On Tue, 2 Jun 2020 14:19:08 +0200
Philippe Mathieu-Daudé wrote:
> +Cleber/Wainer
>
> On 5/11/20 2:27 PM, Lukas Straub wrote:
> > Add high-level test relying on the colo resource-agent to test
> > all failover cases while checking guest network connectivity.
> >
>
chive/html/qemu-block/2020-05/msg00634.html
>
> Berto
>
Hi,
Did you have a look at my series? There it's fixed properly:
https://lore.kernel.org/qemu-devel/9df6b3723ec30cb749ceaa555d82a29a6d79496d.1589199922.git.lukasstra...@web.de/
Regards,
Lukas Straub
pgpOedZXmirGq.pgp
Description: OpenPGP digital signature
Register yank functions on sockets to shut them down.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
migration/channel.c | 12
migration/migration.c | 18 +-
migration/multifd.c | 10 ++
migration/qemu-file
can be queried by a 'query-yank' oob command.
Signed-off-by: Lukas Straub
---
qapi/misc.json | 45 +
yank.c | 174 +
yank.h | 67 +++
3 files changed, 286 insertions(+)
create mode 100644 yan
Register a yank function to shutdown the socket on yank.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
chardev/char-socket.c | 24
2 files changed, 25 insertions(+)
diff --git a/Makefile.objs b/Makefile.objs
index 8e403b81f3..5582f4eda9 100644
--- a
s register callbacks which get
executed with the yank command. For example the callback can shutdown() a
socket. This is intended for the colo use-case, but it can be used for other
things too of course.
Regards,
Lukas Straub
v4:
-fix build errors...
v3:
-don't touch softmmu/vl.c, use
Register a yank function which shuts down the socket and sets
s->state = NBD_CLIENT_QUIT. This is the same behaviour as if an
error occured.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
block/nbd.c | 101 --
2 files changed,
Register yank functions on sockets to shut them down.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
migration/channel.c | 12
migration/migration.c | 18 +-
migration/multifd.c | 10 ++
migration/qemu-file
Register a yank function to shutdown the socket on yank.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
chardev/char-socket.c | 24
2 files changed, 25 insertions(+)
diff --git a/Makefile.objs b/Makefile.objs
index 8e403b81f3..5582f4eda9 100644
--- a
can be queried by a 'query-yank' oob command.
Signed-off-by: Lukas Straub
---
qapi/misc.json | 45 +
yank.c | 174 +
yank.h | 67 +++
3 files changed, 286 insertions(+)
create mode 100644 yan
Register a yank function which shuts down the socket and sets
s->state = NBD_CLIENT_QUIT. This is the same behaviour as if an
error occured.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
block/nbd.c | 101 --
2 files changed,
s register callbacks which get
executed with the yank command. For example the callback can shutdown() a
socket. This is intended for the colo use-case, but it can be used for other
things too of course.
Regards,
Lukas Straub
v3:
-don't touch softmmu/vl.c, use __contructor__ attribute
iodic mode) and to use multiple colo-compare
for multiple network interfaces.
Signed-off-by: Lukas Straub
Tested-by: Lukas Straub
---
net/colo-compare.c | 35 +--
1 file changed, 29 insertions(+), 6 deletions(-)
diff --git a/net/colo-compare.c b/net/colo-comp
full. This leads to a deadlock because both event loops
get blocked.
Fix this by converting compare_chr_send to a coroutine and
putting the packets in a send queue.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Tested-by: Zhang Chen
---
net/colo-compare.c
In colo_compare_complete, insert CompareState into net_compares
only after everything has been initialized.
In colo_compare_finalize, remove CompareState from net_compares
before anything is deinitialized.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
---
net/colo-compare.c | 45
This will be needed in the next patch so compare_chr_send can be
converted to a coroutine.
Signed-off-by: Lukas Straub
Reviewed-by: Marc-André Lureau
Reviewed-by: Zhang Chen
Reviewed-by: Philippe Mathieu-Daudé
---
chardev/char.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion
fix the crashes. This is safe, because
the bh already takes the appropriate locks.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Reviewed-by: Derek Su
Tested-by: Derek Su
---
net/colo-compare.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/colo-compare.c b/net
Hello Everyone,
Here are fixes for bugs that I found in my tests. I have tested this with my
test suite and everything works fine.
Regards,
Lukas Straub
Version changes:
v6:
-don't touch softmmu/vl.c, use __contructor__ attribute instead
-better wording
v5:
-initialize and use notify_s
Else the log will be flooded if there is a lot of network
traffic.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Reviewed-by: Philippe Mathieu-Daudé
Tested-by: Philippe Mathieu-Daudé
---
net/colo-compare.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a
On Thu, 21 May 2020 01:30:48 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Jason Wang
> > Sent: Wednesday, May 20, 2020 8:23 PM
> > To: Zhang, Chen ; qemu-devel@nongnu.org; Lukas
> > Straub
> > Cc: zhangc...@gmail.com
> &
On Wed, 20 May 2020 23:05:50 +0200
Lukas Straub wrote:
> Register yank functions on sockets to shut them down.
>
> Signed-off-by: Lukas Straub
Don't review this commit for now, I'll have to revamp it anyway.
Regards,
Lukas Straub
pgpNmGVoC9Qjd.pgp
Description: OpenPGP digital signature
On Thu, 21 May 2020 16:03:35 +0100
Stefan Hajnoczi wrote:
> On Wed, May 20, 2020 at 11:05:39PM +0200, Lukas Straub wrote:
> > +void yank_generic_iochannel(void *opaque)
> > +{
> > +QIOChannel *ioc = QIO_CHANNEL(opaque);
> > +
> > +qio_channel_shutdown(ioc
Register yank functions on sockets to shut them down.
Signed-off-by: Lukas Straub
---
migration/migration.c | 9 +
migration/qemu-file-channel.c | 6 ++
migration/socket.c| 11 +++
3 files changed, 26 insertions(+)
diff --git a/migration/migration.c b
Register a yank function which shuts down the socket and sets
s->state = NBD_CLIENT_QUIT. This is the same behaviour as if an
error occured.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
block/nbd.c | 101 --
2 files changed,
Register a yank function to shutdown the socket on yank.
Signed-off-by: Lukas Straub
---
chardev/char-socket.c | 24
1 file changed, 24 insertions(+)
diff --git a/chardev/char-socket.c b/chardev/char-socket.c
index 185fe38dda..d5c6cd2153 100644
--- a/chardev/char
can be queried by a 'query-yank' oob command.
Signed-off-by: Lukas Straub
---
qapi/misc.json | 45 +
softmmu/vl.c | 2 +
yank.c | 174 +
yank.h | 69
4 files changed, 290 insertions
s register callbacks which get
executed with the yank command. For example the callback can shutdown() a
socket. This is intended for the colo use-case, but it can be used for other
things too of course.
Regards,
Lukas Straub
v2:
-don't touch io/ code anymore
-always register yank functio
migration_rate_limit will erroneously ratelimit a shutdown socket,
which causes the migration thread to hang in ram_save_host_page
if the socket is shutdown.
Fix this by explicitly testing if the socket has errors or was
shutdown in migration_rate_limit.
Signed-off-by: Lukas Straub
If the secondary has to failover during checkpointing, it still is
in the old state (i.e. different state than primary). Thus we can't
expose the primary state until after the checkpoint is sent.
This fixes sporadic connection reset of client connections during
failover.
Signed-off-by:
er on colo_incoming_sem.
Fix this by relaunching failover even if there was an error. Also,
to make this work properly, set vmstate_loading to false when
returning during the vmstate_loading section.
Signed-off-by: Lukas Straub
Reviewed-by: zhanghailiang
---
migration/colo.c | 17 -
1 file ch
If we suceed in receiving ram state, but fail receiving the device
state, there will be a mismatch between the two.
Fix this by flushing the ram cache only after the vmstate has been
received.
Signed-off-by: Lukas Straub
Reviewed-by: zhanghailiang
---
migration/colo.c | 1 +
migration/ram.c
On Tue, 19 May 2020 15:50:20 +0100
"Dr. David Alan Gilbert" wrote:
> * Lukas Straub (lukasstra...@web.de) wrote:
> > On Mon, 18 May 2020 12:55:34 +0100
> > "Dr. David Alan Gilbert" wrote:
> >
> > > * Zhanghailiang (zhang.zhanghaili...@huawei
Hello Everyone,
Here are fixes for bugs that I found in my tests.
Regards,
Lukas Straub
v2:
-Don't touch qemu_file_rate_limit, solve the problem in migration_rate_limit
Lukas Straub (6):
migration/colo.c: Use event instead of semaphore
migration/colo.c: Use cpu_synchronize_all_s
the
checkpoint event is sent to colo-compare (which releases the
miscompared packets).
Benchmark results (iperf3):
Client-to-server tcp:
without patch: ~66 Mbit/s
with patch: ~61 Mbit/s
Server-to-client tcp:
without patch: ~702 Kbit/s
with patch: ~16 Mbit/s
Signed-off-by: Lukas Straub
Reviewed-by
using cpu_synchronize_all_states(), which initializes
CPUState from the current cpu registers additionally to marking
the vcpus as dirty.
Signed-off-by: Lukas Straub
Reviewed-by: Dr. David Alan Gilbert
---
migration/colo.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a
ork traffic anyway.
> > >
> >
> > I'm not quite sure if this modification can take side effect on normal
> > migration process or not,
> > There are several places calling it.
> >
> > Maybe Juan and Dave can help ;)
> >
> > >
On Fri, 15 May 2020 11:26:13 +0100
Daniel P. Berrangé wrote:
> On Fri, May 15, 2020 at 12:14:47PM +0200, Lukas Straub wrote:
> > On Fri, 15 May 2020 11:04:13 +0100
> > Daniel P. Berrangé wrote:
> >
> > > On Fri, May 15, 2020 at 11:48:18AM +0200, Lukas Straub w
Else the log will be flooded if there is a lot of network
traffic.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Reviewed-by: Philippe Mathieu-Daudé
Tested-by: Philippe Mathieu-Daudé
---
net/colo-compare.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a
iodic mode) and to use multiple colo-compare
for multiple network interfaces.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 35 +--
net/colo-compare.h | 1 +
softmmu/vl.c | 2 ++
3 files changed, 32 insertions(+), 6 deletions(-)
diff --git a/net
In colo_compare_complete, insert CompareState into net_compares
only after everything has been initialized.
In colo_compare_finalize, remove CompareState from net_compares
before anything is deinitialized.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
---
net/colo-compare.c | 45
full. This leads to a deadlock because both event loops
get blocked.
Fix this by converting compare_chr_send to a coroutine and
putting the packets in a send queue.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Tested-by: Zhang Chen
---
net/colo-compare.c
Hello Everyone,
Here are fixes for bugs that I found in my tests.
Regards,
Lukas Straub
Version changes:
v5:
-initialize and use notify_sendco only if notify_dev is set
v4:
-fix potential deadlock with notify_remote_frame
-avoid malloc and memcpy in many cases
v3:
-fix checkpatch.pl error
fix the crashes. This is safe, because
the bh already takes the appropriate locks.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Reviewed-by: Derek Su
Tested-by: Derek Su
---
net/colo-compare.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/colo-compare.c b/net
This will be needed in the next patch.
Signed-off-by: Lukas Straub
Reviewed-by: Marc-André Lureau
Reviewed-by: Zhang Chen
---
chardev/char.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/chardev/char.c b/chardev/char.c
index e77564060d..5c8014199f 100644
--- a
On Fri, 15 May 2020 11:04:13 +0100
Daniel P. Berrangé wrote:
> On Fri, May 15, 2020 at 11:48:18AM +0200, Lukas Straub wrote:
> > On Tue, 12 May 2020 09:54:58 +0100
> > Daniel P. Berrangé wrote:
> >
> > > On Mon, May 11, 2020 at 07:05:24PM +0200, Lukas Straub w
On Tue, 12 May 2020 09:54:58 +0100
Daniel P. Berrangé wrote:
> On Mon, May 11, 2020 at 07:05:24PM +0200, Lukas Straub wrote:
> > On Mon, 11 May 2020 17:19:09 +0100
> > "Dr. David Alan Gilbert" wrote:
> >
> > > * Lukas Straub (lukasstra...@web.de) wrot
aused by Hailiang's optimize patch:
Hmm, I'm on v5.0.0 (where that commit is in) and I don't have this issue in my
testing.
Regards,
Lukas Straub
> From 0393031a16735835a441b6d6e0495a1bd14adb90 Mon Sep 17 00:00:00 2001
> From: zhanghailiang
> Date: Mon, 24 Feb 2020 14:
5. The client responds with ACK to the SYN-ACK packet.
6. Because it doesn't know the connection, the secondary responds with RST,
connection reset.
Regards,
Lukas Straub
> > This fixes sporadic connection reset of client connections during failover.
> >
> > Signed-
On Wed, 13 May 2020 10:47:02 +0100
"Dr. David Alan Gilbert" wrote:
> * Lukas Straub (lukasstra...@web.de) wrote:
> > cpu_synchronize_all_pre_loadvm() marks all vcpus as dirty, so the
> > registers are loaded from CPUState before we continue running
> > the vm
; address, migration could register a yank
function to kill that external command on yank (Won't be implemented yet
though).
Regards,
Lukas Straub
pgpIjV1bZXPyk.pgp
Description: OpenPGP digital signature
On Tue, 12 May 2020 10:43:37 +0100
Daniel P. Berrangé wrote:
> On Tue, May 12, 2020 at 11:32:06AM +0200, Lukas Straub wrote:
> > ...
> >
> > Good Idea. We could name the connections (/yank callbacks) in the
> > form "nbd:", "chardev:" and &qu
ist of connection
names as you suggest and silently ignore connections that don't exist. And
maybe even add a 'query-yank' oob command returning a list of registered
connections so the management application can do pattern matching if it wants.
Comments?
Regards,
Lukas Str
On Mon, 11 May 2020 12:51:46 +0100
Daniel P. Berrangé wrote:
> On Mon, May 11, 2020 at 01:14:41PM +0200, Lukas Straub wrote:
> > Add qio_channel_set_yank function to channel and to channel-socket,
> > which will register a yank function. The yank function calls
> > shu
On Mon, 11 May 2020 13:17:14 +0100
Daniel P. Berrangé wrote:
> On Mon, May 11, 2020 at 01:07:18PM +0100, Dr. David Alan Gilbert wrote:
> > * Daniel P. Berrangé (berra...@redhat.com) wrote:
> > > On Mon, May 11, 2020 at 01:14:34PM +0200, Lukas Straub wrote:
> > > &
On Mon, 11 May 2020 12:49:47 +0100
Daniel P. Berrangé wrote:
> On Mon, May 11, 2020 at 01:14:34PM +0200, Lukas Straub wrote:
> > Hello Everyone,
> > In many cases, if qemu has a network connection (qmp, migration, chardev,
> > etc.)
> > to some other server and that
On Mon, 11 May 2020 17:19:09 +0100
"Dr. David Alan Gilbert" wrote:
> * Lukas Straub (lukasstra...@web.de) wrote:
> > Add yank option, pass it to the socket-channel and register a yank
> > function which sets s->state = NBD_CLIENT_QUIT. This is the same
> >
On Mon, 11 May 2020 10:11:20 +0100
Peter Maydell wrote:
> On Sat, 9 May 2020 at 14:18, Lukas Straub wrote:
> > Hi,
> > Why not simply add slow tag to the test. Like:
> > :avocado: tags=slow
> >
> > The slow tests can then be skipped with
> > $ make
Add high-level test relying on the colo resource-agent to test
all failover cases while checking guest network connectivity.
Signed-off-by: Lukas Straub
---
scripts/colo-resource-agent/crm_master | 44 ++
scripts/colo-resource-agent/crm_resource | 12 +
tests/acceptance/colo.py
While I'm not going to have much time for this, I'll still
try to test and review patches.
Signed-off-by: Lukas Straub
---
MAINTAINERS | 6 ++
1 file changed, 6 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 8cbc1fac2b..4c623a96e1 100644
--- a/MAINTAINERS
+++ b/M
Introduce a resource agent which can be used to manage qemu COLO
in a pacemaker cluster.
Signed-off-by: Lukas Straub
---
scripts/colo-resource-agent/colo | 1429 ++
1 file changed, 1429 insertions(+)
create mode 100755 scripts/colo-resource-agent/colo
diff --git a
Optionally install the resouce-agent so it gets picked up by
pacemaker.
Signed-off-by: Lukas Straub
---
Makefile | 5 +
configure | 10 ++
2 files changed, 15 insertions(+)
diff --git a/Makefile b/Makefile
index 8a9113e666..2ebffc4465 100644
--- a/Makefile
+++ b/Makefile
If we remove the child with the highest index from the quorum,
decrement s->next_child_index. This way we get stable children
names as long as we only remove the last child.
Signed-off-by: Lukas Straub
---
block/quorum.c | 20 ++--
1 file changed, 14 insertions(+), 6 deleti
eir own
CI?
Regards,
Lukas Straub
Lukas Straub (5):
block/quorum.c: stable children names
colo: Introduce resource agent
colo: Introduce high-level test suite
configure,Makefile: Install colo resource-agent
MAINTAINERS: Add myself as maintainer for COLO resource agent
M
Add yank option, pass it to the socket-channel and register a yank
function which sets s->state = NBD_CLIENT_QUIT. This is the same
behaviour as if an error occured.
Signed-off-by: Lukas Straub
---
Makefile.objs| 1 +
block/nbd.c |
Add qio_channel_set_yank function to channel and to channel-socket,
which will register a yank function. The yank function calls
shutdown() on the socket.
Signed-off-by: Lukas Straub
---
Makefile.objs | 1 +
include/io/channel-socket.h | 1 +
include/io/channel.h| 12
If the secondary has to failover during checkpointing, it still is
in the old state (i.e. different state than primary). Thus we can't
expose the primary state until after the checkpoint is sent.
This fixes sporadic connection reset of client connections during
failover.
Signed-off-by:
Add yank option which is passed to the socket-channel.
Signed-off-by: Lukas Straub
---
migration/channel.c | 2 ++
migration/migration.c | 11 +++
qapi/migration.json | 17 ++---
3 files changed, 27 insertions(+), 3 deletions(-)
diff --git a/migration/channel.c b
This causes the migration thread to hang if we failover during
checkpoint. A shutdown fd won't cause network traffic anyway.
Signed-off-by: Lukas Straub
---
migration/qemu-file.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/migration/qemu-file.c b/migration/qemu-f
er on colo_incoming_sem.
Fix this by relaunching failover even if there was an error. Also,
to make this work properly, set vmstate_loading to false when
returning during the vmstate_loading section.
Signed-off-by: Lukas Straub
---
migration/colo.c | 17 -
1 file changed, 12 insertions(
Add yank option which is passed to the socket-channel.
Signed-off-by: Lukas Straub
---
chardev/char-socket.c | 8
chardev/char.c| 3 +++
qapi/char.json| 5 -
3 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/chardev/char-socket.c b/chardev/char-socket.c
If we suceed in receiving ram state, but fail receiving the device
state, there will be a mismatch between the two.
Fix this by flushing the ram cache only after the vmstate has been
received.
Signed-off-by: Lukas Straub
---
migration/colo.c | 1 +
migration/ram.c | 5 +
migration/ram.h
using cpu_synchronize_all_states(), which initializes
CPUState from the current cpu registers additionally to marking
the vcpus as dirty.
Signed-off-by: Lukas Straub
---
migration/colo.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/migration/colo.c b/migration/colo.c
index
s register callbacks which get
executed with the yank command. For example the callback can shutdown() a
socket. This is intended for the colo use-case, but it can be used for other
things too of course.
Regards,
Lukas Straub
Lukas Straub (5):
Introduce yank feature
io/channel.c,io/channel-soc
The yank feature allows to recover from hanging qemu by "yanking"
at various parts. Other qemu systems can register yank functions
which will be called by the 'yank' out-of-band qmp command.
Signed-off-by: Lukas Straub
---
qapi/misc.json | 15 ++
softmmu/vl
the
checkpoint event is sent to colo-compare (which releases the
miscompared packets).
Benchmark results (iperf3):
Client-to-server tcp:
without patch: ~66 Mbit/s
with patch: ~61 Mbit/s
Server-to-client tcp:
without patch: ~702 Kbit/s
with patch: ~16 Mbit/s
Signed-off-by: Lukas Straub
Hello Everyone,
Here are fixes for bugs that I found in my tests.
Regards,
Lukas Straub
Lukas Straub (6):
migration/colo.c: Use event instead of semaphore
migration/colo.c: Use cpu_synchronize_all_states()
migration/colo.c: Flush ram cache only after receiving device state
migration
kpoint() won't touch
the job.
Signed-off-by: Lukas Straub
---
block/replication.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/block/replication.c b/block/replication.c
index da013c2041..33f2f62a44 100644
--- a/block/replication.c
+++ b/block/replication.c
@@ -398,6 +398,8 @@
; def test_virt_tcg(self):
> """
> :avocado: tags=accel:tcg
Hi,
Why not simply add slow tag to the test. Like:
:avocado: tags=slow
The slow tests can then be skipped with
$ make check-acceptance AVOCADO_TAGS='-t -slow'
Regards,
Lukas Straub
pgp6hhth052VR.pgp
Description: OpenPGP digital signature
On Fri, 8 May 2020 06:50:39 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Friday, May 8, 2020 2:11 PM
> > To: Zhang, Chen
> > Cc: qemu-devel ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lureau
On Fri, 8 May 2020 06:28:45 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Friday, May 8, 2020 2:08 PM
> > To: Zhang, Chen
> > Cc: qemu-devel ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lureau
On Fri, 8 May 2020 02:26:21 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Thursday, May 7, 2020 11:54 PM
> > To: Zhang, Chen
> > Cc: qemu-devel ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lure
performance drops in my tests:
> > Client-to-server tcp:
> > without patch: ~66 Mbit/s
> > with patch: ~59 Mbit/s
> > Server-to-client tcp:
> > without patch: ~702 Kbit/s
> > with patch: ~328 Kbit/s
>
> Oh, a big performance drop, is t
On Thu, 7 May 2020 13:26:11 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Monday, May 4, 2020 6:28 PM
> > To: qemu-devel
> > Cc: Zhang, Chen ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lureau
On Thu, 7 May 2020 11:38:04 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Monday, May 4, 2020 6:28 PM
> > To: qemu-devel
> > Cc: Zhang, Chen ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lureau
On Thu, 7 May 2020 11:00:26 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Monday, May 4, 2020 6:28 PM
> > To: qemu-devel
> > Cc: Zhang, Chen ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lur
inux, so you can use
all the tools and interfaces to manage processes for qemu as well.
Keywords: top, nice, cgroups, cpuset, numactl
Regards,
Lukas Straub
pgpJPc9LEiXQT.pgp
Description: OpenPGP digital signature
Else the log will be flooded if there is a lot of network
traffic.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
---
net/colo-compare.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/net/colo-compare.c b/net/colo-compare.c
index 2a4e7f7c4e..56db3d3bfc
iodic mode) and to use multiple colo-compare
for multiple network interfaces.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 35 +--
net/colo-compare.h | 1 +
softmmu/vl.c | 2 ++
3 files changed, 32 insertions(+), 6 deletions(-)
diff --git a/net
This will be needed in the next patch.
Signed-off-by: Lukas Straub
Reviewed-by: Marc-André Lureau
Reviewed-by: Zhang Chen
---
chardev/char.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/chardev/char.c b/chardev/char.c
index e77564060d..5c8014199f 100644
--- a
full. This leads to a deadlock because both event loops
get blocked.
Fix this by converting compare_chr_send to a coroutine and
putting the packets in a send queue.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 187 ++---
net/colo.c | 7 ++
In colo_compare_complete, insert CompareState into net_compares
only after everything has been initialized.
In colo_compare_finalize, remove CompareState from net_compares
before anything is deinitialized.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 45
Hello Everyone,
Here are fixes for bugs that I found in my tests.
Benchmark results:
Client-to-server tcp:
without patch: ~63 Mbit/s
with patch: ~66 Mbit/s
Server-to-client tcp:
without patch: ~771 Kbit/s
with patch: ~702 Kbit/s
Regards,
Lukas Straub
Version changes:
v4:
-fix potential
fix the crashes. This is safe, because
the bh already takes the appropriate locks.
Signed-off-by: Lukas Straub
Reviewed-by: Zhang Chen
Reviewed-by: Derek Su
Tested-by: Derek Su
---
net/colo-compare.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/colo-compare.c b/net
On Wed, 29 Apr 2020 05:37:17 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Monday, April 27, 2020 3:22 PM
> > To: Zhang, Chen
> > Cc: qemu-devel ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lure
On Mon, 27 Apr 2020 11:34:32 +0100
"Dr. David Alan Gilbert" wrote:
> * Lukas Straub (lukasstra...@web.de) wrote:
> > On Sat, 11 Apr 2020 19:16:54 +0200
> > Lukas Straub wrote:
> >
> > > Hello Everyone,
> > > I did some Benchmarking with iperf3
On Mon, 27 Apr 2020 03:36:57 +
"Zhang, Chen" wrote:
> > -Original Message-
> > From: Lukas Straub
> > Sent: Monday, April 27, 2020 5:19 AM
> > To: qemu-devel
> > Cc: Zhang, Chen ; Li Zhijian
> > ; Jason Wang ; Marc-
> > André Lu
Else the log will be flooded if there is a lot of network
traffic.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/net/colo-compare.c b/net/colo-compare.c
index ff6a740284..6634911770 100644
--- a/net/colo
In colo_compare_complete, insert CompareState into net_compares
only after everything has been initialized.
In colo_compare_finalize, remove CompareState from net_compares
before anything is deinitialized.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 47
full. This leads to a deadlock because both event loops
get blocked.
Fix this by converting compare_chr_send to a coroutine and
putting the packets in a send queue. Also create a new
function notify_chr_send, since that should be independend.
Signed-off-by: Lukas Straub
---
net/colo-compare.c
iodic mode) and to use multiple colo-compare
for multiple network interfaces.
Signed-off-by: Lukas Straub
---
net/colo-compare.c | 35 +--
net/colo-compare.h | 1 +
softmmu/vl.c | 2 ++
3 files changed, 32 insertions(+), 6 deletions(-)
diff --git a/net
401 - 500 of 601 matches
Mail list logo