On Tue, Mar 12, 2013 at 09:41:04AM +0100, Peter Lieven wrote:
>
> Am 12.03.2013 um 09:35 schrieb Stefan Hajnoczi :
>
> > On Mon, Mar 11, 2013 at 02:44:03PM +0100, Peter Lieven wrote:
> >> I ever since had a few VMs which are very hard to migrate because of a lot
> >> of memory I/O. I found that
On 12 March 2013 15:08, KONRAD Frédéric wrote:
> On 12/03/2013 15:42, Peter Maydell wrote:
>>
>> Yes, I see your problem there, but DEFINE_VIRTIO_BLK_SCSI_PROPERTY
>> and DEFINE_DATA_PLANE_PROPERTIES are just convenience macros, not
>> ones that are expected to be used by other code, right? So you
On Tue, Mar 12, 2013 at 09:47:57PM +0800, harryxiyou wrote:
> Could anyone give me some suggestions to submit HLFS driver patches for
> QEMU to our QEMU community. Thanks a lot in advance ;-)
http://wiki.qemu.org/Contribute/SubmitAPatch
Stefan
The following changes since commit fe3cc14fd83e0c8f376d849ccd0fc3433388442d:
Merge remote-tracking branch 'quintela/migration.next' into staging
(2013-03-11 08:30:34 -0500)
are available in the git repository at:
git://repo.or.cz/qemu/kevin.git for-anthony
for you to fetch changes up to 6
From: Stefan Hajnoczi
The update_cluster_refcount() function increments/decrements a cluster's
refcount and then returns the new refcount value.
There is no need to flush since both update_cluster_refcount() callers
already take care of this:
1. qcow2_alloc_bytes() calls update_cluster_refcount
On 12/03/2013 16:12, Peter Maydell wrote:
On 12 March 2013 15:08, KONRAD Frédéric wrote:
On 12/03/2013 15:42, Peter Maydell wrote:
Yes, I see your problem there, but DEFINE_VIRTIO_BLK_SCSI_PROPERTY
and DEFINE_DATA_PLANE_PROPERTIES are just convenience macros, not
ones that are expected to be u
From: Stefan Hajnoczi
Since qcow2 metadata is cached we need to flush the caches, not just the
underlying file. Use bdrv_flush(bs) instead of bdrv_flush(bs->file).
Also add the error return path when bdrv_flush() fails and move the
flush after checking for qcow2_alloc_clusters() failure so that
On Tue, 2013-03-12 at 16:46 +0200, Michael S. Tsirkin wrote:
> On Mon, Mar 11, 2013 at 03:18:49PM -0600, Alex Williamson wrote:
> > When creating capabilities devices need to know what kind of bus
> > they're on. If we're on an express bus without a parent_dev, then
> > we're on the root complex a
From: Stefan Hajnoczi
Compressed writes use qcow2_alloc_bytes() to allocate space with byte
granularity. The affected clusters' refcounts will be incremented but
we do not need to flush yet.
Set a L2 cache dependency on the refcount block cache, so that the
refcounts get written out before the
On Tue, Mar 12, 2013 at 04:30:41PM +0800, Wenchao Xia wrote:
> 于 2013-1-15 15:03, Wenchao Xia 写道:
> >于 2013-1-14 18:06, Stefan Hajnoczi 写道:
> >>On Mon, Jan 14, 2013 at 10:56:30AM +0800, Wenchao Xia wrote:
> >>>于 2013-1-11 17:12, Stefan Hajnoczi 写道:
> On Fri, Jan 11, 2013 at 02:22:28PM +0800, We
I will sent a list of optimizations that have been developed during
the last days. Your feedback is much appreciated.
@Paolo: Can you run your tests that you have made for your migration
optimzations that are now upstream against these patches.
Thanks,
Peter
Peter Lieven (9):
move vector defi
vector optimizations will now be used outside is_dup_page() in arch_init.c
Signed-off-by: Peter Lieven
---
arch_init.c | 20
include/qemu-common.h | 23 +++
2 files changed, 23 insertions(+), 20 deletions(-)
diff --git a/arch_init.c b/arch
Signed-off-by: Peter Lieven
---
include/qemu-common.h |1 +
util/cutils.c | 30 ++
2 files changed, 31 insertions(+)
diff --git a/include/qemu-common.h b/include/qemu-common.h
index a3555d0..c22c41e 100644
--- a/include/qemu-common.h
+++ b/include/qemu-
Signed-off-by: Peter Lieven
---
arch_init.c |3 +++
1 file changed, 3 insertions(+)
diff --git a/arch_init.c b/arch_init.c
index 1b71912..d48b914 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -326,6 +326,7 @@ static ram_addr_t last_offset;
static unsigned long *migration_bitmap;
static ui
On Tue, Mar 12, 2013 at 09:36:49AM -0600, Alex Williamson wrote:
> On Tue, 2013-03-12 at 16:46 +0200, Michael S. Tsirkin wrote:
> > On Mon, Mar 11, 2013 at 03:18:49PM -0600, Alex Williamson wrote:
> > > When creating capabilities devices need to know what kind of bus
> > > they're on. If we're on
at the beginning of migration all pages are marked dirty and
in the first round a bulk migration of all pages is performed.
currently all these pages are copied to the page cache regardless
if there are frequently updated or not. this doesn't make sense
since most of these pages are never transfe
this patch adds the usage of buffer_find_nonzero_offset()
to skip large areas of zeroes.
compared to loop unrolling this adds another 50% performance
benefit for skipping large areas of zeroes.
Signed-off-by: Peter Lieven
---
util/bitops.c | 23 ---
1 file changed, 20 ins
if the dup_page is a zero_page do not send it at all.
the memory at the destination reads as zero anyway.
Signed-off-by: Peter Lieven
---
arch_init.c |6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/arch_init.c b/arch_init.c
index 04c82e4..abd54a7 100644
--- a/arch_i
avoid searching for dirty pages just increment the
page offset. all pages are dirty anyway.
Signed-off-by: Peter Lieven
---
arch_init.c | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/arch_init.c b/arch_init.c
index abd54a7..45fcf4d 100644
--- a/arch_init.c
+++
performance gain on SSE2 is approx. 20-25%. altivec
is not tested. performance for unsigned long arithmetic
is unchanged.
Signed-off-by: Peter Lieven
---
util/cutils.c |5 +
1 file changed, 5 insertions(+)
diff --git a/util/cutils.c b/util/cutils.c
index a09d8e8..23f0cd6 100644
--- a/u
pages are dup (most likely zero) from the very beginning.
unless there is an application running inside the VM
zeroing out large areas of memory its wasted time
for checking for dup pages after the bulk stage.
Signed-off-by: Peter Lieven
---
arch_init.c |2 +-
1 file changed, 1 insertion(+)
On 03/12/2013 09:48 AM, Peter Lieven wrote:
> Signed-off-by: Peter Lieven
> ---
> include/qemu-common.h |1 +
> util/cutils.c | 30 ++
> 2 files changed, 31 insertions(+)
>
>
> /*
> + * Searches for an area with non-zero content in a buffer
> + *
> +
On 03/12/2013 09:50 AM, Peter Lieven wrote:
> performance gain on SSE2 is approx. 20-25%. altivec
> is not tested. performance for unsigned long arithmetic
> is unchanged.
>
> Signed-off-by: Peter Lieven
> ---
> util/cutils.c |5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/util
Am 12.03.2013 um 17:01 schrieb Eric Blake :
> On 03/12/2013 09:50 AM, Peter Lieven wrote:
>> performance gain on SSE2 is approx. 20-25%. altivec
>> is not tested. performance for unsigned long arithmetic
>> is unchanged.
>>
>> Signed-off-by: Peter Lieven
>> ---
>> util/cutils.c |5 +
>>
On 03/12/2013 09:51 AM, Peter Lieven wrote:
> at the beginning of migration all pages are marked dirty and
> in the first round a bulk migration of all pages is performed.
>
> currently all these pages are copied to the page cache regardless
> if there are frequently updated or not. this doesn't m
On 03/12/2013 09:52 AM, Peter Lieven wrote:
> this patch adds the usage of buffer_find_nonzero_offset()
> to skip large areas of zeroes.
>
> compared to loop unrolling this adds another 50% performance
> benefit for skipping large areas of zeroes.
>
> Signed-off-by: Peter Lieven
> ---
> util/bi
On 03/12/2013 09:47 AM, Peter Lieven wrote:
> I will sent a list of optimizations that have been developed during
> the last days. Your feedback is much appreciated.
>
> @Paolo: Can you run your tests that you have made for your migration
> optimzations that are now upstream against these patches.
Am 12.03.2013 um 16:56 schrieb Eric Blake :
> On 03/12/2013 09:47 AM, Peter Lieven wrote:
>> I will sent a list of optimizations that have been developed during
>> the last days. Your feedback is much appreciated.
>>
>> @Paolo: Can you run your tests that you have made for your migration
>> opti
On 03/12/2013 10:03 AM, Peter Lieven wrote:
>>> +if (((uintptr_t) buf) % sizeof(VECTYPE) == 0
>>> + && len % 8*sizeof(VECTYPE) == 0) {
>>
>> Space around binary operators. Use CHAR_BITS instead of a magic number
>> 8.
> the 8 is not BITS_PER_BYTE or CHAR_BITS its the number of
> vec
Your patch series is not properly threaded. Please use git-send-email.
On 03/11/2013 11:01 PM, Wenchao Xia wrote:
> 于 2013-3-12 1:49, Eric Blake 写道:
>> On 03/11/2013 05:23 AM, Wenchao Xia wrote:
>>>This patch adds block/snapshot.c and then moves the function
>>> there. It also fixes small code style errors reported by check script.
>>>
>>> + * This work is licens
On 03/12/2013 04:07 AM, Stefan Hajnoczi wrote:
> On Mon, Mar 11, 2013 at 07:23:02PM +0800, Wenchao Xia wrote:
>> In the use of snapshot a way to retrieve related info at runtime is needed,
>> so this serial of patches will merge some code for qemu and qemu-img, and add
>> following interfaces for
On 03/12/2013 10:15 AM, Eric Blake wrote:
> As far as I can tell, there is nothing wrong with leaving the file as
> BSD licensed instead of trying to insist that it be LGPL. The block
> layer will still be [L]GPL because of other files linked together, but
> there is nothing inherently wrong with
On Tue, 2013-03-12 at 17:50 +0200, Michael S. Tsirkin wrote:
> On Tue, Mar 12, 2013 at 09:36:49AM -0600, Alex Williamson wrote:
> > On Tue, 2013-03-12 at 16:46 +0200, Michael S. Tsirkin wrote:
> > > On Mon, Mar 11, 2013 at 03:18:49PM -0600, Alex Williamson wrote:
> > > > When creating capabilities
On Tue, Mar 12, 2013 at 10:25:20AM -0600, Alex Williamson wrote:
> On Tue, 2013-03-12 at 17:50 +0200, Michael S. Tsirkin wrote:
> > On Tue, Mar 12, 2013 at 09:36:49AM -0600, Alex Williamson wrote:
> > > On Tue, 2013-03-12 at 16:46 +0200, Michael S. Tsirkin wrote:
> > > > On Mon, Mar 11, 2013 at 03:
On Tue, 12 Mar 2013 16:22:22 +0100
KONRAD Frédéric wrote:
> On 12/03/2013 16:12, Peter Maydell wrote:
> > On 12 March 2013 15:08, KONRAD Frédéric wrote:
> >> On 12/03/2013 15:42, Peter Maydell wrote:
> >>> Yes, I see your problem there, but DEFINE_VIRTIO_BLK_SCSI_PROPERTY
> >>> and DEFINE_DATA_P
Hi there:
I've read this an article discribe the thread model in Qemu.
http://blog.vmsplice.net/2011/03/qemu-internals-overall-architecture-and.html
Is there really one thread per vcpu in qemu?
When I run the qemu using "-smp 16", I can see only two threads using
"ps -xH".
So I want to know that
v3:
- split 'Add GDB qAttached support' and 'New option -gdb-opts'
Fabien Chouteau (3):
Add GDB qAttached support
Revert "gdbstub: Do not kill target in system emulation mode"
New option -gdb-opts
gdbstub.c | 40 +---
include/exec/gdbstu
The requirements described in this patch are implemented by "Add GDB
qAttached support".
This reverts commit 00e94dbc7fd0110b0555d59592b004333adfb4b8.
Signed-off-by: Fabien Chouteau
---
gdbstub.c |2 --
1 file changed, 2 deletions(-)
diff --git a/gdbstub.c b/gdbstub.c
index 70d54ce..0e9431
We introduce a new command line option. It's a generic option to
customize the gdb server:
-gdb-opts [attached=on|off]
The only parameter for now is "attached".
Signed-off-by: Fabien Chouteau
---
gdbstub.c | 28 +++-
include/exec/gdbstub.h |2 ++
qemu
On 03/12/2013 02:56 AM, Gerd Hoffmann wrote:
> Signed-off-by: Gerd Hoffmann
> ---
> qemu-char.c |3 +++
> 1 file changed, 3 insertions(+)
Did something go wrong in your rebase? The subject line suggest you
intended to squash this to another patch.
>
> diff --git a/qemu-char.c b/qemu-char.
With this patch QEMU handles qAttached request from gdb. When QEMU
replies 1, GDB sends a "detach" command at the end of a debugging
session otherwise GDB sends "kill".
The default value for qAttached is 1 on system emulation and 0 on user
emulation.
This patch implements the requirement describe
On 03/11/2013 05:23 AM, Wenchao Xia wrote:
> This patch adds block/qapi.c and moves the functions there. To avoid
> conflict and tip better, macro in header file is BLOCK_QAPI_H instead
> of QAPI_H. The moving is for making review easier, those functions
> will be modified and renamed later.
>
>
Am 12.03.2013 um 13:02 schrieb Paolo Bonzini :
> Il 12/03/2013 12:51, Peter Lieven ha scritto:
buffer_is_zero is used in somewhat special cases (block
streaming/copy-on-read) where throughput doesn't really matter, unlike
is_dup_page/find_zero_bit which are used in migration. But
On 03/12/2013 09:54 AM, Peter Lieven wrote:
> if the dup_page is a zero_page do not send it at all.
> the memory at the destination reads as zero anyway.
It might help to comment that this is only true for the bulk migration
stage. If the guest is actively zeroing pages, then later stages might
e
On 03/11/2013 05:23 AM, Wenchao Xia wrote:
> This patch adds a parameter to tell whether return valid snapshots
> for whole VM only.
>
> Signed-off-by: Wenchao Xia
> ---
> block/qapi.c | 42 --
> include/block/qapi.h |1 +
> qemu-img.c
On Sat, 2013-03-09 at 01:52 -0600, Vijay Mohan Pandarathil wrote:
> Add support for error containment when a VFIO device assigned to a KVM
> guest encounters an error. This is for PCIe devices/drivers that support AER
> functionality. When the host OS is notified of an error in a device either
> th
On 03/11/2013 05:23 AM, Wenchao Xia wrote:
> This interface returns info of valid internal snapshots for whole vm.
>
> Signed-off-by: Wenchao Xia
> ---
> block/qapi.c | 18 +
> qapi-schema.json | 14 +
> qmp-commands.hx | 55
> +++
On 03/11/2013 05:23 AM, Wenchao Xia wrote:
> This mirror function will return image info including snapshots,
> and if specified backing image's info will also be returned. Now
> Qemu have both query-images and query-block interfaces.
In the middle of a sentence, we prefer QEMU or qemu, not Qemu
Applied. Thanks.
Regards,
Anthony Liguori
Public bug reported:
I am running daily automated tests that involve booting a NetBSD 6.0.1
guest in qemu freshly built from git. The tests are scripted using
pexpect, which interacts with the NetBSD guest over the emulated
serial console. Recently, the tests stopped working; the guest boots
and
Applied. Thanks.
Regards,
Anthony Liguori
Applied. Thanks.
Regards,
Anthony Liguori
On 03/12/13 10:34, 解壁伟 wrote:
> Hi there:
>
> I've read this an article discribe the thread model in Qemu.
> http://blog.vmsplice.net/2011/03/qemu-internals-overall-architecture-and.html
> Is there really one thread per vcpu in qemu?
>
> When I run the qemu using "-smp 16", I can see only two thr
This patch reorganizes qemu file operations to be in their own source file
instead of being lumped in savevm.c. Besides being more logical for maintenance
it also makes it easier for future users of the file functions to add tests.
v3 forward port to resolve conflicts
v2 forward port to resolve c
This is the patch to introduce the query-cpu-max QMP command to get
the maximum number of CPUs supported by the currently running emulator
instance. This may differ machine from machine as defined by -machine
settings and max_cpus member of QEMUMachine structure.
It's been tested both using QMP/qm
On Tue, Mar 12, 2013 at 11:16 PM, Stefan Hajnoczi wrote:
> On Tue, Mar 12, 2013 at 09:47:57PM +0800, harryxiyou wrote:
>> Could anyone give me some suggestions to submit HLFS driver patches for
>> QEMU to our QEMU community. Thanks a lot in advance ;-)
>
> http://wiki.qemu.org/Contribute/SubmitAPa
Oops. Screwed this one up so as to break compile for
CONFIG_USER_ONLY. Fixed version below.
From 0fd64364fcc0af95b23e17fa8ba58c28685a3a02 Mon Sep 17 00:00:00 2001
From: David Gibson
Date: Wed, 13 Mar 2013 11:40:33 +1100
Subject: [PATCH] target-ppc: Use QOM method dispatch for MMU fault handling
On Tue, Mar 12, 2013 at 4:55 PM, Paolo Bonzini wrote:
> Il 07/03/2013 03:53, Liu Ping Fan ha scritto:
>> From: Liu Ping Fan
>>
>> Introduce nc->send_lock, it shield off the race of nc->peer's reader and
>> deleter. With it, after deleter finish, no new qemu_send_packet_xx()
>> can reach ->send_qu
于 2013-3-12 23:43, Stefan Hajnoczi 写道:
> On Tue, Mar 12, 2013 at 04:30:41PM +0800, Wenchao Xia wrote:
>> 于 2013-1-15 15:03, Wenchao Xia 写道:
>>> 于 2013-1-14 18:06, Stefan Hajnoczi 写道:
On Mon, Jan 14, 2013 at 10:56:30AM +0800, Wenchao Xia wrote:
> 于 2013-1-11 17:12, Stefan Hajnoczi 写道:
>
于 2013-3-13 0:22, Eric Blake 写道:
On 03/12/2013 10:15 AM, Eric Blake wrote:
As far as I can tell, there is nothing wrong with leaving the file as
BSD licensed instead of trying to insist that it be LGPL. The block
layer will still be [L]GPL because of other files linked together, but
there is no
On Tue, Mar 12, 2013 at 4:50 PM, Paolo Bonzini wrote:
> Il 07/03/2013 03:53, Liu Ping Fan ha scritto:
>> From: Liu Ping Fan
>>
>> Forward packet to other hub ports by their AioContext.
>>
>> Signed-off-by: Liu Ping Fan
>> ---
>> hw/qdev-properties-system.c |1 +
>> include/block/aio.h
On Tue, Mar 12, 2013 at 09:20:24AM +0100, Stefan Hajnoczi wrote:
> On Tue, Mar 12, 2013 at 02:29:42PM +0800, Asias He wrote:
> > diff --git a/hw/virtio-pci.c b/hw/virtio-pci.c
> > index 39c1966..4a97ca1 100644
> > --- a/hw/virtio-pci.c
> > +++ b/hw/virtio-pci.c
>
> These changes break the build fo
于 2013-3-13 0:16, Eric Blake 写道:
On 03/12/2013 04:07 AM, Stefan Hajnoczi wrote:
On Mon, Mar 11, 2013 at 07:23:02PM +0800, Wenchao Xia wrote:
In the use of snapshot a way to retrieve related info at runtime is needed,
so this serial of patches will merge some code for qemu and qemu-img, and a
These patches implement asn1 ber visitors for encoding and decoding data.
changed since v1:
Moved .c files into qapi directory
Moved .h files into include/qapi
Added sized buffer code
cleaned up Makefile changes to play nicer
Broke out patches for output/input visitors
New tests in tests/test-vis
Add a sized buffer interface to qapi.
Cc: Michael Tsirkin
Signed-off-by: Stefan Berger
Signed-off-by: Joel Schopp
---
include/qapi/visitor-impl.h |2 ++
include/qapi/visitor.h |2 ++
qapi/qapi-visit-core.c |8
3 files changed, 12 insertions(+)
diff --git a/inclu
Add a 3 very short file wrapper functions to make code that follows more
readable. Also export an existing function that is currently static.
Cc: Michael Tsirkin
Signed-off-by: Stefan Berger
Signed-off-by: Joel Schopp
---
include/migration/qemu-file.h |3 +++
qemu-file.c
This patch reorganizes qemu file operations to be in their own source file
instead of being lumped in savevm.c. Besides being more logical for maintenance
it also makes it easier for future users of the file functions to add tests.
Cc: Michael Tsirkin
Signed-off-by: Stefan Berger
Signed-off-by:
Forward ported Mike's previously sent patch
(see http://lists.gnu.org/archive/html/qemu-devel/2012-10/msg05782.html ) in my
series since it impelments a qapi array interface the ASN.1 BER visitor needs.
Generally these will be serialized into lists, but the
representation can be of any form so lon
Since I'm throwing all this code out there I'm also signing up to maintain it.
Send me your bug reports.
Cc: Michael Tsirkin
Cc: Stefan Berger
Signed-off-by: Joel Schopp
---
MAINTAINERS |8
1 file changed, 8 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 0ca7e1d..8a2f
This patch adds support functions for operating on in memory sized file buffers.
Signed-off-by: Stefan Berger
Signed-off-by: Joel Schopp
---
include/migration/qemu-file.h | 12 +++
include/qemu-common.h | 15
qemu-file.c | 184
Implement an input visitor for ASN.1 BER encoding.
Cc: Michael Tsirkin
Signed-off-by: Stefan Berger
Signed-off-by: Joel Schopp
---
include/qapi/ber-input-visitor.h | 30 ++
qapi/Makefile.objs |2 +-
qapi/ber-input-visitor.c | 1073 +++
Implement an output visitor for ASN.1 BER encoding.
Cc: Michael Tsirkin
Signed-off-by: Stefan Berger
Signed-off-by: Joel Schopp
---
configure |2 +-
include/qapi/ber-output-visitor.h | 28 ++
include/qapi/ber.h| 107 +++
qapi/Makefile.objs
Add BER Visitor hooks to test-visitor-serialization
Cc: Michael Tsirkin
Cc: Stefan Berger
Signed-off-by: Joel Schopp
---
tests/Makefile |2 +-
tests/test-visitor-serialization.c | 72
2 files changed, 73 insertions(+), 1 deletion(-
These series aim to port network backend onto glib, and
prepare for moving towards making network layer mutlit-thread.
The brief of the whole aim and plan is documented on
http://wiki.qemu.org/Features/network_reentrant
In these series, attach each NetClientState with a GSource
At the first, I us
Bind each NetClientState with a GSource(ie,NetClientSource). Currently,
these GSource attached with default context, but in future, after
resolving the race between handlers and the interface exposed by NetClientInfo
and other re-entrant issue, we can run NetClientState on different GMainContext
From: Liu Ping Fan
Attach each hubport with a GSource. Currently the GSource is attached
to default main context, and the hub still run on iothread, but in
future, after making the whole layer thread-safe, we will admit ports
to run on different thread(GMainContext).
Signed-off-by: Liu Ping Fan
On 03/12/13 18:42, Eric Blake wrote:
> On 03/12/2013 02:56 AM, Gerd Hoffmann wrote:
>> Signed-off-by: Gerd Hoffmann --- qemu-char.c
>> |3 +++ 1 file changed, 3 insertions(+)
>
> Did something go wrong in your rebase? The subject line suggest
> you intended to squash this to another patch.
Y
201 - 278 of 278 matches
Mail list logo
