makes it possible to test the existing chardev-udp
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Marc-André Lureau <marcandre.lur...@redhat.com>
---
tests/test-char.c | 57 ++-
1 file changed, 40 inser
qemu_chr_fe_get_driver() is unsafe, frontends with hotswap support
should not access CharDriver ptr directly as CharDriver might change.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
Reviewed-by: Marc-
So hotswap also has to be protected so write functions won't access
a backend being replaced.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
include/chardev/char-fe.h | 17 ++
include/chardev/c
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Dr. David Alan Gilbert <dgilb...@redhat.com>
---
include/chardev/char.h | 10 ++
hmp.h | 1 +
chardev/char.c
parse function will be used by the following patch
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
chardev/char.c | 81 --
1 file changed, 51 insertions(+), 30 deletions(-)
diff --git a/chardev/char.c b/chardev/char.c
for a backend change, a number of ioctls has to be replayed to sync
the current setup of a frontend to a backend tty. This is hopefully
enough so we don't have to track, store and replay the whole original
control byte sequence.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Re
Frontends should have an interface to setup the handler of a backend change.
The interface will be used in the next commits
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
Reviewed-by: Marc-André Lureau &l
frontends should avoid accessing CharDriver struct where possible
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
Reviewed-by: Marc-André Lureau <marcandre.lur...@redhat.com>
---
include/chard
connection to a live VM from VM with WinDbg.
Anton Nefedov (13):
char: move QemuOpts->ChardevBackend translation to a separate func
char: add backend hotswap handler
char: chardevice hotswap
char: forbid direct chardevice access for hotswap devices
char: avoid chardevice direct access
t
On 06/13/2017 03:32 PM, Marc-André Lureau wrote:
Hi
On Tue, Jun 13, 2017 at 3:53 PM Anton Nefedov
<anton.nefe...@virtuozzo.com <mailto:anton.nefe...@virtuozzo.com>> wrote:
> The existing chr_write_lock belongs to Chardev.
> For the hotswap case,
On 06/09/2017 05:53 PM, Marc-André Lureau wrote:
On Thu, Jun 1, 2017 at 3:23 PM Anton Nefedov
<anton.nefe...@virtuozzo.com <mailto:anton.nefe...@virtuozzo.com>> wrote:
On 05/31/2017 10:20 PM, Marc-André Lureau wrote:
> Hi
>
> On Tue, May 30,
On 06/01/2017 10:54 PM, Eric Blake wrote:
On 06/01/2017 02:49 PM, Eric Blake wrote:
On 06/01/2017 10:14 AM, Anton Nefedov wrote:
Current write_zeroes implementation is good enough to satisfy this flag too
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/file-posix
On 06/01/2017 10:50 PM, Eric Blake wrote:
On 06/01/2017 10:14 AM, Anton Nefedov wrote:
Support the flag if the underlying BDS supports it
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/blkdebug.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
Sho
On 06/01/2017 10:07 PM, Eric Blake wrote:
On 06/01/2017 10:14 AM, Anton Nefedov wrote:
The flag is supposed to indicate that the region of the disk image has
to be sufficiently allocated so it reads as zeroes. The call with the flag
set has to return -ENOTSUP if allocation cannot be done
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
tests/qemu-iotests/046 | 38 +-
tests/qemu-iotests/046.out | 23 +++
2 files changed, 60 insertions(+), 1 deletion(-)
diff --git a/tests/qemu-iotests/046 b/test
to
unallocated clusters with no backing data:
HDD fio over xfs iodepth=4:
seqwrite 4k: 18400 -> 22800 IOPS ( x1.24 )
seqwrite 68k: 1600 -> 2300 IOPS ( x1.44 )
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qco
- assert the alignment on return if the allocation has to stop
(at the start of a running allocation)
- make use of const specifiers for local variables
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qco
From: "Denis V. Lunev" <d...@openvz.org>
This could be done after calculation of the end of data and metadata in
the qcow2 image.
Signed-off-by: Denis V. Lunev <d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2-cluster.c
Avoid complicated nested conditions; return or continue asap instead.
The logic is not changed.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qcow2-cluster.c | 45 ++---
1
quential write:
seqwrite 4k, size=4g, iodepth=4
SSD : 7000 -> 18000 IOPS ( x2.6 )
- numbers are similar to qemu-io tests, slightly less improvement
(damped by fs?)
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qcow2-cluster.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index ed65961..3d
Qcow2State and BlockDriverState flags have to be in sync
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/block/qcow2.c b/block/qcow2.c
index c6fb714..b090833 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -
to commit messages (see patch 3, 12)
Anton Nefedov (11):
block: introduce BDRV_REQ_ALLOCATE flag
file-posix: support BDRV_REQ_ALLOCATE
blkdebug: support BDRV_REQ_ALLOCATE
qcow2: do not COW the empty areas
qcow2: set inactive flag
qcow2: handle_prealloc(): find out if area zeroed by earlier
<d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2-cache.c| 3 +
block/qcow2-cluster.c | 5 ++
block/qcow2-refcount.c | 14 +
block/qcow2.c | 147 +
block/qcow2.h |
;
Signed-off-by: Denis V. Lunev <d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 31 +++
tests/qemu-iotests/026.out | 104 -
tests/qemu-iotes
Support the flag if the underlying BDS supports it
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/blkdebug.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/block/blkdebug.c b/block/blkdebug.c
index a5196e8..8b1401b 100644
--- a/block/blkdebug.c
Current write_zeroes implementation is good enough to satisfy this flag too
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/file-posix.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/block/file-posix.c b/block/file-posix.c
index de2d3a2..1
The flag is supposed to indicate that the region of the disk image has
to be sufficiently allocated so it reads as zeroes. The call with the flag
set has to return -ENOTSUP if allocation cannot be done efficiently
(i.e. without falling back to writing actual buffers)
Signed-off-by: Anton Nefedov
enis V. Lunev <d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 24 +++-
1 file changed, 23 insertions(+), 1 deletion(-)
diff --git a/block/qcow2.c b/block/qcow2.c
index b3ba5da..cd5efba 100644
--- a/block/qcow2.c
++
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qcow2-cluster.c | 2 ++
block/qcow2.c | 8 +++-
block/qcow2.h | 4
3 files changed, 13 insertions(+), 1 deletion(-)
diff --git a/block/qcow2
On 06/01/2017 10:12 AM, Marc-André Lureau wrote:
Hi
On Tue, May 30, 2017 at 6:08 PM Anton Nefedov
<anton.nefe...@virtuozzo.com <mailto:anton.nefe...@virtuozzo.com>> wrote:
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com
<mailto:anton.nefe...@virtuozzo.
On 05/31/2017 10:20 PM, Marc-André Lureau wrote:
Hi
On Tue, May 30, 2017 at 6:12 PM Anton Nefedov
<anton.nefe...@virtuozzo.com <mailto:anton.nefe...@virtuozzo.com>> wrote:
This patch adds a possibility to change a char device without a frontend
removal.
1. Ideal
On 05/26/2017 11:11 AM, Kevin Wolf wrote:
Am 19.05.2017 um 11:34 hat Anton Nefedov geschrieben:
Qcow2State and BlockDriverState flags have to be in sync
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 1 +
1 file changed, 1 insertion(+)
diff --git a
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
tests/test-char.c | 68 +++
1 file changed, 68 insertions(+)
diff --git a/tests/test-char.c b/tests/test-char.c
index ed6b18f..cd54f88 100644
--- a/tests/test-char.c
+++ b
makes it possible to test the existing chardev-file
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
tests/test-char.c | 127 +-
1 file changed, 79 insertions(+), 48 deletions(-)
diff --git a/tests/test-char.c b/tests/test-
qemu_chr_fe_get_driver() is unsafe, frontends with hotswap support
should not access CharDriver ptr directly as CharDriver might change.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c
So hotswap also has to be protected so write functions won't access
a backend being replaced.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c| 126 +++
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Amit Shah <a...@kernel.org>
---
hw/char/virtio-console.c | 35 ++-
1 file changed, 30 insertions(+), 5 deletions(-
parse function will be used by the following patch
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c | 70 --
1 file changed, 4
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Dr. David Alan Gilbert <dgilb...@redhat.com>
---
chardev/char.c| 2 +-
hmp-commands.hx | 16
hmp.c
a frontend device removal.
Yet isa-serial and virtio-serial frontends are supported.
The feature can be helpful for e.g. Windows debug allowing to
establish connection to a live VM from VM with WinDbg.
Anton Nefedov (13):
char: move QemuOpts->ChardevBackend translation to a separate func
char:
this is only not a problem if the test is last in a suite,
otherwise it makes the following main_loop() calls to fail
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
tests/test-char.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tests/test-char.c b/tests/test-char.c
for a backend change, a number of ioctls has to be replayed to sync
the current setup of a frontend to a backend tty. This is hopefully
enough so we don't have to track, store and replay the whole original
control byte sequence.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Re
makes it possible to test the existing chardev-udp
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
tests/test-char.c | 58 +++
1 file changed, 41 insertions(+), 17 deletions(-)
diff --git a/tests/test-char.c b/tests/test-
will be used by the following patch
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Michael S. Tsirkin <m...@redhat.com>
CC: Paolo Bonzini <pbonz...@redhat.com>
Reviewed-by: Marc-And
frontends should avoid accessing CharDriver struct where possible
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c | 5 +
gdbstub.c | 2 +-
hw/arm
Frontends should have an interface to setup the handler of a backend change.
The interface will be used in the next commits
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
Reviewed-by: Marc-André Lureau &l
On 05/26/2017 04:08 PM, Alberto Garcia wrote:
On Fri 26 May 2017 02:47:55 PM CEST, Anton Nefedov wrote:
Tried the another machine; about 10% improvement here
[...]
[root@localhost ~]# fio --name=randwrite --blocksize=4k
--filename=/dev/sdb --rw=randwrite --direct=1 --ioengine=libaio
--size
On 05/26/2017 01:17 PM, Kevin Wolf wrote:
Am 24.05.2017 um 18:09 hat Anton Nefedov geschrieben:
I agree; as mentioned we have similar patches and they don't conflict much.
We noticed a performance regression on HDD though, for the
presumably optimized case (random 4k write over a large backed
On 05/26/2017 11:57 AM, Denis V. Lunev wrote:
On 05/26/2017 11:11 AM, Kevin Wolf wrote:
Am 19.05.2017 um 11:34 hat Anton Nefedov geschrieben:
From: "Denis V. Lunev" <d...@openvz.org>
Currently each single write operation can result in 3 write operations
if guest offsets
Hi
Hi Marc-André, thanks for reviewing, pls see answers below
On Fri, May 19, 2017 at 4:51 PM Anton Nefedov <address@hidden>
wrote:
This patch adds a possibility to change a char device without a frontend
removal.
1. Ideally, it would have to happen transparently to a fronten
On 05/22/2017 10:29 PM, Eric Blake wrote:
On 05/19/2017 04:34 AM, Anton Nefedov wrote:
From: "Denis V. Lunev" <d...@openvz.org>
This patch adds image preallocation at expand to provide better locality
of QCOW2 image file and optimize this procedure for some distribute
If the guest tries to write data that results on the allocation of a
new cluster, instead of writing the guest data first and then the data
from the COW regions, write everything together using one single I/O
operation.
This can improve the write performance by 25% or more, depending on
several
On 05/24/2017 07:20 PM, Alberto Garcia wrote:
On Wed 24 May 2017 06:09:42 PM CEST, Anton Nefedov wrote:
I agree; as mentioned we have similar patches and they don't conflict
much. We noticed a performance regression on HDD though, for the
presumably optimized case (random 4k write over
On 05/24/2017 05:20 PM, Alberto Garcia wrote:
On Tue 23 May 2017 04:36:52 PM CEST, Eric Blake wrote:
here's a patch series that rewrites the copy-on-write code in the
qcow2 driver to reduce the number of I/O operations.
And it competes with Denis and Anton's patches:
On 05/22/2017 10:14 PM, Eric Blake wrote:
On 05/22/2017 02:12 PM, Eric Blake wrote:
+++ b/block/qcow2.c
@@ -2482,7 +2482,7 @@ static bool is_zero_sectors(BlockDriverState *bs, int64_t
start,
int64_t res;
if (start + count > bs->total_sectors) {
-count = bs->total_sectors
On 05/22/2017 10:24 PM, Eric Blake wrote:
On 05/19/2017 04:34 AM, Anton Nefedov wrote:
If COW area of the newly allocated cluster is zeroes, there is no reason
to write zero sectors in perform_cow() again now as whole clusters are
zeroed out in single chunks by handle_alloc_space
On 05/22/2017 10:00 PM, Eric Blake wrote:
On 05/19/2017 04:34 AM, Anton Nefedov wrote:
From: "Denis V. Lunev" <d...@openvz.org>
Currently each single write operation can result in 3 write operations
if guest offsets are not cluster aligned. One write is performed for the
real
will be used by the following patch
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Michael S. Tsirkin <m...@redhat.com>
CC: Paolo Bonzini <pbonz...@redhat.com>
--
frontends should avoid accessing CharDriver struct where possible
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c | 5 +
gdbstub.c | 2 +-
hw/arm
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Dr. David Alan Gilbert <dgilb...@redhat.com>
---
chardev/char.c| 4 ++--
hmp-commands.hx | 16
hmp.c
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
CC: Amit Shah <a...@kernel.org>
---
hw/char/virtio-console.c | 35 ++-
1 file changed, 30 insertions(+), 5 deletions(-
Frontends should have an interface to setup the handler of a backend change.
The interface will be used in the next commits
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
backends/rng-egd.c
me of the errors
synchronously - but most of them we can)
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c| 147 ++
include/sysemu/char.h
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c | 72 +-
1 file changed, 46 insertions(+), 26 deletions(-)
diff --git a/char
for a backend change, a number of ioctls has to be replayed to sync
the current setup of a frontend to a backend tty. This is hopefully
enough so we don't have to track, store and replay the whole original
control byte sequence.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Re
a frontend device removal.
Yet isa-serial and virtio-serial frontends are supported.
The feature can be helpful for e.g. Windows debug allowing to
establish connection to a live VM from VM with WinDbg.
Anton Nefedov (9):
char: move QemuOpts->ChardevBackend translation to a separate func
c
qemu_chr_fe_get_driver() is unsafe, frontends with hotswap support
should not access CharDriver ptr directly as CharDriver might change.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
chardev/char.c
- assert the alignment on return if the allocation has to stop
(at the start of a running allocation)
- make use of const specifiers for local variables
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qco
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qcow2-cluster.c | 2 ++
block/qcow2.c | 8 +++-
block/qcow2.h | 4
3 files changed, 13 insertions(+), 1 deletion(-)
diff --git a/block/qcow2
From: "Denis V. Lunev" <d...@openvz.org>
This could be done after calculation of the end of data and metadata in
the qcow2 image.
Signed-off-by: Denis V. Lunev <d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2-cluster.c
to
unallocated clusters with no backing data:
HDD fio over xfs iodepth=4:
seqwrite 4k: 18400 -> 22800 IOPS ( x1.24 )
seqwrite 68k: 1600 -> 2300 IOPS ( x1.44 )
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qco
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
tests/qemu-iotests/046 | 38 +-
tests/qemu-iotests/046.out | 23 +++
2 files changed, 60 insertions(+), 1 deletion(-)
diff --git a/tests/qemu-iotests/046 b/test
in such case, bdrv_get_block_status() shall return 0, *nr == 0
iotest 154 updated accordingly: write-zeroes tail alignment can be detected
as zeroes now, so pwrite_zeroes succeeds
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 6 --
test
Avoid complicated nested conditions; return or continue asap instead.
The logic is not changed.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qcow2-cluster.c | 45 ++---
1
Qcow2State and BlockDriverState flags have to be in sync
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/block/qcow2.c b/block/qcow2.c
index 6e7ce96..07c1706 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -
;
Signed-off-by: Denis V. Lunev <d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 31 +++
tests/qemu-iotests/026.out | 104 -
tests/qemu-iotes
enis V. Lunev <d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2.c | 32 +++-
1 file changed, 31 insertions(+), 1 deletion(-)
diff --git a/block/qcow2.c b/block/qcow2.c
index a8d61f0..2e6a0ec 100644
--- a/block
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d...@openvz.org>
---
block/qcow2-cluster.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index 25210cd..42
<d...@openvz.org>
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
block/qcow2-cache.c| 3 +
block/qcow2-cluster.c | 5 ++
block/qcow2-refcount.c | 14 +
block/qcow2.c | 151 +
block/qcow2.h |
quential write:
seqwrite 4k, size=4g, iodepth=4
SSD : 7000 -> 18000 IOPS ( x2.6 )
- numbers are similar to qemu-io tests, slightly less improvement
(damped by fs?)
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
Signed-off-by: Denis V. Lunev <d
to commit messages (see patch 3, 12)
Anton Nefedov (9):
qcow2: is_zero_sectors(): return true if area is outside of backing
file
qcow2: do not COW the empty areas
qcow2: set inactive flag
qcow2: handle_prealloc(): find out if area zeroed by earlier
preallocation
qcow2: fix misleading
On error path (like i/o error in one of the coroutines), it's required to
- wait for coroutines completion before cleaning the common structures
- reenter dependent coroutines so they ever finish
Introduced in 2d9187bc65.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
On 04/25/2017 12:55 PM, Peter Lieven wrote:
Am 24.04.2017 um 22:13 schrieb Anton Nefedov:
On 24/04/2017 21:16, Peter Lieven wrote:
Am 24.04.2017 um 18:27 schrieb Anton Nefedov
<anton.nefe...@virtuozzo.com>:
On 04/21/2017 03:37 PM, Peter Lieven wrote:
Am 21.04.2017 um 14:19 s
On 24/04/2017 21:16, Peter Lieven wrote:
Am 24.04.2017 um 18:27 schrieb Anton Nefedov <anton.nefe...@virtuozzo.com>:
On 04/21/2017 03:37 PM, Peter Lieven wrote:
Am 21.04.2017 um 14:19 schrieb Anton Nefedov:
On 04/21/2017 01:44 PM, Peter Lieven wrote:
Am 21.04.2017 um 12:04 schrieb
On 04/21/2017 03:37 PM, Peter Lieven wrote:
Am 21.04.2017 um 14:19 schrieb Anton Nefedov:
On 04/21/2017 01:44 PM, Peter Lieven wrote:
Am 21.04.2017 um 12:04 schrieb Anton Nefedov:
On error path (like i/o error in one of the coroutines), it's required to
- wait for coroutines completion
On error path (like i/o error in one of the coroutines), it's required to
- wait for coroutines completion before cleaning the common structures
- reenter dependent coroutines so they ever finish
Introduced in 2d9187bc65.
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
On 04/21/2017 01:44 PM, Peter Lieven wrote:
Am 21.04.2017 um 12:04 schrieb Anton Nefedov:
On error path (like i/o error in one of the coroutines), it's required to
- wait for coroutines completion before cleaning the common structures
- reenter dependent coroutines so they ever finish
On 04/21/2017 12:18 PM, Peter Lieven wrote:
Am 18.04.2017 um 12:27 schrieb Denis V. Lunev:
From: Anton Nefedov <anton.nefe...@virtuozzo.com>
We should wait for other coroutines on error path, i.e. one of coroutines
terminates with i/o error, before cleaning the common structures. In the
On 02/28/2017 03:07 PM, Daniel P. Berrange wrote:
The current websockets protocol handshake code is very relaxed, just
doing crude string searching across the HTTP header data. This causes
it to both reject valid connections and fail to reject invalid
connections. For example, according to the
hi,
On 02/21/2017 01:42 PM, Paolo Bonzini wrote:
qemu_system_guest_panicked was already using current_cpu implicitly,
so it makes sense for it to receive a CPUState. This lets the
function call cpu_get_crash_info and free the result.
my motivation against this was that as
This patch is a follow-up for:
http://lists.nongnu.org/archive/html/qemu-devel/2017-02/msg03647.html
also remove a useless NULL check in the event reporting code
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
qapi/event.json | 4 ++--
vl.c| 22 ++
2 files changed, 12 insertions(+), 14 deletions(-)
diff --git a/qapi/event.json b/qapi/even
Signed-off-by: Anton Nefedov <anton.nefe...@virtuozzo.com>
---
qapi/event.json | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/qapi/event.json b/qapi/event.json
index 970ff02..e02852c 100644
--- a/qapi/event.json
+++ b/qapi/event.json
@@ -488,9 +488,9 @@
#
# @
On 01/30/2017 08:33 PM, Denis V. Lunev wrote:
On 01/30/2017 08:16 PM, Eric Blake wrote:
On 01/30/2017 04:22 AM, Denis V. Lunev wrote:
If explicit zeroing out before mirroring is required for the target image,
it moves the block job offset counter to EOF, then offset and len counters
count the
401 - 495 of 495 matches
Mail list logo
