Re: qcow2 slowdown in kvm-78,79

ons 2008-11-19 klockan 12:26 -0600 skrev Anthony Liguori: > Ryan Harper wrote: > > * Henrik Holst <[EMAIL PROTECTED]> [2008-11-18 08:42]: > > > >> The bundled qemu in kvm-78 and kvm-79 slows down disk i/o with qcow2 > >> images by an order of 10. If one got 60MB/s before, one gets around > >> 6M

Re: [PATCH][RFC] Use writeback caching by default with qcow2

On Wed, 19 Nov 2008 14:24:53 -0600, Anthony Liguori wrote: > Right now, qcow2 isn't a > reliable format regardless of the type of cache your using because > metadata is not updated in the correct order. so you don't advise to use qcow2 as a VBD or what do you mean with "isn't reliable"? or co

[PATCH] KVM: IA64: Using CFLAGS_vcpu.o to compile vcpu.c

Hi, Avi Please apply the patch and queue it as 2.6.28 fix. Thanks! Xiantao >From d759572033e1bf0e4ca6b7d8f7be8e8387fb4bbb Mon Sep 17 00:00:00 2001 From: Xiantao Zhang <[EMAIL PROTECTED]> Date: Thu, 20 Nov 2008 10:35:35 +0800 Subject: [PATCH] KVM: IA64: Using CFLAGS_vcpu.o to compile vcpu.

Re: qcow2 slowdown in kvm-78,79

Would this affect 'qemu-img commit'? It took over an hour to commit a fresh install of Vista on a qcow2 virtual disk. (The base image expanded from 3GB to 11GB, which is a lot of committing, though.) Yes. Regards, Anthony Liguori -- To unsubscribe from this list: send the line "unsubscri

Re: qcow2 slowdown in kvm-78,79

Anthony Liguori wrote: Ryan Harper wrote: * Henrik Holst <[EMAIL PROTECTED]> [2008-11-18 08:42]: The bundled qemu in kvm-78 and kvm-79 slows down disk i/o with qcow2 images by an order of 10. If one got 60MB/s before, one gets around 6MB/s with 78 and 79 (measured with dd) dd command? what's

Re: Hangs

Thanks for the responses, I'm not sure if my problem is the same as Roland's, but it definitely sounds plausible. I had been running ntpdate in the host to synchronize time every hour (in a cron job), so it sounds as if we could be seeing the same issue. I'll spend some time trying to reproduce

Re: qcow2 slowdown in kvm-78,79

Thomas Mueller wrote: On Wed, 19 Nov 2008 12:26:41 -0600, Anthony Liguori wrote: This is probably the change to cache=writethrough. I bet if you set cache=writeback then you'll see this go away. this was it. cache=writeback and the throughput goes from 10mb/s to 80/mb/ s with if=scsi

Re: Hangs

Actually it just happenend again with the host running kvm-79. Host CPU is at 100% but I'm still able to login (it recovered from the first hang). But I'm not able to start e.g. top. Writing to disk works (e.g. dd /dev/zero to /tmp/test.file with 1MB, 1G already caused the instance to hang) In the

Re: [PATCH] KVM: VMX: Fix race between pending IRQ and NMI

Avi Kivity wrote: Jan Kiszka wrote: Jiajun kindly provided me a RHEL kernel and initrd (2.6.18-53-el5) which I ran for a while (or booted a few times) to trigger the hang. Basically you need high IRQ load (preferably via LAPIC, to exploit that un-acked IRQs will block low-prio IRQs as well) + hi

Re: [PATCH] always assign userspace_addr

Glauber Costa wrote: On Wed, Nov 19, 2008 at 12:51:00PM -0600, Anthony Liguori wrote: But the problem still exists even with this code. I checked. So if you have something working without modifying the kernel, can you post it? Regards, Anthony Liguori Ok, how do you feel about t

Re: [PATCH] always assign userspace_addr

On Wed, Nov 19, 2008 at 12:51:00PM -0600, Anthony Liguori wrote: > > But the problem still exists even with this code. I checked. > > So if you have something working without modifying the kernel, can you > post it? > > Regards, > > Anthony Liguori Ok, how do you feel about this one? My propos

[PATCH] e1000 VLAN offload emulation

We're currently ignoring the e1000 VLAN tagging, stripping and filtering features in the e1000 emulation. This patch adds backing for the relevant registers and provides a software implementation of the acceleration, such that a guest can make use of VLANs. This is mostly (only?) useful for a gue

[PATCH][RFC] Use writeback caching by default with qcow2

qcow2 writes a cluster reference count on every cluster update. This causes performance to crater when using anything but cache=writeback. This is most noticeable when using savevm. Right now, qcow2 isn't a reliable format regardless of the type of cache your using because metadata is not up

Re: qcow2 slowdown in kvm-78,79

On Wed, 19 Nov 2008 12:26:41 -0600, Anthony Liguori wrote: > Ryan Harper wrote: >> * Henrik Holst <[EMAIL PROTECTED]> [2008-11-18 08:42]: >> >>> The bundled qemu in kvm-78 and kvm-79 slows down disk i/o with qcow2 >>> images by an order of 10. If one got 60MB/s before, one gets around >>> 6MB/s

Re: virtio_net hang

On Wed, 2008-11-19 at 13:13 +, Mark McLoughlin wrote: > On Tue, 2008-11-18 at 19:37 +0100, Emmanuel Lacour wrote: > > On Fri, Nov 14, 2008 at 06:26:44PM +, Mark McLoughlin wrote: > > > > > > Right, the tap device tx queue is full because kvm-userspace isn't > > > reading packets from it. >

Re: [PATCH] always assign userspace_addr

Glauber Costa wrote: On Wed, Nov 19, 2008 at 09:55:10AM -0600, Anthony Liguori wrote: Glauber Costa wrote: Currently, kvm only sets new.userspace_addr in slots that were just allocated. This is not the intended behaviour, and actually breaks when we try to use the slots to implement ali

Re: [PATCH] always assign userspace_addr

On Wed, Nov 19, 2008 at 09:55:10AM -0600, Anthony Liguori wrote: > Glauber Costa wrote: >> Currently, kvm only sets new.userspace_addr in slots >> that were just allocated. This is not the intended behaviour, >> and actually breaks when we try to use the slots to implement >> aliases, for example.

Re: qcow2 slowdown in kvm-78,79

Ryan Harper wrote: * Henrik Holst <[EMAIL PROTECTED]> [2008-11-18 08:42]: The bundled qemu in kvm-78 and kvm-79 slows down disk i/o with qcow2 images by an order of 10. If one got 60MB/s before, one gets around 6MB/s with 78 and 79 (measured with dd) dd command? what's your -drive par

Re: qcow2 slowdown in kvm-78,79

* Henrik Holst <[EMAIL PROTECTED]> [2008-11-18 08:42]: > The bundled qemu in kvm-78 and kvm-79 slows down disk i/o with qcow2 > images by an order of 10. If one got 60MB/s before, one gets around > 6MB/s with 78 and 79 (measured with dd) dd command? what's your -drive parameters look like, specif

Re: qcow2 slowdown in kvm-78,79

On Tue, 18 Nov 2008 15:39:52 +0100, Henrik Holst wrote: > The bundled qemu in kvm-78 and kvm-79 slows down disk i/o with qcow2 > images by an order of 10. If one got 60MB/s before, one gets around > 6MB/s with 78 and 79 (measured with dd) so i'm not the only one facing this problem. - Thomas -

Re: [PATCH] KVM: VMX: Fix race between pending IRQ and NMI

Jan Kiszka wrote: Jiajun kindly provided me a RHEL kernel and initrd (2.6.18-53-el5) which I ran for a while (or booted a few times) to trigger the hang. Basically you need high IRQ load (preferably via LAPIC, to exploit that un-acked IRQs will block low-prio IRQs as well) + high NMI load (e.g. v

Re: [PATCH] always assign userspace_addr

Glauber Costa wrote: Currently, kvm only sets new.userspace_addr in slots that were just allocated. This is not the intended behaviour, and actually breaks when we try to use the slots to implement aliases, for example. Cirrus VGA aliases maps and address to a userspace address, and then keep ma

Re: can we hope a stable version in the near future?

Farkas Levente wrote: Dor Laor wrote: on the other hand the real question are you plan to somehow stabilize any of the following release in the near future? in the last 1.5 years we wait for this. or you currently not recommend and not plan to use kvm in production? it's also an option but wo

kvm unresponsive under io work

hi scenario: - host: 2.6.26, kvm-79, Intel Xeon E5420, Disk is a hardware RAID-10 (4 wd velociraptor's) - host: /tmp is lvm volume with XFS mounted with option nobarriers - host: kvm-img create -f qcow2 /tmp/10gb 10G - host: kvm -drive fiel=/tmp/10gb,if=scsi,boot=on \ -monitor sockt

Re: Unsupported delivery mode 7

Chris Webb wrote: > We're running kvm-78 in production on Linux 2.6.27 x86_64 on dual quad-core > Opteron 'Barcelona' machines. Our kvm modules are built from the kvm-78 > sources rather than the older version bundled with the kernel, and we're > using the NPT features of the processors. > > For t

Unsupported delivery mode 7

We're running kvm-78 in production on Linux 2.6.27 x86_64 on dual quad-core Opteron 'Barcelona' machines. Our kvm modules are built from the kvm-78 sources rather than the older version bundled with the kernel, and we're using the NPT features of the processors. For the most part, everything is pe

Re: virtio_net hang

On Tue, 2008-11-18 at 19:37 +0100, Emmanuel Lacour wrote: > On Fri, Nov 14, 2008 at 06:26:44PM +, Mark McLoughlin wrote: > > > > Right, the tap device tx queue is full because kvm-userspace isn't > > reading packets from it. > > > > This could be because kvm-userspace has just stopped noticin

[ kvm-Bugs-2217430 ] Host crash in kvm:unsync_walk_fn with kvm-78

Bugs item #2217430, was opened at 2008-11-02 19:01 Message generated for change (Comment added) made by johnrrousseau You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=893831&aid=2217430&group_id=180599 Please note that this message will contain a full copy of the com

[ kvm-Bugs-2314737 ] 100% host CPU usage with Win2k guest

Bugs item #2314737, was opened at 2008-11-19 13:53 Message generated for change (Comment added) made by krwi You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=893831&aid=2314737&group_id=180599 Please note that this message will contain a full copy of the comment thre

[ kvm-Bugs-2314737 ] 100% host CPU usage with Win2k guest

Bugs item #2314737, was opened at 2008-11-19 13:53 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=893831&aid=2314737&group_id=180599 Please note that this message will contain a full copy

[PATCH 11/11] KVM: MSI to INTx translate

Now we use MSI as default one, and translate MSI to INTx when guest need INTx rather than MSI. For legacy device, we provide support for non-sharing host IRQ. Provide a parameter msi2intx for this method. The value is true by default. Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- virt/kvm/kv

[PATCH 04/11] KVM: Clean up assigned_device_update_irq

Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- virt/kvm/kvm_main.c | 10 -- 1 files changed, 4 insertions(+), 6 deletions(-) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 638de47..2089f8b 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -180,11 +180,11 @@

[PATCH 09/11] KVM: Add assigned_device_msi_dispatch()

The function is used to dispatch MSI to lapic according to MSI message address and message data. Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- virt/kvm/kvm_main.c | 54 +++ 1 files changed, 54 insertions(+), 0 deletions(-) diff --git a/virt/k

[PATCH 10/11] KVM: Enable MSI for device assignment

We enable guest MSI and host MSI support in this patch. The userspace want to enable MSI should set KVM_DEV_IRQ_ASSIGN_ENABLE_MSI in the assigned_irq's flag. Function would return -ENOTTY if can't enable MSI, userspace shouldn't set MSI Enable bit when KVM_ASSIGN_IRQ return -ENOTTY with KVM_DEV_IRQ

[PATCH 06/11] KVM: Export ioapic_get_delivery_bitmask

It would be used for MSI in device assignment, for MSI dispatch. Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- virt/kvm/ioapic.c |7 --- virt/kvm/ioapic.h |2 ++ 2 files changed, 6 insertions(+), 3 deletions(-) diff --git a/virt/kvm/ioapic.c b/virt/kvm/ioapic.c index c8f939c..23b

[PATCH 05/11] KVM: Add fields for MSI device assignment

Prepared for kvm_arch_assigned_device_msi_dispatch(). Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- include/linux/kvm.h |7 +++ include/linux/kvm_host.h |4 2 files changed, 11 insertions(+), 0 deletions(-) diff --git a/include/linux/kvm.h b/include/linux/kvm.h index 44

[PATCH 08/11] x86: Add MSI delivery mode value

Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- arch/x86/include/asm/msidef.h |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/arch/x86/include/asm/msidef.h b/arch/x86/include/asm/msidef.h index 988cb27..412f94e 100644 --- a/arch/x86/include/asm/msidef.h +++ b/arch/x86/

[PATCH 02/11] KVM: Separate update irq to a single function

Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- virt/kvm/kvm_main.c | 68 -- 1 files changed, 38 insertions(+), 30 deletions(-) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 8966fd1..ef2f03c 100644 --- a/virt/kvm/kvm_main.c +++ b

[PATCH 01/11] KVM: Move ack notifier register and IRQ sourcd ID request

Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- virt/kvm/kvm_main.c | 30 +++--- 1 files changed, 19 insertions(+), 11 deletions(-) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 4727c08..8966fd1 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c

[PATCH 0/11][v5] Enable MSI for KVM

Hi Here is v5 for MSI on assigned devices, sorry for the delay. Due to the bad weather, I got a badly cold recently, indeed feeling bad... :( v4->v5 Addressed all the comments from Avi. I would work on the generic MSI solution later as well. The most important one is a new modules paramter "msi2

[PATCH 03/11] KVM: Replace irq_requested with more generic irq_requested_type

Separate guest irq type and host irq type, for we can support guest using INTx with host using MSI (but not opposite combination). Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- include/linux/kvm_host.h |4 +++- virt/kvm/kvm_main.c |9 + 2 files changed, 8 insertions(+), 5

[PATCH 07/11] x86: Rename MSI macro name

Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- arch/x86/include/asm/msidef.h |4 ++-- arch/x86/kernel/io_apic.c |4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/arch/x86/include/asm/msidef.h b/arch/x86/include/asm/msidef.h index 6706b30..988cb27 100644 --- a/

Re: Hangs

Sorry to repost under a different topic again, but it fits far better here. I saw similar issues when running from a debian lenny 2.6.26-1-amd64 64bit kvm host (which is kvm72 on currently) and the guests are debian lenny 2.6.26-1-486 32bit. So the setup is similar to the Ubuntu setup. I have co

Re: [PATCH 1/1] KVM: VMX: Fix base_ptes for EPT

On Wednesday 19 November 2008 18:25:32 Avi Kivity wrote: > Sheng Yang wrote: > > After support MTRR for EPT, memory type WB shouldn't take as default in > > base_ptes. (Something wrong during the merging...) > > Yes, my fault. > > But don't we still have a big hole here? The guest can set RAM to >

Re: [PATCH 1/1] KVM: VMX: Fix base_ptes for EPT

Sheng Yang wrote: After support MTRR for EPT, memory type WB shouldn't take as default in base_ptes. (Something wrong during the merging...) Yes, my fault. But don't we still have a big hole here? The guest can set RAM to various attributes, even without device assignment, and create in

Re: can we hope a stable version in the near future?

Dor Laor wrote: on the other hand the real question are you plan to somehow stabilize any of the following release in the near future? in the last 1.5 years we wait for this. or you currently not recommend and not plan to use kvm in production? it's also an option but would

Re: can we hope a stable version in the near future?

Farkas Levente wrote: Avi Kivity wrote: Farkas Levente wrote: There is the maint/ series on git.kernel.org. It doesn't have formal releases though. do you plan any formal release? and it'd be nice to see the relationship between the current devel tree and the stable tree

[PATCH 1/1] KVM: VMX: Fix base_ptes for EPT

After support MTRR for EPT, memory type WB shouldn't take as default in base_ptes. (Something wrong during the merging...) Signed-off-by: Sheng Yang <[EMAIL PROTECTED]> --- arch/x86/kvm/vmx.c |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kv