Anthony Liguori wrote:
On 06/23/2010 04:09 PM, Andre Przywara wrote:
Hi,
these three patches add basic NUMA pinning to KVM. According to a user
provided assignment parts of the guest's memory will be bound to different
host nodes. This should increase performance in large virtual machines
and o
Bugs item #2014059, was opened at 2008-07-09 10:30
Message generated for change (Comment added) made by jessorensen
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=893831&aid=2014059&group_id=180599
Please note that this message will contain a full copy of the comme
Note to qemu-devel: this issue is qemu-kvm only.
"Hao, Xudong" writes:
> When assign one PCI device, qemu fail to parse the command line:
> qemu-system_x86 -smp 2 -m 1024 -hda /path/to/img -pcidevice host=00:19.0
> Error:
> qemu-system-x86_64: Parameter 'id' expects an identifier
> Identifiers c
On Wed, Jun 23, 2010 at 11:12 PM, Anthony Liguori wrote:
> Shouldn't it be possible to just drop the lock before invoking
> virtqueue_kick() and reacquire it afterwards? There's nothing in that
> virtqueue_kick() path that the lock is protecting AFAICT.
No, that would lead to a race condition be
On 06/23/2010 11:56 PM, Chris Lalancette wrote:
Did you forget to push out a tree with this patch applied, or are you doing
some testing before doing so? I don't see this patch in the current kvm
tree (git head is a63e16c655f9e68d49d6fae4275ffda16b1888b2).
No and yes. Under testing patch
On Wed, 2010-06-23 at 08:51 +0300, Avi Kivity wrote:
> On 06/23/2010 06:12 AM, Zhang, Yanmin wrote:
> >>>
> >>> This design is to deal with a task context perf collection in guest os.
> >>> Scenario 1:
> >>> 1) guest os starts to collect statistics of process A on vcpu 0;
> >>> 2) process A is sche
With the availablility of a sysfs device attribute for examining disk serial
numbers the ioctl is no longer needed. The user-space changes for this aren't
upstream yet so we don't have any users to worry about.
Signed-off-by: Ryan Harper
---
drivers/block/virtio_blk.c | 10 --
1 files
Create a new attribute for virtio-blk devices that will fetch the serial number
of the block device. This attribute can be used by udev to create disk/by-id
symlinks for devices that don't have a UUID (filesystem) associated with them.
ATA_IDENTIFY strings are special in that they can be up to 20
Using Rusty's suggestion I've respun the patch removing the special copy
function. I've tested this patch in a guest kernel with and without qemu
supplying serial numbers for the block devices and it's working as expected.
When qemu supplies serial numbers, the correct value is supplied to
/sys/bl
When assign one PCI device, qemu fail to parse the command line:
qemu-system_x86 -smp 2 -m 1024 -hda /path/to/img -pcidevice host=00:19.0
Error:
qemu-system-x86_64: Parameter 'id' expects an identifier
Identifiers consist of letters, digits, '-', '.', '_', starting with a letter.
pcidevice argument
On Wed, Jun 23, 2010 at 12:00:47PM +0300, Avi Kivity wrote:
> On 06/23/2010 12:01 PM, Takuya Yoshikawa wrote:
> >(2010/06/23 17:48), Avi Kivity wrote:
> >
> >>>diff --git a/arch/powerpc/kvm/book3s.c b/arch/powerpc/kvm/book3s.c
> >>>index 801d9f3..bea6f7c 100644
> >>>--- a/arch/powerpc/kvm/book3s.c
On 24.06.2010, at 00:21, Anthony Liguori wrote:
> On 06/23/2010 04:09 PM, Andre Przywara wrote:
>> Hi,
>>
>> these three patches add basic NUMA pinning to KVM. According to a user
>> provided assignment parts of the guest's memory will be bound to different
>> host nodes. This should increase pe
On 06/23/2010 04:09 PM, Andre Przywara wrote:
Hi,
these three patches add basic NUMA pinning to KVM. According to a user
provided assignment parts of the guest's memory will be bound to different
host nodes. This should increase performance in large virtual machines
and on loaded hosts.
These pa
On 06/23/2010 04:24 PM, Stefan Hajnoczi wrote:
The virtio block device holds a lock during I/O request processing.
Kicking the virtqueue while the lock is held results in long lock hold
times and increases contention for the lock.
This patch modifies virtqueue_kick() to optionally release a lock
On 06/23/2010 08:12 AM, Avi Kivity wrote:
On 06/14/2010 06:53 PM, Anthony Liguori wrote:
index 000..e0a7b98
--- /dev/null
+++ b/contrib/ivshmem-server/ivshmem_server.c
There's no licensing here. I don't think this belongs in the qemu
tree either to be honest.
I asked for this, to sim
According to the user-provided assignment bind the respective part
of the guest's memory to the given host node. This uses Linux'
libnuma interface to realize the pinning right after the allocation.
Failures are not fatal, but produce a warning.
Signed-off-by: Andre Przywara
---
hw/pc.c | 51 +
Introduce another variant of QEMU's -numa option to allow host node
pinning. This was separated from the guest relevant configuration
to make it cleaner to use, especially for management applications.
The syntax is -numa pin,nodeid=n,host=m to assign the guest node n
to host node m.
Signed-off-by:
Hi,
these three patches add basic NUMA pinning to KVM. According to a user
provided assignment parts of the guest's memory will be bound to different
host nodes. This should increase performance in large virtual machines
and on loaded hosts.
These patches are quite basic (but work) and I send them
The virtio block device holds a lock during I/O request processing.
Kicking the virtqueue while the lock is held results in long lock hold
times and increases contention for the lock.
This patch modifies virtqueue_kick() to optionally release a lock while
notifying the host. Virtio block is modif
Add detection of libnuma (mostly contained in the numactl package)
to the configure script. Currently this is Linux only, but can be
extended later should the need for other interfaces come up.
Can be enabled or disabled on the command line, default is use if
available.
Signed-off-by: Andre Przywa
Hi All,
In the previous mail I did not mention about my machine setup hence I
am rewriting it. I am using Ubuntu Karmic 64 bit on Intel Core2 Duo
with Intel-VT extensions. Actually KVM was working fine last week but
I see only a blank screen now (either directly or using vnc). If I
replace KVM wit
Hi guys
In case you have missed, ConVirt Enterprise Beta is now available. Check out
details here.
http://www.convirture.com/blog/2010/general/convirt-2-0-enterprise-beta-is-underway/
If you are interested kindly use the link given in the post to register.
Thanks
/Jd
--
To unsub
The code to remove the migration files was one indentation
block deeper than it should - my mistake. Now, we ensure
the files are being cleaned up at the end of the test.
Signed-off-by: Lucas Meneghel Rodrigues
---
client/tests/kvm/kvm_test_utils.py |6 --
1 files changed, 4 insertions(+
resent for some printf macros and one addition
Support an inter-vm shared memory device that maps a shared-memory object as a
PCI device in the guest. This patch also supports interrupts between guest by
communicating over a unix domain socket. This patch applies to the qemu-kvm
repository.
Changing subject,
Hi,
I am not sure if this is the right forum for my questions -
I am using Qemu+KVM on centOS to run my freebsd based virtual OS. The Qemu
emulates one of my nics - yukon -II 88E8022 chip. Now, my yukon driver
exists in kernel as well as user space(and I will use the one which
Hi,
I am not sure if this is the right forum for my questions -
I am using Qemu+KVM on centOS to run my freebsd based virtual OS. The Qemu
emulates one of my nics - yukon -II 88E8022 chip. Now, my yukon driver
exists in kernel as well as user space(and I will use the one which I want
based on som
[ please use reply-all when posting to mailing lists ]
Tom Shoes wrote:
> On Wed, Jun 23, 2010 at 3:16 AM, Jan Kiszka wrote:
>
> Thanks Jan for the response.
>
>> Why 2.6.26, specifically if you seem to have no customized bits?
>> Starting with a pre-historic kernel for a new project is hardly
On 06/23/2010 05:29 PM, Tom Shoes wrote:
On Wed, Jun 23, 2010 at 2:36 AM, Avi Kivity wrote:
Thanks Avi for your response.
Re-adding mailing list to cc list.
Request for more clarification on the following:
Is having no VGA adapter and using serial port as console an issue?
When
Tom Shoes wrote:
> Hi there,
>
> I am looking at using KVM for an embedded product. I am also new
> to Virtualization so pardon if
> I ask dumb questions. This is my first time posting to this forum.
>
>The embedded product that need to run KVM has:
>
> a. Intel processor with V
On 06/23/2010 04:29 PM, Jan Kiszka wrote:
No, and this is not surprising - looks like reset is completely bypassed.
You mean there is no system reset during init in kvm mode? Well, then...
Yes. The cpu executes until it hits 0xa, then stops.
--
error compiling committee.c: to
Avi Kivity wrote:
> On 06/23/2010 04:19 PM, Jan Kiszka wrote:
I'm testing uq/master, and I see this breaks it. qemu freezes
immediately on startup.
(might be due to a rebase?)
>>> The symptoms, btw, are a vcpu started from 0:0 instead of 0xf000:0xfff0.
>>>
>>>
On 06/23/2010 04:19 PM, Jan Kiszka wrote:
I'm testing uq/master, and I see this breaks it. qemu freezes
immediately on startup.
(might be due to a rebase?)
The symptoms, btw, are a vcpu started from 0:0 instead of 0xf000:0xfff0.
Might be unrelated, still: Does [1] make any d
Avi Kivity wrote:
> On 06/23/2010 03:29 PM, Avi Kivity wrote:
>> On 03/23/2010 06:37 PM, Marcelo Tosatti wrote:
>>> This is now done via the initialization's qemu_system_reset call.
>>>
>>>
>>> Index: qemu-uq/kvm-all.c
>>> ===
>>> ---
On 06/14/2010 06:53 PM, Anthony Liguori wrote:
index 000..e0a7b98
--- /dev/null
+++ b/contrib/ivshmem-server/ivshmem_server.c
There's no licensing here. I don't think this belongs in the qemu
tree either to be honest.
I asked for this, to simplify life for people trying this out.
If
On 06/17/2010 08:51 PM, Alex Williamson wrote:
An 82576 physical function assigned to a Windows 7 guest currently
doesn't work because the driver seems to gratuitiously disable
MSI and MSIX interrupts. When it does this, we blindly deassign
the current interrupt setup, leaving the device with no
On 06/23/2010 03:29 PM, Avi Kivity wrote:
On 03/23/2010 06:37 PM, Marcelo Tosatti wrote:
This is now done via the initialization's qemu_system_reset call.
Index: qemu-uq/kvm-all.c
===
--- qemu-uq.orig/kvm-all.c
+++ qemu-uq/kvm-all
On 03/23/2010 06:37 PM, Marcelo Tosatti wrote:
This is now done via the initialization's qemu_system_reset call.
Index: qemu-uq/kvm-all.c
===
--- qemu-uq.orig/kvm-all.c
+++ qemu-uq/kvm-all.c
@@ -209,7 +209,6 @@ int kvm_init_vcpu(CP
On 06/23/2010 03:07 PM, Nadav Har'El wrote:
On Wed, Jun 23, 2010, Avi Kivity wrote about "Re: [PATCH 5/24] Introduce vmcs12: a
VMCS structure for L1":
kmap() really should be avoided when possible. It is for when we don't
have a pte pointing to the page (for example, accessing a user page
On Wed, Jun 23, 2010, Avi Kivity wrote about "Re: [PATCH 5/24] Introduce
vmcs12: a VMCS structure for L1":
> kmap() really should be avoided when possible. It is for when we don't
> have a pte pointing to the page (for example, accessing a user page from
> outside its process context).
I'm afr
On 06/21/2010 06:29 PM, Chris Lalancette wrote:
Older versions of 32-bit linux have a "Checking 'hlt' instruction"
test where they repeatedly call the 'hlt' instruction, and then
expect a timer interrupt to kick the CPU out of halt. This happens
before any LAPIC or IOAPIC setup happens, which me
No need to reload the mmu in between two different vcpu->requests checks.
kvm_mmu_reload() may trigger KVM_REQ_TRIPLE_FAULT, but that will be caught
during atomic guest entry later.
Signed-off-by: Avi Kivity
---
arch/x86/kvm/x86.c | 12 +---
1 files changed, 5 insertions(+), 7 deletio
On Wed, Jun 23, 2010 at 04:09:40PM +0800, Dong, Eddie wrote:
>
> Xiaohui & Herbert:
> Mixing copy of head & 0-copy of bulk data imposes additional challange
> to find the guest buffer. The backend driver may be unable to find a spare
> guest buffer from virtqueue at that time which may bloc
On Wed, Jun 23, 2010 at 12:16:46PM +0200, Jan Kiszka wrote:
> Gleb Natapov wrote:
> > To prevent reentering vcpu after IO completion it is not enough
> > to set env->stopped since it is checked only in main loop but control
> > will not get there until next non-IO exit since kvm_run() will reenter
Gleb Natapov wrote:
> To prevent reentering vcpu after IO completion it is not enough
> to set env->stopped since it is checked only in main loop but control
> will not get there until next non-IO exit since kvm_run() will reenter
> vcpu to complete IO instruction. Solve this by sending self-signal
Tom Shoes wrote:
> Hi there,
>
> I am looking at using KVM for an embedded product. I am also new
> to Virtualization so pardon if
> I ask dumb questions. This is my first time posting to this forum.
>
>The embedded product that need to run KVM has:
>
> a. Intel processor with V
Herbert Xu wrote:
> On Wed, Jun 23, 2010 at 04:09:40PM +0800, Dong, Eddie wrote:
>>
>> Xiaohui & Herbert:
>> Mixing copy of head & 0-copy of bulk data imposes additional
>> challange to find the guest buffer. The backend driver may be
>> unable to find a spare guest buffer from virtqueue
On 06/23/2010 04:22 AM, Tom Shoes wrote:
Need help to understand if KVM can be installed, configured and run
with c,d,e and f configuration
mentioned above?
Also can I know
a. Is KVM code in Linux kernel 2.6.26 reliable?
It may be reliable or not, but it's certainly not su
On 06/23/2010 12:15 PM, Alexander Graf wrote:
Alex, Joerg, well gvmcb_{read,write}{32,64}() work for nsvm? All that kmapping
is incredibly annoying.
I'm sceptical that we can actually get that to be as fast as a direct kmap, but
apart from that I don't see an obvious reason why it wou
On 23.06.2010, at 09:57, Avi Kivity wrote:
> On 06/22/2010 05:54 PM, Nadav Har'El wrote:
>>
>>> I'm not sure what we need to do with vmcs that is not in RAM. It may
>>> simplify things to return the error_page to the caller and set
>>> KVM_REQ_TRIPLE_FAULT, so we don't have to deal with error h
On 06/23/2010 12:01 PM, Takuya Yoshikawa wrote:
(2010/06/23 17:48), Avi Kivity wrote:
diff --git a/arch/powerpc/kvm/book3s.c b/arch/powerpc/kvm/book3s.c
index 801d9f3..bea6f7c 100644
--- a/arch/powerpc/kvm/book3s.c
+++ b/arch/powerpc/kvm/book3s.c
@@ -1185,28 +1185,43 @@ int kvm_vm_ioctl_get_dir
(2010/06/23 17:48), Avi Kivity wrote:
diff --git a/arch/powerpc/kvm/book3s.c b/arch/powerpc/kvm/book3s.c
index 801d9f3..bea6f7c 100644
--- a/arch/powerpc/kvm/book3s.c
+++ b/arch/powerpc/kvm/book3s.c
@@ -1185,28 +1185,43 @@ int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
struct kvm_memory_slot *m
On 06/23/2010 08:58 AM, Takuya Yoshikawa wrote:
This patch series is for making dirty logging development, and of course
maintenance, easier. Please see individual patches for details.
Applied, thanks.
--
error compiling committee.c: too many arguments to function
--
To unsubscribe from t
On 06/23/2010 09:01 AM, Takuya Yoshikawa wrote:
kvm_get_dirty_log() is a helper function for kvm_vm_ioctl_get_dirty_log() which
is currently used by ia64 and ppc and the following is what it is doing:
- sanity checks
- bitmap scan to check if the slot is dirty
- copy_to_user()
Consider
Signed-off-by: Avi Kivity
---
Documentation/kvm/review-checklist.txt | 25 +
1 files changed, 25 insertions(+), 0 deletions(-)
create mode 100644 Documentation/kvm/review-checklist.txt
diff --git a/Documentation/kvm/review-checklist.txt
b/Documentation/kvm/review-chec
On 06/23/2010 11:13 AM, Nadav Har'El wrote:
On Mon, Jun 14, 2010, Avi Kivity wrote about "Re: [PATCH 6/24] Implement reading and
writing of VMX MSRs":
On 06/13/2010 03:25 PM, Nadav Har'El wrote:
When the guest can use VMX instructions (when the "nested" module option is
on), it shoul
On Wed, Jun 23, 2010 at 11:22:15AM +0300, Avi Kivity wrote:
> On 06/21/2010 12:01 PM, Gleb Natapov wrote:
> >To prevent reentering vcpu after IO completion it is not enough
> >to set env->stopped since it is checked only in main loop but control
> >will not get there until next non-IO exit since kv
On 06/21/2010 12:01 PM, Gleb Natapov wrote:
To prevent reentering vcpu after IO completion it is not enough
to set env->stopped since it is checked only in main loop but control
will not get there until next non-IO exit since kvm_run() will reenter
vcpu to complete IO instruction. Solve this by s
On 06/23/10 03:13, Zhang, Yanmin wrote:
> On Tue, 2010-06-22 at 09:58 +0200, Jes Sorensen wrote:
>> Exposing the counters read-only would save a lot of overhead for sure.
Question is if it is safe to drop overflow support?
>>> Not safe. One of PMU hardware design objectives is to use interrupt
On Mon, Jun 14, 2010, Avi Kivity wrote about "Re: [PATCH 6/24] Implement
reading and writing of VMX MSRs":
> On 06/13/2010 03:25 PM, Nadav Har'El wrote:
> >When the guest can use VMX instructions (when the "nested" module option is
> >on), it should also be able to read and write VMX MSRs, e.g., t
> 3) As I have mentioned above, with this idea, netdev_alloc_skb() will
> allocate
> as usual, the data pointed by skb->data will be copied into the first
> guest buffer.
> That means we should reserve sufficient room in guest buffer. For PS
> mode
> supported driver (for example ixgbe), the ro
On 06/23/2010 04:43 AM, Zhang, Yanmin wrote:
A reset does not involve destroying a vm; you have to clean up as part
of the rest process.
What does 'reset' here mean? Is it a reboot or halt? If it's a halt, it involves
destroying a vm. If a host user just kills the qemu process, is it a r
- Original Message -
> On Tuesday, June 22, 2010, 12:42:27, --[ UxBoD ]-- wrote:
>
> > I am attempting to get sound working on a W2K server and failing
> > horribly :( The sound device was picked up correctly as a ES1370 and
> > shows as being functional; yet when I rdesktop -r sound:remot
On 06/22/2010 07:53 PM, Nadav Har'El wrote:
On Tue, Jun 22, 2010, Nadav Har'El wrote about "Re: [PATCH 5/24] Introduce vmcs12: a
VMCS structure for L1":
Note that this structure becomes an ABI, it cannot change except in a
backward compatible way due to the need for live migration. So I'd
On 06/22/2010 05:54 PM, Nadav Har'El wrote:
I'm not sure what we need to do with vmcs that is not in RAM. It may
simplify things to return the error_page to the caller and set
KVM_REQ_TRIPLE_FAULT, so we don't have to deal with error handling later on.
This is a very good point. The app
64 matches
Mail list logo