Hi,
some time a go I stumbled across a regression in the KVM Module that has been
introduced somewhere
between 3.17 and 3.19.
I have a rather old openSUSE guest with an XFS filesystem which realiably
crashes after some live migrations.
I originally believed that the issue might be related to
On 05/08/15 11:53, Pavel Fedin wrote:
Before commit 662d9715840aef44dcb573b0f9fab9e8319c868a is was possible to
compile the kernel without vGIC and vTimer support. Commit message says
about possibility to detect vGIC support in runtine, but this has never
been implemented.
This patch
Hello!
This is completely Linux-specific, unfortunately.
Yes. But better than nothing.
And it relies on
userpace to expose a modified DT, so you need to be able to report back
to userspace that you can't deal with the virtual timer.
Easy. If KVM_CAP_IRQCHIP == 0, then we apparently
2015-08-13 16:53+0200, Paolo Bonzini:
On 13/08/2015 15:46, Radim Krčmář wrote:
1) IOAPIC inject a vector from i8254
2) guest reconfigures that vector's VCPU and therefore eoi_exit_bitmap
on original VCPU gets cleared
3) guest's handler for the vector does EOI
4) KVM's EOI handler
2015-08-13 12:03+0200, Christian Borntraeger:
Am 13.08.2015 um 11:29 schrieb Paolo Bonzini:
On 13/08/2015 11:11, Radim Krčmář wrote:
for the new interface. maybe we can rename kvm_check_request in a
separate patch somewhen.
I wonder why haven't we copied the naming convention from bit
Currently, the offset in OperationRegion is limited to 32 bit, extend it
to 64 bit so that we can switch SSDT to 64 bit in later patch
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/acpi/aml-build.c | 2 +-
include/hw/acpi/aml-build.h | 2 +-
2 files changed, 2
NFIT is defined in ACPI 6.0: 5.2.25 NVDIMM Firmware Interface Table (NFIT)
Currently, we only support PMEM mode. Each device has 3 tables:
- SPA table, define the PMEM region info
- MEM DEV table, it has the @handle which is used to associate specified
ACPI NVDIMM device we will introduce in
Implement CreateField term which are used by NVDIMM _DSM method in later patch
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/acpi/aml-build.c | 14 ++
include/hw/acpi/aml-build.h | 1 +
2 files changed, 15 insertions(+)
diff --git a/hw/acpi/aml-build.c
Implement DeRefOf term which is used by NVDIMM _DSM method in later patch
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/acpi/aml-build.c | 8
include/hw/acpi/aml-build.h | 1 +
2 files changed, 9 insertions(+)
diff --git a/hw/acpi/aml-build.c
NVDIMM reserves all the free range above 4G to do:
- Persistent Memory (PMEM) mapping
- implement NVDIMM ACPI device _DSM method
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/i386/pc.c | 12 ++--
hw/mem/nvdimm/pc-nvdimm.c | 13 +
Introduce pc-nvdimm device and it has two parameters:
- @file, which is the backed memory file for NVDIMM device
- @configdata, specify if we need to reserve 128k at the end of
@file for nvdimm device's config data. Default is false
If @configdata is false, Qemu will build a static and
The parameter @file is used as backed memory for NVDIMM which is
divided into two parts if @dataconfig is true:
- first parts is (0, size - 128K], which is used as PMEM (Persistent
Memory)
- 128K at the end of the file, which is used as Config Data Area, it's
used to store Label namespace data
Function 6 is used to set Namespace Label Data
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/mem/nvdimm/acpi.c | 40
1 file changed, 40 insertions(+)
diff --git a/hw/mem/nvdimm/acpi.c b/hw/mem/nvdimm/acpi.c
index 517d710..283228d
Function 4 is used to get Namespace lable size
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/mem/nvdimm/acpi.c | 70
1 file changed, 70 insertions(+)
diff --git a/hw/mem/nvdimm/acpi.c b/hw/mem/nvdimm/acpi.c
index
Function 5 is used to get Namespace Label Data
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/mem/nvdimm/acpi.c | 32
1 file changed, 32 insertions(+)
diff --git a/hw/mem/nvdimm/acpi.c b/hw/mem/nvdimm/acpi.c
index 0a5f2c2..517d710 100644
---
Add NVDIMM maintainer
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
MAINTAINERS | 6 ++
1 file changed, 6 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 978b717..86786e6 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -793,6 +793,12 @@ M: Jiri Pirko
If @configdata is false, Qemu will build a static and readonly
namespace in memory and use it serveing for
DSM GET_CONFIG_SIZE/GET_CONFIG_DATA requests
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/mem/Makefile.objs | 3 +-
hw/mem/nvdimm/acpi.c | 10 ++
NVDIMM devices is defined in ACPI 6.0 9.20 NVDIMM Devices
This is a root device under \_SB and specified NVDIMM device are under the
root device. Each NVDIMM device has _ADR which return its handle used to
associate MEMDEV table in NFIT
We reserve handle 0 for root device. In this patch, we save
Check if the function (Arg2) has additional input info (arg3) and save
the info if needed
We only do the save on NVDIMM device since we are not going to support any
function on root device
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/mem/nvdimm/acpi.c | 73
__DSM is defined in ACPI 6.0: 9.14.1 _DSM (Device Specific Method)
Function 0 is a query function. We do not support any function on root
device and only 3 functions are support for NVDIMM device,
NFIT_CMD_GET_CONFIG_SIZE, NFIT_CMD_GET_CONFIG_DATA and
NFIT_CMD_SET_CONFIG_DATA, that means we
This memory range is used to transfer data between ACPI in guest and Qemu,
it occupies two pages:
- one is RAM-based used to save the input info of _DSM method and Qemu reuse
it store output info
- another one is MMIO-based, ACPI write data to this page to transfer the
control to Qemu
Only 512M is left for MMIO below 4G and that are used by PCI, BIOS etc.
Other components also reserve regions from their internal usage, e.g,
[0xFED0, 0xFED0 + 0x400) is reserved for HPET
Switch SSDT to 64 bit to use the huge free room above 4G. In the later
patches, we will dynamical
Changlog:
- Use litten endian for DSM method, thanks for Stefan's suggestion
- introduce a new parameter, @configdata, if it's false, Qemu will
build a static and readonly namespace in memory and use it serveing
for DSM GET_CONFIG_SIZE/GET_CONFIG_DATA requests. In this case, no
reserved
Implement SizeOf term which is used by NVDIMM _DSM method in later patch
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
hw/acpi/aml-build.c | 8
include/hw/acpi/aml-build.h | 1 +
2 files changed, 9 insertions(+)
diff --git a/hw/acpi/aml-build.c
Am 14.08.2015 um 15:01 schrieb Paolo Bonzini:
- Original Message -
From: Peter Lieven p...@kamp.de
To: qemu-de...@nongnu.org, kvm@vger.kernel.org
Cc: Paolo Bonzini pbonz...@redhat.com
Sent: Friday, August 14, 2015 1:11:34 PM
Subject: Help debugging a regression in KVM Module
Hi,
Peter Lieven p...@kamp.de writes:
Hi,
some time a go I stumbled across a regression in the KVM Module that has been
introduced somewhere
between 3.17 and 3.19.
I have a rather old openSUSE guest with an XFS filesystem which realiably
crashes after some live migrations.
I originally
Hello,
I am making some research abut the x2APIC and number of max vcpus.
According to Intel if we use APIC we can only have 2^8 vcpu and if we
use x2APIC it allows 2^10 vcpu because it has 32 bit address space
for APIC_ID. My question is: is it possible theoretically or any one
tried it? and if
On Fri, Aug 14, 2015 at 01:57:42PM -0700, Andy Lutomirski wrote:
Don't bother testing yet. I'm waffling between trying something like
this and adding SA_SAVE_SS. I have partially written patches for the
latter.
ok, ping me if anything
--
To unsubscribe from this list: send the line
On 13/08/2015 22:18, Andy Lutomirski wrote:
VMX encodes access rights differently from LAR, and the latter is
most likely what x86 people think of when they think of access
rights.
Rename them to avoid confusion.
Good idea, I've gone ahead and applied it for 4.3.
Cc: kvm@vger.kernel.org
On 14/08/2015 10:38, Radim Krčmář wrote:
How do you reproduce the bug?
I run rhel4 (2.6.9) kernel on 2 VCPUs and frequently alternate
smp_affinity of timer. The bug is hit within seconds.
Nice, I'll try to make a unit test for it on the plane. :)
Paolo
--
To unsubscribe from this list:
Linus,
The following changes since commit fc1a8126bf8095b10f5a79893f2d2b19227f88f2:
KVM: MTRR: Use default type for non-MTRR-covered gfn before WARN_ON
(2015-08-05 11:57:57 +0200)
are available in the git repository at:
git://git.kernel.org/pub/scm/virt/kvm/kvm.git tags/for-linus
for you
On Thu, Aug 13, 2015 at 01:18:50PM -0700, Andy Lutomirski wrote:
This adds two new uc_flags flags. UC_SAVED_SS will be set for all
64-bit signals (including x32). It indicates that the saved SS field
is valid and that the kernel understands UC_RESTORE_SS.
The kernel will *not* set
Linus,
the merge window will likely coincide with my two-week vacation
across August and September.
Rather than hoping for an -rc8, I'm sending now what I have. The
PPC and ARM parts might come a few days after the official end
of the merge window.
The uncommon name for the tag (you said you
On 15/08/2015 01:03, Ozgur O Kilic wrote:
My question is: is it possible theoretically or any one
tried it? and if it is I cheched my pc's hardware is support x2APIC
whıch version of KVM should I use for that?
KVM doesn't support more than 256 VCPUs, even with x2APIC enabled in the
guest.
On Fri, Aug 14, 2015 at 1:55 PM, Cyrill Gorcunov gorcu...@gmail.com wrote:
On Thu, Aug 13, 2015 at 01:18:50PM -0700, Andy Lutomirski wrote:
This adds two new uc_flags flags. UC_SAVED_SS will be set for all
64-bit signals (including x32). It indicates that the saved SS field
is valid and that
On 14/08/2015 12:08, Radim Krčmář wrote:
v3:
* acked by Christian [1/5]
* use ioctl argument directly (unsigned long as flags) [4/5]
* precisely #ifdef arch-specific ioctls [5/5]
v2:
* move request_exits debug counter patch right after introduction of
KVM_REQ_EXIT [3/5]
* use
On 07/10/2015 04:21 PM, Andre Przywara wrote:
As the actual LPI number in a guest can be quite high, but is mostly
assigned using a very sparse allocation scheme, bitmaps and arrays
for storing the virtual interrupt status are a waste of memory.
We use our equivalent of the Interrupt
On 07/10/2015 04:21 PM, Andre Przywara wrote:
The LPI configuration and pending tables of the GICv3 LPIs are held
in tables in (guest) memory. To achieve reasonable performance, we
cache this data in our own data structures, so we need to sync those
two views from time to time. This behaviour
Hello! Thank you for quick response.
This is fairly unreadable. Please use a switch statement instead.
Christoffer disliked it in v1, so i thought a bit and changed it. Ok, will
change it back.
And here, we're going to assume that the arch timer still usable. We
definitely need a way to
On 08/14/2015 01:58 PM, Eric Auger wrote:
On 07/10/2015 04:21 PM, Andre Przywara wrote:
The LPI configuration and pending tables of the GICv3 LPIs are held
in tables in (guest) memory. To achieve reasonable performance, we
cache this data in our own data structures, so we need to sync those
- Original Message -
From: Peter Lieven p...@kamp.de
To: qemu-de...@nongnu.org, kvm@vger.kernel.org
Cc: Paolo Bonzini pbonz...@redhat.com
Sent: Friday, August 14, 2015 1:11:34 PM
Subject: Help debugging a regression in KVM Module
Hi,
some time a go I stumbled across a
PING
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
-Original Message-
From: kvm-ow...@vger.kernel.org [mailto:kvm-ow...@vger.kernel.org] On Behalf
Of Pavel Fedin
Sent: Wednesday, August 05, 2015 1:54 PM
To: kvm...@lists.cs.columbia.edu;
The guest can use KVM_USER_EXIT instead of a signal-based exiting to
userspace. Availability depends on KVM_CAP_USER_EXIT.
Only x86 is implemented so far.
Signed-off-by: Radim Krčmář rkrc...@redhat.com
---
v3:
* use ioctl argument directly (unsigned long as flags) [Paolo]
v2:
* use vcpu
When userspace wants KVM to exit to userspace, it sends a signal.
This has a disadvantage of requiring a change to the signal mask because
the signal needs to be blocked in userspace to stay pending when sending
to self.
Using a request flag allows us to shave 200-300 cycles from every
userspace
v3:
* acked by Christian [1/5]
* use ioctl argument directly (unsigned long as flags) [4/5]
* precisely #ifdef arch-specific ioctls [5/5]
v2:
* move request_exits debug counter patch right after introduction of
KVM_REQ_EXIT [3/5]
* use vcpu ioctl instead of vm one [4/5]
* shrink
We want to have requests abstracted from bit operations.
Acked-by: Christian Borntraeger borntrae...@de.ibm.com
Signed-off-by: Radim Krčmář rkrc...@redhat.com
---
v3: acked by Christian
arch/x86/kvm/vmx.c | 2 +-
include/linux/kvm_host.h | 7 ++-
2 files changed, 7 insertions(+), 2
We are still interested in the amount of exits userspace requested and
signal_exits doesn't cover that anymore.
Signed-off-by: Radim Krčmář rkrc...@redhat.com
---
v2: move request_exits debug counter patch right after introduction of
KVM_REQ_EXIT
arch/x86/include/asm/kvm_host.h | 1 +
https://bugzilla.kernel.org/show_bug.cgi?id=102651
Huaitong Han oen...@gmail.com changed:
What|Removed |Added
CC||oen...@gmail.com
---
I find the switch easier to read and modify.
Signed-off-by: Radim Krčmář rkrc...@redhat.com
---
v3: precisely #ifdef arch-specific ioctls [Christian]
v2: new
virt/kvm/kvm_main.c | 12
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/virt/kvm/kvm_main.c
On 14/08/15 13:26, Pavel Fedin wrote:
Hello! Thank you for quick response.
This is fairly unreadable. Please use a switch statement instead.
Christoffer disliked it in v1, so i thought a bit and changed it. Ok, will
change it back.
And here, we're going to assume that the arch timer
50 matches
Mail list logo
