On 21/05/2024 14.46, Fabiano Rosas wrote:
Alex Bennée writes:
Juan Quintela writes:
From: Fabiano Rosas
Add a smoke test that migrates to a file and gives it to the
script. It should catch the most annoying errors such as changes in
the ram flags.
After code has been merged it becomes
On Fri, May 17, 2024 at 03:01:59PM +0200, Yu Zhang wrote:
> Hello Michael and Peter,
Hi,
>
> Exactly, not so compelling, as I did it first only on servers widely
> used for production in our data center. The network adapters are
>
> Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme
Alex Bennée writes:
> Juan Quintela writes:
>
>> From: Fabiano Rosas
>>
>> Add a smoke test that migrates to a file and gives it to the
>> script. It should catch the most annoying errors such as changes in
>> the ram flags.
>>
>> After code has been merged it becomes way harder to figure out
Juan Quintela writes:
> From: Fabiano Rosas
>
> Add a smoke test that migrates to a file and gives it to the
> script. It should catch the most annoying errors such as changes in
> the ram flags.
>
> After code has been merged it becomes way harder to figure out what is
> causing the script to
Changes from v3 (discussion here [3]):
* Improve/fix QAPI documentation.
Changes from v2 (discussion here [2]):
* Cluster size caveats only apply to non-COW diff image, adapt the
cluster size check and documentation accordingly.
* In the IO test, use backing files (rather than stand-alone diff
From: John Snow
for the mirror job. The bitmap's granularity is used as the job's
granularity.
The new @bitmap parameter is marked unstable in the QAPI and can
currently only be used for @sync=full mode.
Clusters initially dirty in the bitmap as well as new writes are
copied to the target.
When using mirror with a bitmap and the target does not do COW and is
is a diff image, i.e. one that should only contain the delta and was
not synced to previously, a too large cluster size for the target can
be problematic. In particular, when the mirror sends data to the
target aligned to the
From: Fabian Grünbichler
heavily based on/practically forked off iotest 257 for bitmap backups,
but:
- no writes to filter node 'mirror-top' between completion and
finalization, as those seem to deadlock?
- extra set of reference/test mirrors to verify that writes in parallel
with active mirror
It is more flexible and is done in preparation to support specifying a
working bitmap for mirror jobs. In particular, this makes it possible
to assert that @sync_mode=full when a bitmap is used. That assertion
is just to be sure, of course the mirror QMP commands will be made to
fail earlier with
Backup supports all modes listed in MirrorSyncMode, while mirror does
not. Introduce BackupSyncMode by copying the current MirrorSyncMode
and drop the variants mirror does not support from MirrorSyncMode as
well as the corresponding manual check in mirror_start().
A consequence is also tighter
This patch adds support for MCQ defined in UFSHCI 4.0. This patch
utilized the legacy I/O codes as much as possible to support MCQ.
MCQ operation & runtime register is placed at 0x1000 offset of UFSHCI
register statically with no spare space among four registers (48B):
UfsMcqSqReg,
UFSHCI 4.0 spec introduced MCQ(Multi-Circular Queue) to support multiple
command queues for UFS controller. To test ufs-mcq path of kernel, MCQ
emulated device would be a good choice to go with.
The first patch added newly introduced fields in UFSHCI 4.0 to support
MCQ. The other one made the
UFSHCI 4.0 spec introduced MCQ(Multi-Circular Queue) to support multiple
command queues for UFS controller. To test ufs-mcq path of kernel, MCQ
emulated device would be a good choice to go with.
The first patch added newly introduced fields in UFSHCI 4.0 to support
MCQ. The other one made the
This patch is a prep patch for the following MCQ support patch for
hw/ufs. This patch updated minimal mandatory fields to support MCQ
based on UFSHCI 4.0.
Signed-off-by: Minwoo Im
---
include/block/ufs.h | 108 +++-
1 file changed, 106 insertions(+), 2
On 5/13/24 08:32, Andrey Drobyshev wrote:
Add a bunch of test cases covering new subclusters behaviour: unmap of
last allocated subclusters; unmap of subclusters within unallocated
cluster; discard of unallocated subclusters within a cluster; regular discard
of subclusters within a cluster;
On 5/13/24 08:32, Andrey Drobyshev wrote:
When zeroizing subclusters within single cluster, detect usage of the
BDRV_REQ_MAY_UNMAP flag and fall through to the subcluster-based discard
operation, much like it's done with the cluster-based discards. That
way subcluster-aligned operations
On 5/13/24 08:32, Andrey Drobyshev wrote:
This commit makes the discard operation work on the subcluster level
rather than cluster level. It introduces discard_l2_subclusters()
function and makes use of it in qcow2 discard implementation, much like
it's done with zero_in_l2_slice() /
On 5/13/24 08:32, Andrey Drobyshev wrote:
When zeroizing the last non-zero subclusters within single cluster, it
makes sense to go zeroize the entire cluster and go down zero_in_l2_slice()
path right away. That way we'd also update the corresponding refcount
table.
Signed-off-by: Andrey
On 5/13/24 08:31, Andrey Drobyshev wrote:
This helper simply obtains the l2 table parameters of the cluster which
contains the given subclusters range. Right now this info is being
obtained and used by zero_l2_subclusters(). As we're about to introduce
the subclusters discard operation,
On 5/13/24 08:31, Andrey Drobyshev wrote:
We basically fill 2 images with identical data and perform discard
operations with and without 'discard-no-unref' enabled. Then we check
that images still read identically, that their disk usage is the same
(i.e.
On 5/13/24 08:31, Andrey Drobyshev wrote:
Move the definition from iotests/250 to common.rc. This is used to
detect real disk usage of sparse files. In particular, we want to use
it for checking subclusters-based discards.
Signed-off-by: Andrey Drobyshev
---
tests/qemu-iotests/250
On 5/13/24 08:31, Andrey Drobyshev wrote:
This would ease debugging of write zeroes and discard operations.
Signed-off-by: Andrey Drobyshev
---
block/file-posix.c | 1 +
block/trace-events | 1 +
2 files changed, 2 insertions(+)
diff --git a/block/file-posix.c b/block/file-posix.c
On 5/13/24 08:31, Andrey Drobyshev wrote:
Normally discard requests are stored in the queue attached to BDRVQcow2State
to be processed later at once. Currently discard-no-unref option handling
causes these requests to be processed straight away. Let's fix that.
Note that when doing regular
On 5/13/24 08:31, Andrey Drobyshev wrote:
Commits 42a2890a and b2b10904 introduce handling of discard-no-unref
option in discard_in_l2_slice() and zero_in_l2_slice(). They add even
more if's when chosing the right l2 entry. What we really need for this
option is the new entry simply to
On 5/13/24 08:31, Andrey Drobyshev wrote:
We are going to need it for discarding separate subclusters. The
function itself doesn't do anything with the refcount tables, it simply
adds a discard request to the queue, so rename it to qcow2_queue_discard().
Signed-off-by: Andrey Drobyshev
25 matches
Mail list logo