hw/core/clock: Add the clock_new helper function

This function creates a clock and parents it to another object with a
given name. It calls clock_setup_canonical_path before returning the
new clock.

This function is useful to create clocks in devices when one doesn't
want to expose it at the qdev level (as an input or an output).

Suggested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Luc Michel <luc@lmichel.fr>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20201010135759.437903-4-luc@lmichel.fr>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

hw/qdev-clock: Display error hint when clock is missing from device

Instead of directly aborting, display a hint to help the developer
figure out the problem (likely trying to connect a clock to a device
pre-dating the Clock API, thus not expecting clocks).

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Luc Michel <luc@lmichel.fr>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-Id: <20201012095804.3335117-4-f4bug@amsat.org>

util/cutils: Introduce freq_to_str() to display Hertz units

Introduce freq_to_str() to convert frequency values in human
friendly units using the SI units for Hertz.

Suggested-by: Luc Michel <luc@lmichel.fr>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Luc Michel <luc@lmichel.fr>
Message-Id: <20201012095804.3335117-2-f4bug@amsat.org>

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

Block layer patches:

- qemu-storage-daemon: Remove QemuOpts from --object parser
- monitor: Fix order in monitor_cleanup()
- Deprecate the sheepdog block driver

# gpg: Signature made Thu 15 Oct 2020 15:48:10 BST
# gpg:                using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg:                issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream:
  block: deprecate the sheepdog block driver
  block: drop moderated sheepdog mailing list from MAINTAINERS file
  monitor: Fix order in monitor_cleanup()
  qemu-storage-daemon: Remove QemuOpts from --object parser
  qom: Add user_creatable_print_help_from_qdict()
  qom: Factor out helpers from user_creatable_print_help()
  keyval: Parse help options
  keyval: Fix parsing of ',' in value of implied key
  test-keyval: Demonstrate misparse of ',' with implied key
  keyval: Fix and clarify grammar

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/kraxel/tags/modules-20201015-pull-request' into staging

modules: misc fixes & tweaks.
modules: build spice chardevs as module.

# gpg: Signature made Thu 15 Oct 2020 13:53:03 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/modules-20201015-pull-request:
  chardev/spice: build spice chardevs as module
  meson: add spice dependency to core spice source files.
  meson: add spice_headers dependency.
  chardev/spice: simplify chardev setup
  chardev/spice: make qemu_chr_open_spice_port static
  ui/spice-app: don't use qemu_chr_open_spice_port directly
  modules: update qom object module comment
  module: silence errors for module_load_qom_all().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/rth/tags/pull-mb-20201014' into staging

Implement rt signal frames for microblaze-linux-user
Adjust linux-user test for musl

# gpg: Signature made Thu 15 Oct 2020 05:50:00 BST
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth/tags/pull-mb-20201014:
  linux-user/microblaze: Remove non-rt signal frames
  linux-user/microblaze: Implement rt signal frames
  tests/tcg/linux-test: Adjust getsockname for musl

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/ehabkost/tags/machine-next-pull-request' into staging

machine + QOM queue, 2020-10-14

* Register some properties as class properties (Eduardo Habkost)
* authz-list-file: Fix crash when filename is not set (Eduardo Habkost)
* can-host-socketcan: Fix crash when 'if' option is not set (Eduardo Habkost)

# gpg: Signature made Wed 14 Oct 2020 15:33:17 BST
# gpg:                using RSA key 5A322FD5ABC4D3DBACCFD1AA2807936F984DC5A6
# gpg:                issuer "ehabkost@redhat.com"
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/machine-next-pull-request:
  can-host-socketcan: Fix crash when 'if' option is not set
  authz-list-file: Fix crash when filename is not set
  vhost-user: Register "chardev" as class property
  vga-pci: Register "big-endian-framebuffer" as class property
  i386: Register most CPU properties as class properties
  input-barrier: Register properties as class properties
  input-linux: Register properties as class properties
  rng: Register "opened" as class property
  rng-random: register "filename" as class property
  rng-egd: Register "chardev" as class property

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

block: deprecate the sheepdog block driver

This thread from a little over a year ago:

  http://lists.wpkg.org/pipermail/sheepdog/2019-March/thread.html

states that sheepdog is no longer actively developed. The only mentioned
users are some companies who are said to have it for legacy reasons with
plans to replace it by Ceph. There is talk about cutting out existing
features to turn it into a simple demo of how to write a distributed
block service. There is no evidence of anyone working on that idea:

  https://github.com/sheepdog/sheepdog/commits/master

No real commits to git since Jan 2018, and before then just some minor
technical debt cleanup.

There is essentially no activity on the mailing list aside from
patches to QEMU that get CC'd due to our MAINTAINERS entry.

Fedora packages for sheepdog failed to build from upstream source
because of the more strict linker that no longer merges duplicate
global symbols. Fedora patches it to add the missing "extern"
annotations and presumably other distros do to, but upstream source
remains broken.

There is only basic compile testing, no functional testing of the
driver.

Since there are no build pre-requisites the sheepdog driver is currently
enabled unconditionally. This would result in configure issuing a
deprecation warning by default for all users. Thus the configure default
is changed to disable it, requiring users to pass --enable-sheepdog to
build the driver.

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20201002113243.2347710-3-berrange@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: drop moderated sheepdog mailing list from MAINTAINERS file

The sheepdog mailing list is setup to stop and queue messages from
non-subscribers, pending moderator approval. Unfortunately it seems
that the moderation queue is not actively dealt with. Even when messages
are approved, the sender is never added to the whitelist, so every
future mail from the same sender continues to get stopped for moderation.

MAINTAINERS entries should be responsive and not unneccessarily block
mails from QEMU contributors, so drop the sheepdog mailing list.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20201002113243.2347710-2-berrange@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

monitor: Fix order in monitor_cleanup()

We can only destroy Monitor objects after we're sure that they are not
in use by the dispatcher coroutine any more. This fixes crashes like the
following where we tried to destroy a monitor mutex while the dispatcher
coroutine still holds it:

 (gdb) bt
 #0  0x00007fe541cf4bc5 in raise () at /lib64/libc.so.6
 #1  0x00007fe541cdd8a4 in abort () at /lib64/libc.so.6
 #2  0x000055c24e965327 in error_exit (err=16, msg=0x55c24eead3a0 <__func__.33> "qemu_mutex_destroy") at ../util/qemu-thread-posix.c:37
 #3  0x000055c24e9654c3 in qemu_mutex_destroy (mutex=0x55c25133e0f0) at ../util/qemu-thread-posix.c:70
 #4  0x000055c24e7cfaf1 in monitor_data_destroy_qmp (mon=0x55c25133dfd0) at ../monitor/qmp.c:439
 #5  0x000055c24e7d23bc in monitor_data_destroy (mon=0x55c25133dfd0) at ../monitor/monitor.c:615
 #6  0x000055c24e7d253a in monitor_cleanup () at ../monitor/monitor.c:644
 #7  0x000055c24e6cb002 in qemu_cleanup () at ../softmmu/vl.c:4549
 #8  0x000055c24e0d259b in main (argc=24, argv=0x7ffff66b0d58, envp=0x7ffff66b0e20) at ../softmmu/main.c:51

Reported-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20201013125027.41003-1-kwolf@redhat.com>
Tested-by: Ben Widawsky <ben.widawsky@intel.com>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-storage-daemon: Remove QemuOpts from --object parser

The command line parser for --object parses the input twice: Once into
QemuOpts just for detecting help options, and then again into a QDict
using the keyval parser for actually creating the object.

Now that the keyval parser can also detect help options, we can simplify
this and remove the QemuOpts part.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201007164903.282198-5-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qom: Add user_creatable_print_help_from_qdict()

This adds a function that, given a QDict of non-help options, prints
help for user creatable objects.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201007164903.282198-4-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qom: Factor out helpers from user_creatable_print_help()

This creates separate helper functions for printing a list of user
creatable object types and for printing a list of properties of a given
type. This will allow using these parts without having a QemuOpts.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201007164903.282198-3-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

keyval: Parse help options

This adds a special meaning for 'help' and '?' as options to the keyval
parser. Instead of being an error (because of a missing value) or a
value for an implied key, they now request help, which is a new boolean
output of the parser in addition to the QDict.

A new parameter 'p_help' is added to keyval_parse() that contains on
return whether help was requested. If NULL is passed, requesting help
results in an error and all other cases work like before.

Turning previous error cases into help is a compatible extension. The
behaviour potentially changes for implied keys: They could previously
get 'help' as their value, which is now interpreted as requesting help.

This is not a problem in practice because 'help' and '?' are not a valid
values for the implied key of any option parsed with keyval_parse():

* audiodev: union Audiodev, implied key "driver" is enum AudiodevDriver,
  "help" and "?" are not among its values

* display: union DisplayOptions, implied key "type" is enum
  DisplayType, "help" and "?" are not among its values

* blockdev: union BlockdevOptions, implied key "driver is enum
  BlockdevDriver, "help" and "?" are not among its values

* export: union BlockExport, implied key "type" is enum BlockExportType,
  "help" and "?" are not among its values

* monitor: struct MonitorOptions, implied key "mode" is enum MonitorMode,
  "help" and "?" are not among its values

* nbd-server: struct NbdServerOptions, no implied key.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201011073505.1185335-5-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

keyval: Fix parsing of ',' in value of implied key

The previous commit demonstrated documentation and code disagree on
parsing of ',' in the value of an implied key.  Fix the code to match
the documentation.

This breaks uses of keyval_parse() that pass an implied key and accept
a value containing ','.  None of the existing uses does:

* audiodev: implied key "driver" is enum AudiodevDriver, none of the
  values contains ','

* display: implied key "type" is enum DisplayType, none of the values
  contains ','

* blockdev: implied key "driver is enum BlockdevDriver, none of the
  values contains ','

* export: implied key "type" is enum BlockExportType, none of the
  values contains ','

* monitor: implied key "mode" is enum MonitorMode, none of the values
  contains ','

* nbd-server: no implied key.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201011073505.1185335-4-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

test-keyval: Demonstrate misparse of ',' with implied key

Add a test for "val,,ue" with implied key.  Documentation says this
should parse as implied key with value "val", then fail.  The code
parses it as implied key with value "val,ue", then succeeds.  The next
commit will fix it.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201011073505.1185335-3-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

keyval: Fix and clarify grammar

The grammar has a few issues:

* key-fragment = / [^=,.]* /

  Prose restricts key fragments: they "must be valid QAPI names or
  consist only of decimal digits".  Technically, '' consists only of
  decimal digits.  The code rejects that.  Fix the grammar.

* val          = { / [^,]* / | ',,' }

  Use + instead of *.  Accepts the same language.

* val-no-key   = / [^=,]* /

  The code rejects an empty value.  Fix the grammar.

* Section "Additional syntax for use with an implied key" is
  confusing.  Rewrite it.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201011073505.1185335-2-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

chardev/spice: build spice chardevs as module

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20201014121120.13482-8-kraxel@redhat.com

meson: add spice dependency to core spice source files.

Right now it happens to work by pure luck because the spice chardevs
add the spice dependency to the softmmu source set.  That'll change
though once we start building spice chardevs as module, so lets fix
it properly.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201014121120.13482-7-kraxel@redhat.com

meson: add spice_headers dependency.

Used for files which (with CONFIG_SPICE=y) depend on spice header files
to pick up some enum, but which do not depend on on the actual spice
shared library.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20201014121120.13482-6-kraxel@redhat.com

chardev/spice: simplify chardev setup

Initialize spice before chardevs.  That allows to register the spice
chardevs directly in the init function and removes the need to maintain
a linked list of chardevs just for registration.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20201014121120.13482-5-kraxel@redhat.com

chardev/spice: make qemu_chr_open_spice_port static

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201014121120.13482-4-kraxel@redhat.com

ui/spice-app: don't use qemu_chr_open_spice_port directly

Save the parent object's open function pointer in the (new)
VCChardevClass struct instead before overwriting it, so we
can look it up when needed.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201014121120.13482-3-kraxel@redhat.com

modules: update qom object module comment

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20200923103728.12026-1-kraxel@redhat.com

module: silence errors for module_load_qom_all().

Add mayfail bool parameter to module loading functions.  Set it to true
for module_load_qom_all() because device modules might not load into all
system emulation variants.  qemu-system-s390x for example will not load
qxl because it lacks vga support.  Makes "make check" less chatty.

Drop module_loaded_qom_all check in module_load_qom_one to make sure we
see errors for explicit load requests, i.e. module_load_qom_one("qxl")
failing will log an error no matter whenever module_load_qom_all() was
called before or not.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200923091217.22662-1-kraxel@redhat.com

linux-user/microblaze: Remove non-rt signal frames

The microblaze kernel does not support these, and uses
only rt style signal frames.

Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/microblaze: Implement rt signal frames

Allows microblaze to pass tests/tcg/multiarch/linux-test.c.

Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tests/tcg/linux-test: Adjust getsockname for musl

Unlike glibc, musl does not use transparent unions to hide
the different structures that overlap struct sockaddr.

Add an explicit cast to work around this.

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

Merge remote-tracking branch 'remotes/kraxel/tags/ui-20201014-pull-request' into staging

ui: fixes for sdl, curses, vnc, input-linux.

# gpg: Signature made Wed 14 Oct 2020 09:21:35 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/ui-20201014-pull-request:
  ui: Fix default window_id value
  input-linux: Reset il->fd handler before closing it
  SDL: enable OpenGL context creation
  vnc-stubs: Allow -vnc none
  configure: Fixes ncursesw detection under msys2/mingw by convert them to meson
  win32: Simplify gmtime_r detection not depends on if _POSIX_C_SOURCE are defined on msys2/mingw
  curses: Fixes curses compiling errors.
  curses: Fixes compiler error that complain don't have langinfo.h on msys2/mingw
  qemu-edid: drop cast

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

ui: Fix default window_id value

./chardev/baum.c expects the default window_id value to be -1, and not 0
which could be confused with a proper window id (when numbered from 0 by
the ui backend).

This fixes getting Braille output with the curses and gtk frontends.

Fixes: f29b3431f62 ("console: move window ID code from baum to sdl")
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200914100637.eeommoflirxrgaeh@function>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

input-linux: Reset il->fd handler before closing it

If object-del input-linux object on-the-fly, instance finalize will
close evdev fd without resetting it. However the main thread is still
trying to lock_acquire/lock_release during ppoll, which leads to a very
high CPU utilization.

Signed-off-by: Colin Xu <colin.xu@intel.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-id: 20200925021808.26471-1-colin.xu@intel.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

SDL: enable OpenGL context creation

We need to specify SDL_WINDOW_OPENGL if we want to create an OpenGL context on it, i.e. when using '-device virtio-gpu-pci,virgl=on'

Signed-off-by: Jan Henrik Weinstock <jan.weinstock@rwth-aachen.de>
Message-id: b2ba98b3-2975-0d4d-1c56-f659923c714d@rwth-aachen.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

vnc-stubs: Allow -vnc none

Currently `-vnc none` is fatal when built with `--disable-vnc`.  Make
vnc_parse accept "none", so QEMU still run without using vnc.

Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Message-id: 20201009014032.3507-1-jandryuk@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

configure: Fixes ncursesw detection under msys2/mingw by convert them to meson

The mingw pkg-config are showing following absolute path and contains : as the separator,

-D_XOPEN_SOURCE=600 -D_POSIX_C_SOURCE=199506L -IC:/CI-Tools/msys64/mingw64/include/ncursesw:-I/usr/include/ncursesw:
-DNCURSES_WIDECHAR -D_XOPEN_SOURCE=600 -D_POSIX_C_SOURCE=199506L -IC -pipe -lncursesw -lgnurx -ltre -lintl -liconv
-DNCURSES_WIDECHAR -D_XOPEN_SOURCE=600 -D_POSIX_C_SOURCE=199506L -IC -lncursesw
-DNCURSES_WIDECHAR -D_XOPEN_SOURCE=600 -D_POSIX_C_SOURCE=199506L -IC -lcursesw
-DNCURSES_WIDECHAR /CI-Tools/msys64/mingw64/include/ncursesw -pipe -lncursesw -lgnurx -ltre -lintl -liconv
-DNCURSES_WIDECHAR /CI-Tools/msys64/mingw64/include/ncursesw -lncursesw
-DNCURSES_WIDECHAR /CI-Tools/msys64/mingw64/include/ncursesw -lcursesw
-DNCURSES_WIDECHAR -I/usr/include/ncursesw -pipe -lncursesw -lgnurx -ltre -lintl -liconv
-DNCURSES_WIDECHAR -I/usr/include/ncursesw -lncursesw
-DNCURSES_WIDECHAR -I/usr/include/ncursesw -lcursesw

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20201012234348.1427-6-luoyonggang@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

win32: Simplify gmtime_r detection not depends on if _POSIX_C_SOURCE are defined on msys2/mingw

We remove the CONFIG_LOCALTIME_R detection option in configure, and move the check
existence of gmtime_r from configure into C header and source directly by using macro
`_POSIX_THREAD_SAFE_FUNCTIONS`.
Before this patch, the configure script are always assume the compiler doesn't define
_POSIX_C_SOURCE macro at all, but that's not true, because thirdparty library such
as ncursesw may define -D_POSIX_C_SOURCE in it's pkg-config file. And that C Flags will
added -D_POSIX_C_SOURCE into each QEMU_CFLAGS. And that's causing the following compiling error:
n file included from C:/work/xemu/qemu/include/qemu/osdep.h:119,
                 from ../softmmu/main.c:25:
C:/work/xemu/qemu/include/sysemu/os-win32.h:53:12: error: redundant redeclaration of 'gmtime_r' [-Werror=redundant-decls]
   53 | struct tm *gmtime_r(const time_t *timep, struct tm *result);
      |            ^~~~~~~~
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:94,
                 from ../softmmu/main.c:25:
C:/CI-Tools/msys64/mingw64/x86_64-w64-mingw32/include/time.h:284:36: note: previous definition of 'gmtime_r' was here
  284 | __forceinline struct tm *__CRTDECL gmtime_r(const time_t *_Time, struct tm *_Tm) {
      |                                    ^~~~~~~~
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:119,
                 from ../softmmu/main.c:25:
C:/work/xemu/qemu/include/sysemu/os-win32.h:55:12: error: redundant redeclaration of 'localtime_r' [-Werror=redundant-decls]
   55 | struct tm *localtime_r(const time_t *timep, struct tm *result);
      |            ^~~~~~~~~~~
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:94,
                 from ../softmmu/main.c:25:
C:/CI-Tools/msys64/mingw64/x86_64-w64-mingw32/include/time.h:281:36: note: previous definition of 'localtime_r' was here
  281 | __forceinline struct tm *__CRTDECL localtime_r(const time_t *_Time, struct tm *_Tm) {
      |                                    ^~~~~~~~~~~
Compiling C object libcommon.fa.p/hw_gpio_zaurus.c.obj
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:119,
                 from ../hw/i2c/smbus_slave.c:16:
C:/work/xemu/qemu/include/sysemu/os-win32.h:53:12: error: redundant redeclaration of 'gmtime_r' [-Werror=redundant-decls]
   53 | struct tm *gmtime_r(const time_t *timep, struct tm *result);
      |            ^~~~~~~~
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:94,
                 from ../hw/i2c/smbus_slave.c:16:
C:/CI-Tools/msys64/mingw64/x86_64-w64-mingw32/include/time.h:284:36: note: previous definition of 'gmtime_r' was here
  284 | __forceinline struct tm *__CRTDECL gmtime_r(const time_t *_Time, struct tm *_Tm) {
      |                                    ^~~~~~~~
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:119,
                 from ../hw/i2c/smbus_slave.c:16:
C:/work/xemu/qemu/include/sysemu/os-win32.h:55:12: error: redundant redeclaration of 'localtime_r' [-Werror=redundant-decls]
   55 | struct tm *localtime_r(const time_t *timep, struct tm *result);
      |            ^~~~~~~~~~~
In file included from C:/work/xemu/qemu/include/qemu/osdep.h:94,
                 from ../hw/i2c/smbus_slave.c:16:
C:/CI-Tools/msys64/mingw64/x86_64-w64-mingw32/include/time.h:281:36: note: previous definition of 'localtime_r' was here
  281 | __forceinline struct tm *__CRTDECL localtime_r(const time_t *_Time, struct tm *_Tm) {
      |                                    ^~~~~~~~~~~
Compiling C object libcommon.fa.p/hw_dma_xilinx_axidma.c.obj

After this patch, whenever ncursesw or other thirdparty libraries tried to define or not
define  _POSIX_C_SOURCE, the source will building properly. Because now, we don't make any
assumption if _POSIX_C_SOURCE are defined. We solely relied on if the macro `_POSIX_THREAD_SAFE_FUNCTIONS`
are defined in msys2/mingw header.

The _POSIX_THREAD_SAFE_FUNCTIONS are defined in mingw header like this:

```
#if defined(_POSIX_C_SOURCE) && !defined(_POSIX_THREAD_SAFE_FUNCTIONS)
#define _POSIX_THREAD_SAFE_FUNCTIONS 200112L
#endif

#ifdef _POSIX_THREAD_SAFE_FUNCTIONS
__forceinline struct tm *__CRTDECL localtime_r(const time_t *_Time, struct tm *_Tm) {
  return localtime_s(_Tm, _Time) ? NULL : _Tm;
}
__forceinline struct tm *__CRTDECL gmtime_r(const time_t *_Time, struct tm *_Tm) {
  return gmtime_s(_Tm, _Time) ? NULL : _Tm;
}
__forceinline char *__CRTDECL ctime_r(const time_t *_Time, char *_Str) {
  return ctime_s(_Str, 0x7fffffff, _Time) ? NULL : _Str;
}
__forceinline char *__CRTDECL asctime_r(const struct tm *_Tm, char * _Str) {
  return asctime_s(_Str, 0x7fffffff, _Tm) ? NULL : _Str;
}
#endif
```

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20201012234348.1427-5-luoyonggang@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

curses: Fixes curses compiling errors.

This is the compiling error:
../ui/curses.c: In function 'curses_refresh':
../ui/curses.c:256:5: error: 'next_maybe_keycode' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  256 |     curses2foo(_curses2keycode, _curseskey2keycode, chr, maybe_keycode)
      |     ^~~~~~~~~~
../ui/curses.c:302:32: note: 'next_maybe_keycode' was declared here
  302 |             enum maybe_keycode next_maybe_keycode;
      |                                ^~~~~~~~~~~~~~~~~~
../ui/curses.c:256:5: error: 'maybe_keycode' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  256 |     curses2foo(_curses2keycode, _curseskey2keycode, chr, maybe_keycode)
      |     ^~~~~~~~~~
../ui/curses.c:265:24: note: 'maybe_keycode' was declared here
  265 |     enum maybe_keycode maybe_keycode;
      |                        ^~~~~~~~~~~~~
cc1.exe: all warnings being treated as errors

gcc version 10.2.0 (Rev1, Built by MSYS2 project)

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20201012234348.1427-4-luoyonggang@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

curses: Fixes compiler error that complain don't have langinfo.h on msys2/mingw

msys2/mingw lacks the POSIX-required langinfo.h.

gcc test.c -DNCURSES_WIDECHAR -I/mingw64/include/ncursesw -pipe -lncursesw -lgnurx -ltre -lintl -liconv
test.c:4:10: fatal error: langinfo.h: No such file or directory
    4 | #include <langinfo.h>
      |          ^~~~~~~~~~~~
compilation terminated.

So we using g_get_codeset instead of nl_langinfo(CODESET)

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201012234348.1427-3-luoyonggang@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

qemu-edid: drop cast

Not needed and makes some compilers error out with:

qemu-edid.c:15:1: error: initializer element is not constant

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20201013091615.14166-1-kraxel@redhat.com

can-host-socketcan: Fix crash when 'if' option is not set

Fix the following crash:

  $ qemu-system-x86_64 -object can-host-socketcan,id=obj0
  Segmentation fault (core dumped)

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Acked-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Message-Id: <20201008202713.1416823-3-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

authz-list-file: Fix crash when filename is not set

Fix the following crash:

  $ qemu-system-x86_64 -object authz-list-file,id=obj0
  qemu-system-x86_64: -object authz-list-file,id=obj0: GLib: g_file_get_contents: assertion 'filename != NULL' failed
  Segmentation fault (core dumped)

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <20201008202713.1416823-2-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

vhost-user: Register "chardev" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20200921221045.699690-6-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

vga-pci: Register "big-endian-framebuffer" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20200921221045.699690-22-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

i386: Register most CPU properties as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20200921221045.699690-14-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

input-barrier: Register properties as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200921221045.699690-11-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

input-linux: Register properties as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200921221045.699690-10-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

rng: Register "opened" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200921221045.699690-8-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

rng-random: register "filename" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200921221045.699690-5-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

rng-egd: Register "chardev" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200921221045.699690-4-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

Merge remote-tracking branch 'remotes/vivier2/tags/trivial-branch-for-5.2-pull-request' into staging

Trivial Patches Pull request 20201013

# gpg: Signature made Tue 13 Oct 2020 12:49:59 BST
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [full]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [full]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [full]
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier2/tags/trivial-branch-for-5.2-pull-request:
  meson.build: drop duplicate 'sparc64' entry
  mingw: fix error __USE_MINGW_ANSI_STDIO redefined
  target/sparc/int32_helper: Remove duplicated 'Tag Overflow' entry
  goldfish_rtc: change MemoryRegionOps endianness to DEVICE_NATIVE_ENDIAN
  hw/char/serial: remove duplicate .class_init in serial_mm_info
  block/blkdebug: fix memory leak
  hw/pci: Fix typo in PCI hot-plug error message
  softmmu/memory: Log invalid memory accesses
  hw/acpi/piix4: Rename piix4_pm_add_propeties() to piix4_pm_add_properties()
  vmdk: fix maybe uninitialized warnings
  tests/test-char: Use a proper fallthrough comment
  hw/block/nvme: Simplify timestamp sum
  target/i386/cpu: Update comment that mentions Texinfo
  qemu-img-cmds.hx: Update comment that mentions Texinfo

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2020-10-13' into staging

* qtest improvements (test for crash found with the fuzzer, increase
  downtime in migration test, less verbose output when running w/o KVM)
* Improve handling of acceptance tests in the Gitlab-CI
* Run checkpatch.pl in the Gitlab-CI
* Improve the gitlab-pipeline-status script
* Misc patches (mark 'moxie' as deprecated, remove stale .gitignore files, ...)

# gpg: Signature made Tue 13 Oct 2020 11:49:06 BST
# 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 <huth@tuxfamily.org>" [full]
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>" [unknown]
# Primary key fingerprint: 27B8 8847 EEE0 2501 18F3  EAB9 2ED9 D774 FE70 2DB5

* remotes/huth-gitlab/tags/pull-request-2020-10-13: (23 commits)
  scripts/ci/gitlab-pipeline-status: wait for pipeline creation
  scripts/ci/gitlab-pipeline-status: use more descriptive exceptions
  scripts/ci/gitlab-pipeline-status: handle keyboard interrupts
  scripts/ci/gitlab-pipeline-status: refactor parser creation
  scripts/ci/gitlab-pipeline-status: give early feedback on running pipelines
  scripts/ci/gitlab-pipeline-status: improve message regarding timeout
  scripts/ci/gitlab-pipeline-status: make branch name configurable
  gitlab: assign python helper files to GitLab maintainers section
  gitlab: add a CI job to validate the DCO sign off
  gitlab: add a CI job for running checkpatch.pl
  configure: fixes indent of $meson setup
  docs/system/deprecated: Mark the 'moxie' CPU as deprecated
  Remove superfluous .gitignore files
  MAINTAINERS: Ignore bios-tables-test in the qtest section
  Add a comment in bios-tables-test.c to clarify the reason behind approach
  softmmu/vl: Be less verbose about missing KVM when running the qtests
  tests/migration: Allow longer timeouts
  qtest: add fuzz test case
  Acceptance tests: show test report on GitLab CI
  Acceptance tests: do not show canceled test logs on GitLab CI
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

meson.build: drop duplicate 'sparc64' entry

CC: Laurent Vivier <laurent@vivier.eu>
CC: qemu-trivial@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
Message-Id: <20201012175719.2573367-1-slyfox@gentoo.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

mingw: fix error __USE_MINGW_ANSI_STDIO redefined

Always put osdep.h first, and remove redundant stdlib.h include.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20201008165953.884599-1-marcandre.lureau@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

target/sparc/int32_helper: Remove duplicated 'Tag Overflow' entry

Commit 0b09be2b2f ("Nicer debug output for exceptions") added
twice the same "Tag Overflow" entry, remove the extra one.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20201011200112.3222822-1-f4bug@amsat.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

goldfish_rtc: change MemoryRegionOps endianness to DEVICE_NATIVE_ENDIAN

The doc [1] doesn't define the endianness, but the kernel driver
uses readl() to access the registers, so we can guess it depends
on the architecture endianness.

As riscv architecture endianness is little it might not change anything
for it.

Moreover, android implementation uses DEVICE_NATIVE_ENDIAN [2]

[1] https://android.googlesource.com/platform/external/qemu/+/master/docs/GOLDFISH-VIRTUAL-HARDWARE.TXT
[2] https://android.googlesource.com/platform/external/qemu/+/refs/heads/emu-master-dev/hw/timer/goldfish_timer.c#177

Fixes: 9a5b40b84279 ("hw: rtc: Add Goldfish RTC device")
Cc: Anup.Patel@wdc.com
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20201009113843.60995-2-lvivier@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/char/serial: remove duplicate .class_init in serial_mm_info

.class_init is already set to serial_mm_class_init.

Remove the duplicate entry.

Fixes: 17fd1a6490b1 ("serial-mm: add "regshift" property")
Cc: marcandre.lureau@redhat.com
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20201009113843.60995-1-lvivier@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

block/blkdebug: fix memory leak

Spotted by PVS-Studio

Signed-off-by: Elena Afanasova <eafanasova@gmail.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <1e903f928eb3da332cc95e2a6f87243bd9fe66e4.camel@gmail.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/pci: Fix typo in PCI hot-plug error message

'occupied' is spelled like 'ocuppied' in the message.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201006133958.600932-1-jusual@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

softmmu/memory: Log invalid memory accesses

Log invalid memory accesses with as GUEST_ERROR.

This is particularly useful since commit 5d971f9e67 which reverted
("memory: accept mismatching sizes in memory_region_access_valid").

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20201005152725.2143444-1-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/acpi/piix4: Rename piix4_pm_add_propeties() to piix4_pm_add_properties()

Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <160165476743.57452.2128307974125615413.stgit@bahia.lan>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

vmdk: fix maybe uninitialized warnings

Fedora 32 gcc 10 seems to give false positives:

Compiling C object libblock.fa.p/block_vmdk.c.o
../block/vmdk.c: In function ‘vmdk_parse_extents’:
../block/vmdk.c:587:5: error: ‘extent’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
  587 |     g_free(extent->l1_table);
      |     ^~~~~~~~~~~~~~~~~~~~~~~~
../block/vmdk.c:754:17: note: ‘extent’ was declared here
  754 |     VmdkExtent *extent;
      |                 ^~~~~~
../block/vmdk.c:620:11: error: ‘extent’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
  620 |     ret = vmdk_init_tables(bs, extent, errp);
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../block/vmdk.c:598:17: note: ‘extent’ was declared here
  598 |     VmdkExtent *extent;
      |                 ^~~~~~
../block/vmdk.c:1178:39: error: ‘extent’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
 1178 |             extent->flat_start_offset = flat_offset << 9;
      |             ~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
../block/vmdk.c: In function ‘vmdk_open_vmdk4’:
../block/vmdk.c:581:22: error: ‘extent’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
  581 |     extent->l2_cache =
      |     ~~~~~~~~~~~~~~~~~^
  582 |         g_malloc(extent->entry_size * extent->l2_size * L2_CACHE_SIZE);
      |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../block/vmdk.c:872:17: note: ‘extent’ was declared here
  872 |     VmdkExtent *extent;
      |                 ^~~~~~
../block/vmdk.c: In function ‘vmdk_open’:
../block/vmdk.c:620:11: error: ‘extent’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
  620 |     ret = vmdk_init_tables(bs, extent, errp);
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../block/vmdk.c:598:17: note: ‘extent’ was declared here
  598 |     VmdkExtent *extent;
      |                 ^~~~~~
cc1: all warnings being treated as errors
make: *** [Makefile.ninja:884: libblock.fa.p/block_vmdk.c.o] Error 1

fix them by assigning a default value.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Fam Zheng <fam@euphon.net>
Message-Id: <20200930155859.303148-2-borntraeger@de.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

tests/test-char: Use a proper fallthrough comment

For being able to compile with -Werror=implicit-fallthrough we need
to use comments that the compiler recognizes. Use "fallthrough" instead
of "no break" here.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201002171343.283426-1-thuth@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

hw/block/nvme: Simplify timestamp sum

As the 'timestamp' variable is declared as a 48-bit bitfield,
we do not need to wrap the sum result.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Message-Id: <20201002075716.1657849-1-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

scripts/ci/gitlab-pipeline-status: wait for pipeline creation

When called in wait mode, this script will also wait for the pipeline
to be get to a "running" state.  Because many more status may be seen
until a pipeline gets to "running", and those need to be handle too.

Reference: https://docs.gitlab.com/ee/api/pipelines.html#list-project-pipelines
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-8-crosa@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

scripts/ci/gitlab-pipeline-status: use more descriptive exceptions

For two very different error conditions.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-7-crosa@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

scripts/ci/gitlab-pipeline-status: handle keyboard interrupts

So that exits based on user requests are handled more gracefully.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-6-crosa@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

scripts/ci/gitlab-pipeline-status: refactor parser creation

Out of the main function.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-5-crosa@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

scripts/ci/gitlab-pipeline-status: give early feedback on running pipelines

When waiting for a pipeline to run and finish, it's better to give
early feedback, and then sleep and wait, than the other wait around.

Specially for the first iteration, it's frustrating to see nothing
while the script is sleeping.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-4-crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

scripts/ci/gitlab-pipeline-status: improve message regarding timeout

The script has its own timeout, which is about how long the script
will wait (when called with --wait) for the pipeline to complete, and
not necessarily for the pipeline to complete.

Hopefully this new wording will be clearer.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-3-crosa@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

scripts/ci/gitlab-pipeline-status: make branch name configurable

With the utility function `get_local_staging_branch_commit()`, the
name of the branch is hard coded (including in the function name).

For extensibility reasons, let's make that configurable.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20200904164258.240278-2-crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

gitlab: assign python helper files to GitLab maintainers section

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200918132903.1848939-4-berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

gitlab: add a CI job to validate the DCO sign off

While checkpatch.pl can validate DCO sign off that job must always be
advisory only since it is expected that certain patches will fail some
code style rules.

We require the DCO sign off to be mandatory for all commits though, so
it benefits from being validated in a standalone job.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200918132903.1848939-3-berrange@redhat.com>
[thuth: Use "stage: build" to let it run earlier]
Signed-off-by: Thomas Huth <thuth@redhat.com>

gitlab: add a CI job for running checkpatch.pl

This job is advisory since it is expected that certain patches will fail
the style checks and checkpatch.pl provides no way to mark exceptions to
the rules.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200918132903.1848939-2-berrange@redhat.com>
[thuth: Use "stage: build" to let it run earlier]
Signed-off-by: Thomas Huth <thuth@redhat.com>

configure: fixes indent of $meson setup

convert these line from tab to space

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20201012234348.1427-2-luoyonggang@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

docs/system/deprecated: Mark the 'moxie' CPU as deprecated

It is currently unclear whether anybody is still using the 'moxie' CPU,
and there are no images for testing available this CPU, so the code has
likely bit-rotten in the course of time. When I asked the maintainer
for information, I did not get a reply within four weeks yet (see
https://lists.gnu.org/archive/html/qemu-devel/2020-08/msg07201.html).
The last Signed-off-by line from Anthony in our repo is from 2013,
so it seems like this code is rather unmaintained. Time to put it onto
the deprecation list to see whether somebody is still interested in this
code or whether we could remove it in a couple of releases.

Message-Id: <20200923171815.97801-1-thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Remove superfluous .gitignore files

Since we are now always doing out-of-tree builds, these gitignore
files should not be necessary anymore.

Message-Id: <20200919133637.72744-1-thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

MAINTAINERS: Ignore bios-tables-test in the qtest section

I'm very often getting CC: on rather large patch series that
modify the ACPI stuff of either ARM or x86, just because the
bios-table-test is often slightly involved here. I can't say
much about ACPI, and the bios-table-test is already covered
by the ACPI section in MAINTAINERS, so I'd rather prefer to
not getting automatically CC-ed on such patch series anymore.
If people want my opinion about qtest-related changes, they
can still put me on CC manually.

Message-Id: <20201001042717.136033-1-thuth@redhat.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Add a comment in bios-tables-test.c to clarify the reason behind approach

A comment is added in bios-tables-test.c that explains the reasoning
behind the process of updating the ACPI table blobs when new tests are added
or old tests are modified or code is committed that affect tests. The
explanation would help future contributors follow the correct process when
making code changes that affect ACPI tables.

Signed-off-by: Ani Sinha <ani@anisinha.ca>
Acked-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20200929142501.1057-1-ani@anisinha.ca>
Signed-off-by: Thomas Huth <thuth@redhat.com>

softmmu/vl: Be less verbose about missing KVM when running the qtests

Some of the qtests use "-accel kvm -accel tcg" to run real guest code.
This causes some error messages when kvm is not available. We do not
really care about these messages since the fallback to tcg is expected
here. So let's silence them to avoid that they spoil the output of
the tests.

Unfortunately, we can not use the qtest_enabled() wrapper in this case,
since the qtest accelerator itself is not initialized. Thus we have to
test for the qtest_chrdev variable instead.

Message-Id: <20200710085020.28222-1-thuth@redhat.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests/migration: Allow longer timeouts

In travis, with gcov and gprof we're seeing timeouts; hopefully fix
this by increasing the test timeouts a bit, but for xbzrle ensure it
really does get a couple of cycles through to test the cache.

I think the problem in travis is we have about 2 host CPU threads,
in the test we have at least 3:
   a) The vCPU thread (100% flat out)
   b) The source migration thread
   c) The destination migration thread

if (b) & (c) are slow for any reason - gcov+gperf or a slow host -
then they're sharing one host CPU thread so limit the migration
bandwidth.

Tested on my laptop with:
   taskset -c 0,1 ./tests/qtest/migration-test -p /x86_64/migration

Reported-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20201008160330.130431-1-dgilbert@redhat.com>
[thuth: Move the #define to the right location]
Signed-off-by: Thomas Huth <thuth@redhat.com>

qtest: add fuzz test case

Currently the device fuzzer finds more and more issues.
For every fuzz case, we need not only the fixes but also
the corresponding test case. We can analysis the reproducer
for every case and find what happened in where and write
a beautiful test case. However the raw data of reproducer is not
friendly to analysis. It will take a very long time, even far more
than the fixes itself. So let's create a new file to hold all of
the fuzz test cases and just use the raw data to act as the test
case. This way nobody will be afraid of writing a test case for
the fuzz reproducer.

This patch adds the issue LP#1878263 test case.

Signed-off-by: Li Qiang <liq3ea@163.com>
Message-Id: <20200921160605.19329-1-liq3ea@163.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alexander Bulekov <alxndr@bu.edu>
[thuth: Slightly adjusted commit message, removed empty lines]
Signed-off-by: Thomas Huth <thuth@redhat.com>

Acceptance tests: show test report on GitLab CI

Avocado will, by default, produce JUnit files.  Let's ask GitLab
to present those in the web UI.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20201009205513.751968-4-crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Acceptance tests: do not show canceled test logs on GitLab CI

Tests resulting in "CANCEL" in Avocado are usually canceled on
purpose, and are almost identical to "SKIP".  The logs for canceled
tests are adding a lot of noise to the logs being shown on GitLab CI,
and causing distraction from real failures.

As a side note, this "after script" is scheduled for removal once the
feature is implemented within Avocado itself.

Reference: https://github.com/avocado-framework/avocado/issues/4266
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20201009205513.751968-3-crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Acceptance tests: bump pycdlib version for easier installation

On with certain versions of "pip", package installations will attempt
to create wheels.  And, on environments without a "complete" Python
installation (as described in the acceptance tests requirements docs),
that will fail.

pycdlib, starting with version 1.11.0, is now being made available
as wheels, so its instalation on those constrained environments is
now possible.

Buglink: https://bugs.launchpad.net/qemu/+bug/1897783
Reported-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Message-Id: <20201009205513.751968-2-crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

gitlab-ci.yml: Only run one test-case per fuzzer

With 1000 runs, there is a non-negligible chance that the fuzzer can
trigger a crash. With this CI job, we care about catching build/runtime
issues in the core fuzzing code. Actual device fuzzing takes place on
oss-fuzz. For these purposes, only running one input should be
sufficient.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Suggested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201002143524.56930-1-alxndr@bu.edu>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests/qtest: Replace magic value by NANOSECONDS_PER_SECOND definition

Use self-explicit NANOSECONDS_PER_SECOND definition instead
of a magic value.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20201011194918.3219195-5-f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

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

* qtest documentation improvements (Eduardo, myself)
* libqtest event buffering (Maxim)
* use RCU for list of children of a bus (Maxim)
* move more files to softmmu/ (myself)
* meson.build cleanups, qemu-storage-daemon fix (Philippe)

# gpg: Signature made Mon 12 Oct 2020 16:55:19 BST
# 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]
# 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-gitlab/tags/for-upstream: (38 commits)
  meson: identify more sections of meson.build
  scsi/scsi_bus: fix races in REPORT LUNS
  virtio-scsi: use scsi_device_get
  scsi/scsi_bus: Add scsi_device_get
  scsi/scsi-bus: scsi_device_find: don't return unrealized devices
  device-core: use atomic_set on .realized property
  scsi: switch to bus->check_address
  device-core: use RCU for list of children of a bus
  device_core: use drain_call_rcu in in qmp_device_add
  scsi/scsi_bus: switch search direction in scsi_device_find
  qdev: add "check if address free" callback for buses
  qemu-iotests, qtest: rewrite test 067 as a qtest
  qtest: check that drives are really appearing and disappearing
  qtest: switch users back to qtest_qmp_receive
  device-plug-test: use qtest_qmp to send the device_del command
  qtest: remove qtest_qmp_receive_success
  qtest: Reintroduce qtest_qmp_receive with QMP event buffering
  qtest: rename qtest_qmp_receive to qtest_qmp_receive_dict
  meson.build: Re-enable KVM support for MIPS
  build-sys: fix git version from -version
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/i386/cpu: Update comment that mentions Texinfo

Missed in commit 41fba1618b "docs/system: convert the documentation of
deprecated features to rST."

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200929075824.1517969-3-armbru@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

qemu-img-cmds.hx: Update comment that mentions Texinfo

Missed in 3c95fdef94 "Update comments in .hx files that mention
Texinfo".

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200929075824.1517969-2-armbru@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

meson: identify more sections of meson.build

Add more headers that clarify code organization.

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

scsi/scsi_bus: fix races in REPORT LUNS

Currently scsi_target_emulate_report_luns iterates over the child device list
twice, and there is no guarantee that this list is the same in both iterations.

The reason for iterating twice is that the first iteration calculates
how much memory to allocate.  However if we use a dynamic array we can
avoid iterating twice, and therefore we avoid this race.

Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1866707
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-10-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-14-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

virtio-scsi: use scsi_device_get

This will help us to avoid the scsi device disappearing
after we took a reference to it.

It doesn't by itself forbid case when we try to access
an unrealized device

Suggested-by: Stefan Hajnoczi <stefanha@gmail.com>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-9-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-13-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

scsi/scsi_bus: Add scsi_device_get

Add scsi_device_get which finds the scsi device
and takes a reference to it.

Suggested-by: Stefan Hajnoczi <stefanha@gmail.com>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20200913160259.32145-8-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-12-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

scsi/scsi-bus: scsi_device_find: don't return unrealized devices

The device core first places a device on the bus and then realizes it.
Make scsi_device_find avoid returing such devices to avoid
races in drivers that use an iothread (currently virtio-scsi)

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1812399

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-7-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-11-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

device-core: use atomic_set on .realized property

Some code might race with placement of new devices on a bus.
We currently first place a (unrealized) device on the bus
and then realize it.

As a workaround, users that scan the child device list, can
check the realized property to see if it is safe to access such a device.
Use an atomic write here too to aid with this.

A separate discussion is what to do with devices that are unrealized:
It looks like for this case we only call the hotplug handler's unplug
callback and its up to it to unrealize the device.
An atomic operation doesn't cause harm for this code path though.

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-6-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-10-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

scsi: switch to bus->check_address

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-6-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

device-core: use RCU for list of children of a bus

This fixes the race between device emulation code that tries to find
a child device to dispatch the request to (e.g a scsi disk),
and hotplug of a new device to that bus.

Note that this doesn't convert all the readers of the list
but only these that might go over that list without BQL held.

This is a very small first step to make this code thread safe.

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-5-mlevitsk@redhat.com>
[Use RCU_READ_LOCK_GUARD in more places, adjust testcase now that
 the delay in DEVICE_DELETED due to RCU is more consistent. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-9-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

device_core: use drain_call_rcu in in qmp_device_add

Soon, a device removal might only happen on RCU callback execution.
This is okay for device-del which provides a DEVICE_DELETED event,
but not for the failure case of device-add.  To avoid changing
monitor semantics, just drain all pending RCU callbacks on error.

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Suggested-by: Stefan Hajnoczi <stefanha@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-4-mlevitsk@redhat.com>
[Don't use it in qmp_device_del. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

scsi/scsi_bus: switch search direction in scsi_device_find

This change will allow us to convert the bus children list to RCU,
while not changing the logic of this function

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200913160259.32145-2-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qdev: add "check if address free" callback for buses

Check if an address is free on the bus before plugging in the
device.  This makes it possible to do the check without any
side effects, and to detect the problem early without having
to do it in the realize callback.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20201006123904.610658-5-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qemu-iotests, qtest: rewrite test 067 as a qtest

Test 067 from qemu-iotests is executing QMP commands to hotplug
and hot-unplug disks, devices and blockdevs.  Because the power
of the text-based test harness is limited, it is actually limiting
the checks that it does, for example by skipping DEVICE_DELETED
events.

tests/qtest already has a similar test, drive_del-test.c.
We can merge them, and even reuse some of the existing code in
drive_del-test.c.  This will improve the quality of the test by
covering DEVICE_DELETED events and testing multiple architectures
(therefore covering multiple PCI hotplug mechanisms as well as s390x
virtio-ccw).

The only difference is that the new test will always use null-co:// for
the medium rather than qcow2 or raw, but this should be irrelevant for
what the test is covering.  For example there are no "qemu-img check"
runs in 067 that would check that the file is properly closed.

The new tests requires PCI hot-plug support, so drive_del-test
is moved from qemu-system-ppc to qemu-system-ppc64.

Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>