block-backend: convert blk_co_pdiscard to int64_t bytes

We updated blk_do_pdiscard() and its wrapper blk_co_pdiscard(). Both
functions are updated so that the parameter type becomes wider, so all
callers should be OK with it.

Look at blk_do_pdiscard(): bytes is passed only to
blk_check_byte_request() and bdrv_co_pdiscard(), which already have
int64_t bytes parameter, so we are OK.

Note that requests exceeding INT_MAX are still restricted by
blk_check_byte_request().

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211006131718.214235-5-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: grammar tweaks]
Signed-off-by: Eric Blake <eblake@redhat.com>

block-backend: convert blk_co_pwritev_part to int64_t bytes

We convert blk_do_pwritev_part() and some wrappers:
blk_co_pwritev_part(), blk_co_pwritev(), blk_co_pwrite_zeroes().

All functions are converted so that the parameter type becomes wider, so
all callers should be OK with it.

Look at blk_do_pwritev_part() body:
bytes is passed to:

 - trace_blk_co_pwritev (we update it here)
 - blk_check_byte_request, throttle_group_co_io_limits_intercept,
   bdrv_co_pwritev_part - all already have int64_t argument.

Note that requests exceeding INT_MAX are still restricted by
blk_check_byte_request().

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211006131718.214235-4-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: grammar tweaks]
Signed-off-by: Eric Blake <eblake@redhat.com>

block-backend: make blk_co_preadv() 64bit

For both updated functions, the type of bytes becomes wider, so all callers
should be OK with it.

blk_co_preadv() only passes its arguments to blk_do_preadv().

blk_do_preadv() passes bytes to:

 - trace_blk_co_preadv, which is updated too
 - blk_check_byte_request, throttle_group_co_io_limits_intercept,
   bdrv_co_preadv, which are already int64_t.

Note that requests exceeding INT_MAX are still restricted by
blk_check_byte_request().

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211006131718.214235-3-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: grammar tweaks]
Signed-off-by: Eric Blake <eblake@redhat.com>

block-backend: blk_check_byte_request(): int64_t bytes

Rename size and make it int64_t to correspond to modern block layer,
which always uses int64_t for offset and bytes (not in blk layer yet,
which is a task for following commits).

All callers pass int or unsigned int.

So, for bytes in [0, INT_MAX] nothing is changed, for negative bytes we
now fail on "bytes < 0" check instead of "bytes > INT_MAX" check.

Note, that blk_check_byte_request() still doesn't allow requests
exceeding INT_MAX.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211006131718.214235-2-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>

qcow2: Silence clang -m32 compiler warning

With -m32, size_t is generally only a uint32_t.  That makes clang
complain that in the assertion

  assert(qiov->size <= INT64_MAX);

the range of the type of qiov->size (size_t) is too small for any of its
values to ever exceed INT64_MAX.

Cast qiov->size to uint64_t to silence clang.

Fixes: f7ef38dd1310d7d9db76d0aa16899cbc5744f36d
       ("block: use int64_t instead of uint64_t in driver read
       handlers")
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20211011155031.149158-1-hreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>

Merge remote-tracking branch 'remotes/kwolf/tags/for-upstream' into staging

qdev: Add JSON -device

- Add a JSON mode to the -device command line option
- net/vhost-{user,vdpa}: Fix device compatibility check
- Minor iotests fixes

# gpg: Signature made Fri 15 Oct 2021 07:41:22 AM PDT
# gpg:                using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg:                issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]

* remotes/kwolf/tags/for-upstream:
  vl: Enable JSON syntax for -device
  qdev: Base object creation on QDict rather than QemuOpts
  virtio-net: Avoid QemuOpts in failover_find_primary_device()
  virtio-net: Store failover primary opts pointer locally
  qdev: Add Error parameter to hide_device() callbacks
  qemu-option: Allow deleting opts during qemu_opts_foreach()
  softmmu/qdev-monitor: add error handling in qdev_set_id
  qdev: Make DeviceState.id independent of QemuOpts
  qdev: Avoid using string visitor for properties
  iotests/051: Fix typo
  iotests/245: Fix type for iothread property
  qom: Reduce use of error_propagate()
  net/vhost-vdpa: Fix device compatibility check
  net/vhost-user: Fix device compatibility check
  net: Introduce NetClientInfo.check_peer_type()

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge remote-tracking branch 'remotes/thuth/tags/pull-request-2021-10-15' into staging

* Check kernel command line size on s390x
* Simplification of one of the SIGP instructions on s390x
* Cornelia stepping down as maintainer in some subsystems
* Update the dtc submodule to a proper release version

# gpg: Signature made Fri 15 Oct 2021 02:11:13 AM PDT
# gpg:                using RSA key 27B88847EEE0250118F3EAB92ED9D774FE702DB5
# gpg:                issuer "thuth@redhat.com"
# gpg: Good signature from "Thomas Huth <th.huth@gmx.de>" [full]
# gpg:                 aka "Thomas Huth <thuth@redhat.com>" [full]
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>" [unknown]
# gpg:                 aka "Thomas Huth <huth@tuxfamily.org>" [full]

* remotes/thuth/tags/pull-request-2021-10-15:
  dtc: Update to version 1.6.1
  s390x virtio-ccw machine: step down as maintainer
  s390x/kvm: step down as maintainer
  vfio-ccw: step down as maintainer
  s390x: sigp: Force Set Architecture to return Invalid Parameter
  s390x/ipl: check kernel command line size

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

vl: Enable JSON syntax for -device

Like we already do for -object, introduce support for JSON syntax in
-device, which can be kept stable in the long term and guarantees that a
single code path with identical behaviour is used for both QMP and the
command line. Compared to the QemuOpts based code, the parser contains
less surprises and has support for non-scalar options (lists and
structs). Switching management tools to JSON means that we can more
easily change the "human" CLI syntax from QemuOpts to the keyval parser
later.

In the QAPI schema, a feature flag is added to the device-add command to
allow management tools to detect support for this.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20211008133442.141332-16-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qdev: Base object creation on QDict rather than QemuOpts

QDicts are both what QMP natively uses and what the keyval parser
produces. Going through QemuOpts isn't useful for either one, so switch
the main device creation function to QDicts. By sharing more code with
the -object/object-add code path, we can even reduce the code size a
bit.

This commit doesn't remove the detour through QemuOpts from any code
path yet, but it allows the following commits to do so.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-15-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

virtio-net: Avoid QemuOpts in failover_find_primary_device()

Don't go through the global QemuOptsList, it is state of the legacy
command line parser and we will create devices that are not contained
in it. It is also just the command line configuration and not
necessarily the current runtime state.

Instead, look at the qdev device tree which has the current state of all
existing devices.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-14-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

virtio-net: Store failover primary opts pointer locally

Instead of accessing the global QemuOptsList, which really belong to the
command line parser and shouldn't be accessed from devices, store a
pointer to the QemuOpts in a new VirtIONet field.

This is not the final state, but just an intermediate step to get rid of
QemuOpts in devices. It will later be replaced with an options QDict.

Before this patch, two "primary" devices could be hidden for the same
standby device, but only one of them would actually be enabled and the
other one would be kept hidden forever, so this doesn't make sense.
After this patch, configuring a second primary device is an error.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-13-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qdev: Add Error parameter to hide_device() callbacks

hide_device() is used for virtio-net failover, where the standby virtio
device delays creation of the primary device. It only makes sense to
have a single primary device for each standby device. Adding a second
one should result in an error instead of hiding it and never using it
afterwards.

Prepare for this by adding an Error parameter to the hide_device()
callback where virtio-net is informed about adding a primary device.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-12-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-option: Allow deleting opts during qemu_opts_foreach()

Use QTAILQ_FOREACH_SAFE() so that the current QemuOpts can be deleted
while iterating through the whole list.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211008133442.141332-11-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

softmmu/qdev-monitor: add error handling in qdev_set_id

qdev_set_id() is mostly used when the user adds a device (using
-device cli option or device_add qmp command). This commit adds
an error parameter to handle the case where the given id is
already taken.

Also document the function and add a return value in order to
be able to capture success/failure: the function now returns the
id in case of success, or NULL in case of failure.

The commit modifies the 2 calling places (qdev-monitor and
xen-legacy-backend) to add the error object parameter.

Note that the id is, right now, guaranteed to be unique because
all ids came from the "device" QemuOptsList where the id is used
as key. This addition is a preparation for a future commit which
will relax the uniqueness.

Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-10-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qdev: Make DeviceState.id independent of QemuOpts

DeviceState.id is a pointer to a string that is stored in the QemuOpts
object DeviceState.opts and freed together with it. We want to create
devices without going through QemuOpts in the future, so make this a
separately allocated string.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211008133442.141332-9-kwolf@redhat.com>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qdev: Avoid using string visitor for properties

The only thing the string visitor adds compared to a keyval visitor is
list support. git grep for 'visit_start_list' and 'visit.*List' shows
that devices don't make use of this.

In a world with a QAPIfied command line interface, the keyval visitor is
used to parse the command line. In order to make sure that no devices
start using this feature that would make backwards compatibility harder,
just switch away from object_property_parse(), which internally uses the
string visitor, to a keyval visitor and object_property_set().

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20211008133442.141332-8-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests/051: Fix typo

The iothread isn't called 'iothread0', but 'thread0'. Depending on the
order that properties are parsed, the error message may change from the
expected one to another one saying that the iothread doesn't exist.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211008133442.141332-7-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests/245: Fix type for iothread property

iothread is a string property, so None (= JSON null) is not a valid
value for it. Pass the empty string instead to get the default iothread.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20211008133442.141332-6-kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qom: Reduce use of error_propagate()

ERRP_GUARD() makes debugging easier by making sure that &error_abort
still fails at the real origin of the error instead of
error_propagate().

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-5-kwolf@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

net/vhost-vdpa: Fix device compatibility check

vhost-vdpa works only with specific devices. At startup, it second
guesses what the command line option handling will do and error out if
it thinks a non-virtio device will attach to them.

This second guessing is not only ugly, it can lead to wrong error
messages ('-device floppy,netdev=foo' should complain about an unknown
property, not about the wrong kind of network device being attached) and
completely ignores hotplugging.

Drop the old checks and implement .check_peer_type() instead to fix
this. As a nice side effect, it also removes one more dependency on the
legacy QemuOpts infrastructure and even reduces the code size.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-4-kwolf@redhat.com>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

net/vhost-user: Fix device compatibility check

vhost-user works only with specific devices. At startup, it second
guesses what the command line option handling will do and error out if
it thinks a non-virtio device will attach to them.

This second guessing is not only ugly, it can lead to wrong error
messages ('-device floppy,netdev=foo' should complain about an unknown
property, not about the wrong kind of network device being attached) and
completely ignores hotplugging.

Drop the old checks and implement .check_peer_type() instead to fix
this. As a nice side effect, it also removes one more dependency on the
legacy QemuOpts infrastructure and even reduces the code size.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-3-kwolf@redhat.com>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

net: Introduce NetClientInfo.check_peer_type()

Some network backends (vhost-user and vhost-vdpa) work only with
specific devices. At startup, they second guess what the command line
option handling will do and error out if they think a non-virtio device
will attach to them.

This second guessing is not only ugly, it can lead to wrong error
messages ('-device floppy,netdev=foo' should complain about an unknown
property, not about the wrong kind of network device being attached) and
completely ignores hotplugging.

Add a callback where backends can check compatibility with a device when
it actually tries to attach, even on hotplug.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20211008133442.141332-2-kwolf@redhat.com>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* Meson conversions + introspection-based command line parser

# gpg: Signature made Thu 14 Oct 2021 12:51:54 AM PDT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]

* remotes/bonzini/tags/for-upstream: (26 commits)
  configure: automatically parse command line for meson -D options
  meson-buildoptions: include list of tracing backends
  configure: prepare for auto-generated option parsing
  configure: accept "internal" for --enable-capstone/slirp/fdt
  configure: remove deprecated --{enable, disable}-git-update
  configure, meson: move more compiler checks to Meson
  configure: remove obsolete Solaris ar check
  configure, meson: move Spice configure handling to meson
  configure, meson: move netmap detection to meson
  configure, meson: move vde detection to meson
  configure, meson: move libaio check to meson.build
  configure, meson: move pthread_setname_np checks to Meson
  configure, meson: move remaining HAVE_* compiler tests to Meson
  meson: HAVE_GDB_BIN is not used by C code
  configure, meson: remove CONFIG_GCOV from config-host.mak
  configure, meson: get HOST_WORDS_BIGENDIAN via the machine object
  configure, meson: move CONFIG_HOST_DSOSUF to Meson
  trace: move configuration from configure to Meson
  trace: simple: pass trace_file unmodified to config-host.h
  configure, meson: move fuzzing configuration to Meson
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

configure: automatically parse command line for meson -D options

Right now meson_options.txt lists about 90 options.  Each option
needs code in configure to parse it and pass the option down to Meson as
a -D command-line argument; in addition the default must be duplicated
between configure and meson_options.txt.  This series tries to remove
the code duplication by generating the case statement for those --enable
and --disable options, as well as the corresponding help text.

About 80% of the options can be handled completely by the new mechanism.
Eight meson options are not of the --enable/--disable kind.  Six more need
to be parsed in configure for various reasons documented in the patch,
but they still have their help automatically generated.

The advantages are:

- less code in configure

- parsing and help is more consistent (for example --enable-blobs was
  not supported)

- options are described entirely in one place, meson_options.txt.
  This make it more attractive to use Meson options instead of
  hand-crafted configure options and config-host.mak

A few options change name: --enable-tcmalloc and --enable-jemalloc
become --enable-malloc={tcmalloc,jemalloc}; --disable-blobs becomes
--disable-install-blobs; --enable-trace-backend becomes
--enable-trace-backends.  However, the old names are allowed
for backwards compatibility.

Message-Id: <20211007130829.632254-19-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson-buildoptions: include list of tracing backends

Manually patch the introspection data to include the tracing backends.
This works around a deficiency in Meson that will be fixed by
https://github.com/mesonbuild/meson/pull/9395.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: prepare for auto-generated option parsing

Prepare the configure script and Makefile for automatically generated
help and parsing.

Because we need to run the script to generate the full help, we
cannot rely on the user supplying the path to a Python interpreter
with --python; therefore, the introspection output is parsed into
shell functions and stored in scripts/.  The converter is written
in Python as standard for QEMU, and this commit contains a stub.

Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20211007130829.632254-18-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: accept "internal" for --enable-capstone/slirp/fdt

Options such as "--enable-capstone=git" do not make much sense when building
from a tarball.  Accept "internal" for consistency with the meson options.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-17-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: remove deprecated --{enable, disable}-git-update

The options were deprecated in 6.0.  That said, we do not really have a
formal deprecation cycle for build-time changes, since they do not affect
users.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-16-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move more compiler checks to Meson

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-15-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: remove obsolete Solaris ar check

Meson already has its own logic to find the "ar" binary, so remove the
Solaris specific check.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20211007130829.632254-14-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move Spice configure handling to meson

Add meson feature options for Spice and Spice protocol, and move
detection logic out of configure.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007102453.978041-1-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20211007130829.632254-13-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move netmap detection to meson

Message-Id: <20211007130829.632254-12-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move vde detection to meson

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-11-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move libaio check to meson.build

Message-Id: <20211007130829.632254-10-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move pthread_setname_np checks to Meson

This makes the pthreads check dead in configure, so remove it
as well.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-9-pbonzini@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move remaining HAVE_* compiler tests to Meson

Remove some special cases by moving them to Meson.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-8-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: HAVE_GDB_BIN is not used by C code

It is only used by the TCG tests, remove it from config-host.h.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: remove CONFIG_GCOV from config-host.mak

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-7-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: get HOST_WORDS_BIGENDIAN via the machine object

No need to pass it in config-host.mak.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007130829.632254-6-pbonzini@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move CONFIG_HOST_DSOSUF to Meson

This is just a constant string, there is no need to pass it in config-host.mak.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007130829.632254-5-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

trace: move configuration from configure to Meson

Cc: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-4-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

trace: simple: pass trace_file unmodified to config-host.h

Add the suffix directly in trace/simple.c, so that quoting is done
properly by Meson.

Cc: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130829.632254-3-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move fuzzing configuration to Meson

Cc: Alexander Oleinik <alxndr@bu.edu>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <20211007130829.632254-2-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

kconfig: split CONFIG_SPARSE_MEM from fuzzing

Pass CONFIG_FUZZ via host_kconfig, and use it to select the
sparse-mem device.

Cc: Alexander Oleinik <alxndr@bu.edu>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <20211007130829.632254-1-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: add command line options for audio drivers

Handle the choice of audio drivers the same as all other dependencies.

Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130630.632028-6-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: define symbols for all available audio drivers

All drivers are now built by default if the corresponding libraries
are available, similar to how all other modules behave;
--audio-drv-list only governs the default choice of the audio driver.

Adjust the CONFIG_AUDIO_* preprocessor symbols so that they are
based on library availability rather than --audio-drv-list, so that
the tests and -audiodev help follow the new logic.

Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20211007130630.632028-5-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: move audio driver detection to Meson

This brings a change that makes audio drivers more similar to all
other modules.  All drivers are built by default, while
--audio-drv-list only governs the default choice of the audio driver.

Meson options are added to disable the drivers, and the next patches
will fix the help messages and command line options, and especially
make the non-default drivers available via -audiodev.

Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130630.632028-4-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

audio: remove CONFIG_AUDIO_WIN_INT

Ever since winwaveaudio was removed in 2015, CONFIG_AUDIO_WIN_INT
is only set if dsound is in use, so use CONFIG_AUDIO_DSOUND directly.

Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130630.632028-3-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: remove --oss-lib

OSS is a kernel API, so the option should not be needed.  The library
is used on NetBSD, where OSS is emulated, so keep the variable.

Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20211007130630.632028-2-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

dtc: Update to version 1.6.1

The dtc submodule is currently pointing to non-release commit. It's nicer
if submodules point to release versions instead and since dtc 1.6.1 is
available now, let's update to that version.

Message-Id: <20210827120901.150276-4-thuth@redhat.com>
Acked-by: Greg Kurz <groug@kaod.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20211013' into staging

Use MO_128 for 16-byte atomic memory operations.
Add cpu_ld/st_mmu memory primitives.
Move helper_ld/st memory helpers out of tcg.h.
Canonicalize alignment flags in MemOp.

# gpg: Signature made Wed 13 Oct 2021 10:48:45 AM PDT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate]

* remotes/rth/tags/pull-tcg-20211013:
  tcg: Canonicalize alignment flags in MemOp
  tcg: Move helper_*_mmu decls to tcg/tcg-ldst.h
  target/arm: Use cpu_*_mmu instead of helper_*_mmu
  target/sparc: Use cpu_*_mmu instead of helper_*_mmu
  target/s390x: Use cpu_*_mmu instead of helper_*_mmu
  target/mips: Use 8-byte memory ops for msa load/store
  target/mips: Use cpu_*_data_ra for msa load/store
  accel/tcg: Move cpu_atomic decls to exec/cpu_ldst.h
  accel/tcg: Add cpu_{ld,st}*_mmu interfaces
  target/hexagon: Implement cpu_mmu_index
  target/s390x: Use MO_128 for 16 byte atomics
  target/ppc: Use MO_128 for 16 byte atomics
  target/i386: Use MO_128 for 16 byte atomics
  target/arm: Use MO_128 for 16 byte atomics
  memory: Log access direction for invalid accesses

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge remote-tracking branch 'remotes/rth/tags/pull-axp-20211013' into staging

Cleanup alpha memory ops prior to prctl PR_SET_UNALIGN

# gpg: Signature made Wed 13 Oct 2021 10:34:10 AM PDT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate]

* remotes/rth/tags/pull-axp-20211013:
  target/alpha: Reorg integer memory operations
  target/alpha: Reorg fp memory operations

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/alpha: Reorg integer memory operations

Pass in the MemOp instead of a callback.
Drop the fp argument; add a locked argument.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/alpha: Reorg fp memory operations

Pass in the context to each mini-helper, instead of an
incorrectly named "flags".  Separate gen_load_fp and
gen_store_fp, away from the integer helpers.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Canonicalize alignment flags in MemOp

Having observed e.g. al8+leq in dumps, canonicalize to al+leq.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Move helper_*_mmu decls to tcg/tcg-ldst.h

These functions have been replaced by cpu_*_mmu as the
most proper interface to use from target code.

Hide these declarations from code that should not use them.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Use cpu_*_mmu instead of helper_*_mmu

The helper_*_mmu functions were the only thing available
when this code was written.  This could have been adjusted
when we added cpu_*_mmuidx_ra, but now we can most easily
use the newest set of interfaces.

Cc: qemu-arm@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/sparc: Use cpu_*_mmu instead of helper_*_mmu

The helper_*_mmu functions were the only thing available
when this code was written.  This could have been adjusted
when we added cpu_*_mmuidx_ra, but now we can most easily
use the newest set of interfaces.

Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/s390x: Use cpu_*_mmu instead of helper_*_mmu

The helper_*_mmu functions were the only thing available
when this code was written.  This could have been adjusted
when we added cpu_*_mmuidx_ra, but now we can most easily
use the newest set of interfaces.

Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/mips: Use 8-byte memory ops for msa load/store

Rather than use 4-16 separate operations, use 2 operations
plus some byte reordering as necessary.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/mips: Use cpu_*_data_ra for msa load/store

We should not have been using the helper_ret_* set of
functions, as they are supposed to be private to tcg.
Nor should we have been using the plain cpu_*_data set
of functions, as they do not handle unwinding properly.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Move cpu_atomic decls to exec/cpu_ldst.h

The previous placement in tcg/tcg.h was not logical.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Add cpu_{ld,st}*_mmu interfaces

These functions are much closer to the softmmu helper
functions, in that they take the complete MemOpIdx,
and from that they may enforce required alignment.

The previous cpu_ldst.h functions did not have alignment info,
and so did not enforce it.  Retain this by adding MO_UNALN to
the MemOp that we create in calling the new functions.

Note that we are not yet enforcing alignment for user-only,
but we now have the information with which to do so.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/hexagon: Implement cpu_mmu_index

The function is trivial for user-only, but still must be present.

Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/s390x: Use MO_128 for 16 byte atomics

Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/ppc: Use MO_128 for 16 byte atomics

Cc: qemu-ppc@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/i386: Use MO_128 for 16 byte atomics

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Use MO_128 for 16 byte atomics

Cc: qemu-arm@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

memory: Log access direction for invalid accesses

In memory_region_access_valid() invalid accesses are logged to help
debugging but the log message does not say if it was a read or write.
Log that too to better identify the access causing the problem.

Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Message-Id: <20211011173616.F1DE0756022@zero.eik.bme.hu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

s390x virtio-ccw machine: step down as maintainer

I currently don't have time to work on the s390x virtio-ccw machine
anymore, so let's step down. (I will, however, continue as a
maintainer for the virtio-ccw *transport*.)

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Acked-by: Halil Pasic <pasic@linux.ibm.com>
Message-Id: <20211012144040.360887-4-cohuck@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

s390x/kvm: step down as maintainer

I'm no longer involved with KVM/s390 on the kernel side, and I don't
have enough resources to work on the s390 KVM cpus support, so I'll
step down.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Acked-by: Halil Pasic <pasic@linux.ibm.com>
Message-Id: <20211012144040.360887-3-cohuck@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

vfio-ccw: step down as maintainer

I currently don't have time to act as vfio-ccw maintainer anymore,
so remove myself there.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20211012144040.360887-2-cohuck@redhat.com>
Acked-by: Matthew Rosato <mjrosato@linux.ibm.com>
Acked-by: Eric Farman <farman@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

s390x: sigp: Force Set Architecture to return Invalid Parameter

According to the Principles of Operation, the SIGP Set Architecture
order will return Incorrect State if some CPUs are not stopped, but
only if the CZAM facility is not present. If it is, the order will
return Invalid Parameter because the architecture mode cannot be
changed.

Since CZAM always exists when S390_FEAT_ZARCH exists, which in turn
exists for every defined CPU model, we can simplify this code.

Fixes: 075e52b81664 ("s390x/cpumodel: we are always in zarchitecture mode")
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Janosch Frank <frankja@linux.ibm.com>
Message-Id: <20211008203811.1980478-2-farman@linux.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

s390x/ipl: check kernel command line size

Check if the provided kernel command line exceeds the maximum size of the s390x
Linux kernel command line size, which is 896 bytes.

Reported-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Message-Id: <20211006092631.20732-1-mhartmay@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
[thuth: Adjusted format specifier for size_t]
Signed-off-by: Thomas Huth <thuth@redhat.com>

Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* SEV and SGX cleanups (Philippe, Dov)
* bugfixes for "check-block"
* bugfix for static build
* ObjectOptions cleanups (Thomas)
* binutils fix for PVH (Cole)
* HVF cleanup (Alex)

# gpg: Signature made Wed 13 Oct 2021 01:47:56 AM PDT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]

* remotes/bonzini/tags/for-upstream: (40 commits)
  ebpf: really include it only in system emulators
  target/i386/sev: Use local variable for kvm_sev_launch_measure
  target/i386/sev: Use local variable for kvm_sev_launch_start
  monitor: Tidy up find_device_state()
  Revert "hw/misc: applesmc: use host osk as default on macs"
  hw/i386/sgx: Move qmp_query_sgx() and hmp_info_sgx() to hw/i386/sgx.c
  hw/i386/sgx: Move qmp_query_sgx_capabilities() to hw/i386/sgx.c
  hw/i386/sgx: Have sgx_epc_get_section() return a boolean
  MAINTAINERS: Cover SGX documentation file with X86/KVM section
  hvf: Determine slot count from struct layout
  tests: tcg: Fix PVH test with binutils 2.36+
  qapi: Make some ObjectTypes depend on the build settings
  MAINTAINERS: Cover SEV-related files with X86/KVM section
  monitor: Reduce hmp_info_sev() declaration
  target/i386/sev: Move qmp_query_sev() & hmp_info_sev() to sev.c
  target/i386/sev: Move qmp_query_sev_launch_measure() to sev.c
  target/i386/sev: Move qmp_query_sev_capabilities() to sev.c
  target/i386/sev: Move qmp_sev_inject_launch_secret() to sev.c
  target/i386/sev: Move qmp_query_sev_attestation_report() to sev.c
  target/i386/sev: Remove stubs by using code elision
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

ebpf: really include it only in system emulators

eBPF libraries are being included in user emulators, which is useless and
also breaks --static compilation if a shared library for libbpf is
present in the system.

Reported-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Use local variable for kvm_sev_launch_measure

The struct kvm_sev_launch_measure has a constant and small size, and
therefore we can use a regular local variable for it instead of
allocating and freeing heap memory for it.

Signed-off-by: Dov Murik <dovmurik@linux.ibm.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20211011173026.2454294-3-dovmurik@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Use local variable for kvm_sev_launch_start

The struct kvm_sev_launch_start has a constant and small size, and
therefore we can use a regular local variable for it instead of
allocating and freeing heap memory for it.

Signed-off-by: Dov Murik <dovmurik@linux.ibm.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20211011173026.2454294-2-dovmurik@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

monitor: Tidy up find_device_state()

Commit 6287d827d4 "monitor: allow device_del to accept QOM paths"
extended find_device_state() to accept QOM paths in addition to qdev
IDs.  This added a checked conversion to TYPE_DEVICE at the end, which
duplicates the check done for the qdev ID case earlier, except it sets
a *different* error: GenericError "ID is not a hotpluggable device"
when passed a QOM path, and DeviceNotFound "Device 'ID' not found"
when passed a qdev ID.  Fortunately, the latter won't happen as long
as we add only devices to /machine/peripheral/.

Earlier, commit b6cc36abb2 "qdev: device_del: Search for to be
unplugged device in 'peripheral' container" rewrote the lookup by qdev
ID to use QOM instead of qdev_find_recursive(), so it can handle
buss-less devices.  It does so by constructing an absolute QOM path.
Works, but object_resolve_path_component() is easier.  Switching to it
also gets rid of the unclean duplication described above.

While there, avoid converting to TYPE_DEVICE twice, first to check
whether it's possible, and then for real.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20210916111707.84999-1-armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Revert "hw/misc: applesmc: use host osk as default on macs"

This reverts commit 93ddefbc3c909bb6c3b76086f1dfc8ad98dd3725.
The commit included code under the APSL 2.0, which is incompatible
with the GPL v2.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

hw/i386/sgx: Move qmp_query_sgx() and hmp_info_sgx() to hw/i386/sgx.c

Move qmp_query_sgx() and hmp_info_sgx() from target/i386/monitor.c
to hw/i386/sgx.c, removing the sgx_get_info() indirection and the
"hw/i386/sgx.h" header.

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007175612.496366-5-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

hw/i386/sgx: Move qmp_query_sgx_capabilities() to hw/i386/sgx.c

Move qmp_query_sgx_capabilities() from target/i386/monitor.c to
hw/i386/sgx.c, removing the sgx_get_capabilities() indirection.

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007175612.496366-4-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

hw/i386/sgx: Have sgx_epc_get_section() return a boolean

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007175612.496366-3-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

MAINTAINERS: Cover SGX documentation file with X86/KVM section

Cc: Yang Zhong <yang.zhong@intel.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007175612.496366-2-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

hvf: Determine slot count from struct layout

We can handle up to a static amount of memory slots, capped by the size of
an internal array.

Let's make sure that array size is the only source of truth for the number
of elements in that array.

Signed-off-by: Alexander Graf <agraf@csgraf.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20211008054616.43828-1-agraf@csgraf.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tests: tcg: Fix PVH test with binutils 2.36+

binutils started adding a .note.gnu.property ELF section which
makes the PVH test fail:

  TEST    hello on x86_64
qemu-system-x86_64: Error loading uncompressed kernel without PVH ELF Note

Discard .note.gnu* while keeping the PVH .note bits intact.

This also strips the build-id note, so drop the related comment.

Signed-off-by: Cole Robinson <crobinso@redhat.com>
Message-Id: <5ab2a54c262c61f64c22dbb49ade3e2db8a740bb.1633708346.git.crobinso@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qapi: Make some ObjectTypes depend on the build settings

Some of the ObjectType entries already depend on CONFIG_* switches.
Some others also only make sense with certain configurations, but
are currently always listed in the ObjectType enum. Let's make them
depend on the correpsonding CONFIG_* switches, too, so that upper
layers (like libvirt) have a better way to determine which features
are available in QEMU.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20210928160232.432980-1-thuth@redhat.com>
[Do the same for MemoryBackendEpcProperties. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

MAINTAINERS: Cover SEV-related files with X86/KVM section

Complete the x86/KVM section with SEV-related files.

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-24-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

monitor: Reduce hmp_info_sev() declaration

While being conditionally used for TARGET_I386 in hmp-commands-info.hx,
hmp_info_sev() is declared for all targets. Reduce its declaration
to target including "monitor/hmp-target.h". This is a minor cleanup.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-23-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Move qmp_query_sev() & hmp_info_sev() to sev.c

Move qmp_query_sev() & hmp_info_sev()() from monitor.c to sev.c
and make sev_get_info() static. We don't need the stub anymore,
remove it. Add a stub for hmp_info_sev().

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-22-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Move qmp_query_sev_launch_measure() to sev.c

Move qmp_query_sev_launch_measure() from monitor.c to sev.c
and make sev_get_launch_measurement() static. We don't need the
stub anymore, remove it.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-21-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Move qmp_query_sev_capabilities() to sev.c

Move qmp_query_sev_capabilities() from monitor.c to sev.c
and make sev_get_capabilities() static. We don't need the
stub anymore, remove it.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-20-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Move qmp_sev_inject_launch_secret() to sev.c

Move qmp_sev_inject_launch_secret() from monitor.c to sev.c
and make sev_inject_launch_secret() static. We don't need the
stub anymore, remove it.

Previously with binaries built without SEV, management layer
was getting an empty response:

  { "execute": "sev-inject-launch-secret",
    "arguments": { "packet-header": "mypkt", "secret": "mypass", "gpa": 4294959104 }
  }
  {
      "return": {
      }
  }

Now the response is explicit, mentioning the feature is disabled:

  { "execute": "sev-inject-launch-secret",
          "arguments": { "packet-header": "mypkt", "secret": "mypass", "gpa": 4294959104 }
  }
  {
      "error": {
          "class": "GenericError",
          "desc": "this feature or command is not currently supported"
      }
  }

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-19-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Move qmp_query_sev_attestation_report() to sev.c

Move qmp_query_sev_attestation_report() from monitor.c to sev.c
and make sev_get_attestation_report() static. We don't need the
stub anymore, remove it.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-18-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Remove stubs by using code elision

Only declare sev_enabled() and sev_es_enabled() when CONFIG_SEV is
set, to allow the compiler to elide unused code. Remove unnecessary
stubs.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20211007161716.453984-17-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Declare system-specific functions in 'sev.h'

"sysemu/sev.h" is only used from x86-specific files. Let's move it
to include/hw/i386, and merge it with target/i386/sev.h.

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-16-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Rename sev_i386.h -> sev.h

SEV is a x86 specific feature, and the "sev_i386.h" header
is already in target/i386/. Rename it as "sev.h" to simplify.

Patch created mechanically using:

  $ git mv target/i386/sev_i386.h target/i386/sev.h
  $ sed -i s/sev_i386.h/sev.h/ $(git grep -l sev_i386.h)

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20211007161716.453984-15-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Restrict SEV to system emulation

SEV is irrelevant on user emulation, so restrict it to sysemu.
Some stubs are still required because used in cpu.c by
x86_register_cpudef_types(), so move the sysemu specific stubs
to sev-sysemu-stub.c instead. This will allow us to simplify
monitor.c (which is not available in user emulation) in the
next commit.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-14-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Use g_autofree in sev_launch_get_measure()

Use g_autofree to remove a pair of g_free/goto.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-13-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: sev_get_attestation_report use g_autofree

Removes a whole bunch of g_free's and a goto.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
Reviewed-by: Brijesh Singh <brijesh.singh@amd.com>
Message-Id: <20210603113017.34922-1-dgilbert@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211007161716.453984-12-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>