Re: Direct access to GPGPU do-able?

Carsten Aulbert wrote: Hi, I'll start with a one-off question here, so please cc me on the reply. We are running a largish cluster and are currently buying GPGPU systems (Tesla and soon Fermi based). We will have at least 2 possibly 4 of these cards per box and have the problem that some cod

Re: 2.6.31.5 x86-64 KVM: 'emulation failed (pagetable) rip 7fb49335da7b 66 0f 7f 07'

On 11/30/2009 11:40 PM, Nix wrote: So I started with -vga std. The emulation failed flood is unaffected (strangely it's not consistent: it happens about half the time). As soon as I issue a 'stop gdm', X exits and the emulation flood halts: but although there are other getties running I can't fi

Re: 2.6.31.5 x86-64 KVM: 'emulation failed (pagetable) rip 7fb49335da7b 66 0f 7f 07'

On 12/01/2009 11:59 AM, Avi Kivity wrote: So it looks like it is X related, and we'll have no choice other than to emulate the instruction. Annoying, since we don't emulate sse instructions yet. I reproduced it and confirmed - the instruction is accessing 0xa, the low memory vga wind

Re: [KVM-AUTOTEST PATCH] Improve kvm subtest AutoIt - add option to download script from remote server

On Tue, Dec 01, 2009 at 02:49:43AM -0500, Michael Goldish wrote: > > - "Chen Cao" wrote: > > > Add an option to let user download the .au3 script from a remote > > server, and keep the original method to prepare the script. > > > > This makes it possible to employ AutoIt tests that involve

Re: [Autotest] [KVM-AUTOTEST] KSM-overcommit test v.2 (python version)

Dne 29.11.2009 17:17, Dor Laor napsal(a): On 11/26/2009 12:11 PM, Lukáš Doktor wrote: Hello Dor, Thank you for your review. I have few questions about your comments: --- snip --- + stat += "Guests memsh = {" + for vm in lvms: + if vm.is_dead(): + logging.info("Trying to get in

[KVM-autotest][RFC] 32/64 bit guest system definition

Hello, In our test (KSM-overcommit) we need to know whether the guest system is 32, 32-PAE, or 64bit system. Nowadays we are using params['image_name'] string which ends with 32 or 64. Can we confine on this 'image_name' parameter string ending or do you think KSM-autotest should define this

[PATCH] KVM test: Add a subtest physical_resources_check

This test will test following parameters of a VM: 1) count of CPU, hard disks and NICs 2) memory size 3) model of hard disks and NICs 4) NICs' mac address 5) UUID and serial number (if defined the command in config file) Signed-off-by: Yolkfull Chow --- client/tests/kvm/kvm_tests.cfg.sample

Re: [Autotest] [PATCH 1/2] Adds a test to verify resources inside a VM

Hi Sudhir and Yolkfull: Thanks for your work on this test! Since Yolkfull's test matches Sudhir's test functionality and extends it, I will go with it. Some points: * A failure on checking a given resource shouldn't prevent us from testing other resources. Hence, instead of TestFail() exceptions

[PATCH] configure: Correct KVM options in help output

Signed-off-by: Pierre Riteau --- configure |8 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/configure b/configure index 376c458..85f7b5e 100755 --- a/configure +++ b/configure @@ -723,10 +723,10 @@ echo " --disable-bluez disable bluez stack connectivity"

[PATCH 0/12]: Fix kdump under KVM

Another version of the patch series to make kdump work inside KVM guests. The current problem with using kdump is that KVM only delivers PIT interrupts to the BSP. While this is technically allowed by the MPS spec, most motherboards deliver timer interrupts to *any* LAPIC in virtual wire mode.

[PATCH 01/12] Fix up some comments around the source tree.

Signed-off-by: Chris Lalancette --- arch/x86/kvm/svm.c |2 +- virt/kvm/ioapic.c |4 ++-- virt/kvm/kvm_main.c |2 +- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 3de0b37..68fe7f2 100644 --- a/arch/x86/kvm/svm.c +++ b/ar

[PATCH 02/12] Make kvm_pic_reset static.

Signed-off-by: Chris Lalancette --- arch/x86/kvm/i8259.c |2 +- arch/x86/kvm/irq.h |2 -- 2 files changed, 1 insertions(+), 3 deletions(-) diff --git a/arch/x86/kvm/i8259.c b/arch/x86/kvm/i8259.c index d057c0c..59a8a68 100644 --- a/arch/x86/kvm/i8259.c +++ b/arch/x86/kvm/i8259.c @@ -23

[PATCH 03/12] Remove KVM_REQ_PENDING_TIMER.

KVM_REQ_PENDING_TIMER is set and cleared in a couple of places, but it never seems to be actually checked. Remove it. Signed-off-by: Chris Lalancette --- arch/x86/kvm/timer.c |5 + arch/x86/kvm/x86.c |1 - include/linux/kvm_host.h |1 - 3 files changed, 1 insertions(+)

[PATCH 06/12] Make the PIC use interrupt safe locking.

Since we want to be able to call kvm_pic_set_irq() in interrupt context, convert the uses of spin_lock() to spin_lock_irqsave() as appropriate in i8259.c Signed-off-by: Chris Lalancette --- arch/x86/kvm/i8259.c | 34 ++ 1 files changed, 22 insertions(+), 12 dele

[PATCH 05/12] Make the IOAPIC lock a spinlock.

In order to be able to call kvm_set_irq from an interrupt context, the IOAPIC lock can't be a (possibly sleeping) mutex. Convert it to a spinlock. Signed-off-by: Chris Lalancette --- virt/kvm/ioapic.c | 40 +--- virt/kvm/ioapic.h |2 +- 2 files changed,

[PATCH 04/12] IOAPIC timer interrupt redirect.

Only redirect IRQ 0 (i.e. timer interrupt) to the BSP if the APIC destination is multiple vcpus. Signed-off-by: Chris Lalancette --- virt/kvm/ioapic.c | 15 +-- 1 files changed, 13 insertions(+), 2 deletions(-) diff --git a/virt/kvm/ioapic.c b/virt/kvm/ioapic.c index 53f3166..fd52

[PATCH 08/12] Remove timer.c

The code in timer.c isn't really similar enough between the i8254 and the lapic to share. Split these into separate functions, and remove timer.c Signed-off-by: Chris Lalancette --- arch/x86/kvm/Makefile|3 +- arch/x86/kvm/i8254.c | 27 - arch/x86/kvm/kvm_t

[PATCH 07/12] Rename kvm_apic_accept_pic_intr

Call it kvm_apic_in_virtual_wire_mode, which is more correct. Also change it to not only operate properly on the boot CPU, but on any CPU. Signed-off-by: Chris Lalancette --- arch/x86/kvm/i8259.c |2 +- arch/x86/kvm/irq.c |4 ++-- arch/x86/kvm/lapic.c | 17 - arch/x8

[PATCH 10/12] Inject i8254 interrupts directly from hrtimer callback.

Signed-off-by: Chris Lalancette --- arch/x86/kvm/i8254.c | 101 +++--- arch/x86/kvm/i8254.h |1 - arch/x86/kvm/irq.c | 10 + arch/x86/kvm/irq.h |2 - 4 files changed, 23 insertions(+), 91 deletions(-) diff --git a/arch/x86/kvm/i8254.

[PATCH 11/12] Allow the PIC to signal CPUs other than BSP.

While generally the PIT is used to signal only the BSP, it's doesn't actually have to do that architecturally. Allow the PIC to signal any VCPU that is in Virtual Wire mode. Signed-off-by: Chris Lalancette --- arch/x86/kvm/i8259.c | 19 ++- 1 files changed, 10 insertions(+), 9

[PATCH 12/12] Kick appropriate CPUs when signalling interrupts.

Make sure that we kick the appropriate vcpu when delivering an interrupt. This makes sure that we wake any idle cpus to cause a "vcpu_run" and an interrupt injection to occur. Signed-off-by: Chris Lalancette --- virt/kvm/irq_comm.c |5 - 1 files changed, 4 insertions(+), 1 deletions(-)

[PATCH 09/12] Fix missing spin_lock in PIT timer.

Changes to the irq_ack variable in the pit_state must be protected by the inject_lock spinlock; otherwise, we can erroneously inject a timer interrupt into a guest. Signed-off-by: Chris Lalancette --- arch/x86/kvm/i8254.c |3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a

[PATCH] If DEBUG_MEMREG is defined, build fails.

If DEBUG_MEMREG is defined, build fails as follows. CCx86_64-softmmu/kvm-all.o cc1: warnings being treated as errors In file included from kvm-all.c:1094: qemu-kvm.c: In function 'set_gsi': qemu-kvm.c:127: error: too few arguments for format qemu-kvm.c: In function 'clear_gsi': qemu-kvm.c:13

Re: [Autotest] [PATCH 2/3] Make a standalone client to be able to use global_config.ini

FYI, this broke server/autotest_unittest: show...@gluon:~/opensrc/autotest$ server/autotest_unittest.py ERROR:root:Could not install autotest from repos ..E == ERROR: test_run (__main__.TestBaseAutotest) --

Re: [Autotest] [PATCH 2/3] Make a standalone client to be able to use global_config.ini

Will look into this, Steve! On Tue, Dec 1, 2009 at 2:31 PM, Steve Howard wrote: > FYI, this broke server/autotest_unittest: > > show...@gluon:~/opensrc/autotest$ server/autotest_unittest.py > ERROR:root:Could not install autotest from repos > ..E >

Re: Direct access to GPGPU do-able?

I analyzed both XEN patches and determined where this patch must be done in KVM. It's just a matter of writing the code. Currently I'm studying for some final exams, so I will take some time off before I continue with this port. And because of hardware related problems, if it's not done in two or

Re: [PATCH 03/12] Remove KVM_REQ_PENDING_TIMER.

Chris Lalancette wrote: > KVM_REQ_PENDING_TIMER is set and cleared in a couple of places, > but it never seems to be actually checked. Remove it. > I would suggest to study the introducing commit 06e05645661211b9eaadaf6344c335d2e80f0ba2. My strong feeling is that this removal is wrong. Jan --

[PATCH] Fix autotest client when checking only client from svn

When the client was made configurable through global_config.ini, the scenario "developer checking out client directory only" wasn't considered, and an exception will be thrown due to the lack of a global_config.ini present. In order to fix this, instead of throwing an exception, just print a warni

Re: [PATCH] Fix autotest client when checking only client from svn

This looks good to me. -- John On Tue, Dec 1, 2009 at 2:37 PM, Lucas Meneghel Rodrigues wrote: > When the client was made configurable through > global_config.ini, the scenario "developer > checking out client directory only" wasn't > considered, and an exception will be thrown > due to the lac

Re: [Autotest] [PATCH] Fix autotest client when checking only client from svn

yup, seems important - lmr, do you want to go ahead and apply this? I'm stuck in a meeting for a while On Tue, Dec 1, 2009 at 2:39 PM, John Admanski wrote: > This looks good to me. > > -- John > > On Tue, Dec 1, 2009 at 2:37 PM, Lucas Meneghel Rodrigues > wrote: >> When the client was made conf

[PATCH] KVM test: Not execute build test by default

Instead of trying to build KVM using one of the methods defined on the control file, use "noinstall" as the default mode, and make "noinstall" entirely skip the build test, as having "noinstall" as one of the options for the build test was making the test to run, calling the kvm preprocessor and ki

Re: [Autotest] [PATCH] Fix autotest client when checking only client from svn

On Tue, Dec 1, 2009 at 8:42 PM, Martin Bligh wrote: > yup, seems important - lmr, do you want to go ahead and apply this? I'm stuck > in a meeting for a while Alright, already applied: http://autotest.kernel.org/changeset/3981 > On Tue, Dec 1, 2009 at 2:39 PM, John Admanski wrote: >> This look

Re: [Autotest] [PATCH 1/2] Adds a test to verify resources inside a VM

On Tue, Dec 01, 2009 at 11:56:43AM -0200, Lucas Meneghel Rodrigues wrote: > Hi Sudhir and Yolkfull: > > Thanks for your work on this test! Since Yolkfull's test matches > Sudhir's test functionality and extends it, I will go with it. Some > points: > > * A failure on checking a given resource sh

[PATCH] KVM test: Fix subtest imports

The KVM test was breaking when trying to import the subtest 'autotest', as a naming clash was happening (client/bin/autotest was being imported instead of the autotest subtest), so to make naming clashes less likely, import test.[type] instead of just [type]. Also, as the kvm test was already inser

Re: [Autotest] [PATCH 1/2] Adds a test to verify resources inside a VM

On Wed, Dec 2, 2009 at 7:51 AM, Yolkfull Chow wrote: > On Tue, Dec 01, 2009 at 11:56:43AM -0200, Lucas Meneghel Rodrigues wrote: >> Hi Sudhir and Yolkfull: >> >> Thanks for your work on this test! Since Yolkfull's test matches >> Sudhir's test functionality and extends it, I will go with it. Some

Re: [Autotest] [PATCH 1/2] Adds a test to verify resources inside a VM

On Wed, 2009-12-02 at 08:59 +0530, sudhir kumar wrote: > On Wed, Dec 2, 2009 at 7:51 AM, Yolkfull Chow wrote: > > > > Looks good for me. Thanks Lucas for improving this test. > > > > Sudhir, what do you think about this? :) > Needs couple of hours before I go through the patch. I will post my > co

Re: [PATCH] KVM test: Not execute build test by default

* Lucas Meneghel Rodrigues [2009-12-01 17:12]: > Instead of trying to build KVM using one of the methods > defined on the control file, use "noinstall" as the > default mode, and make "noinstall" entirely skip the > build test, as having "noinstall" as one of the options > for the build test was m

Re: [Autotest] [PATCH] KVM test: Not execute build test by default

On Wed, Dec 2, 2009 at 9:39 AM, Ryan Harper wrote: > * Lucas Meneghel Rodrigues [2009-12-01 17:12]: >> Instead of trying to build KVM using one of the methods >> defined on the control file, use "noinstall" as the >> default mode, and make "noinstall" entirely skip the >> build test, as having "n

Re: [PATCH 04/12] IOAPIC timer interrupt redirect.

On Tue, Dec 01, 2009 at 03:36:33PM +0100, Chris Lalancette wrote: > Only redirect IRQ 0 (i.e. timer interrupt) to the BSP if > the APIC destination is multiple vcpus. > > Signed-off-by: Chris Lalancette > --- > virt/kvm/ioapic.c | 15 +-- > 1 files changed, 13 insertions(+), 2 dele

Re: [PATCH 08/12] Remove timer.c

On Tue, Dec 01, 2009 at 03:36:37PM +0100, Chris Lalancette wrote: > The code in timer.c isn't really similar enough between > the i8254 and the lapic to share. Split these into > separate functions, and remove timer.c > > Signed-off-by: Chris Lalancette There should be one function. Abstract th

Re: [PATCH 03/12] Remove KVM_REQ_PENDING_TIMER.

On Tue, Dec 01, 2009 at 08:04:15PM +0100, Jan Kiszka wrote: > Chris Lalancette wrote: > > KVM_REQ_PENDING_TIMER is set and cleared in a couple of places, > > but it never seems to be actually checked. Remove it. > > > > I would suggest to study the introducing commit > 06e05645661211b9eaadaf6344

[PATCH] Add a server-side test: kvm_migration

This patch will add a server-side test namely kvm_migration. Currently, it will use existing KVM client test framework and add a new file kvm_migration.py to help judge executing routine: source machine or dest machine. * Things need to be improved: 1) a method/mechanism to let source machine that