunlap
2006-10-19 18:14       ` Avi Kivity
2006-10-19 18:30         ` Randy.Dunlap
2006-10-21 16:16     ` Arnd Bergmann [this message]
2006-10-22  8:37       ` Avi Kivity
2006-10-22 15:23         ` Arnd Bergmann
2006-10-22 16:18           ` Avi Kivity
2006-10-22 16:51             ` Arnd Bergmann
2006-10-22 17:01               ` Avi Kivity
2006-10-22 17:06                 ` Arnd Bergmann
2006-10-22 17:41                   ` Avi Kivity
2006-10-22 17:47                     ` Arnd Bergmann
2006-10-22 17:56                 ` Christoph Hellwig
2006-10-22 18:00                   ` Avi Kivity
2006-10-22 18:36                     ` Arnd Bergmann
2006-10-22 18:41                       ` Avi Kivity
2006-10-22 18:49                         ` Arnd Bergmann
2006-10-22 18:55                           ` Avi Kivity
2006-10-22 22:26                     ` Andi Kleen
2006-10-23 22:29                       ` Jeremy Fitzhardinge
2006-10-22 20:01                   ` Alan Cox
2006-10-22 20:45                   ` Roland Dreier
2006-10-23  0:29                   ` Anthony Liguori
2006-10-25 16:42                   ` Pavel Machek
2006-10-22 19:59               ` Alan Cox
2006-10-22 22:28                 ` Andi Kleen
2006-10-23  0:27                   ` Roland Dreier
2006-10-23  0:39                     ` Andi Kleen
2006-10-23  0:51                       ` Roland Dreier
2006-10-22 17:39         ` Anthony Liguori
2006-10-22 17:53           ` Arnd Bergmann
2006-10-22 19:56         ` Alan Cox
2006-10-23  7:42           ` Avi Kivity
2006-10-24 21:38       ` kvm_create() (was Re: [PATCH 0/7] KVM: Kernel-based Virtual Machine) Andy Isaacson
2006-10-19 18:55   ` [PATCH 0/7] KVM: Kernel-based Virtual Machine Anthony Liguori

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200610211816.27964.arnd@arndb.de \
    --to=arnd@arndb.de \
    --cc=avi@qumranet.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=muli@il.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.