On 2/3/21 1:15 AM, Eric Blake wrote:
> On 1/28/21 11:21 AM, Vladimir Sementsov-Ogievskiy wrote:
>> 28.01.2021 20:13, Denis V. Lunev wrote:
>>> Original specification says that l1 table size if 64 * l1_size, which
>>> is obviously wrong. The size of the l1 entry is 64 _bi
with the later text.
Signed-off-by: Denis V. Lunev
CC: Stefan Hajnoczi
CC: Vladimir Sementsov-Ogievskiy
---
docs/interop/parallels.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/docs/interop/parallels.txt b/docs/interop/parallels.txt
index e9271eba5d..f15bf35bd1 100644
edundant after Denis's 0b46160 - but
> I don't know the snapshot code much;
>
> Denis - do you agree this is correct?
For me it looks too that the code becomes redundant, thus
Reviewed-by: Denis V. Lunev
> Dave
>
>> ---
>> migration/savevm.c | 10 ++
>> 1
On 9/17/20 3:41 PM, Stefan Hajnoczi wrote:
> On Wed, Sep 16, 2020 at 5:43 PM Denis V. Lunev wrote:
>> On 9/16/20 5:07 PM, Denis V. Lunev wrote:
>>> I will make a check today.
>>>
>>> Talking about our performance measurements, we have not
>>> seen
On 9/17/20 4:56 PM, Max Reitz wrote:
> On 10.08.20 12:14, Denis V. Lunev wrote:
>> There are severe delays with IO requests processing if QEMU is running in
>> virtual machine or over software defined storage. Such delays potentially
>> results in unpredictable guest behavior
On 9/16/20 5:07 PM, Denis V. Lunev wrote:
> On 9/16/20 4:32 PM, Stefan Hajnoczi wrote:
>> On Thu, Aug 27, 2020 at 3:24 PM Stefan Hajnoczi wrote:
>>> Hi Denis,
>>> A performance regression was found after the virtio-blk queue-size
>>> property was increased f
On 9/16/20 4:32 PM, Stefan Hajnoczi wrote:
> On Thu, Aug 27, 2020 at 3:24 PM Stefan Hajnoczi wrote:
>> Hi Denis,
>> A performance regression was found after the virtio-blk queue-size
>> property was increased from 128 to 256 in QEMU 5.0 in commit
>> c9b7d9ec21dfca716f0bb3b68dee75660d86629c
On 8/27/20 3:58 PM, Daniel P. Berrangé wrote:
> On Thu, Jul 09, 2020 at 04:26:42PM +0300, Denis V. Lunev wrote:
>> Right now bdrv_fclose() is just calling bdrv_flush().
>>
>> The problem is that migration code is working inefficiently from block
>> layer terms and are
On 8/20/20 10:42 AM, Denis V. Lunev wrote:
> On 7/9/20 4:26 PM, Denis V. Lunev wrote:
>> This series do standard basic things:
>> - it creates intermediate buffer for all writes from QEMU migration code
>> to QCOW2 image,
>> - this buffer is sent to disk asynchronous
On 8/20/20 11:03 AM, David Edmondson wrote:
> On Monday, 2020-08-10 at 13:14:46 +03, Denis V. Lunev wrote:
>
>> +strftime(buf, sizeof(buf), "%m-%d %H:%M:%S",
> "%F %T" would include the year, which can be useful.
ok
>
>> + localtime_
On 7/9/20 4:26 PM, Denis V. Lunev wrote:
> This series do standard basic things:
> - it creates intermediate buffer for all writes from QEMU migration code
> to QCOW2 image,
> - this buffer is sent to disk asynchronously, allowing several writes to
> run in parallel.
>
>
On 8/12/20 5:00 PM, Stefan Hajnoczi wrote:
> On Mon, Aug 10, 2020 at 01:14:44PM +0300, Denis V. Lunev wrote:
>> There are severe delays with IO requests processing if QEMU is running in
>> virtual machine or over software defined storage. Such delays potentially
>> results i
On 8/10/20 5:40 PM, Kevin Wolf wrote:
> Am 10.08.2020 um 16:02 hat Vladimir Sementsov-Ogievskiy geschrieben:
>> 10.08.2020 15:35, Denis V. Lunev wrote:
>>> The problem this patch is trying to address is libguestfs behavior on the
>>> appliance startup. It starts
On 8/10/20 3:52 PM, Richard W.M. Jones wrote:
> On Mon, Aug 10, 2020 at 03:35:55PM +0300, Denis V. Lunev wrote:
>> The problem this patch is trying to address is libguestfs behavior on the
>> appliance startup. It starts supporting to use root=UUID definition in
>> the kerne
much more intrusive.
Signed-off-by: Denis V. Lunev
CC: Andrey Shinkevich
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Richard W.M. Jones
---
P.S. Patch to libguestfs will follow.
qemu-io-cmds.c | 17 +
1 file changed, 13 insertions(+), 4 deletions(-)
than 10 seconds.
Such reports are very complex to process. Some good starting point for this
seems quite reasonable. This patch provides one. It adds logging of such
potentially dangerous long IO operations.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Stefan
Latency threshold is set to 10 seconds following guest request timeout
on legacy storage controller.
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Stefan Hajnoczi
CC: Kevin Wolf
CC: Max Reitz
---
blockdev.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion
timeout selection to patch 3
Changes from v1:
- fixed conversions using macros suggested by Stefan
- fixed option declaration
- enabled by default with patch 3
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
Right now BlockAcctStats is always reside on BlockBackend. This structure
is not used in any other place. Thus we are able to create a converter
from one pointer to another.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Stefan Hajnoczi
CC: Kevin Wolf
CC: Max
On 8/5/20 4:51 PM, Philippe Mathieu-Daudé wrote:
> On 8/5/20 12:08 PM, Denis V. Lunev wrote:
>> There are severe delays with IO requests processing if QEMU is running in
>> virtual machine or over software defined storage. Such delays potentially
>> results in unpredi
Right now BlockAcctStats is always reside on BlockBackend. This structure
is not used in any other place. Thus we are able to create a converter
from one pointer to another.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Stefan Hajnoczi
CC: Kevin Wolf
CC: Max
Latency threshold is set to 10 seconds following guest request timeout
on legacy storage controller.
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Stefan Hajnoczi
CC: Kevin Wolf
CC: Max Reitz
---
blockdev.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff
than 10 seconds.
Such reports are very complex to process. Some good starting point for this
seems quite reasonable. This patch provides one. It adds logging of such
potentially dangerous long IO operations.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Stefan
declaration
- enabled by default with patch 3
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
On 7/31/20 1:25 PM, Stefan Hajnoczi wrote:
> On Fri, Jul 10, 2020 at 08:27:09PM +0300, Denis V. Lunev wrote:
>> There are severe delays with IO requests processing if QEMU is running in
>> virtual machine or over software defined storage. Such delays potentially
>> results i
On 7/29/20 1:30 PM, Stefan Hajnoczi wrote:
> On Fri, Jul 10, 2020 at 08:27:11PM +0300, Denis V. Lunev wrote:
>> +static void block_acct_report_long(BlockAcctStats *stats,
>> + BlockAcctCookie *cookie, int64_t
>> latency_ns)
>> +{
than 10 seconds.
Such reports are very complex to process. Some good starting point for this
seems quite reasonable. This patch provides one. It adds logging of such
potentially dangerous long IO operations.
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max
than 10 seconds.
Such reports are very complex to process. Some good starting point for this
seems quite reasonable. This patch provides one. It adds logging of such
potentially dangerous long IO operations.
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max
Right now BlockAcctStats is always reside on BlockBackend. This structure
is not used in any other place. Thus we are able to create a converter
from one pointer to another.
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
---
block/block-backend.c
):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
Fedora 31 VM
over NVME storage are the following:
original fixed
cached: 1.84s 1.16s
non-cached: 12.74s 1.27s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max
gested with waiting fixes from me
Changes from v1:
- patchew warning fixed
- fixed validation that only 1 waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sem
-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
bdrv_save_vmstate() returns either error with negative return value or
size. Thus this check is useless.
Signed-off-by: Denis V. Lunev
Suggested-by: Eric Blake
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC
It is not used outside the module.
Actually there are 2 kind of waiters:
- for a slot and
- for all tasks to finish
This patch limits external API to listed types.
Signed-off-by: Denis V. Lunev
Suggested-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
is used from
qemu-io only.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Denis Plotnikov
---
block/block-backend.c | 6 +-
block/io.c
On 7/9/20 11:29 AM, Daniel P. Berrangé wrote:
> On Wed, Jul 08, 2020 at 10:15:39PM +0300, Vladimir Sementsov-Ogievskiy wrote:
>> Keep-alive won't hurt, let's try to enable it even if not requested by
>> user.
> Keep-alive intentionally breaks TCP connections earlier than normal
> in face of
On 7/6/20 7:03 PM, Daniel P. Berrangé wrote:
> On Mon, Jul 06, 2020 at 05:50:11PM +0200, Kevin Wolf wrote:
>> Am 06.07.2020 um 17:29 hat Daniel P. Berrangé geschrieben:
>>> On Mon, Jul 06, 2020 at 05:27:01PM +0200, Kevin Wolf wrote:
>>>> Am 03.07.2020 um 19:29
On 7/6/20 5:28 PM, Daniel P. Berrangé wrote:
> On Fri, Jul 03, 2020 at 08:29:08PM +0300, Denis V. Lunev wrote:
>> On 7/3/20 8:22 PM, Daniel P. Berrangé wrote:
>>> On Fri, Jul 03, 2020 at 08:15:44PM +0300, Denis V. Lunev wrote:
>>>> On 7/2/20 8:57 PM, Daniel P.
waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
bdrv_save_vmstate() returns either error with negative return value or
size. Thus this check is useless.
Signed-off-by: Denis V. Lunev
Suggested-by: Eric Blake
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC
Fedora 31 VM
over NVME storage are the following:
original fixed
cached: 1.84s 1.16s
non-cached: 12.74s 1.27s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max
On 6/11/20 8:16 PM, Denis V. Lunev wrote:
> Nowaday SCSI drivers in guests are able to align UNMAP requests before
> sending to the device. Right now QEMU provides an ability to set
> this via "discard_granularity" property of the block device which could
> be used by manage
-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
qemu_fclose() could return error, f.e. if bdrv_co_flush() will return
the error.
This validation will become more important once we will start waiting of
asynchronous IO operations, started from bdrv_write_vmstate(), which are
coming soon.
Signed-off-by: Denis V. Lunev
Reviewed-by: "Dr.
is used from
qemu-io only.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Denis Plotnikov
---
block/block-backend.c | 6 +-
block/io.c
It is not used outside the module.
Actually there are 2 kind of waiters:
- for a slot and
- for all tasks to finish
This patch limits external API to listed types.
Signed-off-by: Denis V. Lunev
Suggested-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
On 7/3/20 8:22 PM, Daniel P. Berrangé wrote:
> On Fri, Jul 03, 2020 at 08:15:44PM +0300, Denis V. Lunev wrote:
>> On 7/2/20 8:57 PM, Daniel P. Berrangé wrote:
>>> When QMP was first introduced some 10+ years ago now, the snapshot
>>> related commands (savevm/loadv
On 7/2/20 8:57 PM, Daniel P. Berrangé wrote:
> savevm, loadvm and delvm are some of the few commands that have never
> been converted to use QMP. The primary reason for this lack of
> conversion is that they block execution of the thread for as long as
> they run.
>
> Despite this downside,
On 7/2/20 8:57 PM, Daniel P. Berrangé wrote:
> When QMP was first introduced some 10+ years ago now, the snapshot
> related commands (savevm/loadvm/delvm) were not converted. This was
> primarily because their implementation causes blocking of the thread
> running the monitor commands. This was
is used from
qemu-io only.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Denis Plotnikov
---
block/block-backend.c | 6 +-
block/io.c
-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
It is not used outside the module.
Actually there are 2 kind of waiters:
- for a slot and
- for all tasks to finish
This patch limits external API to listed types.
Signed-off-by: Denis V. Lunev
Suggested-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
Fedora 31 VM
over NVME storage are the following:
original fixed
cached: 1.84s 1.16s
non-cached: 12.74s 1.27s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max
bdrv_save_vmstate() returns either error with negative return value or
size. Thus this check is useless.
Signed-off-by: Denis V. Lunev
Suggested-by: Eric Blake
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC
qemu_fclose() could return error, f.e. if bdrv_co_flush() will return
the error.
This validation will become more important once we will start waiting of
asynchronous IO operations, started from bdrv_write_vmstate(), which are
coming soon.
Signed-off-by: Denis V. Lunev
Reviewed-by: "Dr.
t;file in bdrv_co_flush_vmstate (fix 267)
- fixed blk_save_vmstate helper
- fixed coroutine wait as Vladimir suggested with waiting fixes from me
Changes from v1:
- patchew warning fixed
- fixed validation that only 1 waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max
On 6/11/20 8:16 PM, Denis V. Lunev wrote:
> Nowaday SCSI drivers in guests are able to align UNMAP requests before
> sending to the device. Right now QEMU provides an ability to set
> this via "discard_granularity" property of the block device which could
> be used by manage
On 6/22/20 8:02 PM, Vladimir Sementsov-Ogievskiy wrote:
> 22.06.2020 11:33, Denis V. Lunev wrote:
>> This patch creates intermediate buffer for reading from block driver
>> state and performs read-ahead to this buffer. Snapshot code performs
>> reads sequentially and thus
On 6/19/20 2:02 PM, no-re...@patchew.org wrote:
> Patchew URL: https://patchew.org/QEMU/20200619100708.30440-1-...@openvz.org/
>
>
>
> Hi,
>
> This series failed the asan build test. Please find the testing commands and
> their output below. If you have Docker installed, you can probably reproduce
Fedora 31 VM
over NVME storage are the following:
original fixed
cached: 1.84s 1.16s
non-cached: 12.74s 1.27s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max
On 6/19/20 7:20 PM, Eduardo Habkost wrote:
> On Thu, Jun 11, 2020 at 08:16:06PM +0300, Denis V. Lunev wrote:
>> Nowaday SCSI drivers in guests are able to align UNMAP requests before
>> sending to the device. Right now QEMU provides an ability to set
>> this via "dis
):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
fixed coroutine wait as Vladimir suggested with waiting fixes from me
Changes from v1:
- patchew warning fixed
- fixed validation that only 1 waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr.
-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
bdrv_save_vmstate() returns either error with negative return value or
size. Thus this check is useless.
Signed-off-by: Denis V. Lunev
Suggested-by: Eric Blake
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC
It is not used outside the module.
Actually there are 2 kind of waiters:
- for a slot and
- for all tasks to finish
This patch limits external API to listed types.
Signed-off-by: Denis V. Lunev
Suggested-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
is used from
qemu-io only.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Denis Plotnikov
---
block/block-backend.c | 6 +-
block/io.c
qemu_fclose() could return error, f.e. if bdrv_co_flush() will return
the error.
This validation will become more important once we will start waiting of
asynchronous IO operations, started from bdrv_write_vmstate(), which are
coming soon.
Signed-off-by: Denis V. Lunev
Reviewed-by: "Dr.
On 6/11/20 8:21 PM, Denis V. Lunev wrote:
> On 6/11/20 8:16 PM, Denis V. Lunev wrote:
>> Nowaday SCSI drivers in guests are able to align UNMAP requests before
>> sending to the device. Right now QEMU provides an ability to set
>> this via "discard_granularity" pro
On 6/18/20 1:56 PM, Vladimir Sementsov-Ogievskiy wrote:
> 16.06.2020 19:20, Denis V. Lunev wrote:
>> This patch does 2 standard basic things:
>> - it creates intermediate buffer for all writes from QEMU migration code
>> to block driver,
>> - this buffer is sent to di
On 6/17/20 12:11 AM, Eric Blake wrote:
> On 6/16/20 11:20 AM, Denis V. Lunev wrote:
>> Right now bdrv_fclose() is just calling bdrv_flush().
>>
>> The problem is that migration code is working inefficently from black
>
> inefficiently, block
>
>> layer term
is used from
qemu-io only.
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/block-backend.c | 6 +-
block/io.c
ile in bdrv_co_flush_vmstate (fix 267)
- fixed blk_save_vmstate helper
- fixed coroutine wait as Vladimir suggested with waiting fixes from me
Changes from v1:
- patchew warning fixed
- fixed validation that only 1 waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Ste
It is not used outside the module.
Actually there are 2 kind of waiters:
- for a slot and
- for all tasks to finish
This patch limits external API to listed types.
Signed-off-by: Denis V. Lunev
Suggested-by: Vladimir Sementsov-Ogievskiy
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC
qemu_fclose() could return error, f.e. if bdrv_co_flush() will return
the error.
This validation will become more important once we will start waiting of
asynchronous IO operations, started from bdrv_write_vmstate(), which are
coming soon.
Signed-off-by: Denis V. Lunev
Reviewed-by: "Dr.
):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gi
-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
On 6/15/20 12:34 PM, Vladimir Sementsov-Ogievskiy wrote:
> 15.06.2020 10:47, Vladimir Sementsov-Ogievskiy wrote:
>> 11.06.2020 20:11, Denis V. Lunev wrote:
>>> From: Vladimir Sementsov-Ogievskiy
>>>
>>> Currently, aio task pool assumes that there is a mai
On 6/15/20 3:17 PM, Dr. David Alan Gilbert wrote:
> * Denis V. Lunev (d...@openvz.org) wrote:
>> This series do standard basic things:
>> - it creates intermediate buffer for all writes from QEMU migration code
>> to QCOW2 image,
>> - this buffer is sent to disk async
On 6/11/20 8:16 PM, Denis V. Lunev wrote:
> Nowaday SCSI drivers in guests are able to align UNMAP requests before
> sending to the device. Right now QEMU provides an ability to set
> this via "discard_granularity" property of the block device which could
> be used by manage
Default discard granularity is set to -2 now.
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Eduardo Habkost
CC: Marcel Apfelbaum
CC: John Snow
CC: Paolo Bonzini
CC: Fam Zheng
---
tests/qemu-iotests/172.out | 106 ++---
1 file changed, 53
nt of QEMU, there is
pdiscard_granularity on the format driver level, f.e. on QCOW2 or iSCSI.
It would be beneficial to pass this value as a default for this
property.
Technically this should reduce the amount of use less UNMAP requests
from the guest to the host. Basic test confirms this.
Signed-off-by: Deni
ng 'fstrim /' on 32 Gb disk has issued 401/415 requests with/without
proper alignment to QEMU.
Changes from v2:
- 172 iotest fixed
Changes from v1:
- fixed typos in description
- added machine type compatibility layer as suggested by Kevin
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max
is used from
qemu-io only.
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/block-backend.c | 6 +-
block/io.c
ate helper
- fixed coroutine wait as Vladimir suggested with waiting fixes from me
Changes from v1:
- patchew warning fixed
- fixed validation that only 1 waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gilbert"
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 21 ++---
1 file changed, 6 insertions(+), 15 deletions(-)
):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Stefan Hajnoczi
CC: Fam Zheng
CC: Juan Quintela
CC: "Dr. David Alan Gi
qemu_fclose() could return error, f.e. if bdrv_co_flush() will return
the error.
This validation will become more important once we will start waiting of
asynchronous IO operations, started from bdrv_write_vmstate(), which are
coming soon.
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max
On 6/11/20 10:36 AM, Vladimir Sementsov-Ogievskiy wrote:
> Currently, aio task pool assumes that there is a main coroutine, which
> creates tasks and wait for them. Let's remove the restriction by using
> CoQueue. Code becomes clearer, interface more obvious.
>
> Signed-off-by: Vladimir
On 6/11/20 11:04 AM, Vladimir Sementsov-Ogievskiy wrote:
> 10.06.2020 22:00, Denis V. Lunev wrote:
>> This patch does 2 standard basic things:
>> - it creates intermediate buffer for all writes from QEMU migration code
>> to QCOW2 image,
>> - this buffer is sent to di
On 6/10/20 9:58 PM, Denis V. Lunev wrote:
> This patch does 2 standard basic things:
> - it creates intermediate buffer for all writes from QEMU migration code
> to QCOW2 image,
> - this buffer is sent to disk asynchronously, allowing several writes to
> run in parallel.
: 1.79s 2.38s 1.27s
non-cached: 3.29s 1.31s 0.81s
Changes from v1:
- patchew warning fixed
- fixed validation that only 1 waiter is allowed in patch 1
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
ake_co != NULL'.
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 17 -
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/block/aio_task.c b/block/aio_task.c
index 88989fa248..5183
On 6/10/20 9:58 PM, Denis V. Lunev wrote:
> The patch preserves the constraint that the only waiter is allowed.
>
> The patch renames AioTaskPool->main_co to wake_co and removes
> AioTaskPool->waiting flag. wake_co keeps coroutine, which is
> waiting for wakeup on worker comp
writes to such new clusters.
Snapshot creation time (2 GB Fedora-31 VM running over NVME storage):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Kevin
ake_co != NULL'.
Signed-off-by: Denis V. Lunev
CC: Kevin Wolf
CC: Max Reitz
CC: Vladimir Sementsov-Ogievskiy
CC: Denis Plotnikov
---
block/aio_task.c | 17 -
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/block/aio_task.c b/block/aio_task.c
index 88989fa248..5183
writes to such new clusters.
Snapshot creation time (2 GB Fedora-31 VM running over NVME storage):
original fixed
cached: 1.79s 1.27s
non-cached: 3.29s 0.81s
The difference over HDD would be more significant :)
Signed-off-by: Denis V. Lunev
CC: Kevin
401 - 500 of 2500 matches
Mail list logo
