* Code is easier to review than bullet points.
Yes. We'd send the code soon.
* Direct I/O has to be safe when page is shared by multiple threads,
and has to be non-blocking since network I/O can take indeterminately
long (think big queue's, tunneling, ...)
In the situation, one queue
On (Tue) Sep 01 2009 [19:18:05], Avi Kivity wrote:
This should really be in the kernel, but there are too many of them.
Signed-off-by: Avi Kivity a...@redhat.com
---
v2: try /dev/cpu/0/msr first
kvm/scripts/vmxcap | 154
Why not
Glauber Costa wrote:
diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c
index fc409e9..3a4e1bd 100644
--- a/arch/x86/kernel/kvmclock.c
+++ b/arch/x86/kernel/kvmclock.c
@@ -27,7 +27,7 @@
#define KVM_SCALE 22
static int kvmclock = 1;
-static unsigned int
On 09/02/2009 09:50 AM, Amit Shah wrote:
On (Tue) Sep 01 2009 [19:18:05], Avi Kivity wrote:
This should really be in the kernel, but there are too many of them.
Signed-off-by: Avi Kivitya...@redhat.com
---
v2: try /dev/cpu/0/msr first
kvm/scripts/vmxcap | 154
On (Wed) Sep 02 2009 [10:54:05], Avi Kivity wrote:
On 09/02/2009 09:50 AM, Amit Shah wrote:
On (Tue) Sep 01 2009 [19:18:05], Avi Kivity wrote:
This should really be in the kernel, but there are too many of them.
Signed-off-by: Avi Kivitya...@redhat.com
---
v2: try /dev/cpu/0/msr first
On 09/02/2009 11:02 AM, Amit Shah wrote:
It's just a hack.
In that case there's danger of this becoming a stable interface
- when eg libvirt picks up support for it. It is useful
functionality.
They'll learn when the first spelling fix comes in. Don't use text APIs.
--
error
On (Wed) Sep 02 2009 [11:18:10], Avi Kivity wrote:
On 09/02/2009 11:02 AM, Amit Shah wrote:
It's just a hack.
In that case there's danger of this becoming a stable interface
- when eg libvirt picks up support for it. It is useful
functionality.
They'll learn when the first
On 09/02/2009 01:24 PM, Amit Shah wrote:
So then is this just meant to be a debug or informational tool?
It's a stopgap until we expose it via /proc/cpuinfo or something similar.
--
error compiling committee.c: too many arguments to function
--
To unsubscribe from this list: send the
The general idea of test case comparison is to compare a given test case
execution with other executions of the same test according to a
predefined
list of attributes.
The comparison result is a list of test executions ordered by attributes
mismatching distance (same idea as hamming distance, just
On 09/02/2009 05:24 AM, Max Laier wrote:
Okay ... I have/something/, but I'm certainly not there yet as I have to
admit that I don't understand your idea completely. In addition it seems that
EPT doesn't have an accessed bit :-\ Any idea for this?
Use the rwx bits as an approximation.
On Wed, Sep 02, 2009 at 08:54:37AM +0200, Chris Lalancette wrote:
Glauber Costa wrote:
diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c
index fc409e9..3a4e1bd 100644
--- a/arch/x86/kernel/kvmclock.c
+++ b/arch/x86/kernel/kvmclock.c
@@ -27,7 +27,7 @@
#define
On 09/02/2009 02:31 PM, Glauber Costa wrote:
I think the overall idea is very interesting, but I also think that it should be
disabled by default. Because of the problems with time in virtualization,
people are already conditioned to run ntpd inside their guests, and this
kvmclock change will
On Wed, Sep 02, 2009 at 02:44:11PM +0300, Avi Kivity wrote:
On 09/01/2009 02:50 PM, Glauber Costa wrote:
KVM clock is great to avoid drifting in guest VMs running ontop of kvm.
However, the current mechanism will not propagate changes in wallclock value
upwards. This effectively means that in
On 09/01/2009 02:50 PM, Glauber Costa wrote:
KVM clock is great to avoid drifting in guest VMs running ontop of kvm.
However, the current mechanism will not propagate changes in wallclock value
upwards. This effectively means that in a large pool of VMs that need accurate
timing,
all of them
On Wed, Sep 02, 2009 at 03:24:26PM +0300, Avi Kivity wrote:
On 09/02/2009 03:21 PM, Glauber Costa wrote:
+static void kvm_sync_wall_clock(struct work_struct *work)
+{
+ struct timespec now;
+
+ kvm_get_wall_ts(now);
What happens if we schedule here?
hummm, I guess
On 09/02/2009 03:48 PM, Glauber Costa wrote:
You can't prevent host preemption. You might read kvmclock again and
repeat if too much time has passed.
But then you can be scheduled after you did settimeofday, but before reading
kvmclock again. Since we're aiming for periodic adjustments
i folks,
In this proposed patch, I am introducing a worker fired by kvmclock that updates
guest wallclock periodically to reflect changes in the host's wallclock. With
this
patch, a large pool of VMs will no longer have to run NTP in all of its guests.
The worker does that at a configurable
This patch introduces a new sysctl called kvm_sync_wallclock.
It controls the behaviour of the worker that updates guest wallclock time.
The worker will fire in periods specified by its value, if it is greater than
zero,
and not fire at all otherwise.
Signed-off-by: Glauber Costa
KVM clock is great to avoid drifting in guest VMs running ontop of kvm.
However, the current mechanism will not propagate changes in wallclock value
upwards. This effectively means that in a large pool of VMs that need accurate
timing,
all of them has to run NTP, instead of just the host doing
Hi Marcelo,
qemu-kvm's installation (0.11-rc as well as git head) is currently
broken /wrt extboot.bin. Please check some proper binary in as
pc-bios/extboot.bin.
Thanks,
Jan
--
Siemens AG, Corporate Technology, CT SE 2
Corporate Competence Center Embedded Linux
--
To unsubscribe from this
On 09/02/2009 05:34 PM, Jan Kiszka wrote:
Hi Marcelo,
qemu-kvm's installation (0.11-rc as well as git head) is currently
broken /wrt extboot.bin. Please check some proper binary in as
pc-bios/extboot.bin.
It's built at runtime. Why check in a binary?
--
error compiling committee.c: too
Avi Kivity wrote:
On 09/02/2009 05:34 PM, Jan Kiszka wrote:
Hi Marcelo,
qemu-kvm's installation (0.11-rc as well as git head) is currently
broken /wrt extboot.bin. Please check some proper binary in as
pc-bios/extboot.bin.
It's built at runtime. Why check in a binary?
First of all,
On 09/02/2009 05:51 PM, Jan Kiszka wrote:
Avi Kivity wrote:
On 09/02/2009 05:34 PM, Jan Kiszka wrote:
Hi Marcelo,
qemu-kvm's installation (0.11-rc as well as git head) is currently
broken /wrt extboot.bin. Please check some proper binary in as
pc-bios/extboot.bin.
It's
Avi Kivity wrote:
On 09/02/2009 05:51 PM, Jan Kiszka wrote:
Avi Kivity wrote:
On 09/02/2009 05:34 PM, Jan Kiszka wrote:
Hi Marcelo,
qemu-kvm's installation (0.11-rc as well as git head) is currently
broken /wrt extboot.bin. Please check some proper binary in as
Jan Kiszka wrote:
Avi Kivity wrote:
On 09/02/2009 05:51 PM, Jan Kiszka wrote:
Avi Kivity wrote:
On 09/02/2009 05:34 PM, Jan Kiszka wrote:
Hi Marcelo,
qemu-kvm's installation (0.11-rc as well as git head) is currently
broken /wrt extboot.bin. Please check some proper binary in
The following patches implement nested VMX support. The patches enable a guest
to use the VMX APIs in order to run its own nested guest (i.e., enable running
other hypervisors which use VMX under KVM). The current patches support running
Linux under a nested KVM using shadow page table (with
From: Orit Wasserman or...@il.ibm.com
---
arch/x86/kvm/vmx.c | 24 +++-
1 files changed, 23 insertions(+), 1 deletions(-)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index abba325..2b1fc3b 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -70,6 +70,8 @@
From: Orit Wasserman or...@il.ibm.com
---
arch/x86/kvm/vmx.c | 533 +++-
1 files changed, 523 insertions(+), 10 deletions(-)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 2b1fc3b..5ab07a0 100644
--- a/arch/x86/kvm/vmx.c
+++
From: Orit Wasserman or...@il.ibm.com
---
arch/x86/kvm/vmx.c | 1142 +++-
1 files changed, 1130 insertions(+), 12 deletions(-)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 62139b5..a7a62df 100644
--- a/arch/x86/kvm/vmx.c
+++
From: Orit Wasserman or...@il.ibm.com
---
arch/x86/kvm/vmx.c | 556 +++-
1 files changed, 554 insertions(+), 2 deletions(-)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 5ab07a0..2453c67 100644
--- a/arch/x86/kvm/vmx.c
+++
On 09/02/2009 06:26 PM, Jan Kiszka wrote:
Well, people may expect qemu-kvm's source tree to build properly also
for non-kvm usage. But this is not that important for now.
It's not tested for non-kvm usage unfortunately.
Ah, I see: some out-of-tree build bug. Will fix.
That too is
On 09/02/2009 06:31 PM, Jan Kiszka wrote:
Ah, I see: some out-of-tree build bug. Will fix.
And this patch is missing in master, isn't it?
IIRC it works in master somehow.
--
error compiling committee.c: too many arguments to function
--
To unsubscribe from this list: send the
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
Makefile |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/Makefile b/Makefile
index d1de788..2217956 100644
--- a/Makefile
+++ b/Makefile
@@ -306,7 +306,7 @@ ifneq ($(BLOBS),)
$(INSTALL_DATA)
Am 02.09.2009 um 17:38 schrieb or...@il.ibm.com:
The following patches implement nested VMX support. The patches
enable a guest
to use the VMX APIs in order to run its own nested guest (i.e.,
enable running
other hypervisors which use VMX under KVM).
Copl! Great job here. I was expecting
Based on Avi's version:
Pick up option roms from the build directory if available. As we do not
ship a extboot.bin binary, this also fixes the installation of this rom.
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
Makefile |3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
Avi Kivity wrote:
On 09/02/2009 06:31 PM, Jan Kiszka wrote:
Ah, I see: some out-of-tree build bug. Will fix.
And this patch is missing in master, isn't it?
IIRC it works in master somehow.
Old extboot.bin still hanging around in your pc-bios folder...? :)
Jan
--
Siemens AG,
On 09/02/2009 07:17 PM, Jan Kiszka wrote:
IIRC it works in master somehow.
Old extboot.bin still hanging around in your pc-bios folder...? :)
I'm sure I deleted them... but can't argue with reality.
--
error compiling committee.c: too many arguments to function
--
To unsubscribe
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
v2: Removed another incorrect reference to $(SRC_PATH)
Makefile |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/Makefile b/Makefile
index d1de788..f2b8885 100644
--- a/Makefile
+++ b/Makefile
@@ -305,8 +305,8 @@
Based on Avi's version:
Pick up option roms from the build directory if available. As we do not
ship a extboot.bin binary, this also fixes the installation of this rom.
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
v2: Removed forgotten reference to $(SRC_PATH)
Makefile |3 +++
1
Hello,
we're having problem one of our kvm guests. According to storage summary mail
Christopher
sent few days ago, I switched my disk model to SCSI which should be the safest
choice.
But now we've stumbled upon the problem that the whole guest crashes when we
run one
specific postgres query
On Tue, Sep 01, 2009 at 04:06:25PM +0300, Avi Kivity wrote:
Only reload debug register 6 if we're running with the guest's
debug registers. Saves around 150 cycles from the guest lightweight
exit path.
dr6 contains a couple of bits that are updated on #DB, so intercept
that unconditionally
On Tue, Sep 01, 2009 at 07:18:05PM +0300, Avi Kivity wrote:
This should really be in the kernel, but there are too many of them.
Signed-off-by: Avi Kivity a...@redhat.com
---
v2: try /dev/cpu/0/msr first
kvm/scripts/vmxcap | 154
Hi All,
I am interested to know how the KVM Synchronize with Qemu. I know KVM maintain
the time in PIT/RTC and Qemu it's own thread for IO devices callback (using
alarm timers). How it is guaranteed that Qemu maintained view of time is not
passing to slow or too fast w.r.t KVM VCPu.
Also,
On Wed, Sep 02, 2009 at 06:27:15PM +0200, Jan Kiszka wrote:
Based on Avi's version:
Pick up option roms from the build directory if available. As we do not
ship a extboot.bin binary, this also fixes the installation of this rom.
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
v2:
On 09/02/2009 01:47 AM, Mohammed Gamal wrote:
This adds tests for pusha/popa to the test harness. Added some new typedefs
as well.
It would be good to test memory contents after pusha, and register
contents after popa.
--
I have a truly marvellous patch that fixes the bug which this
On 09/02/2009 08:23 PM, Nikola Ciprich wrote:
Hello,
we're having problem one of our kvm guests. According to storage summary mail
Christopher
sent few days ago, I switched my disk model to SCSI which should be the safest
choice.
But now we've stumbled upon the problem that the whole guest
Hi there,
This is probably minor, but I'm getting the following segfault when
trying to run qemu-kvm w/o kvm modules loaded:
sudo ./qemu-kvm -hda fedora-11-kratos-i386.img -m 1G -snapshot
open /dev/kvm: No such file or directory
Could not initialize KVM, will disable KVM support
Segmentation
On 09/02/2009 06:38 PM, or...@il.ibm.com wrote:
From: Orit Wassermanor...@il.ibm.com
---
arch/x86/kvm/svm.c |3 -
arch/x86/kvm/vmx.c | 187 +++-
arch/x86/kvm/x86.c |6 ++-
arch/x86/kvm/x86.h |2 +
4 files changed, 192
On 09/02/2009 06:38 PM, or...@il.ibm.com wrote:
From: Orit Wassermanor...@il.ibm.com
---
arch/x86/kvm/vmx.c | 24 +++-
1 files changed, 23 insertions(+), 1 deletions(-)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index abba325..2b1fc3b 100644
---
On 09/02/2009 06:38 PM, or...@il.ibm.com wrote:
+struct __attribute__ ((__packed__)) level_state {
+ struct shadow_vmcs *shadow_vmcs;
+
+ u16 vpid;
+ u64 shadow_efer;
+ unsigned long cr2;
+ unsigned long cr3;
+ unsigned long cr4;
+ unsigned long cr8;
+
+
Thanks for the reply. Can keeping device models and KVM pit synchronize with
the host clock result in any issues with booting up OSes?
-Abhishek
-Original Message-
From: Avi Kivity [mailto:a...@redhat.com]
Sent: Wednesday, September 02, 2009 12:16 PM
To: Saksena, Abhishek
Cc:
On 09/02/2009 06:38 PM, or...@il.ibm.com wrote:
+
+static void init_vmcs_field_to_offset_table(void)
+{
+ memset(vmcs_field_to_offset_table,0xff,
+ sizeof(vmcs_field_to_offset_table));
+
+ vmcs_field_to_offset_table[VIRTUAL_PROCESSOR_ID] =
+ offsetof(struct
Correction to question:-
Can keeping device models and KVM pit synchronize with the host clock
separately result in any issues with booting up OSes?
-Abhishek
-Original Message-
From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On Behalf Of
Saksena, Abhishek
Sent:
The current KVM build test contains a workaround for
older repository formats: If some roms are not found
under the pc-bios directory, copy them to that directory.
However the workaround might mask a problem on the
build scripts, since one of the purposes of the test
is to verify that the KVM
Hello Avi, thnaks for reply.
scsi is hardly tested, so this isn't surprising.
well, but if it's still the most safe way to go, then there's not much other
choices, is it?
AMD or Intel? Uni or smp guests?
intel, SMP guests. sorry for not reporting this right away..
The host crash is of
On 09/02/2009 06:38 PM, or...@il.ibm.com wrote:
-struct nested_vmx {
- /* Has the level1 guest done vmon? */
+struct nested_vmx {/* Has the level1 guest done vmon? */
A \n died here.
bool vmon;
/* Has the level1 guest done vmclear? */
bool vmclear;
+
+
On 09/03/2009 12:14 AM, Nikola Ciprich wrote:
Hello Avi, thnaks for reply.
scsi is hardly tested, so this isn't surprising.
well, but if it's still the most safe way to go, then there's not much other
choices, is it?
If it crashes, it isn't the safest way.
AMD or Intel?
On 09/02/2009 06:38 PM, or...@il.ibm.com wrote:
The following patches implement nested VMX support. The patches enable a guest
to use the VMX APIs in order to run its own nested guest (i.e., enable running
other hypervisors which use VMX under KVM). The current patches support running
Linux
On 02.09.2009, at 07:40, Benjamin Herrenschmidt wrote:
On Fri, 2009-07-24 at 18:30 +0200, Alexander Graf wrote:
KVM for PowerPC only supports embedded cores at the moment.
While it makes sense to virtualize on small machines, it's even
more fun
to do so on big boxes. So I figured we need
Well, we need mm_alloc to reserve a VA range. And since that happens
in module code, it needs to be exported.
Hrm... I've missed where you call it. Can't you directly allocate
a context ID instead of a whole mm ?
As far as switch_slb goes, I think with the new always switch the
full SLB
On 02.09.2009, at 08:23, Benjamin Herrenschmidt wrote:
Well, we need mm_alloc to reserve a VA range. And since that happens
in module code, it needs to be exported.
Hrm... I've missed where you call it. Can't you directly allocate
a context ID instead of a whole mm ?
It's in the vcpu init
On Wed, 2009-09-02 at 08:34 +0200, Alexander Graf wrote:
On 02.09.2009, at 08:23, Benjamin Herrenschmidt wrote:
Well, we need mm_alloc to reserve a VA range. And since that happens
in module code, it needs to be exported.
Hrm... I've missed where you call it. Can't you directly allocate
On Wed, 2009-09-02 at 15:40 +1000, Benjamin Herrenschmidt wrote:
On Fri, 2009-07-24 at 18:30 +0200, Alexander Graf wrote:
KVM for PowerPC only supports embedded cores at the moment.
While it makes sense to virtualize on small machines, it's even more fun
to do so on big boxes. So I
63 matches
Mail list logo