On 10/12/21 2:41 PM, John Snow wrote:
The following changes since commit bfd9a76f9c143d450ab5545dedfa74364b39fc56:
Merge remote-tracking branch 'remotes/stsquad/tags/pull-for-6.2-121021-2'
into staging (2021-10-12 06:16:25 -0700)
are available in the Git repository at:
To use the AQMP backend, Machine just needs to be a little more diligent
about what happens when closing a QMP connection. The operation is no
longer a freebie in the async world; it may return errors encountered in
the async bottom half on incoming message receipt, etc.
(AQMP's disconnect,
Swap out the synchronous QEMUMonitorProtocol from qemu.qmp with the sync
wrapper from qemu.aqmp instead.
Add an escape hatch in the form of the environment variable
QEMU_PYTHON_LEGACY_QMP which allows you to cajole QEMUMachine into using
the old implementatin, proving that both implementations
This is a wrapper around the async QMPClient that mimics the old,
synchronous QEMUMonitorProtocol class. It is designed to be
interchangeable with the old implementation.
It does not, however, attempt to mimic Exception compatibility.
Signed-off-by: John Snow
---
python/qemu/aqmp/legacy.py |
AQMP likes to be very chatty about errors it encounters. In general,
this is good because it allows us to get good diagnostic information for
otherwise complex async failures.
For example, during a failed QMP connection attempt, we might see:
+ERROR:qemu.aqmp.qmp_client.qemub-2536319:Negotiation
Based-on: <20211012214152.802483-1-js...@redhat.com>
[PULL 00/10] Python patches
GitLab: https://gitlab.com/jsnow/qemu/-/commits/python-aqmp-iotest-wrapper
CI: https://gitlab.com/jsnow/qemu/-/pipelines/387210591
Hiya,
This series continues where the last two AQMP series left off and
The scary message interferes with the iotests output. Coincidentally, if
iotests works by removing this, then it's good evidence that we don't
really need to scare people away from using it.
Signed-off-by: John Snow
---
python/qemu/aqmp/__init__.py | 12
1 file changed, 12
(But continue to support the old ones for now, too.)
There are very few cases of any user of QEMUMachine or a subclass
thereof relying on a QMP Exception type. If you'd like to check for
yourself, you want to grep for all of the derivatives of QMPError,
excluding 'AQMPError' and its derivatives.
If we spy on the QMP commands instead, we don't need callers to remember
to pass it. Seems like a fair trade-off.
The one slightly weird bit is overloading this instance variable for
wait(), where we use it to mean "don't issue the qmp 'quit'
command". This means that wait() will "fail" if the
It's not used anymore, now.
Signed-off-by: John Snow
Reviewed-by: Hanna Reitz
Reviewed-by: Paolo Bonzini
Message-id: 20210923004938.363-11-js...@redhat.com
Signed-off-by: John Snow
---
tests/qemu-iotests/socket_scm_helper.c | 136 -
python/qemu/machine/machine.py
13.10.2021 00:37, Eric Blake wrote:
On Tue, Oct 12, 2021 at 07:13:23PM +0300, Vladimir Sementsov-Ogievskiy wrote:
@@ -1530,6 +1531,7 @@ BlockAIOCB *blk_aio_preadv(BlockBackend *blk, int64_t
offset,
QEMUIOVector *qiov, BdrvRequestFlags flags,
AQMP is a library, and ideally it should not print error diagnostics
unless a user opts into seeing them. By default, Python will print all
WARNING, ERROR or CRITICAL messages to screen if no logging
configuration has been created by a client application.
In AQMP's case, ERROR logging statements
It turns out you can do this directly from Python ... and because of
this, you don't need to worry about setting the inheritability of the
fds or spawning another process.
Doing this is helpful because it allows QEMUMonitorProtocol to keep its
file descriptor and socket object as private
All callers in the tree *already* clear the events after a call to
get_events(). Do it automatically instead and update callsites to remove
the manual clear call.
These semantics are quite a bit easier to emulate with async QMP, and
nobody appears to be abusing some emergent properties of what
When we encounter an EOFError, we don't know if it's an "error" in the
perspective of the user of the library yet. Therefore, we should not log
it as an error. Reduce the severity of this logging message to "INFO" to
indicate that it's something that we expect to occur during the normal
operation
Synchronous clients may want to know if they're about to block waiting
for an event or not. A method such as this is necessary to implement a
compatible interface for the old QEMUMonitorProtocol using the new async
internals.
Signed-off-by: John Snow
Reviewed-by: Hanna Reitz
Reviewed-by: Paolo
The iotests interface expects to return the greeting as a dict; AQMP
offers it as a rich object.
Signed-off-by: John Snow
Reviewed-by: Paolo Bonzini
Reviewed-by: Eric Blake
Message-id: 20210923004938.363-6-js...@redhat.com
Signed-off-by: John Snow
---
python/qemu/aqmp/models.py | 13
Add an implementation for send_fd_scm to the async QMP implementation.
Like socket_scm_helper mentions, a non-empty payload is required for
QEMU to process the ancillary data. A space is most useful because it
does not disturb the parsing of subsequent JSON objects.
A note on "voiding the
This serves two purposes:
(1) It is now possible to discern whether or not clear() removed any
event(s) from the queue with absolute certainty, and
(2) It is now very easy to get a List of all pending events in one
chunk, which is useful for the sync bridge.
Signed-off-by: John Snow
Expose the greeting as a read-only property of QMPClient so it can be
retrieved at-will.
Signed-off-by: John Snow
Reviewed-by: Hanna Reitz
Reviewed-by: Paolo Bonzini
Message-id: 20210923004938.363-2-js...@redhat.com
Signed-off-by: John Snow
---
python/qemu/aqmp/qmp_client.py | 5 +
1
The following changes since commit bfd9a76f9c143d450ab5545dedfa74364b39fc56:
Merge remote-tracking branch 'remotes/stsquad/tags/pull-for-6.2-121021-2'
into staging (2021-10-12 06:16:25 -0700)
are available in the Git repository at:
https://gitlab.com/jsnow/qemu.git tags/python-pull-request
On Tue, Oct 12, 2021 at 07:13:23PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> > > @@ -1530,6 +1531,7 @@ BlockAIOCB *blk_aio_preadv(BlockBackend *blk,
> > > int64_t offset,
> > > QEMUIOVector *qiov, BdrvRequestFlags flags,
> > >
On Tue, Oct 12, 2021 at 12:06 PM Hanna Reitz wrote:
> On 23.09.21 02:49, John Snow wrote:
> > (But continue to support the old ones for now, too.)
> >
> > There are very few cases of any user of QEMUMachine or a subclass
> > thereof relying on a QMP Exception type. If you'd like to check for
> >
On Tue, Oct 12, 2021 at 03:25:12AM -0400, Michael S. Tsirkin wrote:
> On Thu, Oct 07, 2021 at 06:23:55PM +0200, Lukasz Maniak wrote:
> > PCIe devices implementing SR-IOV may need to perform certain actions
> > before the VFs are unrealized or vice versa.
> >
> > Signed-off-by: Lukasz Maniak
>
>
10/6/21 23:29, Eric Blake wrote:
On Wed, Oct 06, 2021 at 03:17:16PM +0200, Vladimir Sementsov-Ogievskiy wrote:
1. Convert bytes in BlkAioEmAIOCB:
aio->bytes is only passed to already int64_t interfaces, and set in
blk_aio_prwv, which is updated here.
2. For all updated functions
On 23.09.21 02:49, John Snow wrote:
(But continue to support the old ones for now, too.)
There are very few cases of any user of QEMUMachine or a subclass
thereof relying on a QMP Exception type. If you'd like to check for
yourself, you want to grep for all of the derivatives of QMPError,
On 07.10.21 18:52, John Snow wrote:
On Wed, Sep 22, 2021 at 8:50 PM John Snow wrote:
To use the AQMP backend, Machine just needs to be a little more
diligent
about what happens when closing a QMP connection. The operation is no
longer a freebie in the async world; it may
On 23.09.21 02:49, John Snow wrote:
If we spy on the QMP commands instead, we don't need callers to remember
to pass it. Seems like a fair trade-off.
The one slightly weird bit is overloading this instance variable for
wait(), where we use it to mean "don't issue the qmp 'quit'
command". This
the qemu rbd driver currently lacks support for bdrv_co_block_status.
This results mainly in incorrect progress during block operations (e.g.
qemu-img convert with an rbd image as source).
This patch utilizes the rbd_diff_iterate2 call from librbd to detect
allocated and unallocated (all zero
On 19/05/2021 12.56, Kevin Wolf wrote:
[...]
We have a bug in vpc_open(), though: It sets the local variable
disk_type correctly, but other functions use s->footer.type instead and
we never check that it actually matches the disk type we think we're
opening.
So I think we need to add this check
On 12/10/21 10:48, Emanuele Giuseppe Esposito wrote:
+/* Common functions that are neither I/O nor Global State */
+
+DriveInfo *blk_legacy_dinfo(BlockBackend *blk);
+int drive_get_max_devs(BlockInterfaceType type);
+
blk_legacy_dinfo should count as global state (blk->legacy_dinfo is
On 12/10/21 10:48, Emanuele Giuseppe Esposito wrote:
We want to be sure that the functions that write the child and
parent list of a bs are under BQL and drain.
BQL prevents from concurrent writings from the GS API, while
drains protect from I/O.
TODO: drains are missing in some functions
On 12/10/21 10:48, Emanuele Giuseppe Esposito wrote:
+if (type != IF_DEFAULT) {
+qemu_opt_set(opts, "if", if_name[type], _abort);
+}
To avoid exporting if_name, this can be replaced by a string argument
directly. But in any case this is okay for this series:
Reviewed-by:
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
job.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/job.c b/job.c
index dbfa67bb0a..94b142684f 100644
--- a/job.c
+++ b/job.c
@@ -380,6 +380,8 @@ void job_ref(Job *job)
void job_unref(Job *job)
{
+
Assertions in the callers of the funciton pointrs are already
added by previous patches.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
include/sysemu/block-backend-common.h | 28 ++-
1 file changed, 23 insertions(+), 5 deletions(-)
diff
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
block.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/block.c b/block.c
index 40c4729b8d..da80e89ad4 100644
--- a/block.c
+++ b/block.c
@@ -1462,6 +1462,7 @@ const BdrvChildClass child_of_bds = {
The job API will be handled separately in another serie.
Signed-off-by: Emanuele Giuseppe Esposito
---
include/qemu/job.h | 16
1 file changed, 16 insertions(+)
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 6e67b6977f..7e9e59f4b8 100644
--- a/include/qemu/job.h
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Paolo Bonzini
Reviewed-by: Stefan Hajnoczi
---
block/block-backend.c | 2 ++
blockdev.c| 12
2 files changed, 14 insertions(+)
diff --git a/block/block-backend.c b/block/block-backend.c
index
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
include/block/block_int-common.h | 51
1 file changed, 32 insertions(+), 19 deletions(-)
diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h
index
Snapshots run also under the BQL lock, so they all are
in the global state API. The aiocontext lock that they hold
is currently an overkill and in future could be removed.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Paolo Bonzini
Reviewed-by: Stefan Hajnoczi
---
block/snapshot.c
block coroutines functions run in different aiocontext, and are
not protected by the BQL. Therefore are I/O.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Paolo Bonzini
Reviewed-by: Stefan Hajnoczi
---
block/coroutines.h | 7 +++
1 file changed, 7 insertions(+)
diff --git
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
block.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/block.c b/block.c
index 94bff5c757..40c4729b8d 100644
--- a/block.c
+++ b/block.c
@@ -1074,6 +1074,7 @@ int
Similar to the header split, also the function pointers in BlockDriver
can be split in I/O and global state.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
include/block/block_int-common.h | 458 ---
1 file changed, 237 insertions(+), 221
copy-before-write functions always run under BQL lock.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Paolo Bonzini
---
block/copy-before-write.c | 2 ++
block/copy-before-write.h | 7 +++
2 files changed, 9 insertions(+)
diff --git a/block/copy-before-write.c
blockdev functions run always under the BQL lock.
Signed-off-by: Emanuele Giuseppe Esposito
---
include/sysemu/blockdev.h | 14 --
1 file changed, 12 insertions(+), 2 deletions(-)
diff --git a/include/sysemu/blockdev.h b/include/sysemu/blockdev.h
index 960b54d320..cc9dfb5181 100644
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
blockjob.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/blockjob.c b/blockjob.c
index fbd6c7d873..4982f6a2b5 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -61,6 +61,7 @@ static bool is_block_job(Job
blockjob functions run always under the BQL lock.
Signed-off-by: Emanuele Giuseppe Esposito
---
include/block/blockjob.h | 9 +
1 file changed, 9 insertions(+)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index d200f33c10..fa0c3f7a47 100644
---
Since the I/O functions are not many, keep a single file.
Also split the function pointers in BlockJobDriver.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
include/block/blockjob_int.h | 28
1 file changed, 28 insertions(+)
diff --git
Similarly to the previous patch, split block_int.h
in block_int-io.h and block_int-global-state.h
block_int-common.h contains the structures shared between
the two headers, and the functions that can't be categorized as
I/O or global state.
Assertions are added in the next patch.
Signed-off-by:
drive_add is only used in softmmu/vl.c, so it can be a static
function there, and drive_def is only a particular use case of
qemu_opts_parse_noisily, so it can be inlined.
Also remove drive_mark_claimed_by_board, as it is only defined
but not implemented (nor used) anywhere.
This also helps
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
blockjob.c | 4
1 file changed, 4 insertions(+)
diff --git a/blockjob.c b/blockjob.c
index 4bad1408cb..fbd6c7d873 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -83,6 +83,7 @@ BlockJob *block_job_get(const char *id)
We want to be sure that the functions that write the child and
parent list of a bs are under BQL and drain.
BQL prevents from concurrent writings from the GS API, while
drains protect from I/O.
TODO: drains are missing in some functions using this assert.
Therefore a proper assertion will fail.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
block.c | 17 +
block/backup.c | 1 +
block/block-backend.c | 3 +++
block/commit.c | 2 ++
block/dirty-bitmap.c| 1 +
Similarly to the previous patches, split block-backend.h
in block-backend-io.h and block-backend-global-state.h
In addition, remove "block/block.h" include as it seems
it is not necessary anymore, together with "qemu/iov.h"
block-backend-common.h contains the structures shared between
the two
All the global state (GS) API functions will check that
qemu_in_main_thread() returns true. If not, it means
that the safety of BQL cannot be guaranteed, and
they need to be moved to I/O.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
block.c| 136
All the global state (GS) API functions will check that
qemu_in_main_thread() returns true. If not, it means
that the safety of BQL cannot be guaranteed, and
they need to be moved to I/O.
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Stefan Hajnoczi
---
block/block-backend.c | 90
Currently, block layer APIs like block-backend.h contain a mix of
functions that are either running in the main loop and under the
BQL, or are thread-safe functions and run in iothreads performing I/O.
The functions running under BQL also take care of modifying the
block graph, by using drain
block.h currently contains a mix of functions:
some of them run under the BQL and modify the block layer graph,
others are instead thread-safe and perform I/O in iothreads.
It is not easy to understand which function is part of which
group (I/O vs GS), and this patch aims to clarify it.
The "GS"
When invoked from the main loop, this function is the same
as qemu_mutex_iothread_locked, and returns true if the BQL is held.
When invoked from iothreads or tests, it returns true only
if the current AioContext is the Main Loop.
This essentially just extends qemu_mutex_iothread_locked to work
The code in vpc.c uses BDRVVPCState->footer.type in various places
to decide whether the image is a fixed-size (VHD_FIXED) or a dynamic
(VHD_DYNAMIC) image. However, we never check that this field really
contains VHD_FIXED if we detected a fixed size image in vpc_open(),
so a wrong value here
On 10/7/21 9:27 AM, Eric Blake wrote:
On Tue, Oct 05, 2021 at 12:45:52PM -0400, Jonah Palmer wrote:
From: Laurent Vivier
This new command shows the information of a VirtQueue element.
Signed-off-by: Jonah Palmer
---
+++ b/qapi/virtio.json
+##
+# @VirtioRingAvail:
+#
+# @flags: VRingAvail
On 10/5/21 5:18 PM, Eric Blake wrote:
On Tue, Oct 05, 2021 at 12:45:46PM -0400, Jonah Palmer wrote:
This patch drops the name parameter for the virtio_init function.
The pair between the numeric device ID and the string device ID
(name) of a virtio device already exists, but not in a way that
On 11.10.21 18:24, Eric Blake wrote:
On Mon, Oct 11, 2021 at 05:50:31PM +0200, Hanna Reitz wrote:
With -m32, size_t is generally only a uint32_t. That makes clang
complain that in the assertion
assert(qiov->size <= INT64_MAX);
the range of the type of qiov->size (size_t) is too small for
On Thu, Oct 07, 2021 at 06:23:55PM +0200, Lukasz Maniak wrote:
> PCIe devices implementing SR-IOV may need to perform certain actions
> before the VFs are unrealized or vice versa.
>
> Signed-off-by: Lukasz Maniak
Callbacks are annoying and easy to misuse though.
VFs are enabled through a
64 matches
Mail list logo
