On Wed, May 14, 2008 at 06:09:42PM +0300, Dor Laor wrote:
Do you have any performance numbers for networking to see how it
compares to the real hardware?
- Linux host (or: real Windows running on that host)
For host you can measure yourself but for Linux guest (to host) it
currently
On Mon, May 05, 2008 at 02:36:23PM -0700, Kay, Allen M wrote:
+ for (j = 0; j npages; j++) {
+ gpa += PAGE_SIZE;
+ page = gfn_to_page(kvm, gpa PAGE_SHIFT);
+ hpa = page_to_phys(page);
+ domain_page_mapping(kvm-arch.domain, gpa, hpa,
On Wed, Apr 30, 2008 at 01:48:38AM +0300, Avi Kivity wrote:
Amit Shah wrote:
+ if (is_error_page(host_page)) {
+ printk(KERN_INFO %s: gfn %p not valid\n,
+ __func__, (void *)page_gfn);
+ r = -1;
r = -1 is not really informative. Better use some
On Tue, Apr 29, 2008 at 02:09:20PM -0500, Anthony Liguori wrote:
This patch allows VMA's that contain no backing page to be used for guest
memory. This is a drop-in replacement for Ben-Ami's first page in his direct
mmio series. Here, we continue to allow mmio pages to be represented in the
On Tue, Apr 29, 2008 at 01:37:29PM +0300, Amit Shah wrote:
dma_alloc_coherent() doesn't call dma_ops-alloc_coherent in case no
IOMMU translations are necessary.
I always thought this was a huge wart in the x86-64 DMA ops. Would
there be strong resistance to fixing it so that alloc_coherent
On Fri, Apr 18, 2008 at 06:50:03PM +0300, Avi Kivity wrote:
[EMAIL PROTECTED] wrote:
From: Ben-Ami Yassour [EMAIL PROTECTED]
Signed-off-by: Ben-Ami Yassour [EMAIL PROTECTED]
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
---
arch/x86/kvm/mmu.c | 59
On Fri, Apr 18, 2008 at 06:56:41PM +0300, Avi Kivity wrote:
[EMAIL PROTECTED] wrote:
From: Ben-Ami Yassour [EMAIL PROTECTED]
Signed-off-by: Ben-Ami Yassour [EMAIL PROTECTED]
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
---
libkvm/libkvm.c | 24
qemu/hw
On Wed, Nov 07, 2007 at 04:21:05PM +0200, Amit Shah wrote:
A guest can call dma_ops-is_pv_device() to find out if a device is
a passthrough'ed device (device passed on to a guest by the
host). If this is true, a hypercall will be made to translate DMA
mapping operations.
Doesn't really
On Wed, Nov 07, 2007 at 04:21:06PM +0200, Amit Shah wrote:
Of all the DMA calls, only dma_alloc_coherent might not actually
call dma_ops-alloc_coherent. We make sure that gets called if the
device that's being worked on is a PV device
I always thougt that's a mess... the reason it's done this
On Mon, Nov 12, 2007 at 05:26:24PM +0530, Amit Shah wrote:
On Monday 12 November 2007 16:20:01 Muli Ben-Yehuda wrote:
On Wed, Nov 07, 2007 at 04:21:04PM +0200, Amit Shah wrote:
We make the dma_mapping_ops structure to point to our structure so
that every DMA access goes through us
On Mon, Nov 12, 2007 at 07:25:27PM +0530, Amit Shah wrote:
Selectively? What happens in the case when some iommu doesn't want
to invoke the prev_op, but the mapping depends on it being called
(eg, the hypercalling op is embedded somewhere in the prev_op chain)
Bad things :-)
There needs to be
On Thu, Oct 11, 2007 at 10:40:47AM +0200, Laurent Vivier wrote:
There is work in progress for pci pass through capability. Besides
PCI it also required to have pv dma or 1-1 mapping between the
guest and the host. Both will be released in the following
month. NIC pass through works but
On Tue, Oct 09, 2007 at 04:31:54PM -0500, Anthony Liguori wrote:
This patch adds a --memory option to kvmctl to allow the memory size of the
guest to be specified.
Signed-off-by: Anthony Liguori [EMAIL PROTECTED]
diff --git a/user/main.c b/user/main.c
index 9a57a24..7fc0924 100644
---
On Tue, Oct 09, 2007 at 05:09:19PM -0500, Anthony Liguori wrote:
+ default:
+ fprintf(stderr,
+ Unrecongized memory suffix: %c\n,
+ *endptr);
+ exit(1);
On Wed, Aug 08, 2007 at 12:03:03AM -0700, Dor Laor wrote:
Actually we'll do both.
Doubly cool. Looking forward to more information.
Cheers,
Muli
-
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log
On Tue, Aug 07, 2007 at 02:53:01PM -0700, Dor Laor wrote:
Actually we're in the middle of developing pci pass through support
for enabling physical device pass through for guests. It's work in
progress but eventually we'll post the code to the list.
Cool. Are you relying on an
On Sat, Jun 16, 2007 at 11:29:34AM +0100, Richard Hughes wrote:
Is there any easy way we can detect if virtual machines are running?
The module use counts will be elevanted and they'll refuse to unload
(or there's a nasty bug somewhere).
Cheers,
Muli
On Sat, Jun 16, 2007 at 07:36:18PM +0200, Luca wrote:
On 6/16/07, Richard Hughes [EMAIL PROTECTED] wrote:
On Sat, 2007-06-16 at 15:30 +0300, Muli Ben-Yehuda wrote:
Is there any easy way we can detect if virtual machines are running?
The module use counts will be elevanted and they'll
On Sun, Jun 10, 2007 at 10:33:57AM +0300, Avi Kivity wrote:
It is worthwhile, when designing virtio, to keep in mind as many
possible users as possible. In addition to block and net, I see at
least the following:
- vmgl (paravirtualized 3D graphics)
On Wed, Jun 13, 2007 at 01:54:26AM +0200, Arnd Bergmann wrote:
On Wednesday 13 June 2007, Caitlin Bestler wrote:
It can be done, but you'd also need a passthrough for the
IOMMU in that case, and you get a potential security hole: if
a malicious guest is smart enough to figure out
Acceptance Notification September 13, 2007
Camera-ready (full papers) October 11, 2007
SYSTOR 2007 Virtualization Workshop October 29, 2007
Organizing committee
* Erez Hadad, IBM Haifa Research Lab
* Muli Ben-Yehuda, IBM Haifa Research Lab
* Eliezer Dekel, IBM
On Sun, May 13, 2007 at 11:49:14AM -0500, Anthony Liguori wrote:
Dor Laor wrote:
Furthermore,
the plan is to completely rearchitect the netback/netfront
protocol for the next Xen release (this effort is referred to
netchannel2).
But isn't Jeremy Fitzhardinge is pushing big
On Sun, Feb 18, 2007 at 11:47:07AM +0200, Avi Kivity wrote:
Ronald Lew wrote:
Can Guest OS access abnormal PCI cards such as a fax card or a card
designed
for analog phones (like Digium cards)? Can the guest OS gain full access
to
the PCI card without going through the
]
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
Index: configure
===
--- configure (revision 4414)
+++ configure (working copy)
@@ -94,7 +94,7 @@
cat EOF config.mak
PREFIX=$prefix
-KERNELDIR=$(readlink -f $kerneldir)
+KERNELDIR
On Thu, Feb 08, 2007 at 03:46:25PM +, Omar Khan wrote:
hi,
make install does not install the modules in the right directory
(actually
it is not making the 'extra' directory). I changed
KERNELDIR ?= /lib/modules/$(shell uname -r)/build
to
KERNELDIR :=
Now it removes config.make, which doesn't exist? I'm guessing it's a
typo and should be config.mak. Alternatively, we may want config.clean
to only clean the make results, and add make distclean to undo the
configure results.
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
diff -r 9312ff753476
check for KVM_GET/SET_REGS ioctl failures
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
diff -r ee6f88782b5f -r 9a3e101daf18 user/kvmctl.c
--- a/user/kvmctl.c Tue Feb 06 10:55:17 2007 +
+++ b/user/kvmctl.c Tue Feb 06 17:05:56 2007 +0200
@@ -235,19 +235,23 @@ static int handle_io
On Tue, Feb 06, 2007 at 01:26:56PM -0600, Anthony Liguori wrote:
Please don't add exit()'s to a library, these are the sort of things
we should be trying to remove.
I agree on general principles, but do we envision libkvm.a being used
in anything outside qemu? in any case, if not exit(), what
On Sun, Feb 04, 2007 at 11:35:41AM +0200, Avi Kivity wrote:
Markus Rechberger wrote:
This patch fixes the mutex, page and kzalloc problems with older kernels.
Signed-off-by: Markus Rechberger [EMAIL PROTECTED]
The page_private() and kzalloc() stuff looks good, I can apply it.
/iommu/calgary/obj.linux M=`pwd` $@ ==
To support building directly from kernel/, we don't complain if we
can't find config.mak.
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
diff -r 0df0a31d8154 kernel/Makefile
--- a/kernel/Makefile Sun Feb 04 08:38:53 2007 +
+++ b/kernel/Makefile Sun
Hi Avi,
You might want to consider setting the reply-to on kvm-commits to
kvm-devel.
On Thu, Feb 01, 2007 at 02:29:04PM +0200, Avi Kivity wrote:
Modified: kvm/trunk/kernel/vmx.c
==
--- kvm/trunk/kernel/vmx.c
Small cleanup: we check (pdpte 1) (pdpte constant). Instead
just check (pdpte (constant | 1)).
Signed-off-by: Muli Ben-Yehuda [EMAIL PROTECTED]
Index: kernel/kvm_main.c
===
--- kernel/kvm_main.c (revision 4379)
+++ kernel
On Thu, Feb 01, 2007 at 03:08:17PM +0200, Avi Kivity wrote:
I noticed that the svm 'vmcall_interception' doesn't set
kvm_run-exit_reason - is the setting here superflous or needed? and
if it's needed, is it also needed in svm?
A 'return 1' means there won't be an exit to userspace, so
On Thu, Feb 01, 2007 at 04:52:18PM +0200, Avi Kivity wrote:
pdpte == 2 is a legal value. in the original code, the (pdpte 1)
== 0 and we don't 'goto out'. In the patched code, (pdpte 0xblah)
== 2 and we 'goto out'.
Yup, I got it now. Thanks.
Cheers,
Muli
On Thu, Feb 01, 2007 at 06:52:47AM -0800, Dor Laor wrote:
What about 'if (pdpte 0xfff001e7ull 0x1)' ?
Although the original is more readable.
The original is better, perhaps with a nice comment :-)
Cheers,
Muli
35 matches
Mail list logo