| Sehrawat Nipun | Using MSR Load/Store Areas
Hi Everyone,
I am trying to use the MSR load and store areas in vmx. I did following:
1. Setting vm_entry_msr_load_count:
vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, 1);
2. Setting up vm_entry_msr_load_addr :
static struct page *vmx_msr_load;
Then in vmx_init() :
vmx_msr_load = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
va=kmap(vmx_msr_load);
memset(va, MSR_P6_PERFCTR0, 32); /*Bits 31:0 contain msr's address*/
memset(va+32, 0x0, 32); /*Bits 63:32 are reserved and shall be all Zeros*/
mems...
| Jul 3, 5:42 pm 2009 |
| Andre Przywara | [RFC] allow multi-core guests: introduce cores= option to -cpu
Hi,
currently SMP guests happen to see <n> vCPUs as <n> different sockets.
Some guests (Windows comes to mind) have license restrictions and refuse
to run on multi-socket machines.
So lets introduce a "cores=" parameter to the -cpu option to let the user
specify the number of _cores_ the guest should see.
This patch has not been tested with all corner cases, so I just want to
hear your comments whether
a) we need such an option and
b) you like this particular approach.
Applying ...
| Jul 3, 10:41 am 2009 |
| Paul Brook |
| Jul 3, 11:46 am 2009 |
| Andre Przywara | Re: [Qemu-devel] [RFC] allow multi-core guests: introduce co...
Sound reasonable on the first glance, but would make it rather
complicated in real life. I suppose multi-core is far more interesting
to most of the people than multi-node, so I would opt for the easier:
-smp 2,cores=2 to specify a dual core guest.
Regards,
Andre.
--
Andre Przywara
AMD-Operating System Research Center (OSRC), Dresden, Germany
Tel: +49 351 488-3567-12
--
| Jul 3, 7:45 pm 2009 |
| Brian Jackson | Re: [RFC] allow multi-core guests: introduce cores= option t...
Personally, I'd like to see it as an extra arg to the -smp option. We've
seen too many people use -cpu incorrectly in #kvm, so we've gotten into
the habit of telling people not to touch that option unless they know
exactly what they are doing. Plus it seems odd to have to use -cpu foo
when you just want more cpus, not a specific cpu.
--
| Jul 3, 11:16 am 2009 |
| Andre Przywara | Re: [RFC] allow multi-core guests: introduce cores= option t...
Ok, I see your point. I simply used -cpu because of technical reasons
(the core topology is reflected in CPUID, which -cpu cares about). But
you are right, it does not belong here, -smp looks like a good candidate
(IMO better than -numa). Or we use an abstract "-topology" for this, but
this seems like overkill.
So what about: "-smp 4,cores=2,threads=2[,sockets=1]" to inject 4 vCPUs
in one package (automatically determined if omitted) with two cores and
two threads/core? All parameters except ...
| Jul 3, 6:52 pm 2009 |
| Samuel Thibault |
| Jul 3, 10:52 am 2009 |
| Andre Przywara | Re: [Qemu-devel] [RFC] allow multi-core guests: introduce co...
No, that meant: 16 vCPUs total with 8 cores per physical packages. I
don't have any notion for threads in the current code, although I agree
that this should be introduced.
AFAIK Windows XP Home is limited to one socket, where XP Pro can use up
to two sockets. Using -smp 4 on a XP Pro (64bit) caused an early BSOD
stating "MULTIPROCESSOR_CONFIGURATION_NOT_SUPPORTED" (because it saw
four sockets instead of four cores).
I think WindowsXP Home should even be worse (though I don't have an
Well, gen...
| Jul 3, 7:28 pm 2009 |
| Samuel Thibault |
| Jul 3, 7:53 pm 2009 |
| Nipun sehrawat | Using MSR load/store areas
Hi Everyone,
I am trying to use the MSR load and store areas in vmx. I did following:
1. Setting vm_entry_msr_load_count:
vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, 1);
2. Setting up vm_entry_msr_load_addr :
static struct page *vmx_msr_load;
Then in vmx_init() :
vmx_msr_load = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
va=kmap(vmx_msr_load);
memset(va, MSR_P6_PERFCTR0, 32); /*Bits 31:0 contain msr's address*/
memset(va+32, 0x0, 32); /*Bits 63:32 are reserved and shall be all Zeros*/
mem...
| Jul 3, 10:04 am 2009 |
| Andre Przywara | [PATCH] handle AMD microcode MSR
Windows 7 tries to update the CPU's microcode on some processors,
so we ignore the MSR write here. The patchlevel register is already handled
(returning 0), because the MSR number is the same as Intel's.
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
---
arch/x86/kvm/x86.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 60b2527..4f9789d 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -870,6 +870,7...
| Jul 3, 10:00 am 2009 |
| Andre Przywara | [PATCH] qemu-kvm: remove "KVM misreports CPUID" hack
This should be no longer necessary.
Effectively reverts 143eb2bd043e82bcf353cf82d33c127f06411d82.
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
---
kvm/libkvm/libkvm-x86.c | 9 ---------
qemu-kvm-x86.c | 9 ---------
2 files changed, 0 insertions(+), 18 deletions(-)
Hi Avi,
I am not sure what caused your problem back in May when you committed this,
but it seems to work without this now. I booted Win7RC 64bit and WinXP 64bit
and saw no problems.
Regards,
And...
| Jul 3, 8:23 am 2009 |
| Kevin Lapagna | passing a tape device through
Hi,
I'm (unsuccessfully) trying to pass-through a tape device, that is
connected to the host, to a kvm guest (Windows 2003). The tape device
is connected over SAS with a PCI card (aacraid).
First I tried to add the device with the following parameter
# ... -drive file=/dev/st0,if=scsi
This leads to errors on the console and a crashes the guest after a
while.
After this, I tried to pass-through the whole PCI card (with aacraid
disabled in the host system)
# ... -pcidevice host=0a:00...
| Jul 3, 6:45 am 2009 |
| Yolkfull Chow | [PATCH] Add a subtest pci_hotplug in kvm test
Signed-off-by: Yolkfull Chow <yzhou@redhat.com>
---
client/tests/kvm/kvm.py | 1 +
client/tests/kvm/kvm_tests.cfg.sample | 65 ++++++++++++++++++++++-
client/tests/kvm/kvm_tests.py | 94 +++++++++++++++++++++++++++++++++
client/tests/kvm/kvm_vm.py | 2 +
4 files changed, 161 insertions(+), 1 deletions(-)
diff --git a/client/tests/kvm/kvm.py b/client/tests/kvm/kvm.py
index b18b643..fc92e10 100644
--- a/client/tests/kvm/kvm.py
+++ b/client/tests/...
| Jul 3, 1:57 am 2009 |
| Yolkfull Chow |
| Jul 3, 2:00 am 2009 |
| Randy Smith | KVM-87 Compile warnings on 64-bit Jaunty
Greetings,
I'm seeing warnings when compiling KVM-87 on 64-bit jaunty. Make and
make install seem to complete and the module loads but networking fails
on VMs.
~$ uname -a
Linux tank 2.6.28-13-server #45-Ubuntu SMP Tue Jun 30 22:56:18 UTC 2009
x86_64 GNU/Linux
MAKE OUTPUT:
CC qemu-nbd.o
CC qemu-tool.o
CC tool-osdep.o
CC cutils.o
CC cache-utils.o
CC qemu-malloc.o
CC qemu-option.o
CC module.o
CC block/cow.o
block/cow.c: In function =E2=80=9...
| Jul 3, 12:38 am 2009 |
| Lucas Meneghel Rodrigues | Re: [Autotest] [PATCH] Add a client-side test qemu_iotests
Ok, I've been trough the test and took a look at the testsuite itself.
Both look good. The testsuite requires a very current qemu-img package,
due to the use of qemu-io (F11 for instance doesn't ship it). So if you
want to try this make sure you use a current version of the kvm
utilities.
+1
Lucas
--
| Jul 2, 8:23 pm 2009 |
| Lucas Meneghel Rodrigues | Re: [Autotest] [PATCH] Add a client-side test qemu_iotests
It just occurred to me some suggestions based on my last comment:
* Make sure we have the qemu, qemu-img and the qemu-io ELF files
available to execute the test suite, and if we doesn't, raise a
error.TestError() exception.
* Make it possible to specify an alternate prefix to look for those
ELFs, and make sure the test suite can find them (ie, appending the path
to os.environ['PATH']).
Yolkfull, please take a look at implementing those suggestions.
Thanks again for your work!
Lucas
--
...
| Jul 3, 1:03 am 2009 |
| Yolkfull Chow |
| Jul 3, 5:18 am 2009 |
| Lucas Meneghel Rodrigues | Re: [Autotest] [PATCH] Add a client-side test qemu_iotests
Sure it's not. My point was just to allow the user to provide an
additional path (say self.srcdir or something like that) that could be
added to the environmental variable PATH. This would be useful if we
build/install the ELFs under a non standard location on a previous test.
No need to walk through the entire filesystem to find the binaries.
--
| Jul 3, 8:37 am 2009 |
| Yolkfull Chow |
| Jul 3, 1:11 am 2009 |
| Yolkfull Chow |
| Jul 2, 10:01 pm 2009 |
| Yaniv Kaul |
| Jul 3, 2:45 am 2009 |
| sudhir kumar | Re: Weird Windows license issue
Hmm... What key are you using. I did not get any issue at all. I am
using a volume key. I did multiple installs and never got such an
issue.
--
Sudhir Kumar
--
| Jul 3, 12:09 am 2009 |
| Charles Duffy | Re: Weird Windows license issue
Last time I had licensing trouble when installing Windows under KVM
(IIRC this was Vista, not long after release), I called up Microsoft,
swore up and down three times that I was using a virtualization product,
and they gave me a code to punch in that resolved the issue.
--
| Jul 2, 10:14 pm 2009 |
| Lars Ellenberg | Re: [DRBD-user] kvm, drbd, elevator, rotational - quite an i...
I'm usually near-realtime letting relevant non-subscriber posts through,
did that improve?
we are about to release 8.3.2 now,
Interessting findings.
What IO scheduler is used on the HOST?
did you try to leave the guests in their default setting,
(rotational=0, elevator=noop)
not much.
disk IO on Primary is usually submitted in the context of the
submitter (vm subsystem, filesystem or the process itself)
whereas on Secondary, IO is naturally submitted just by the
that should not happe...
| Jul 3, 4:45 am 2009 |
| Javier Guerra | Re: [DRBD-user] kvm, drbd, elevator, rotational - quite an i...
just like with KVM itself, using several worker threads against a single IO device makes performance heavily dependent on a sensible elevator algorithm. ideally, there should be only one worker thread for each thread/process originating the initial write. unfortunately DRBD, being a block/level protocol, might have a hard time unraveling which writes belong to which process. maybe just merging adjacent (in block address space, not in time) write operations would keep most of the relationships.
--
...
| Jul 3, 9:06 am 2009 |
| Lars Ellenberg | Re: [DRBD-user] kvm, drbd, elevator, rotational - quite an i...
again, on the DRBD secondary, _all_ IO is naturally submitted
by _one_ thread (the receiver thread), appart from just a few
"internally" generated writes, which have to happen from a second
thread, the (one) drbd worker thread.
no, we do not do thread pooling (yet).
the elevator of the lower level block device (in this case,
the kvm virtual block device, or the host real block device)
is responsible for merging requests adjacent in block address space.
you have to choose a sensible elevator th...
| Jul 3, 10:00 am 2009 |
| Javier Guerra |
| Jul 3, 6:59 pm 2009 |
| G | Re: KVM crashes when using certain USB device
Hello again,
I've continued my attempts to get the HASP dongle working, but with no success:
Downloaded kvm-72.tar.gz through kvm-87.tar.gz to find out when the
problem first appear, as kvm-72 is working. Unfortunately, kvm-72
through kvm-82 fails to compile on my Debian system with kernel 2.6.30
and gcc 4.3, and kvm-83 crashes in the same way that kvm-87 crashes.
I have also tested two other USB devices: a C-Pen ( www.cpen.com) and a
GEM card reader, both successfully. So it seems it is the HAS...
| Jul 3, 5:18 am 2009 |
| Jim Paris |
| Jul 3, 12:18 pm 2009 |
| Liu Yu-B13201 |
| Jul 2, 9:35 pm 2009 |
| Ingo Molnar |
| Jul 3, 4:29 am 2009 |
| Matty |
| Jul 3, 6:41 am 2009 |
| Mark McLoughlin |
| Jul 3, 11:43 am 2009 |
| Erik Jacobson | Re: slow guest performance with build load, looking for ideas
Thanks Mark.
I have not tried this yet.
To be honest, I wasn't fully understanding some of Avi's last comments and
was waiting for one of my co-workers to be available to help me parse them.
I plan to perform Avi's suggestions, plus this rotational idea, early next
week. I'll update the BZs too.
By the way, a request to work-order a system for a few weeks to play with this
stuff was approved. This means I'll have easier access to a nicely configured
multi-socket Nehalem system for a whi...
| Jul 3, 12:28 pm 2009 |
