> 2024年5月14日 21:58，Raphael Norwitz 写道：
>
> Code looks good. Just a question on the error case you're trying to fix.
>
> On Tue, May 14, 2024 at 2:12 AM Li Feng wrote:
>>
>> When the vhost-user is reconnecting to the backend, and if the vhost-user
>> fails
>> at the get_features in
> 2024年5月14日 21:58，Raphael Norwitz 写道：
>
> The code for these two patches looks fine. Just some questions on the
> failure case you're trying to fix.
>
>
> On Tue, May 14, 2024 at 2:12 AM Li Feng wrote:
>>
>> This reverts commit f02a4b8e6431598612466f76aac64ab492849abf.
>>
>> Since the
When iterating all_sections, this is helpful to be able to distinguish
"members" from "features"; the only other way to do so is to
cross-reference these sections against QAPIDoc.args or QAPIDoc.features,
but if the desired end goal for QAPIDoc is to remove everything except
all_sections, we need
Transactions have the only instance of an Errors section that isn't a
rST list; turn it into one.
Signed-off-by: John Snow
---
qapi/transaction.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/qapi/transaction.json b/qapi/transaction.json
index 5749c133d4a..07afc269d54
Rephrase this paragraph so that it can apply to any commands that
inherit from this object.
Signed-off-by: John Snow
---
qapi/block-export.json | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/qapi/block-export.json b/qapi/block-export.json
index
These aren't ready for upstream inclusion, because they do not properly
manage version dependencies, execution environment and so on. These are
just the tools I use in my Own Special Environment :tm: for testing and
debugging.
They've been tested only on Fedora 38 for right now, which means:
If a comment immediately follows a doc block, the parser doesn't ignore
that token appropriately. Fix that.
Signed-off-by: John Snow
---
scripts/qapi/parser.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/scripts/qapi/parser.py b/scripts/qapi/parser.py
index
If we parse all examples as QMP, we need them to conform to a standard
so that they render correctly. Once the QMP lexer is active for
examples, these will produce warning messages and fail the build.
The QMP lexer still supports elisions, but they must be represented as
the value "...", so two
In the coming patches, it's helpful to have a linting baseline. However,
there's no need to shuffle around the deck chairs too much, because most
of this code will be removed once the new qapidoc generator (the
"transmogrifier") is in place.
To ease my pain: just turn off the black auto-formatter
We do not need a dedicated section for notes. By eliminating a specially
parsed section, these notes can be treated as normal rST paragraphs in
the new QMP reference manual, and can be placed and styled much more
flexibly.
Update the QAPI parser to now prohibit special "Note" sections while
This is part of a project to overhaul the QMP reference manual. One goal
of this overhaul is to "inline" inherited argument sections into command
reference sections. A consequence of this design decision is that
inherited doc block sections need to be merged with the inheritor's
sections.
When
Sphinx does not like sections without titles, because it wants to
convert every section into a reference. When there is no title, it
struggles to do this and transforms the tree inproperly.
Depending on the rST used, this may result in an assertion error deep in
the docutils HTMLWriter.
When
This just makes it easier to do something like:
for var in variants:
...
Instead of the more cumbersome and repetitive:
for var in variants.variants:
...
Especially in conjunction with entities that aren't guaranteed to have
variants. Compare:
for var in variants.variants if variants
Rewrite the StatsFilter intro paragraph to be more meaningful to
end-users when it is inlined in generated documentation.
Signed-off-by: John Snow
---
qapi/stats.json | 8 ++--
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/qapi/stats.json b/qapi/stats.json
index
This helps simplify the doc generator if it doesn't have to check for
undocumented members.
Signed-off-by: John Snow
---
scripts/qapi/parser.py | 20 ++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/scripts/qapi/parser.py b/scripts/qapi/parser.py
index
Prior to this patch, a section like this:
@name: lorem ipsum
dolor sit amet
consectetur adipiscing elit
would be parsed as:
"lorem ipsum\ndolor sit amet\n consectetur adipiscing elit"
We want to preserve the indentation for even the first body line so that
the entire block can be
Instead of using the info object for the doc block as a whole, update
the info pointer for each call to ensure_untagged_section when the
existing section is otherwise empty. This way, Sphinx error information
will match precisely to where the text actually starts.
Signed-off-by: John Snow
---
Add a semantic tag to paragraphs that appear *before* tagged
sections/members/features and those that appear after. This will control
how they are inlined when doc sections are merged and flattened.
Signed-off-by: John Snow
---
scripts/qapi/parser.py | 22 +-
1 file changed,
This is for the sake of the new rST generator (the "transmogrifier") so
we can advance multiple lines on occasion while keeping the
generated<-->source mappings accurate.
next_line now simply takes an optional n parameter which chooses the
number of lines to advance.
Signed-off-by: John Snow
Eliminate the "Example" sections in QAPI doc blocks, converting them
into QMP example code blocks. This is generally done by converting
"Example:" or "Examples:" lines into ".. code-block:: QMP" lines.
This patch does also allow for the use of the rST syntax "Example::" by
exempting double-colon
The intent here is to mark only certain definitions as visible in the
end-user docs.
All commands and events are inherently visible. Everything else is
visible only if it is a member (or a branch member) of a type that is
visible, or if it is named as a return type for a command.
Notably, this
Fix minor irritants to pylint/flake8 et al.
(Yes, these need to be guarded by the Python tests. That's a work in
progress, a series that's quite likely to follow once I finish this
Sphinx project. Please pardon the temporary irritation.)
Signed-off-by: John Snow
---
scripts/qapi/introspect.py
Howdy - this patch series is the first batch of patches meant to prepare
the QAPI documentation for a new Sphinx module that adds
cross-references, an index, improved inlining, elision of types unseen
on the wire, and other goodies.
This series addresses just existing code and documentation that
Hello Peter and all,
I did a comparison of the VM live-migration speeds between RDMA and
TCP/IP on our servers
and plotted the results to get an initial impression. Unfortunately,
the Ethernet NICs are not the
recent ones, therefore, it may not make much sense. I can do it on
servers with more
Code looks good. Just a question on the error case you're trying to fix.
On Tue, May 14, 2024 at 2:12 AM Li Feng wrote:
>
> When the vhost-user is reconnecting to the backend, and if the vhost-user
> fails
> at the get_features in vhost_dev_init(), then the reconnect will fail
> and it will not
The code for these two patches looks fine. Just some questions on the
failure case you're trying to fix.
On Tue, May 14, 2024 at 2:12 AM Li Feng wrote:
>
> This reverts commit f02a4b8e6431598612466f76aac64ab492849abf.
>
> Since the current patch cannot completely fix the lost reconnect
>
Fiona Ebner writes:
> Am 26.03.24 um 10:06 schrieb Markus Armbruster:
>>> @@ -365,7 +368,13 @@ BlockCopyState *block_copy_state_new(BdrvChild
>>> *source, BdrvChild *target,
>>>
>>> GLOBAL_STATE_CODE();
>>>
>>> -cluster_size = block_copy_calculate_cluster_size(target->bs, errp);
When the vhost-user is reconnecting to the backend, and if the vhost-user fails
at the get_features in vhost_dev_init(), then the reconnect will fail
and it will not be retriggered forever.
The reason is:
When the vhost-user fail at get_features, the vhost_dev_cleanup will be called
immediately.
This reverts commit f02a4b8e6431598612466f76aac64ab492849abf.
Since the current patch cannot completely fix the lost reconnect
problem, there is a scenario that is not considered:
- When the virtio-blk driver is removed from the guest os,
s->connected has no chance to be set to false, resulting
29 matches
Mail list logo
