qemu.git
19 months agoconfigure: Enable test and libs for zstd
Juan Quintela [Tue, 17 Dec 2019 20:15:24 +0000 (21:15 +0100)] 
configure: Enable test and libs for zstd

Add it to several build systems to make testing good.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
19 months agomultifd: Add zlib compression multifd support
Juan Quintela [Fri, 4 Jan 2019 14:30:06 +0000 (15:30 +0100)] 
multifd: Add zlib compression multifd support

Signed-off-by: Juan Quintela <quintela@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
19 months agomultifd: Add multifd-zlib-level parameter
Juan Quintela [Thu, 23 Jan 2020 16:08:52 +0000 (17:08 +0100)] 
multifd: Add multifd-zlib-level parameter

This parameter specifies the zlib compression level. The next patch
will put it to use.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
19 months agomultifd: Make no compression operations into its own structure
Juan Quintela [Wed, 15 May 2019 11:37:46 +0000 (13:37 +0200)] 
multifd: Make no compression operations into its own structure

It will be used later.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---

No comp value needs to be zero.

19 months agomigration: Add support for modules
Juan Quintela [Wed, 22 Jan 2020 17:55:02 +0000 (18:55 +0100)] 
migration: Add support for modules

So we don't have to compile everything in, or have ifdefs

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
19 months agomultifd: Add multifd-compression parameter
Juan Quintela [Wed, 16 Jan 2019 09:35:55 +0000 (10:35 +0100)] 
multifd: Add multifd-compression parameter

This will store the compression method to use.  We start with none.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---

Rename multifd-method to multifd-compression

19 months agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Thu, 27 Feb 2020 19:15:14 +0000 (19:15 +0000)] 
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

virtio, pc: fixes, features

New virtio iommu.
Unrealize memory leaks.
In-band kick/call support.
Bugfixes, documentation all over the place.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Thu 27 Feb 2020 08:46:33 GMT
# gpg:                using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469
# gpg:                issuer "mst@redhat.com"
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [full]
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>" [full]
# 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: (30 commits)
  Fixed assert in vhost_user_set_mem_table_postcopy
  vhost-user: only set slave channel for first vq
  acpi: cpuhp: document CPHP_GET_CPU_ID_CMD command
  libvhost-user: implement in-band notifications
  docs: vhost-user: add in-band kick/call messages
  libvhost-user: handle NOFD flag in call/kick/err better
  libvhost-user-glib: use g_main_context_get_thread_default()
  libvhost-user-glib: fix VugDev main fd cleanup
  libvhost-user: implement VHOST_USER_PROTOCOL_F_REPLY_ACK
  MAINTAINERS: add virtio-iommu related files
  hw/arm/virt: Add the virtio-iommu device tree mappings
  virtio-iommu-pci: Add virtio iommu pci support
  virtio-iommu: Support migration
  virtio-iommu: Implement fault reporting
  virtio-iommu: Implement translate
  virtio-iommu: Implement map/unmap
  virtio-iommu: Implement attach/detach command
  virtio-iommu: Decode the command payload
  virtio-iommu: Add skeleton
  virtio: gracefully handle invalid region caches
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agoMerge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2020-02-26' into staging
Peter Maydell [Thu, 27 Feb 2020 18:16:03 +0000 (18:16 +0000)] 
Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2020-02-26' into staging

nbd patches for 2020-02-26

- ensure multiple meta contexts work
- allow leading / in export names
- fix a failure path memory leak

# gpg: Signature made Thu 27 Feb 2020 01:53:03 GMT
# gpg:                using RSA key 71C2CC22B1C4602927D2F3AAA7A16B4A2527436A
# gpg: Good signature from "Eric Blake <eblake@redhat.com>" [full]
# gpg:                 aka "Eric Blake (Free Software Programmer) <ebb9@byu.net>" [full]
# gpg:                 aka "[jpeg image of size 6874]" [full]
# Primary key fingerprint: 71C2 CC22 B1C4 6029 27D2  F3AA A7A1 6B4A 2527 436A

* remotes/ericb/tags/pull-nbd-2020-02-26:
  block/nbd: fix memory leak in nbd_open()
  block/nbd: extract the common cleanup code
  nbd-client: Support leading / in NBD URI
  nbd: Fix regression with multiple meta contexts

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agoMerge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-plugins-250220...
Peter Maydell [Thu, 27 Feb 2020 17:12:31 +0000 (17:12 +0000)] 
Merge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-plugins-250220-1' into staging

Testing and plugin updates:

 - fix pauth TCG tests
 - tweak away rcutorture failures
 - various Travis updates
 - relax iotest size check a little
 - fix for -trace/-D clash
 - fix cross compile detection for tcg tests
 - document plugin query lifetime
 - fix missing break in plugin core
 - fix some plugin warnings
 - better progressive instruction decode
 - avoid trampling vaddr in plugins

# gpg: Signature made Tue 25 Feb 2020 20:21:56 GMT
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-testing-and-plugins-250220-1:
  tests/tcg: take into account expected clashes pauth-4
  tests/tcg: fix typo in configure.sh test for v8.3
  tcg: save vaddr temp for plugin usage
  tests/tcg: give debug builds a little bit longer
  tests/plugins: make howvec clean-up after itself.
  target/riscv: progressively load the instruction during decode
  qemu/bitops.h: Add extract8 and extract16
  tests/plugin: prevent uninitialized warning
  plugins/core: add missing break in cb_to_tcg_flags
  docs/devel: document query handle lifetimes
  tracing: only allow -trace to override -D if set
  tests/iotests: be a little more forgiving on the size test
  travis.yml: single-thread build-tcg stages
  travis.yml: Fix Travis YAML configuration warnings
  travis.yml: Test the s390-ccw build, too
  tests/rcutorture: mild documenting refactor of update thread
  tests/rcutorture: better document locking of stats
  tests/rcutorture: update usage hint
  tests/tcg: include a skip runner for pauth3 with plugins

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agosoftmmu/vl.c: fix too slow TCG regression
Igor Mammedov [Thu, 27 Feb 2020 16:14:54 +0000 (11:14 -0500)] 
softmmu/vl.c: fix too slow TCG regression

Commit a1b18df9a4 moved -m option parsing after configure_accelerators()
that broke TCG accelerator initialization which accesses global ram_size
from size_code_gen_buffer() which is equal to 0 at that moment.

Partially revert a1b18df9a4, by returning set_memory_options() to its
original location and only keep 32-bit host VA check and 'memory-backend'
size check introduced by fe64d06afc at current place.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agoFixed assert in vhost_user_set_mem_table_postcopy
Raphael Norwitz [Thu, 16 Jan 2020 02:57:04 +0000 (21:57 -0500)] 
Fixed assert in vhost_user_set_mem_table_postcopy

The current vhost_user_set_mem_table_postcopy() implementation
populates each region of the VHOST_USER_SET_MEM_TABLE message without
first checking if there are more than VHOST_MEMORY_MAX_NREGIONS already
populated. This can cause memory corruption if too many regions are
added to the message during the postcopy step.

This change moves an existing assert up such that attempting to
construct a VHOST_USER_SET_MEM_TABLE message with too many memory
regions will gracefully bring down qemu instead of corrupting memory.

Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Signed-off-by: Peter Turschmid <peter.turschm@nutanix.com>
Message-Id: <1579143426-18305-2-git-send-email-raphael.norwitz@nutanix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovhost-user: only set slave channel for first vq
Adrian Moreno [Tue, 21 Jan 2020 21:45:53 +0000 (22:45 +0100)] 
vhost-user: only set slave channel for first vq

When multiqueue is enabled, a vhost_dev is created for each queue
pair. However, only one slave channel is needed.

Fixes: 4bbeeba023f2 (vhost-user: add slave-req-fd support)
Cc: marcandre.lureau@redhat.com
Signed-off-by: Adrian Moreno <amorenoz@redhat.com>
Message-Id: <20200121214553.28459-1-amorenoz@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agoacpi: cpuhp: document CPHP_GET_CPU_ID_CMD command
Igor Mammedov [Wed, 29 Jan 2020 14:06:21 +0000 (15:06 +0100)] 
acpi: cpuhp: document CPHP_GET_CPU_ID_CMD command

Commit 3a61c8db9d25 introduced CPHP_GET_CPU_ID_CMD command but
did not sufficiently describe it. Fix it by adding missing command
documentation.

Fixes: 3a61c8db9d25 ("acpi: cpuhp: add CPHP_GET_CPU_ID_CMD command")
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <1580306781-228371-1-git-send-email-imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agolibvhost-user: implement in-band notifications
Johannes Berg [Thu, 23 Jan 2020 08:17:08 +0000 (09:17 +0100)] 
libvhost-user: implement in-band notifications

Add support for VHOST_USER_PROTOCOL_F_IN_BAND_NOTIFICATIONS, but
as it's not desired by default, don't enable it unless the device
implementation opts in by returning it from its protocol features
callback.

Note that I updated vu_set_vring_err_exec(), but didn't add any
sending of the VHOST_USER_SLAVE_VRING_ERR message as there's no
write to the err_fd today either.

This also adds vu_queue_notify_sync() which can be used to force
a synchronous notification if inband notifications are supported.
Previously, I had left out the slave->master direction handling
of F_REPLY_ACK, this now adds some code to support it as well.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Message-Id: <20200123081708.7817-7-johannes@sipsolutions.net>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agodocs: vhost-user: add in-band kick/call messages
Johannes Berg [Thu, 23 Jan 2020 08:17:07 +0000 (09:17 +0100)] 
docs: vhost-user: add in-band kick/call messages

For good reason, vhost-user is currently built asynchronously, that
way better performance can be obtained. However, for certain use
cases such as simulation, this is problematic.

Consider an event-based simulation in which both the device and CPU
have scheduled according to a simulation "calendar". Now, consider
the CPU sending I/O to the device, over a vring in the vhost-user
protocol. In this case, the CPU must wait for the vring interrupt
to have been processed by the device, so that the device is able to
put an entry onto the simulation calendar to obtain time to handle
the interrupt. Note that this doesn't mean the I/O is actually done
at this time, it just means that the handling of it is scheduled
before the CPU can continue running.

This cannot be done with the asynchronous eventfd based vring kick
and call design.

Extend the protocol slightly, so that a message can be used for kick
and call instead, if VHOST_USER_PROTOCOL_F_INBAND_NOTIFICATIONS is
negotiated. This in itself doesn't guarantee synchronisation, but both
sides can also negotiate VHOST_USER_PROTOCOL_F_REPLY_ACK and thus get
a reply to this message by setting the need_reply flag, and ensure
synchronisation this way.

To really use it in both directions, VHOST_USER_PROTOCOL_F_SLAVE_REQ
is also needed.

Since it is used for simulation purposes and too many messages on
the socket can lock up the virtual machine, document that this should
only be used together with the mentioned features.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Message-Id: <20200123081708.7817-6-johannes@sipsolutions.net>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agolibvhost-user: handle NOFD flag in call/kick/err better
Johannes Berg [Thu, 23 Jan 2020 08:17:06 +0000 (09:17 +0100)] 
libvhost-user: handle NOFD flag in call/kick/err better

The code here is odd, for example will it print out invalid
file descriptor numbers that were never sent in the message.

Clean that up a bit so it's actually possible to implement
a device that uses polling.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Message-Id: <20200123081708.7817-5-johannes@sipsolutions.net>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agolibvhost-user-glib: use g_main_context_get_thread_default()
Johannes Berg [Thu, 23 Jan 2020 08:17:05 +0000 (09:17 +0100)] 
libvhost-user-glib: use g_main_context_get_thread_default()

If we use NULL, we just get the main program default mainloop
here. Using g_main_context_get_thread_default() has basically
the same effect, but it lets us start different devices in
different threads with different mainloops, which can be useful.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Message-Id: <20200123081708.7817-4-johannes@sipsolutions.net>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agolibvhost-user-glib: fix VugDev main fd cleanup
Johannes Berg [Thu, 23 Jan 2020 08:17:04 +0000 (09:17 +0100)] 
libvhost-user-glib: fix VugDev main fd cleanup

If you try to make a device implementation that can handle multiple
connections and allow disconnections (which requires overriding the
VHOST_USER_NONE handling), then glib will warn that we remove a src
while it's still on the mainloop, and will poll() an FD that doesn't
exist anymore.

Fix this by making vug_source_new() require pairing with the new
vug_source_destroy() so we can keep the GSource referenced in the
meantime.

Note that this requires calling the new API in vhost-user-input.
vhost-user-gpu also uses vug_source_new(), but never seems to free
the result at all, so I haven't changed anything there.

Fixes: 8bb7ddb78a1c ("libvhost-user: add glib source helper")
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Message-Id: <20200123081708.7817-3-johannes@sipsolutions.net>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agolibvhost-user: implement VHOST_USER_PROTOCOL_F_REPLY_ACK
Johannes Berg [Thu, 23 Jan 2020 08:17:03 +0000 (09:17 +0100)] 
libvhost-user: implement VHOST_USER_PROTOCOL_F_REPLY_ACK

This is really simple, since we know whether a response is
already requested or not, so we can just send a (successful)
response when there isn't one already.

Given that, it's not all _that_ useful but the master can at
least be sure the message was processed, and we can exercise
more code paths using the example code.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Message-Id: <20200123081708.7817-2-johannes@sipsolutions.net>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agoMAINTAINERS: add virtio-iommu related files
Eric Auger [Fri, 14 Feb 2020 13:27:45 +0000 (14:27 +0100)] 
MAINTAINERS: add virtio-iommu related files

Add a new "virtio-iommu" section with the new files
related to this device.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Message-Id: <20200214132745.23392-11-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
19 months agohw/arm/virt: Add the virtio-iommu device tree mappings
Eric Auger [Fri, 14 Feb 2020 13:27:44 +0000 (14:27 +0100)] 
hw/arm/virt: Add the virtio-iommu device tree mappings

Adds the "virtio,pci-iommu" node in the host bridge node and
the RID mapping, excluding the IOMMU RID.

This is done in the virtio-iommu-pci hotplug handler which
gets called only if no firmware is loaded or if -no-acpi is
passed on the command line. As non DT integration is
not yet supported by the kernel we must make sure we
are in DT mode. This limitation will be removed as soon
as the topology description feature gets supported.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Message-Id: <20200214132745.23392-10-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
19 months agovirtio-iommu-pci: Add virtio iommu pci support
Eric Auger [Fri, 14 Feb 2020 13:27:43 +0000 (14:27 +0100)] 
virtio-iommu-pci: Add virtio iommu pci support

This patch adds virtio-iommu-pci, which is the pci proxy for
the virtio-iommu device.

Currently non DT integration is not yet supported by the kernel.
So the machine must implement a hotplug handler for the
virtio-iommu-pci device that creates the device tree iommu-map
bindings as documented in kernel documentation:

Documentation/devicetree/bindings/virtio/iommu.txt

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-9-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Support migration
Eric Auger [Fri, 14 Feb 2020 13:27:42 +0000 (14:27 +0100)] 
virtio-iommu: Support migration

Add Migration support. We rely on recently added gtree and qlist
migration. We only migrate the domain gtree. The endpoint gtree
is re-constructed in a post-load operation.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-8-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Implement fault reporting
Eric Auger [Fri, 14 Feb 2020 13:27:41 +0000 (14:27 +0100)] 
virtio-iommu: Implement fault reporting

The event queue allows to report asynchronous errors.
The translate function now injects faults when relevant.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-7-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Implement translate
Eric Auger [Fri, 14 Feb 2020 13:27:40 +0000 (14:27 +0100)] 
virtio-iommu: Implement translate

This patch implements the translate callback

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-6-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Implement map/unmap
Eric Auger [Fri, 14 Feb 2020 13:27:39 +0000 (14:27 +0100)] 
virtio-iommu: Implement map/unmap

This patch implements virtio_iommu_map/unmap.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-5-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Implement attach/detach command
Eric Auger [Fri, 14 Feb 2020 13:27:38 +0000 (14:27 +0100)] 
virtio-iommu: Implement attach/detach command

This patch implements the endpoint attach/detach to/from
a domain.

Domain and endpoint internal datatypes are introduced.
Both are stored in RB trees. The domain owns a list of
endpoints attached to it. Also helpers to get/put
end points and domains are introduced.

As for the IOMMU memory regions, a callback is called on
PCI bus enumeration that initializes for a given device
on the bus hierarchy an IOMMU memory region. The PCI bus
hierarchy is stored locally in IOMMUPciBus and IOMMUDevice
objects.

At the time of the enumeration, the bus number may not be
computed yet.

So operations that will need to retrieve the IOMMUdevice
and its IOMMU memory region from the bus number and devfn,
once the bus number is garanteed to be frozen, use an array
of IOMMUPciBus, lazily populated.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-4-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Decode the command payload
Eric Auger [Fri, 14 Feb 2020 13:27:37 +0000 (14:27 +0100)] 
virtio-iommu: Decode the command payload

This patch adds the command payload decoding and
introduces the functions that will do the actual
command handling. Those functions are not yet implemented.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-3-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-iommu: Add skeleton
Eric Auger [Fri, 14 Feb 2020 13:27:36 +0000 (14:27 +0100)] 
virtio-iommu: Add skeleton

This patchs adds the skeleton for the virtio-iommu device.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20200214132745.23392-2-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio: gracefully handle invalid region caches
Stefan Hajnoczi [Fri, 7 Feb 2020 10:46:19 +0000 (10:46 +0000)] 
virtio: gracefully handle invalid region caches

The virtqueue code sets up MemoryRegionCaches to access the virtqueue
guest RAM data structures.  The code currently assumes that
VRingMemoryRegionCaches is initialized before device emulation code
accesses the virtqueue.  An assertion will fail in
vring_get_region_caches() when this is not true.  Device fuzzing found a
case where this assumption is false (see below).

Virtqueue guest RAM addresses can also be changed from a vCPU thread
while an IOThread is accessing the virtqueue.  This breaks the same
assumption but this time the caches could become invalid partway through
the virtqueue code.  The code fetches the caches RCU pointer multiple
times so we will need to validate the pointer every time it is fetched.

Add checks each time we call vring_get_region_caches() and treat invalid
caches as a nop: memory stores are ignored and memory reads return 0.

The fuzz test failure is as follows:

  $ qemu -M pc -device virtio-blk-pci,id=drv0,drive=drive0,addr=4.0 \
         -drive if=none,id=drive0,file=null-co://,format=raw,auto-read-only=off \
         -drive if=none,id=drive1,file=null-co://,file.read-zeroes=on,format=raw \
         -display none \
         -qtest stdio
  endianness
  outl 0xcf8 0x80002020
  outl 0xcfc 0xe0000000
  outl 0xcf8 0x80002004
  outw 0xcfc 0x7
  write 0xe0000000 0x24 0x00ffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffab5cffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffabffffffab0000000001
  inb 0x4
  writew 0xe000001c 0x1
  write 0xe0000014 0x1 0x0d

The following error message is produced:

  qemu-system-x86_64: /home/stefanha/qemu/hw/virtio/virtio.c:286: vring_get_region_caches: Assertion `caches != NULL' failed.

The backtrace looks like this:

  #0  0x00007ffff5520625 in raise () at /lib64/libc.so.6
  #1  0x00007ffff55098d9 in abort () at /lib64/libc.so.6
  #2  0x00007ffff55097a9 in _nl_load_domain.cold () at /lib64/libc.so.6
  #3  0x00007ffff5518a66 in annobin_assert.c_end () at /lib64/libc.so.6
  #4  0x00005555559073da in vring_get_region_caches (vq=<optimized out>) at qemu/hw/virtio/virtio.c:286
  #5  vring_get_region_caches (vq=<optimized out>) at qemu/hw/virtio/virtio.c:283
  #6  0x000055555590818d in vring_used_flags_set_bit (mask=1, vq=0x5555575ceea0) at qemu/hw/virtio/virtio.c:398
  #7  virtio_queue_split_set_notification (enable=0, vq=0x5555575ceea0) at qemu/hw/virtio/virtio.c:398
  #8  virtio_queue_set_notification (vq=vq@entry=0x5555575ceea0, enable=enable@entry=0) at qemu/hw/virtio/virtio.c:451
  #9  0x0000555555908512 in virtio_queue_set_notification (vq=vq@entry=0x5555575ceea0, enable=enable@entry=0) at qemu/hw/virtio/virtio.c:444
  #10 0x00005555558c697a in virtio_blk_handle_vq (s=0x5555575c57e0, vq=0x5555575ceea0) at qemu/hw/block/virtio-blk.c:775
  #11 0x0000555555907836 in virtio_queue_notify_aio_vq (vq=0x5555575ceea0) at qemu/hw/virtio/virtio.c:2244
  #12 0x0000555555cb5dd7 in aio_dispatch_handlers (ctx=ctx@entry=0x55555671a420) at util/aio-posix.c:429
  #13 0x0000555555cb67a8 in aio_dispatch (ctx=0x55555671a420) at util/aio-posix.c:460
  #14 0x0000555555cb307e in aio_ctx_dispatch (source=<optimized out>, callback=<optimized out>, user_data=<optimized out>) at util/async.c:260
  #15 0x00007ffff7bbc510 in g_main_context_dispatch () at /lib64/libglib-2.0.so.0
  #16 0x0000555555cb5848 in glib_pollfds_poll () at util/main-loop.c:219
  #17 os_host_main_loop_wait (timeout=<optimized out>) at util/main-loop.c:242
  #18 main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:518
  #19 0x00005555559b20c9 in main_loop () at vl.c:1683
  #20 0x0000555555838115 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at vl.c:4441

Reported-by: Alexander Bulekov <alxndr@bu.edu>
Cc: Michael Tsirkin <mst@redhat.com>
Cc: Cornelia Huck <cohuck@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200207104619.164892-1-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agoblock/nbd: fix memory leak in nbd_open()
Pan Nengyuan [Thu, 5 Dec 2019 03:45:28 +0000 (11:45 +0800)] 
block/nbd: fix memory leak in nbd_open()

In currently implementation there will be a memory leak when
nbd_client_connect() returns error status. Here is an easy way to
reproduce:

1. run qemu-iotests as follow and check the result with asan:
    ./check -raw 143

Following is the asan output backtrack:
Direct leak of 40 byte(s) in 1 object(s) allocated from:
    #0 0x7f629688a560 in calloc (/usr/lib64/libasan.so.3+0xc7560)
    #1 0x7f6295e7e015 in g_malloc0  (/usr/lib64/libglib-2.0.so.0+0x50015)
    #2 0x56281dab4642 in qobject_input_start_struct  /mnt/sdb/qemu-4.2.0-rc0/qapi/qobject-input-visitor.c:295
    #3 0x56281dab1a04 in visit_start_struct  /mnt/sdb/qemu-4.2.0-rc0/qapi/qapi-visit-core.c:49
    #4 0x56281dad1827 in visit_type_SocketAddress  qapi/qapi-visit-sockets.c:386
    #5 0x56281da8062f in nbd_config   /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1716
    #6 0x56281da8062f in nbd_process_options /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1829
    #7 0x56281da8062f in nbd_open /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1873

Direct leak of 15 byte(s) in 1 object(s) allocated from:
    #0 0x7f629688a3a0 in malloc (/usr/lib64/libasan.so.3+0xc73a0)
    #1 0x7f6295e7dfbd in g_malloc (/usr/lib64/libglib-2.0.so.0+0x4ffbd)
    #2 0x7f6295e96ace in g_strdup (/usr/lib64/libglib-2.0.so.0+0x68ace)
    #3 0x56281da804ac in nbd_process_options /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1834
    #4 0x56281da804ac in nbd_open /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1873

Indirect leak of 24 byte(s) in 1 object(s) allocated from:
    #0 0x7f629688a3a0 in malloc (/usr/lib64/libasan.so.3+0xc73a0)
    #1 0x7f6295e7dfbd in g_malloc (/usr/lib64/libglib-2.0.so.0+0x4ffbd)
    #2 0x7f6295e96ace in g_strdup (/usr/lib64/libglib-2.0.so.0+0x68ace)
    #3 0x56281dab41a3 in qobject_input_type_str_keyval /mnt/sdb/qemu-4.2.0-rc0/qapi/qobject-input-visitor.c:536
    #4 0x56281dab2ee9 in visit_type_str /mnt/sdb/qemu-4.2.0-rc0/qapi/qapi-visit-core.c:297
    #5 0x56281dad0fa1 in visit_type_UnixSocketAddress_members qapi/qapi-visit-sockets.c:141
    #6 0x56281dad17b6 in visit_type_SocketAddress_members qapi/qapi-visit-sockets.c:366
    #7 0x56281dad186a in visit_type_SocketAddress qapi/qapi-visit-sockets.c:393
    #8 0x56281da8062f in nbd_config /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1716
    #9 0x56281da8062f in nbd_process_options /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1829
    #10 0x56281da8062f in nbd_open /mnt/sdb/qemu-4.2.0-rc0/block/nbd.c:1873

Fixes: 8f071c9db506e03ab
Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Cc: qemu-stable <qemu-stable@nongnu.org>
Cc: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <1575517528-44312-3-git-send-email-pannengyuan@huawei.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
19 months agoblock/nbd: extract the common cleanup code
Pan Nengyuan [Thu, 5 Dec 2019 03:45:27 +0000 (11:45 +0800)] 
block/nbd: extract the common cleanup code

The BDRVNBDState cleanup code is common in two places, add
nbd_clear_bdrvstate() function to do these cleanups.

Suggested-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <1575517528-44312-2-git-send-email-pannengyuan@huawei.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: fix compilation error and commit message]
Signed-off-by: Eric Blake <eblake@redhat.com>
19 months agonbd-client: Support leading / in NBD URI
Eric Blake [Wed, 12 Feb 2020 02:31:01 +0000 (20:31 -0600)] 
nbd-client: Support leading / in NBD URI

The NBD URI specification [1] states that only one leading slash at
the beginning of the URI path component is stripped, not all such
slashes.  This becomes important to a patch I just proposed to nbdkit
[2], which would allow the exportname to select a file embedded within
an ext2 image: ext2fs demands an absolute pathname beginning with '/',
and because qemu was inadvertantly stripping it, my nbdkit patch had
to work around the behavior.

[1] https://github.com/NetworkBlockDevice/nbd/blob/master/doc/uri.md
[2] https://www.redhat.com/archives/libguestfs/2020-February/msg00109.html

Note that the qemu bug only affects handling of URIs such as
nbd://host:port//abs/path (where '/abs/path' should be the export
name); it is still possible to use --image-opts and pass the desired
export name with a leading slash directly through JSON even without
this patch.

Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200212023101.1162686-1-eblake@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
19 months agonbd: Fix regression with multiple meta contexts
Eric Blake [Thu, 6 Feb 2020 17:38:32 +0000 (11:38 -0600)] 
nbd: Fix regression with multiple meta contexts

Detected by a hang in the libnbd testsuite.  If a client requests
multiple meta contexts (both base:allocation and qemu:dirty-bitmap:x)
at the same time, our attempt to silence a false-positive warning
about a potential uninitialized variable introduced botched logic: we
were short-circuiting the second context, and never sending the
NBD_REPLY_FLAG_DONE.  Combining two 'if' into one 'if/else' in
bdf200a55 was wrong (I'm a bit embarrassed that such a change was my
initial suggestion after the v1 patch, then I did not review the v2
patch that actually got committed). Revert that, and instead silence
the false positive warning by replacing 'return ret' with 'return 0'
(the value it always has at that point in the code, even though it
eluded the deduction abilities of the robot that reported the false
positive).

Fixes: bdf200a5535
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200206173832.130004-1-eblake@redhat.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
19 months agotests/tcg: take into account expected clashes pauth-4
Alex Bennée [Tue, 25 Feb 2020 12:47:10 +0000 (12:47 +0000)] 
tests/tcg: take into account expected clashes pauth-4

Pointer authentication isn't perfect so measure the percentage of
failed checks. As we want to vary the pointer we work through a bunch
of different addresses.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200225124710.14152-20-alex.bennee@linaro.org>

19 months agotests/tcg: fix typo in configure.sh test for v8.3
Alex Bennée [Tue, 25 Feb 2020 12:47:09 +0000 (12:47 +0000)] 
tests/tcg: fix typo in configure.sh test for v8.3

Although most people use the docker images this can trip up on
developer systems with actual valid cross-compilers!

Fixes: bb516dfc5b3
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200225124710.14152-19-alex.bennee@linaro.org>

19 months agotcg: save vaddr temp for plugin usage
Alex Bennée [Tue, 25 Feb 2020 17:49:08 +0000 (17:49 +0000)] 
tcg: save vaddr temp for plugin usage

While do_gen_mem_cb does copy (via extu_tl_i64) vaddr into a new temp
this won't help if the vaddr temp gets clobbered by the actual
load/store op. To avoid this clobbering we explicitly copy vaddr
before the op to ensure it is live my the time we do the
instrumentation.

Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Emilio G. Cota <cota@braap.org>
Cc: qemu-stable@nongnu.org
Message-Id: <20200225124710.14152-18-alex.bennee@linaro.org>

19 months agotests/tcg: give debug builds a little bit longer
Alex Bennée [Tue, 25 Feb 2020 12:47:07 +0000 (12:47 +0000)] 
tests/tcg: give debug builds a little bit longer

When combined with heavy plugins we occasionally hit the timeouts.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200225124710.14152-17-alex.bennee@linaro.org>

19 months agotests/plugins: make howvec clean-up after itself.
Alex Bennée [Tue, 25 Feb 2020 12:47:06 +0000 (12:47 +0000)] 
tests/plugins: make howvec clean-up after itself.

TCG plugins are responsible for their own memory usage and although
the plugin_exit is tied to the end of execution in this case it is
still poor practice. Ensure we delete the hash table and related data
when we are done to be a good plugin citizen.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200225124710.14152-16-alex.bennee@linaro.org>

19 months agotarget/riscv: progressively load the instruction during decode
Alex Bennée [Tue, 25 Feb 2020 12:47:05 +0000 (12:47 +0000)] 
target/riscv: progressively load the instruction during decode

The plugin system would throw up a harmless warning when it detected
that a disassembly of an instruction didn't use all it's bytes. Fix
the riscv decoder to only load the instruction bytes it needs as it
needs them.

This drops opcode from the ctx in favour if passing the appropriately
sized opcode down a few levels of the decode.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Message-Id: <20200225124710.14152-15-alex.bennee@linaro.org>

19 months agoqemu/bitops.h: Add extract8 and extract16
Yoshinori Sato [Tue, 25 Feb 2020 12:47:04 +0000 (12:47 +0000)] 
qemu/bitops.h: Add extract8 and extract16

Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200212130311.127515-3-ysato@users.sourceforge.jp>
Message-Id: <20200225124710.14152-14-alex.bennee@linaro.org>

19 months agotests/plugin: prevent uninitialized warning
Chen Qun [Tue, 25 Feb 2020 12:47:03 +0000 (12:47 +0000)] 
tests/plugin: prevent uninitialized warning

According to the glibc function requirements, we need initialise
 the variable. Otherwise there will be compilation warnings:

glib-autocleanups.h:28:3: warning: ‘out’ may be
used uninitialized in this function [-Wmaybe-uninitialized]
   g_free (*pp);
   ^~~~~~~~~~~~

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Chen Qun <kuhn.chenqun@huawei.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20200206093238.203984-1-kuhn.chenqun@huawei.com>
[AJB: uses Thomas's single line allocation]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200225124710.14152-13-alex.bennee@linaro.org>

19 months agoplugins/core: add missing break in cb_to_tcg_flags
Emilio G. Cota [Tue, 25 Feb 2020 12:47:02 +0000 (12:47 +0000)] 
plugins/core: add missing break in cb_to_tcg_flags

Fixes: 54cb65d8588
Reported-by: Robert Henry <robhenry@microsoft.com>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200105072940.32204-1-cota@braap.org>
Cc: qemu-stable@nongnu.org
Message-Id: <20200225124710.14152-12-alex.bennee@linaro.org>

19 months agodocs/devel: document query handle lifetimes
Alex Bennée [Tue, 25 Feb 2020 12:47:01 +0000 (12:47 +0000)] 
docs/devel: document query handle lifetimes

I forgot to document the lifetime of handles in the developer
documentation. Do so now.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Message-Id: <20200225124710.14152-11-alex.bennee@linaro.org>

19 months agotracing: only allow -trace to override -D if set
Alex Bennée [Tue, 25 Feb 2020 12:47:00 +0000 (12:47 +0000)] 
tracing: only allow -trace to override -D if set

Otherwise any -D settings the user may have made get ignored.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Message-Id: <20200225124710.14152-10-alex.bennee@linaro.org>

19 months agotests/iotests: be a little more forgiving on the size test
Alex Bennée [Tue, 25 Feb 2020 20:20:09 +0000 (20:20 +0000)] 
tests/iotests: be a little more forgiving on the size test

At least on ZFS this was failing as 512 was less than or equal to 512.
I suspect the reason is additional compression done by ZFS and however
qemu-img gets the actual size.

Loosen the criteria to make sure after is not bigger than before and
also dump the values in the report.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200225124710.14152-9-alex.bennee@linaro.org>

19 months agotravis.yml: single-thread build-tcg stages
Alex Bennée [Tue, 25 Feb 2020 12:46:58 +0000 (12:46 +0000)] 
travis.yml: single-thread build-tcg stages

This still seems to be a problem for Travis.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200225124710.14152-8-alex.bennee@linaro.org>

19 months agotravis.yml: Fix Travis YAML configuration warnings
Wainer dos Santos Moschetta [Tue, 25 Feb 2020 12:46:57 +0000 (12:46 +0000)] 
travis.yml: Fix Travis YAML configuration warnings

This fixes the following warnings Travis has detected on the
YAML configuration:

- 'on root: missing os, using the default "linux"'
- 'on root: the key matrix is an alias for jobs, using jobs'
- 'on jobs.include.python: unexpected sequence, using the first value (3.5)'
- 'on jobs.include.python: unexpected sequence, using the first value (3.6)'

Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200207210124.141119-2-wainersm@redhat.com>
Message-Id: <20200225124710.14152-7-alex.bennee@linaro.org>

19 months agotravis.yml: Test the s390-ccw build, too
Thomas Huth [Tue, 25 Feb 2020 12:46:56 +0000 (12:46 +0000)] 
travis.yml: Test the s390-ccw build, too

Since we can now use a s390x host on Travis, we can also build and
test the s390-ccw bios images there. For this we have to make sure
that roms/SLOF is checked out, too, and then move the generated *.img
files to the right location before running the tests.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20200206202543.7085-1-thuth@redhat.com>
Message-Id: <20200225124710.14152-6-alex.bennee@linaro.org>

19 months agotests/rcutorture: mild documenting refactor of update thread
Alex Bennée [Tue, 25 Feb 2020 12:46:55 +0000 (12:46 +0000)] 
tests/rcutorture: mild documenting refactor of update thread

This is mainly to help with reasoning what the test is trying to do.
We can move rcu_stress_idx to a local variable as there is only ever
one updater thread. I've also added an assert to catch the case where
we end up updating the current structure to itself which is the only
way I can see the mberror cases we are seeing on Travis.

We shall see if the rcutorture test failures go away now.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200225124710.14152-5-alex.bennee@linaro.org>

19 months agotests/rcutorture: better document locking of stats
Alex Bennée [Tue, 25 Feb 2020 12:46:54 +0000 (12:46 +0000)] 
tests/rcutorture: better document locking of stats

This is pure code motion with no functional effect.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200225124710.14152-4-alex.bennee@linaro.org>

19 months agotests/rcutorture: update usage hint
Alex Bennée [Tue, 25 Feb 2020 12:46:53 +0000 (12:46 +0000)] 
tests/rcutorture: update usage hint

Although documented in the comments we don't display all the various
invocations we can in the usage.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200225124710.14152-3-alex.bennee@linaro.org>

19 months agotests/tcg: include a skip runner for pauth3 with plugins
Alex Bennée [Tue, 25 Feb 2020 12:46:52 +0000 (12:46 +0000)] 
tests/tcg: include a skip runner for pauth3 with plugins

If we have plugins enabled we still need to have built the test to be
able to run it.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Robert Foley <robert.foley@linaro.org>
Message-Id: <20200225124710.14152-2-alex.bennee@linaro.org>

19 months agovhost-user-blk: convert to new virtio_delete_queue
Pan Nengyuan [Mon, 24 Feb 2020 04:13:36 +0000 (12:13 +0800)] 
vhost-user-blk: convert to new virtio_delete_queue

use the new virtio_delete_queue function to cleanup.

Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Message-Id: <20200224041336.30790-3-pannengyuan@huawei.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovhost-user-blk: delete virtioqueues in unrealize to fix memleaks
Pan Nengyuan [Mon, 24 Feb 2020 04:13:35 +0000 (12:13 +0800)] 
vhost-user-blk: delete virtioqueues in unrealize to fix memleaks

virtio queues forgot to delete in unrealize, and aslo error path in
realize, this patch fix these memleaks, the leak stack is as follow:

Direct leak of 114688 byte(s) in 16 object(s) allocated from:
    #0 0x7f24024fdbf0 in calloc (/lib64/libasan.so.3+0xcabf0)
    #1 0x7f2401642015 in g_malloc0 (/lib64/libglib-2.0.so.0+0x50015)
    #2 0x55ad175a6447 in virtio_add_queue /mnt/sdb/qemu/hw/virtio/virtio.c:2327
    #3 0x55ad17570cf9 in vhost_user_blk_device_realize /mnt/sdb/qemu/hw/block/vhost-user-blk.c:419
    #4 0x55ad175a3707 in virtio_device_realize /mnt/sdb/qemu/hw/virtio/virtio.c:3509
    #5 0x55ad176ad0d1 in device_set_realized /mnt/sdb/qemu/hw/core/qdev.c:876
    #6 0x55ad1781ff9d in property_set_bool /mnt/sdb/qemu/qom/object.c:2080
    #7 0x55ad178245ae in object_property_set_qobject /mnt/sdb/qemu/qom/qom-qobject.c:26
    #8 0x55ad17821eb4 in object_property_set_bool /mnt/sdb/qemu/qom/object.c:1338
    #9 0x55ad177aeed7 in virtio_pci_realize /mnt/sdb/qemu/hw/virtio/virtio-pci.c:1801

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200224041336.30790-2-pannengyuan@huawei.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-crypto: do delete ctrl_vq in virtio_crypto_device_unrealize
Pan Nengyuan [Tue, 25 Feb 2020 07:55:54 +0000 (15:55 +0800)] 
virtio-crypto: do delete ctrl_vq in virtio_crypto_device_unrealize

Similar to other virtio-deivces, ctrl_vq forgot to delete in virtio_crypto_device_unrealize, this patch fix it.
This device has aleardy maintained vq pointers. Thus, we use the new virtio_delete_queue function directly to do the cleanup.

The leak stack:
Direct leak of 10752 byte(s) in 3 object(s) allocated from:
    #0 0x7f4c024b1970 in __interceptor_calloc (/lib64/libasan.so.5+0xef970)
    #1 0x7f4c018be49d in g_malloc0 (/lib64/libglib-2.0.so.0+0x5249d)
    #2 0x55a2f8017279 in virtio_add_queue /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/virtio.c:2333
    #3 0x55a2f8057035 in virtio_crypto_device_realize /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/virtio-crypto.c:814
    #4 0x55a2f8005d80 in virtio_device_realize /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/virtio.c:3531
    #5 0x55a2f8497d1b in device_set_realized /mnt/sdb/qemu-new/qemu_test/qemu/hw/core/qdev.c:891
    #6 0x55a2f8b48595 in property_set_bool /mnt/sdb/qemu-new/qemu_test/qemu/qom/object.c:2238
    #7 0x55a2f8b54fad in object_property_set_qobject /mnt/sdb/qemu-new/qemu_test/qemu/qom/qom-qobject.c:26
    #8 0x55a2f8b4de2c in object_property_set_bool /mnt/sdb/qemu-new/qemu_test/qemu/qom/object.c:1390
    #9 0x55a2f80609c9 in virtio_crypto_pci_realize /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/virtio-crypto-pci.c:58

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Cc: "Gonglei (Arei)" <arei.gonglei@huawei.com>
Message-Id: <20200225075554.10835-5-pannengyuan@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovirtio-pmem: do delete rq_vq in virtio_pmem_unrealize
Pan Nengyuan [Tue, 25 Feb 2020 07:55:53 +0000 (15:55 +0800)] 
virtio-pmem: do delete rq_vq in virtio_pmem_unrealize

Similar to other virtio-devices, rq_vq forgot to delete in
virtio_pmem_unrealize, this patch fix it.  This device has already
maintained a vq pointer, thus we use the new virtio_delete_queue
function directly to do the cleanup.

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Message-Id: <20200225075554.10835-4-pannengyuan@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovhost-user-fs: convert to the new virtio_delete_queue function
Pan Nengyuan [Tue, 25 Feb 2020 07:55:52 +0000 (15:55 +0800)] 
vhost-user-fs: convert to the new virtio_delete_queue function

use the new virtio_delete_queue function to cleanup.

Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200225075554.10835-3-pannengyuan@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agovhost-user-fs: do delete virtio_queues in unrealize
Pan Nengyuan [Tue, 25 Feb 2020 07:55:51 +0000 (15:55 +0800)] 
vhost-user-fs: do delete virtio_queues in unrealize

Similar to other virtio device(https://patchwork.kernel.org/patch/11399237/), virtio queues forgot to delete in unrealize, and aslo error path in realize, this patch fix these memleaks, the leak stack is as follow:
Direct leak of 57344 byte(s) in 1 object(s) allocated from:
    #0 0x7f15784fb970 in __interceptor_calloc (/lib64/libasan.so.5+0xef970)
    #1 0x7f157790849d in g_malloc0 (/lib64/libglib-2.0.so.0+0x5249d)
    #2 0x55587a1bf859 in virtio_add_queue /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/virtio.c:2333
    #3 0x55587a2071d5 in vuf_device_realize /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/vhost-user-fs.c:212
    #4 0x55587a1ae360 in virtio_device_realize /mnt/sdb/qemu-new/qemu_test/qemu/hw/virtio/virtio.c:3531
    #5 0x55587a63fb7b in device_set_realized /mnt/sdb/qemu-new/qemu_test/qemu/hw/core/qdev.c:891
    #6 0x55587acf03f5 in property_set_bool /mnt/sdb/qemu-new/qemu_test/qemu/qom/object.c:2238
    #7 0x55587acfce0d in object_property_set_qobject /mnt/sdb/qemu-new/qemu_test/qemu/qom/qom-qobject.c:26
    #8 0x55587acf5c8c in object_property_set_bool /mnt/sdb/qemu-new/qemu_test/qemu/qom/object.c:1390
    #9 0x55587a8e22a2 in pci_qdev_realize /mnt/sdb/qemu-new/qemu_test/qemu/hw/pci/pci.c:2095
    #10 0x55587a63fb7b in device_set_realized /mnt/sdb/qemu-new/qemu_test/qemu/hw/core/qdev.c:891
    #11 0x55587acf03f5 in property_set_bool /mnt/sdb/qemu-new/qemu_test/qemu/qom/object.c:2238
    #12 0x55587acfce0d in object_property_set_qobject /mnt/sdb/qemu-new/qemu_test/qemu/qom/qom-qobject.c:26
    #13 0x55587acf5c8c in object_property_set_bool /mnt/sdb/qemu-new/qemu_test/qemu/qom/object.c:1390
    #14 0x55587a496d65 in qdev_device_add /mnt/sdb/qemu-new/qemu_test/qemu/qdev-monitor.c:679

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200225075554.10835-2-pannengyuan@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agoMerge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
Peter Maydell [Tue, 25 Feb 2020 13:31:16 +0000 (13:31 +0000)] 
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* device_del fix (Julia)
* FXAM fix (myself)
* memdev refactoring (Igor)
* memory region API cleanups (Peter, Philippe)
* ioeventfd optimization (Stefan)
* new WHPX maintainer (Sunil)
* Large guest startup optimizations (Chen)

# gpg: Signature made Tue 25 Feb 2020 12:42:24 GMT
# gpg:                using RSA key BFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream: (104 commits)
  WHPX: Assigning maintainer for Windows Hypervisor Platform
  accel/kvm: Check ioctl(KVM_SET_USER_MEMORY_REGION) return value
  target/i386: check for empty register in FXAM
  qdev-monitor: Forbid repeated device_del
  mem-prealloc: optimize large guest startup
  memory: batch allocate ioeventfds[] in address_space_update_ioeventfds()
  Avoid cpu_physical_memory_rw() with a constant is_write argument
  Let cpu_[physical]_memory() calls pass a boolean 'is_write' argument
  exec: Let cpu_[physical]_memory API use a boolean 'is_write' argument
  Avoid address_space_rw() with a constant is_write argument
  Let address_space_rw() calls pass a boolean 'is_write' argument
  exec: Let address_space_unmap() use a boolean 'is_write' argument
  hw/virtio: Let vhost_memory_map() use a boolean 'is_write' argument
  hw/virtio: Let virtqueue_map_iovec() use a boolean 'is_write' argument
  hw/ide: Let the DMAIntFunc prototype use a boolean 'is_write' argument
  hw/ide/internal: Remove unused DMARestartFunc typedef
  Remove unnecessary cast when using the cpu_[physical]_memory API
  exec: Let the cpu_[physical]_memory API use void pointer arguments
  Remove unnecessary cast when using the address_space API
  hw/net: Avoid casting non-const pointer, use address_space_write()
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agorebuild-expected-aml.sh: remind about the process
Michael S. Tsirkin [Wed, 22 Jan 2020 08:02:43 +0000 (03:02 -0500)] 
rebuild-expected-aml.sh: remind about the process

Remind users of rebuild-expected-aml.sh about the process
to follow. Suppress the warning if allowed file list exists -
that's a big hint user is already aware of the process.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agobios-tables-test: default diff command
Michael S. Tsirkin [Wed, 22 Jan 2020 07:24:54 +0000 (02:24 -0500)] 
bios-tables-test: default diff command

Most people probably just want diff -u. So let's use that
as the default.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agobios-tables-test: fix up DIFF generation
Michael S. Tsirkin [Wed, 22 Jan 2020 07:21:18 +0000 (02:21 -0500)] 
bios-tables-test: fix up DIFF generation

Turns out it goes to stdout which is suppressed even with V=1.
Force DIFF output to stderr to make it visible.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agobios-tables-test: tell people how to update
Michael S. Tsirkin [Wed, 22 Jan 2020 07:20:36 +0000 (02:20 -0500)] 
bios-tables-test: tell people how to update

For now just a pointer to the source file.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
19 months agoMerge branch 'exec_rw_const_v4' of https://github.com/philmd/qemu into HEAD
Paolo Bonzini [Tue, 25 Feb 2020 12:41:48 +0000 (13:41 +0100)] 
Merge branch 'exec_rw_const_v4' of https://github.com/philmd/qemu into HEAD

19 months agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-docs-20200225' into staging
Peter Maydell [Tue, 25 Feb 2020 11:03:46 +0000 (11:03 +0000)] 
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-docs-20200225' into staging

 * create a new 'tools' manual, and move the documentation for
   qemu-img, qemu-nbd, virtfs-proxy-helper, qemu-trace-stap and
   virtiofsd to it

# gpg: Signature made Tue 25 Feb 2020 11:00:16 GMT
# gpg:                using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg:                issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [ultimate]
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-docs-20200225:
  docs: Move tools documentation to tools manual
  docs: Move qemu-option-trace.rst.inc to docs/
  docs: Create new 'tools' manual

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agodocs: Move tools documentation to tools manual
Peter Maydell [Tue, 25 Feb 2020 10:45:14 +0000 (10:45 +0000)] 
docs: Move tools documentation to tools manual

Move the following tools documentation files to the new tools manual:

 docs/interop/qemu-img.rst
 docs/interop/qemu-nbd.rst
 docs/interop/virtfs-proxy-helper.rst
 docs/interop/qemu-trace-stap.rst
 docs/interop/virtiofsd.rst

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200217155415.30949-4-peter.maydell@linaro.org

19 months agodocs: Move qemu-option-trace.rst.inc to docs/
Peter Maydell [Tue, 25 Feb 2020 10:45:13 +0000 (10:45 +0000)] 
docs: Move qemu-option-trace.rst.inc to docs/

The qemu-option-trace.rst.inc file contains a rST documentation
fragment which describes trace options common to qemu-nbd and
qemu-img.  We put this file into interop/, but we'd like to move the
qemu-nbd and qemu-img files into the tools/ manual.  We could move
the .rst.inc file along with them, but we're eventually going to want
to use it for the main QEMU binary options documentation too, and
that will be in system/.  So move qemu-option-trace.rst.inc to the
top-level docs/ directory, where all these files can include it via

   .. include:: ../qemu-option-trace.rst.inc

This does have the slight downside that we now need to explicitly
tell Make which manuals use this file rather than relying on
a wildcard for all .rst.inc in the manual.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200217155415.30949-3-peter.maydell@linaro.org

19 months agodocs: Create new 'tools' manual
Peter Maydell [Tue, 25 Feb 2020 10:45:13 +0000 (10:45 +0000)] 
docs: Create new 'tools' manual

Some of the documentation for QEMU "tools" which are standalone
binaries like qemu-img is an awkward fit in our current 5-manual
split. We've put it into "interop", but they're not really
about interoperability.

Create a new top level manual "tools" which will be a better
home for this documentation. This commit creates an empty
initial manual; we will move the relevant documentation
files in a subsequent commit.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200217155415.30949-2-peter.maydell@linaro.org

19 months agoMerge tag 'patchew/20200219160953.13771-1-imammedo@redhat.com' of https://github...
Paolo Bonzini [Tue, 25 Feb 2020 08:19:00 +0000 (09:19 +0100)] 
Merge tag 'patchew/20200219160953.13771-1-imammedo@redhat.com' of https://github.com/patchew-project/qemu into HEAD

This series removes ad hoc RAM allocation API (memory_region_allocate_system_memory)
and consolidates it around hostmem backend. It allows to

* resolve conflicts between global -mem-prealloc and hostmem's "policy" option,
  fixing premature allocation before binding policy is applied

* simplify complicated memory allocation routines which had to deal with 2 ways
  to allocate RAM.

* reuse hostmem backends of a choice for main RAM without adding extra CLI
  options to duplicate hostmem features.  A recent case was -mem-shared, to
  enable vhost-user on targets that don't support hostmem backends [1] (ex: s390)

* move RAM allocation from individual boards into generic machine code and
  provide them with prepared MemoryRegion.

* clean up deprecated NUMA features which were tied to the old API (see patches)
  - "numa: remove deprecated -mem-path fallback to anonymous RAM"
  - (POSTPONED, waiting on libvirt side) "forbid '-numa node,mem' for 5.0 and newer machine types"
  - (POSTPONED) "numa: remove deprecated implicit RAM distribution between nodes"

Introduce a new machine.memory-backend property and wrapper code that aliases
global -mem-path and -mem-alloc into automatically created hostmem backend
properties (provided memory-backend was not set explicitly given by user).
A bulk of trivial patches then follow to incrementally convert individual
boards to using machine.memory-backend provided MemoryRegion.

Board conversion typically involves:

* providing MachineClass::default_ram_size and MachineClass::default_ram_id
  so generic code could create default backend if user didn't explicitly provide
  memory-backend or -m options

* dropping memory_region_allocate_system_memory() call

* using convenience MachineState::ram MemoryRegion, which points to MemoryRegion
   allocated by ram-memdev

On top of that for some boards:

* missing ram_size checks are added (typically it were boards with fixed ram size)

* ram_size fixups are replaced by checks and hard errors, forcing user to
  provide correct "-m" values instead of ignoring it and continuing running.

After all boards are converted, the old API is removed and memory allocation
routines are cleaned up.

19 months agoWHPX: Assigning maintainer for Windows Hypervisor Platform
Sunil Muthuswamy [Tue, 18 Feb 2020 20:38:32 +0000 (20:38 +0000)] 
WHPX: Assigning maintainer for Windows Hypervisor Platform

Signed-off-by: Sunil Muthuswamy <sunilmut@microsoft.com>
Message-Id: <SN4PR2101MB0880E245954826FD91C9D67DC0110@SN4PR2101MB0880.namprd21.prod.outlook.com>
Reviewed-by: Justin Terry (VM) <juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
19 months agoaccel/kvm: Check ioctl(KVM_SET_USER_MEMORY_REGION) return value
Philippe Mathieu-Daudé [Fri, 21 Feb 2020 16:33:36 +0000 (17:33 +0100)] 
accel/kvm: Check ioctl(KVM_SET_USER_MEMORY_REGION) return value

kvm_vm_ioctl() can fail, check its return value, and log an error
when it failed. This fixes Coverity CID 1412229:

  Unchecked return value (CHECKED_RETURN)

  check_return: Calling kvm_vm_ioctl without checking return value

Reported-by: Coverity (CID 1412229)
Fixes: 235e8982ad3 ("support using KVM_MEM_READONLY flag for regions")
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Message-Id: <20200221163336.2362-1-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
19 months agotarget/i386: check for empty register in FXAM
Paolo Bonzini [Fri, 21 Feb 2020 16:20:17 +0000 (17:20 +0100)] 
target/i386: check for empty register in FXAM

The fxam instruction returns the wrong result after fdecstp or after
an underflow.  Check fptags to handle this.

Reported-by: <chengang@emindsoft.com.cn>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
19 months agoqdev-monitor: Forbid repeated device_del
Julia Suvorova [Thu, 20 Feb 2020 16:55:56 +0000 (17:55 +0100)] 
qdev-monitor: Forbid repeated device_del

Device unplug can be done asynchronously. Thus, sending the second
device_del before the previous unplug is complete may lead to
unexpected results. On PCIe devices, this cancels the hot-unplug
process.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200220165556.39388-1-jusual@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
19 months agomem-prealloc: optimize large guest startup
bauerchen [Tue, 11 Feb 2020 09:10:35 +0000 (17:10 +0800)] 
mem-prealloc: optimize large guest startup

[desc]:
    Large memory VM starts slowly when using -mem-prealloc, and
    there are some areas to optimize in current method;

    1、mmap will be used to alloc threads stack during create page
    clearing threads, and it will attempt mm->mmap_sem for write
    lock, but clearing threads have hold read lock, this competition
    will cause threads createion very slow;

    2、methods of calcuating pages for per threads is not well;if we use
    64 threads to split 160 hugepage,63 threads clear 2page,1 thread
    clear 34 page,so the entire speed is very slow;

    to solve the first problem,we add a mutex in thread function,and
    start all threads when all threads finished createion;
    and the second problem, we spread remainder to other threads,in
    situation that 160 hugepage and 64 threads, there are 32 threads
    clear 3 pages,and 32 threads clear 2 pages.

[test]:
    320G 84c VM start time can be reduced to 10s
    680G 84c VM start time can be reduced to 18s

Signed-off-by: bauerchen <bauerchen@tencent.com>
Reviewed-by: Pan Rui <ruippan@tencent.com>
Reviewed-by: Ivan Ren <ivanren@tencent.com>
[Simplify computation of the number of pages per thread. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
19 months agomemory: batch allocate ioeventfds[] in address_space_update_ioeventfds()
Stefan Hajnoczi [Tue, 18 Feb 2020 18:22:26 +0000 (18:22 +0000)] 
memory: batch allocate ioeventfds[] in address_space_update_ioeventfds()

Reallocing the ioeventfds[] array each time an element is added is very
expensive as the number of ioeventfds increases.  Batch allocate instead
to amortize the cost of realloc.

This patch reduces Linux guest boot times from 362s to 140s when there
are 2 virtio-blk devices with 1 virtqueue and 99 virtio-blk devices with
32 virtqueues.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200218182226.913977-1-stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
19 months agoMerge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
Peter Maydell [Mon, 24 Feb 2020 11:38:54 +0000 (11:38 +0000)] 
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

Pull request

This pull request contains a virtio-blk/scsi performance optimization, event
loop scalability improvements, and a qtest-based device fuzzing framework.  I
am including the fuzzing patches because I have reviewed them and Thomas Huth
is currently away on leave.

# gpg: Signature made Sat 22 Feb 2020 08:50:05 GMT
# gpg:                using RSA key 8695A8BFD3F97CDAAC35775A9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>" [full]
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>" [full]
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/block-pull-request: (31 commits)
  fuzz: add documentation to docs/devel/
  fuzz: add virtio-scsi fuzz target
  fuzz: add virtio-net fuzz target
  fuzz: add i440fx fuzz targets
  fuzz: add configure flag --enable-fuzzing
  fuzz: add target/fuzz makefile rules
  fuzz: add support for qos-assisted fuzz targets
  fuzz: support for fork-based fuzzing.
  main: keep rcu_atfork callback enabled for qtest
  exec: keep ram block across fork when using qtest
  fuzz: add fuzzer skeleton
  libqos: move useful qos-test funcs to qos_external
  libqos: split qos-test and libqos makefile vars
  libqos: rename i2c_send and i2c_recv
  qtest: add in-process incoming command handler
  libqtest: make bufwrite rely on the TransportOps
  libqtest: add a layer of abstraction to send/recv
  qtest: add qtest_server_send abstraction
  fuzz: add FUZZ_TARGET module type
  module: check module wasn't already initialized
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
19 months agofuzz: add documentation to docs/devel/
Alexander Bulekov [Thu, 20 Feb 2020 04:11:18 +0000 (23:11 -0500)] 
fuzz: add documentation to docs/devel/

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-23-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add virtio-scsi fuzz target
Alexander Bulekov [Thu, 20 Feb 2020 04:11:17 +0000 (23:11 -0500)] 
fuzz: add virtio-scsi fuzz target

The virtio-scsi fuzz target sets up and fuzzes the available virtio-scsi
queues. After an element is placed on a queue, the fuzzer can select
whether to perform a kick, or continue adding elements.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-22-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add virtio-net fuzz target
Alexander Bulekov [Thu, 20 Feb 2020 04:11:16 +0000 (23:11 -0500)] 
fuzz: add virtio-net fuzz target

The virtio-net fuzz target feeds inputs to all three virtio-net
virtqueues, and uses forking to avoid leaking state between fuzz runs.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-21-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add i440fx fuzz targets
Alexander Bulekov [Thu, 20 Feb 2020 04:11:15 +0000 (23:11 -0500)] 
fuzz: add i440fx fuzz targets

These three targets should simply fuzz reads/writes to a couple ioports,
but they mostly serve as examples of different ways to write targets.
They demonstrate using qtest and qos for fuzzing, as well as using
rebooting and forking to reset state, or not resetting it at all.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-20-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add configure flag --enable-fuzzing
Alexander Bulekov [Thu, 20 Feb 2020 04:11:14 +0000 (23:11 -0500)] 
fuzz: add configure flag --enable-fuzzing

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-19-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add target/fuzz makefile rules
Alexander Bulekov [Thu, 20 Feb 2020 04:11:13 +0000 (23:11 -0500)] 
fuzz: add target/fuzz makefile rules

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20200220041118.23264-18-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add support for qos-assisted fuzz targets
Alexander Bulekov [Thu, 20 Feb 2020 04:11:12 +0000 (23:11 -0500)] 
fuzz: add support for qos-assisted fuzz targets

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-17-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: support for fork-based fuzzing.
Alexander Bulekov [Thu, 20 Feb 2020 04:11:11 +0000 (23:11 -0500)] 
fuzz: support for fork-based fuzzing.

fork() is a simple way to ensure that state does not leak in between
fuzzing runs. Unfortunately, the fuzzer mutation engine relies on
bitmaps which contain coverage information for each fuzzing run, and
these bitmaps should be copied from the child to the parent(where the
mutation occurs). These bitmaps are created through compile-time
instrumentation and they are not shared with fork()-ed processes, by
default. To address this, we create a shared memory region, adjust its
size and map it _over_ the counter region. Furthermore, libfuzzer
doesn't generally expose the globals that specify the location of the
counters/coverage bitmap. As a workaround, we rely on a custom linker
script which forces all of the bitmaps we care about to be placed in a
contiguous region, which is easy to locate and mmap over.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-16-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agomain: keep rcu_atfork callback enabled for qtest
Alexander Bulekov [Thu, 20 Feb 2020 04:11:10 +0000 (23:11 -0500)] 
main: keep rcu_atfork callback enabled for qtest

The qtest-based fuzzer makes use of forking to reset-state between
tests. Keep the callback enabled, so the call_rcu thread gets created
within the child process.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20200220041118.23264-15-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agoexec: keep ram block across fork when using qtest
Alexander Bulekov [Thu, 20 Feb 2020 04:11:09 +0000 (23:11 -0500)] 
exec: keep ram block across fork when using qtest

Ram blocks were marked MADV_DONTFORK breaking fuzzing-tests which
execute each test-input in a forked process.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-14-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add fuzzer skeleton
Alexander Bulekov [Thu, 20 Feb 2020 04:11:08 +0000 (23:11 -0500)] 
fuzz: add fuzzer skeleton

tests/fuzz/fuzz.c serves as the entry point for the virtual-device
fuzzer. Namely, libfuzzer invokes the LLVMFuzzerInitialize and
LLVMFuzzerTestOneInput functions, both of which are defined in this
file. This change adds a "FuzzTarget" struct, along with the
fuzz_add_target function, which should be used to define new fuzz
targets.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-13-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agolibqos: move useful qos-test funcs to qos_external
Alexander Bulekov [Thu, 20 Feb 2020 04:11:07 +0000 (23:11 -0500)] 
libqos: move useful qos-test funcs to qos_external

The moved functions are not specific to qos-test and might be useful
elsewhere. For example the virtual-device fuzzer makes use of them for
qos-assisted fuzz-targets.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-12-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agolibqos: split qos-test and libqos makefile vars
Alexander Bulekov [Thu, 20 Feb 2020 04:11:06 +0000 (23:11 -0500)] 
libqos: split qos-test and libqos makefile vars

Most qos-related objects were specified in the qos-test-obj-y variable.
qos-test-obj-y also included qos-test.o which defines a main().
This made it difficult to repurpose qos-test-obj-y to link anything
beside tests/qos-test against libqos. This change separates objects that
are libqos-specific and ones that are qos-test specific into different
variables.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20200220041118.23264-11-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agolibqos: rename i2c_send and i2c_recv
Alexander Bulekov [Thu, 20 Feb 2020 04:11:05 +0000 (23:11 -0500)] 
libqos: rename i2c_send and i2c_recv

The names i2c_send and i2c_recv collide with functions defined in
hw/i2c/core.c. This causes an error when linking against libqos and
softmmu simultaneously (for example when using qtest inproc). Rename the
libqos functions to avoid this.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Acked-by: Thomas Huth <thuth@redhat.com>
Message-id: 20200220041118.23264-10-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agoqtest: add in-process incoming command handler
Alexander Bulekov [Thu, 20 Feb 2020 04:11:04 +0000 (23:11 -0500)] 
qtest: add in-process incoming command handler

The handler allows a qtest client to send commands to the server by
directly calling a function, rather than using a file/CharBackend

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-9-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agolibqtest: make bufwrite rely on the TransportOps
Alexander Bulekov [Thu, 20 Feb 2020 04:11:03 +0000 (23:11 -0500)] 
libqtest: make bufwrite rely on the TransportOps

When using qtest "in-process" communication, qtest_sendf directly calls
a function in the server (qtest.c). Previously, bufwrite used
socket_send, which bypasses the TransportOps enabling the call into
qtest.c. This change replaces the socket_send calls with ops->send,
maintaining the benefits of the direct socket_send call, while adding
support for in-process qtest calls.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-8-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agolibqtest: add a layer of abstraction to send/recv
Alexander Bulekov [Thu, 20 Feb 2020 04:11:02 +0000 (23:11 -0500)] 
libqtest: add a layer of abstraction to send/recv

This makes it simple to swap the transport functions for qtest commands
to and from the qtest client. For example, now it is possible to
directly pass qtest commands to a server handler that exists within the
same process, without the standard way of writing to a file descriptor.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-7-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agoqtest: add qtest_server_send abstraction
Alexander Bulekov [Thu, 20 Feb 2020 04:11:01 +0000 (23:11 -0500)] 
qtest: add qtest_server_send abstraction

qtest_server_send is a function pointer specifying the handler used to
transmit data to the qtest client. In the standard configuration, this
calls the CharBackend handler, but now it is possible for other types of
handlers, e.g direct-function calls if the qtest client and server
exist within the same process (inproc)

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Acked-by: Thomas Huth <thuth@redhat.com>
Message-id: 20200220041118.23264-6-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agofuzz: add FUZZ_TARGET module type
Alexander Bulekov [Thu, 20 Feb 2020 04:11:00 +0000 (23:11 -0500)] 
fuzz: add FUZZ_TARGET module type

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-5-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agomodule: check module wasn't already initialized
Alexander Bulekov [Thu, 20 Feb 2020 04:10:59 +0000 (23:10 -0500)] 
module: check module wasn't already initialized

The virtual-device fuzzer must initialize QOM, prior to running
vl:qemu_init, so that it can use the qos_graph to identify the arguments
required to initialize a guest for libqos-assisted fuzzing. This change
prevents errors when vl:qemu_init tries to (re)initialize the previously
initialized QOM module.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20200220041118.23264-4-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agosoftmmu: split off vl.c:main() into main.c
Alexander Bulekov [Thu, 20 Feb 2020 04:10:58 +0000 (23:10 -0500)] 
softmmu: split off vl.c:main() into main.c

A program might rely on functions implemented in vl.c, but implement its
own main(). By placing main into a separate source file, there are no
complaints about duplicate main()s when linking against vl.o. For
example, the virtual-device fuzzer uses a main() provided by libfuzzer,
and needs to perform some initialization before running the softmmu
initialization. Now, main simply calls three vl.c functions which
handle the guest initialization, main loop and cleanup.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-3-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agosoftmmu: move vl.c to softmmu/
Alexander Bulekov [Thu, 20 Feb 2020 04:10:57 +0000 (23:10 -0500)] 
softmmu: move vl.c to softmmu/

Move vl.c to a separate directory, similar to linux-user/
Update the chechpatch and get_maintainer scripts, since they relied on
/vl.c for top_of_tree checks.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-id: 20200220041118.23264-2-alxndr@bu.edu
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
19 months agoaio-posix: make AioHandler dispatch O(1) with epoll
Stefan Hajnoczi [Fri, 14 Feb 2020 17:17:12 +0000 (17:17 +0000)] 
aio-posix: make AioHandler dispatch O(1) with epoll

File descriptor monitoring is O(1) with epoll(7), but
aio_dispatch_handlers() still scans all AioHandlers instead of
dispatching just those that are ready.  This makes aio_poll() O(n) with
respect to the total number of registered handlers.

Add a local ready_list to aio_poll() so that each nested aio_poll()
builds a list of handlers ready to be dispatched.  Since file descriptor
polling is level-triggered, nested aio_poll() calls also see fds that
were ready in the parent but not yet dispatched.  This guarantees that
nested aio_poll() invocations will dispatch all fds, even those that
became ready before the nested invocation.

Since only handlers ready to be dispatched are placed onto the
ready_list, the new aio_dispatch_ready_handlers() function provides O(1)
dispatch.

Note that AioContext polling is still O(n) and currently cannot be fully
disabled.  This still needs to be fixed before aio_poll() is fully O(1).

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Sergio Lopez <slp@redhat.com>
Message-id: 20200214171712.541358-6-stefanha@redhat.com
[Fix compilation error on macOS where there is no epoll(87).  The
aio_epoll() prototype was out of date and aio_add_ready_list() needed to
be moved outside the ifdef.
--Stefan]
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>