aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2017-03-16qapi: Improve error message on @NAME: in free-form docMarkus Armbruster
Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-34-git-send-email-armbru@redhat.com>
2017-03-16qapi: Move detection of doc / expression name mismatchMarkus Armbruster
Move the check whether the doc matches the expression name from check_definition_doc() to check_exprs(). This changes the error location from the comment to the expression. Makes sense as the message talks about the expression: "Definition of '%s' follows documentation for '%s'". It's also a step towards getting rid of check_docs(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-33-git-send-email-armbru@redhat.com>
2017-03-16qapi: Fix detection of doc / expression mismatchMarkus Armbruster
This fixes the errors uncovered by the previous commit. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-32-git-send-email-armbru@redhat.com>
2017-03-16tests/qapi-schema: Improve doc / expression mismatch coverageMarkus Armbruster
New tests doc-before-include.json and doc-before-pragma.json show we fail to reject a misplaced expression comment. New test doc-no-symbol.json shows a bad error message. Signed-off-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1489582656-31133-31-git-send-email-armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
2017-03-16qapi2texi: Use category "Object" for all object typesMarkus Armbruster
At the protocol level, the distinction between struct, flat union and simple union is meaningless, they are all JSON objects. Document them that way. Example change (qemu-qmp-ref.txt): - -- Simple Union: InputEvent + -- Object: InputEvent Input event union. This also fixes the completely broken headings for flat and simple unions in qemu-qmp-ref.7 and qemu-ga-ref.7, by sidestepping a bug in texi2pod.pl. For instance, it mistranslates "@deftp {Simple Union} InputEvent" to "B<Union> (Simple)", but translates "@deftp Object InputEvent" to "B<SocketAddress> (Object)". Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-30-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Generate descriptions for simple union tagsMarkus Armbruster
Simple union tags carry no type information, because their type is implicit. Their description should make up for it, but many have none. Generate one automatically then. Example change (qemu-qmp-ref.txt): -- Simple Union: ImageInfoSpecific A discriminated record of image format specific information structures. Members: 'type' - Not documented + One of "qcow2", "vmdk", "luks" 'data: ImageInfoSpecificQCow2' when 'type' is "qcow2" 'data: ImageInfoSpecificVmdk' when 'type' is "vmdk" 'data: QCryptoBlockInfoLUKS' when 'type' is "luks" Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-29-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Generate documentation for variant membersMarkus Armbruster
A flat union's branch brings in the members of another type. Generate a suitable reference to that type. Example change (qemu-qmp-ref.txt): -- Flat Union: QCryptoBlockOpenOptions The options that are available for all encryption formats when opening an existing volume Members: The members of 'QCryptoBlockOptionsBase' + The members of 'QCryptoBlockOptionsQCow' when 'format' is "qcow" + The members of 'QCryptoBlockOptionsLUKS' when 'format' is "luks" Since: 2.6 A simple union's branch adds a member 'data' of some other type. Generate documentation for that member. Example change (qemu-qmp-ref.txt): -- Simple Union: SocketAddress Captures the address of a socket, which could also be a named file descriptor Members: 'type' Not documented + 'data: InetSocketAddress' when 'type' is "inet" + 'data: UnixSocketAddress' when 'type' is "unix" + 'data: VsockSocketAddress' when 'type' is "vsock" + 'data: String' when 'type' is "fd" Since: 1.3 Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-28-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Generate reference to base type membersMarkus Armbruster
The generated documentation doesn't mention object type members inherited from a base type. Fix that. Example change (qemu-qmp-ref.txt): -- Struct: VncServerInfo The network connection information for server Members: 'auth' (optional) authentication method used for the plain (non-websocket) VNC server + The members of 'VncBasicInfo' Since: 2.1 Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-27-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Include member type in generated documentationMarkus Armbruster
The recent merge of docs/qmp-commands.txt and docs/qmp-events.txt into the schema lost type information. Fix this documentation regression. Example change (qemu-qmp-ref.txt): -- Struct: InputKeyEvent Keyboard input event. Members: - 'button' + 'button: InputButton' Which button this event is for. - 'down' + 'down: boolean' True for key-down and false for key-up events. Since: 2.0 Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-26-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Implement boxed argument documentationMarkus Armbruster
This replaces manual references like "For the arguments, see the documentation of ..." by a generated reference "Arguments: the members of ...". Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-25-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Don't hide undocumented members and argumentsMarkus Armbruster
Show undocumented object, alternate type members and command, event arguments exactly like undocumented enumeration type values. Example change (qemu-qmp-ref.txt): -- Command: query-rocker Return rocker switch information. + Arguments: + 'name' + Not documented + Returns: 'Rocker' information Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-24-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Explain enum value undocumentedness more clearlyMarkus Armbruster
Instead of not saying anything when we have no documentation, say "Not documented". Example change (qemu-qmp-ref.txt): -- Enum: GuestPanicAction An enumeration of the actions taken when guest OS panic is detected Values: 'pause' system pauses 'poweroff' + Not documented Since: 2.1 (poweroff since 2.8) Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-23-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Present the table of members more clearlyMarkus Armbruster
The table of members follows the main descriptive text immediately. Makes it hard to see what it is about. Start a new paragraph, and lead with a line "Members:" for object and alternate types, "Values:" for enumeration types, and "Arguments:" for commands and events. Example change (qemu-qmp-ref.txt): -- Command: set_link Sets the link status of a virtual network adapter. + + Arguments: 'name' the device name of the virtual network adapter 'up' true to set the link status to be up Returns: Nothing on success If 'name' is not a valid network device, DeviceNotFound Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-22-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Plainer enum value and member name formattingMarkus Armbruster
Use @code{%s} instead of @code{'%s'}. Impact, using @id as example: * Texinfo -@item @code{'id'} +@item @code{id} * HTML -<dt><code>'id'</code></dt> +<dt><code>id</code></dt> * POD (for manual pages): -=item C<'id'> +=item C<id> * Formatted manual pages: -'id' +"id" * Plain text: - ''id'' + 'id' Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-21-git-send-email-armbru@redhat.com>
2017-03-16qapi: Prefer single-quoted strings more consistentlyMarkus Armbruster
PEP 8 advises: In Python, single-quoted strings and double-quoted strings are the same. This PEP does not make a recommendation for this. Pick a rule and stick to it. When a string contains single or double quote characters, however, use the other one to avoid backslashes in the string. It improves readability. The QAPI generators succeed at picking a rule, but fail at sticking to it. Convert a bunch of double-quoted strings to single-quoted ones. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-20-git-send-email-armbru@redhat.com>
2017-03-16qapi: Use raw strings for regular expressions consistentlyMarkus Armbruster
Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-19-git-send-email-armbru@redhat.com>
2017-03-16qapi: The #optional tag is redundant, dropMarkus Armbruster
We traditionally mark optional members #optional in the doc comment. Before commit 3313b61, this was entirely manual. Commit 3313b61 added some automation because its qapi2texi.py relied on #optional to determine whether a member is optional. This is no longer the case since the previous commit: the only thing qapi2texi.py still does with #optional is stripping it out. We still reject bogus qapi-schema.json and six places for qga/qapi-schema.json. Thus, you can't actually rely on #optional to see whether something is optional. Yet we still make people add it manually. That's just busy-work. Drop the code to check, fix up and strip out #optional, along with all instances of #optional. To keep it out, add code to reject it, to be dropped again once the dust settles. No change to generated documentation. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-18-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Convert to QAPISchemaVisitorMarkus Armbruster
qapi2texi works with schema expression trees. Such a tight coupling to schema language syntax is not a good idea. Convert it to the visitor interface the other generators use. No change to generated documentation. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-17-git-send-email-armbru@redhat.com>
2017-03-16qapi: Conjure up QAPIDoc.ArgSection for undocumented membersMarkus Armbruster
qapi2texi.py already conjures up ArgSections for undocumented enumeration values, in texi_enum. Drop that, and conjure them up for all kinds of "arguments" (enumeration values, object and alternate type members) in qapi.py instead. Take care to keep generated documentation exactly the same for now. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-16-git-send-email-armbru@redhat.com>
2017-03-16qapi: Prepare for requiring more complete documentationMarkus Armbruster
We currently neglect to check all enumeration values, common members of object types and members of alternate types are documented. Unsurprisingly, many aren't. Add the necessary plumbing to find undocumented ones, except for variant members of object types. Don't enforce anything just yet, but connect each QAPIDoc.ArgSection to its QAPISchemaMember. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-15-git-send-email-armbru@redhat.com>
2017-03-16qapi: Fix QAPISchemaEnumType.is_implicit() for 'QType'Markus Armbruster
Missed in commit 7264f5c. Harmless, because nothing checks whether an enumeration type is implicit so far. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-14-git-send-email-armbru@redhat.com>
2017-03-16qapi/rocker: Fix up doc comment notes on optional membersMarkus Armbruster
Talking about #optional like this # Note: fields are marked #optional to indicate that they may or may # not appear ... doesn't work so well in generated documentation, because the #optional tag is not visible there. Replace by # Note: optional members may or may not appear ... Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Message-Id: <1489582656-31133-13-git-send-email-armbru@redhat.com>
2017-03-16qapi: Avoid unwanted blank lines in QAPIDocMarkus Armbruster
We silently fix missing #optional tags for QAPIDoc by appending a line "#optional" to the section's .content. However, this interferes with .__repr__ stripping trailing blank lines from .content. Use new ArgSection instance variable .optional instead, and leave .content alone. To permit testing .optional in texi_body(), clean up texi_enum()'s hack to add empty documentation for undocumented enum values: add an ArgSection instead of ''. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Message-Id: <1489582656-31133-12-git-send-email-armbru@redhat.com>
2017-03-16qapi2texi: Fix up output around #optionalMarkus Armbruster
We use tag #optional to mark optional members, like this: # @name: #optional The name of the guest texi_body() strips #optional, but not whitespace around it. For the above, we get in qemu-qmp-qapi.texi @item @code{'name'} (optional) The name of the guest @end table The extra space can lead to artifacts in output, e.g in qemu-qmp-ref.7.pod =item C<'name'> (optional) The name of the guest and then in qemu-qmp-ref.7 .IX Item "name (optional)" .Vb 1 \& The name of the guest .Ve instead of intended plain .IX Item "name (optional)" The name of the guest Get rid of these artifacts by removing whitespace around #optional along with it. This turns three minus signs in qapi-schema.json into markup, because they're now at the beginning of the line. Drop them, they're unwanted there. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Message-Id: <1489582656-31133-11-git-send-email-armbru@redhat.com>
2017-03-16qapi: Fix to reject empty union base gracefullyMarkus Armbruster
Common Python pitfall: 'assert base_members' fires on [] in addition to None. Correct to 'assert base_members is not None'. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-10-git-send-email-armbru@redhat.com>
2017-03-16tests/qapi-schema: Cover empty union baseMarkus Armbruster
The new test case shows off qapi.py choking on an empty union base. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-9-git-send-email-armbru@redhat.com>
2017-03-16qapi: Clean up build of generated documentationMarkus Armbruster
Rename intermediate qemu-qapi.texi to qemu-qmp-qapi.texi to match its user qemu-qmp-ref.texi, just like qemu-ga-qapi.texi matches qemu-ga-ref.texi. Build the intermediate .texi next to the sources and the final output in docs/ instead of dumping them into the build root. Fix version.texi dependencies so that only the targets that actually need it depend on it. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-8-git-send-email-armbru@redhat.com>
2017-03-16qapi: Have each QAPI schema declare its name rule violationsMarkus Armbruster
qapi.py has a hardcoded white-list of type names that may violate the rule on use of upper and lower case. Add a new pragma directive 'name-case-whitelist', and use it to replace the hard-coded white-list. Signed-off-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1489582656-31133-7-git-send-email-armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
2017-03-16qapi: Have each QAPI schema declare its returns white-listMarkus Armbruster
qapi.py has a hardcoded white-list of command names that may violate the rules on permitted return types. Add a new pragma directive 'returns-whitelist', and use it to replace the hard-coded white-list. Signed-off-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1489582656-31133-6-git-send-email-armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
2017-03-16docs/qapi-code-gen.txt: Drop confusing reference to 'gen'Markus Armbruster
Section "Commands" qualifies its rules on permitted argument and return types "with one exception noted below when 'gen' is used". The note went away in commit 2d21291. Clean up the dangling references. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Message-Id: <1489582656-31133-5-git-send-email-armbru@redhat.com>
2017-03-16qapi: Back out doc comments added just to please qapi.pyMarkus Armbruster
This reverts commit 3313b61's changes to tests/qapi-schema/, except for tests/qapi-schema/doc-*. We could keep some of these doc comments to serve as positive test cases. However, they don't actually add to what we get from doc comment use in actual schemas, as we we don't test output matches expectations, and don't systematically cover doc comment features. Proper positive test coverage would be nice. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1489582656-31133-4-git-send-email-armbru@redhat.com>
2017-03-16qapi: Make doc comments optional where we don't need themMarkus Armbruster
Since we added the documentation generator in commit 3313b61, doc comments are mandatory. That's a very good idea for a schema that needs to be documented, but has proven to be annoying for testing. Make doc comments optional again, but add a new directive { 'pragma': { 'doc-required': true } } to let a QAPI schema require them. Add test cases for the new pragma directive. While there, plug a minor hole in includ directive test coverage. Require documentation in the schemas we actually want documented: qapi-schema.json and qga/qapi-schema.json. We could probably make qapi2texi.py cope with incomplete documentation, but for now, simply make it refuse to run unless the schema has 'doc-required': true. Signed-off-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1489582656-31133-3-git-send-email-armbru@redhat.com> [qapi-code-gen.txt wording tweaked] Reviewed-by: Eric Blake <eblake@redhat.com>
2017-03-16qapi: Factor QAPISchemaParser._include() out of .__init__()Markus Armbruster
Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Message-Id: <1489582656-31133-2-git-send-email-armbru@redhat.com>
2017-03-16qmp: allow setting properties to empty string in qmp-shellDaniel P. Berrange
The qmp-shell property parser currently rejects attempts to set string properties to the empty string eg (QEMU) migrate-set-parameters tls-hostname= Error while parsing command line: Expected a key=value pair, got 'tls-hostname=' command format: <command-name> [arg-name1=arg1] ... [arg-nameN=argN] This is caused by checking the wrong condition after splitting the parameter on '='. The "partition" method will return "" for the separator field, if the seperator was not present, so that is the correct thing to check for malformed syntax. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <20170302122429.7737-1-berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: John Snow <jsnow@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-03-16qapi2texi: change texi formattersMarc-André Lureau
STRUCT_FMT is generic enough, rename it to TYPE_FMT, use it for unions. Rename COMMAND_FMT to MSG_FMT, since it applies to both commands and events. Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com> Message-Id: <20170125130308.16104-2-marcandre.lureau@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-03-15Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into stagingPeter Maydell
virtio, pc: fixes Some fixes to fallback from using virtio caching, pls a minor vm gen id fix. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> # gpg: Signature made Wed 15 Mar 2017 17:59:25 GMT # gpg: using RSA key 0x281F0DB8D28D5469 # gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" # gpg: aka "Michael S. Tsirkin <mst@redhat.com>" # Primary key fingerprint: 0270 606B 6F3C DF3D 0B17 0970 C350 3912 AFBE 8E67 # Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA 8A0D 281F 0DB8 D28D 5469 * remotes/mst/tags/for_upstream: virtio-pci: reset modern vq meta data Revert "virtio: unbreak virtio-pci with IOMMU after caching ring translations" pci: introduce a bus master container virtio: validate address space cache during init virtio: destroy region cache during reset virtio: guard against NULL pfn Bugfix: Handle error if VM Generation ID device not present Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2017-03-15virtio-pci: reset modern vq meta dataJason Wang
We don't reset proxy->vqs[].{num|desc[]|avail[]|used[]}. This means if a driver enable the vq without setting vq address after reset. The old addresses were leaked. Fixing this by resetting modern vq meta data during device reset. Cc: qemu-stable@nongnu.org Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2017-03-15Revert "virtio: unbreak virtio-pci with IOMMU after caching ring translations"Jason Wang
This reverts commit 96a8821d21411f10d77ea994af369c6e5c35a2cc. Previous patch is a better solution which does not require a strict order between virtio and IOMMU. CC: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
2017-03-15Merge remote-tracking branch 'remotes/armbru/tags/pull-misc-2017-03-15' into ↵Peter Maydell
staging Miscellaneous patches for 2017-03-15 # gpg: Signature made Wed 15 Mar 2017 13:12:35 GMT # gpg: using RSA key 0x3870B400EB918653 # gpg: Good signature from "Markus Armbruster <armbru@redhat.com>" # gpg: aka "Markus Armbruster <armbru@pond.sub.org>" # Primary key fingerprint: 354B C8B3 D7EB 2A6B 6867 4E5F 3870 B400 EB91 8653 * remotes/armbru/tags/pull-misc-2017-03-15: coverity-model: model address_space_read/write tests: Use error_free_or_abort() where appropriate Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2017-03-15pci: introduce a bus master containerJason Wang
96a8821d2141 ("virtio: unbreak virtio-pci with IOMMU after caching ring translations") tries to make IOMMU works with virtio memory region cache, but it requires IOMMU to be created before any virtio devices. This is sub optimal, fixing this by introduce a bus master container to make sure address space can be initialized during device registering, and then we can safely set alias and make bus_master_enable_region as its subregion during bus master initialization. Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2017-03-15virtio: validate address space cache during initJason Wang
We don't check the return value of address_space_cache_init(), this may lead buggy driver use incorrect region caches. Instead of triggering an assert, catch and warn this early in virtio_init_region_cache(). Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2017-03-15virtio: destroy region cache during resetJason Wang
We don't destroy region cache during reset which can make the maps of previous driver leaked to a buggy or malicious driver that don't set vring address before starting to use the device. Fix this by destroy the region cache during reset and validate it before trying to see them. Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2017-03-15virtio: guard against NULL pfnJason Wang
To avoid access stale memory region cache after reset, this patch check the existence of virtqueue pfn for all exported virtqueue access helpers before trying to use them. Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2017-03-15Bugfix: Handle error if VM Generation ID device not presentBen Warren
This was crashing due to NULL-pointer dereference QMP Test case: ============== (QEMU) query-vm-generation-id {"error": {"class": "GenericError", "desc": "VM Generation ID device not found"}} HMP Test case: ============== virsh # qemu-monitor-command --hmp 3 info vm-generation-id VM Generation ID device not found Signed-off-by: Ben Warren <ben@skyportsystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
2017-03-15Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into ↵Peter Maydell
staging # gpg: Signature made Wed 15 Mar 2017 05:05:04 GMT # gpg: using RSA key 0x9CA4ABB381AB73C8 # gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>" # gpg: aka "Stefan Hajnoczi <stefanha@gmail.com>" # Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35 775A 9CA4 ABB3 81AB 73C8 * remotes/stefanha/tags/block-pull-request: os: don't corrupt pre-existing memory-backend data with prealloc Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2017-03-15Merge remote-tracking branch 'remotes/ehabkost/tags/machine-pull-request' ↵Peter Maydell
into staging Fix global property and -cpu handling bug This bug fix was supposed to be applied just after 2.8.0 was released, but it slipped through the cracks. Sending it now for the next -rc. # gpg: Signature made Tue 14 Mar 2017 20:04:50 GMT # gpg: using RSA key 0x2807936F984DC5A6 # gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>" # Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF D1AA 2807 936F 984D C5A6 * remotes/ehabkost/tags/machine-pull-request: machine: Convert abstract typename on compat_props to subclass names Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2017-03-15coverity-model: model address_space_read/writePaolo Bonzini
Commit eb7eeb8 ("memory: split address_space_read and address_space_write", 2015-12-17) made address_space_rw dispatch to one of address_space_read or address_space_write, rather than vice versa. For callers of address_space_read and address_space_write this causes false positive defects when Coverity sees a length-8 write in address_space_read and a length-4 (e.g. int*) buffer to read into. As long as the size of the buffer is okay, this is a false positive. Reflect the code change into the model. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20170315081641.20588-1-pbonzini@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-03-15tests: Use error_free_or_abort() where appropriateMarkus Armbruster
Done with this Coccinelle semantic patch: @@ expression E; @@ - g_assert(E); - error_free(E); + error_free_or_abort(&E); Signed-off-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1487362554-5688-1-git-send-email-armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
2017-03-15os: don't corrupt pre-existing memory-backend data with preallocDaniel P. Berrange
When using a memory-backend object with prealloc turned on, QEMU will memset() the first byte in every memory page to zero. While this might have been acceptable for memory backends associated with RAM, this corrupts application data for NVDIMMs. Instead of setting every page to zero, read the current byte value and then just write that same value back, so we are not corrupting the original data. Directly write the value instead of memset()ing it, since there's no benefit to memset for a single byte write. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Andrea Arcangeli <aarcange@redhat.com> Message-id: 20170303113255.28262-1-berrange@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2017-03-14machine: Convert abstract typename on compat_props to subclass namesEduardo Habkost
Original problem description by Greg Kurz: > Since commit "9a4c0e220d8a hw/virtio-pci: fix virtio > behaviour", passing -device virtio-blk-pci.disable-modern=off > has no effect on 2.6 machine types because the internal > virtio-pci.disable-modern=on compat property always prevail. The same bug also affects other abstract type names mentioned on compat_props by machine-types: apic-common, i386-cpu, pci-device, powerpc64-cpu, s390-skeys, spapr-pci-host-bridge, usb-device, virtio-pci, x86_64-cpu. The right fix for this problem is to make sure compat_props and -global options are always applied in the order they are registered, instead of reordering them based on the type hierarchy. But changing the ordering rules of -global is risky and might break existing configurations, so we shouldn't do that on a stable branch. This is a temporary hack that will work around the bug when registering compat_props properties: if we find an abstract class on compat_props, register properties for all its non-abstract subtypes instead. This will make sure -global won't be overridden by compat_props, while keeping the existing ordering rules on -global options. Note that there's one case that won't be fixed by this hack: "-global spapr-pci-vfio-host-bridge.<option>=<value>" won't be able to override compat_props, because spapr-pci-host-bridge is not an abstract class. Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> Message-Id: <1481575745-26120-1-git-send-email-ehabkost@redhat.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Reviewed-by: Halil Pasic <pasic@linux.vnet.ibm.com> Reviewed-by: Greg Kurz <groug@kaod.org> Tested-by: Greg Kurz <groug@kaod.org> Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
generated by cgit v1.2.3 (git 2.26.2) at 2025-02-03 02:59:49 +0000