Warner Losh [Fri, 6 Aug 2021 00:15:47 +0000 (18:15 -0600)]
bsd-user: Update mapping to handle reserved and starting conditions
Update the reserved base based on what platform we're on, as well as the
start of the mmap range. Update routines that find va ranges to interact
with the reserved ranges as well as properly align the mapping (this is
especially important for targets whose page size does not match the
host's). Loop where appropriate when the initial address space offered
by mmap does not meet the contraints.
This has 18e80c55bb6 from linux-user folded in to the upstream
bsd-user code as well.
Signed-off-by: Mikaël Urankar <mikael.urankar@gmail.com> Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Acked-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Kyle Evans <kevans@FreeBSD.org>
Colin Percival [Sat, 7 Aug 2021 20:34:21 +0000 (14:34 -0600)]
bsd-user: Add '-0 argv0' option to bsd-user/main.c
Previously it was impossible to emulate a program with a file name
different from its argv[0]. With this change, you can run
qemu -0 fakename realname args
which runs the program "realname" with an argv of "fakename args".
Signed-off-by: Colin Percival <cperciva@tarsnap.com> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Sat, 7 Aug 2021 20:22:34 +0000 (14:22 -0600)]
bsd-user: Implement interlock for atomic operations
Implement the internlock in fork_start() and fork_end() to properly cope
with atomic operations and to safely keep state for parent and child
processes.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Fri, 6 Aug 2021 20:20:16 +0000 (14:20 -0600)]
bsd-user: add stubbed out core dump support
Add a stubbed-out version of the bsd-user fork's core dump support. This
allows elfload.c to be almost the same between what's upstream and
what's in qemu-project upstream w/o the burden of reviewing the core
dump support.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Fri, 6 Aug 2021 20:05:41 +0000 (14:05 -0600)]
bsd-user: Add target_os_user.h to capture the user/kernel structures
This file evolved over the years to capture the user/kernel interfaces,
including those that changed over time.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Michal Meloun <mmel@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Acked-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Kyle Evans <kevans@FreeBSD.org>
Warner Losh [Fri, 6 Aug 2021 20:02:10 +0000 (14:02 -0600)]
bsd-user: Add target_arch_reg to describe a target's register set
target_reg_t is the normal register. target_fpreg_t is the floating
point registers. target_copy_regs copies the registers out of CPU
context for things like core dumps.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Thu, 5 Aug 2021 19:36:41 +0000 (13:36 -0600)]
bsd-user: Rewrite target system call definintion glue
Rewrite target definnitions to interface with the FreeBSD system calls.
This covers basic types (time_t, iovec, umtx_time, timespec, timeval,
rusage, rwusage) and basic defines (mmap, rusage). Also included are
FreeBSD version-specific variations.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Acked-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Kyle Evans <kevans@FreeBSD.org>
Warner Losh [Wed, 4 Aug 2021 23:13:24 +0000 (17:13 -0600)]
bsd-user: elf cleanup
Move OS-dependent defines into target_os_elf.h. Move the architectural
dependent stuff into target_arch_elf.h. Adjust elfload.c to use
target_create_elf_tables instead of create_elf_tables.
Signed-off-by: Warner Losh <imp@bsdimp.com> Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Kyle Evans <kevans@FreeBSD.org> Signed-off-by: Justin Hibbits <chmeeedalf@gmail.com> Signed-off-by: Alexander Kabaev <kan@FreeBSD.ORG> Acked-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Kyle Evans <kevans@FreeBSD.org>
Warner Losh [Wed, 4 Aug 2021 19:34:05 +0000 (13:34 -0600)]
bsd-user: Add architecture specific signal tramp code
Add a stubbed out version of setup_sigtramp. This is not yet used for
x86, but is used for other architectures. This will be connected in
future commits.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Thu, 2 Sep 2021 22:52:45 +0000 (16:52 -0600)]
bsd-user: Implement --seed and initialize random state
Copy --seed implementation (translated from linux-user's newer command
line scheme to the older one bsd-user still uses). Initialize the
randomness with the glib if a specific seed is specified or use the
qcrypto library if not.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Thu, 5 Aug 2021 19:41:52 +0000 (13:41 -0600)]
bsd-user: *BSD specific siginfo defintions
Add FreeBSD, NetBSD and OpenBSD values for the various signal info types
and defines to decode different signals to discover more information
about the specific signal types.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Acked-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Kyle Evans <kevans@FreeBSD.org>
Warner Losh [Sat, 7 Aug 2021 00:48:37 +0000 (18:48 -0600)]
bsd-user: Add system independent stack, data and text limiting
Eliminate the x86 specific stack stuff in favor of more generic control
over the process size:
target_maxtsiz max text size
target_dfldsiz initial data size limit
target_maxdsiz max data size
target_dflssiz initial stack size limit
target_maxssiz max stack size
target_sgrowsiz amount to grow stack
These can be set on a per-arch basis, and the stack size can be set
on the command line. Adjust the stack size parameters at startup.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Wed, 4 Aug 2021 04:04:20 +0000 (22:04 -0600)]
bsd-user: pull in target_arch_thread.h update target_arch_elf.h
Update target_arch_elf.h to remove thread_init. Move its contents to
target_arch_thread.h and rename to target_thread_init(). Update
elfload.c to call it. Create thread_os_thread.h to hold the os specific
parts of the thread and threat manipulation routines. Currently, it just
includes target_arch_thread.h. target_arch_thread.h contains the at the
moment unused target_thread_set_upcall which will be used in the future
when creating actual thread (i386 has this stubbed, but other
architectures in the bsd-user tree have real ones). FreeBSD doesn't do
AT_HWCAP, so remove that code. Linux does, and this code came from there.
These changes are all interrelated and could be brokend down, but seem
to represent a reviewable changeset since most of the change is boiler
plate.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Kyle Evans <kevans@FreeBSD.org>
Warner Losh [Tue, 3 Aug 2021 18:34:52 +0000 (12:34 -0600)]
bsd-user: assume pthreads and support of __thread
All compilers for some time have supported this. Follow linux-user and
eliminate the #define THREAD and unconditionally insert __thread where
needed. Please insert: "(see 24cb36a61c6: "configure: Make NPTL
non-optional")"
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Warner Losh [Fri, 30 Apr 2021 14:17:23 +0000 (08:17 -0600)]
bsd-user: remove a.out support
Remove still-born a.out support. The BSDs switched from a.out to ELF 20+ years
ago. It's out of scope for bsd-user, and what little support there was would
simply wind up at a not-implemented message. Simplify the whole mess by removing
it entirely. Should future support be required, it would be better to start from
scratch.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Warner Losh [Fri, 30 Apr 2021 04:25:36 +0000 (22:25 -0600)]
bsd-user: Eliminate elf personality
The linux kernel supports a number of different ELF binaries. The Linux userland
emulator inheritted some of that. And we inheritted it from there. However, for
BSD there's only one kind of ELF file supported per platform, so there's no need
to cope with historical quirks. Simply the code as a result.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Fri, 30 Apr 2021 01:34:34 +0000 (19:34 -0600)]
bsd-user: implement path searching
Use the PATH to find the executable given a bare argument. We need to do
this so we can implement mixing native and emulated binaries (e.g.,
execing a x86 native binary from an emulated arm binary to optimize
parts of the build). By finding the binary, we will know how to exec it.
Signed-off-by: Stacey Son <sson@FreeBSD.org> Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Fri, 30 Apr 2021 00:47:51 +0000 (18:47 -0600)]
bsd-user: Fix calculation of size to allocate
It was incorrect to subtract off the size of an unsigned int here. In
bsd-user fork, this change was made when moving the arch specific items
to specific files. The size in BSD that's available for the arguments
does not need a return address subtracted from it.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Fri, 30 Apr 2021 00:45:13 +0000 (18:45 -0600)]
bsd-user: pass the bsd_param into loader_exec
Pass the bsd_param into loader_exec, and adjust. We use it to track the
inital stack allocation and to set stack, open files, and other state
shared between bsdload.c and elfload.c
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Warner Losh [Thu, 29 Apr 2021 16:41:29 +0000 (10:41 -0600)]
bsd-user: move arch specific defines out of elfload.c
Move the architecture specific defines to target_arch_elf.h and delete
them from elfload.c. Only retain ifdefs appropriate for i386 and x86_64.
Add the copyright/license comments, and guard ifdefs.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Thu, 29 Apr 2021 16:10:02 +0000 (10:10 -0600)]
bsd-user: Remove all non-x86 code from elfload.c
bsd-user only builds x86 at the moment. Remove all non x86 code from
elfload.c. We'll move the x86 code to {i386,x86_64}/target_arch_elf.h
and bring it that support code from the forked bsd-user when the time
comes.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Warner Losh [Tue, 3 Aug 2021 22:50:22 +0000 (16:50 -0600)]
bsd-user: remove sparc and sparc64
These are broken here and in the bsd-user fork. They won't be fixed as
FreeBSD has dropped support for sparc. If people wish to support this in
other BSDs, you're better off starting over than starting from these
files.
Signed-off-by: Warner Losh <imp@bsdimp.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Michael Tokarev [Wed, 1 Sep 2021 13:16:24 +0000 (16:16 +0300)]
qemu-sockets: fix unix socket path copy (again)
Commit 4cfd970ec188558daa6214f26203fe553fb1e01f added an
assert which ensures the path within an address of a unix
socket returned from the kernel is at least one byte and
does not exceed sun_path buffer. Both of this constraints
are wrong:
A unix socket can be unnamed, in this case the path is
completely empty (not even \0)
And some implementations (notable linux) can add extra
trailing byte (\0) _after_ the sun_path buffer if we
passed buffer larger than it (and we do).
So remove the assertion (since it causes real-life breakage)
but at the same time fix the usage of sun_path. Namely,
we should not access sun_path[0] if kernel did not return
it at all (this is the case for unnamed sockets),
and use the returned salen when copyig actual path as an
upper constraint for the amount of bytes to copy - this
will ensure we wont exceed the information provided by
the kernel, regardless whenever there is a trailing \0
or not. This also helps with unnamed sockets.
Note the case of abstract socket, the sun_path is actually
a blob and can contain \0 characters, - it should not be
passed to g_strndup and the like, it should be accessed by
memcpy-like functions.
Fixes: 4cfd970ec188558daa6214f26203fe553fb1e01f Fixes: http://bugs.debian.org/993145 Signed-off-by: Michael Tokarev <mjt@tls.msk.ru> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> CC: qemu-stable@nongnu.org
Peter Maydell [Mon, 6 Sep 2021 11:38:07 +0000 (12:38 +0100)]
Merge remote-tracking branch 'remotes/thuth-gitlab/tags/pull-request-2021-09-06' into staging
* Add definitions of terms for CI/testing
* Fix g_setenv problem discovered by Coverity
* Gitlab CI improvements
* Build system improvements (configure script + meson.build)
* Removal of the show-fixed-bugs.sh script
* Clean up of the sdl and curses options
* remotes/thuth-gitlab/tags/pull-request-2021-09-06:
softmmu/vl: Deprecate the -sdl and -curses option
softmmu/vl: Deprecate the old grab options
softmmu/vl: Add a "grab-mod" parameter to the -display sdl option
scripts: Remove the "show-fixed-bugs.sh" file
configure / meson: Move the GBM handling to meson.build
meson.build: Don't use internal libfdt if the user requested the system libfdt
meson.build: Fix the check for a usable libfdt
gitlab-ci: Don't try to use the system libfdt in the debian job
libqtest: check for g_setenv() failure
docs: add definitions of terms for CI/testing
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Thomas Huth [Wed, 25 Aug 2021 09:20:23 +0000 (11:20 +0200)]
softmmu/vl: Deprecate the -sdl and -curses option
It's not that much complicated to type "-display sdl" or "-display curses",
so we should not clutter our main option name space with such simple
wrapper options and rather present the users with a concise interface
instead. Thus let's deprecate the "-sdl" and "-curses" wrapper options now.
Message-Id: <20210825092023.81396-4-thuth@redhat.com> Reviewed-by: Gerd Hoffmann <kraxel@redhat.com> Acked-by: Peter Krempa <pkrempa@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Wed, 25 Aug 2021 09:20:22 +0000 (11:20 +0200)]
softmmu/vl: Deprecate the old grab options
The alt_grab and ctrl_grab parameter of the -display sdl option prevent
the QAPIfication of the "sdl" part of the -display option, so we should
eventually remove them. And since this feature is also rather niche anyway,
we should not clutter the top-level option list with these, so let's
also deprecate the "-alt-grab" and the "-ctrl-grab" options while we're
at it.
Once the deprecation period of "alt_grab" and "ctrl_grab" is over, we
then can finally switch the -display sdl option to use QAPI internally,
too.
Message-Id: <20210825092023.81396-3-thuth@redhat.com> Reviewed-by: Gerd Hoffmann <kraxel@redhat.com> Acked-by: Peter Krempa <pkrempa@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Wed, 25 Aug 2021 09:20:21 +0000 (11:20 +0200)]
softmmu/vl: Add a "grab-mod" parameter to the -display sdl option
The -display sdl option is not using QAPI internally yet, and uses hand-
crafted parsing instead (see parse_display() in vl.c), which is quite
ugly, since most of the other code is using the QAPIfied DisplayOption
already. Unfortunately, the "alt_grab" and "ctrl_grab" use underscores in
their names which has recently been forbidden in new QAPI code, so
a straight conversion is not possible. While we could add some exceptions
to the QAPI schema parser for this, the way these parameters have been
designed was maybe a bad idea anyway: First, it's not possible to enable
both parameters at the same time, thus instead of two boolean parameters
it would be better to have only one multi-choice parameter instead.
Second, the naming is also somewhat unfortunate since the "alt_grab"
parameter is not about the ALT key, but rather about the left SHIFT key
that has to be used additionally when the parameter is enabled.
So instead of trying to QAPIfy "alt_grab" and "ctrl_grab", let's rather
introduce an alternative to these parameters instead, a new parameter
called "grab-mod" which can either be set to "lshift-lctrl-lalt" or to
"rctrl". In case we ever want to support additional modes later, we can
then also simply extend the list of supported strings here.
Message-Id: <20210825092023.81396-2-thuth@redhat.com> Reviewed-by: Gerd Hoffmann <kraxel@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Tue, 13 Jul 2021 11:09:02 +0000 (13:09 +0200)]
configure / meson: Move the GBM handling to meson.build
The GBM library detection does not need to be in the configure script,
since it does not have any user-facing options (there are no
--enable-gbm or --disable-gbm switches). Let's move it to meson.build
instead, so we don't have to clutter config-host.mak with the related
switches.
Additionally, only check for GBM if it is really required, i.e. if we
either compile with OpenGL or with virglrenderer support.
Message-Id: <20210714085045.797168-1-thuth@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Fri, 27 Aug 2021 12:09:00 +0000 (14:09 +0200)]
meson.build: Don't use internal libfdt if the user requested the system libfdt
If the users ran configure with --enable-libfdt=system, they likely did
that on purpose. We should not silently fall back to the internal libfdt
if the system libfdt is not usable, but report the problem with a proper
message instead.
Message-Id: <20210827120901.150276-3-thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Fri, 27 Aug 2021 12:08:59 +0000 (14:08 +0200)]
meson.build: Fix the check for a usable libfdt
The check for libfdt currently has a flaw: If there is a system libfdt, the
meson.build code initialized the fdt variable with fdt = cc.find_library(...).
However, if this libfdt is too old and there is no internal dtc module
available, it continues with "fdt" pointing to the old and unusable version.
The check later in the file that tries to detect whether libfdt is necessary
then fails to trigger:
if not fdt.found() and fdt_required.length() > 0
error('fdt not available but required by targets ' + ', '.join(fdt_required))
endif
The build fails then during compilation instead, which is of course bad
since this is quite confusing and already wasted quite some time of the user.
Thus if libfdt is not usable, we should unset the "fdt" variable immediately
again, so that the build already fails during the configuration phase.
Thomas Huth [Fri, 27 Aug 2021 15:17:18 +0000 (17:17 +0200)]
gitlab-ci: Don't try to use the system libfdt in the debian job
libfdt in Debian is too old to be usable for QEMU. So far we were
silently falling back to the internal dtc submodule, but since
this is wrong, let's remove the --enable-fdt=system switch here now.
Message-Id: <20210827151718.178988-1-thuth@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Acked-by: Philippe Mathieu-Daudé <philmd@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Peter Maydell [Fri, 20 Aug 2021 16:37:50 +0000 (17:37 +0100)]
libqtest: check for g_setenv() failure
g_setenv() can fail; check for it when starting a QEMU process
when we set the QEMU_AUDIO_DRV environment variable.
Because this happens after fork() reporting an exact message
via printf() is a bad idea; just exit(1), as we already do
for the case of execlp() failure.
Fixes: Coverity CID 1460117 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210820163750.9106-1-peter.maydell@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Willian Rampazzo [Tue, 31 Aug 2021 15:29:39 +0000 (12:29 -0300)]
docs: add definitions of terms for CI/testing
To understand the current state of QEMU CI/testing and have a base to
discuss the plans for the future, it is important to define some usual
terms. This patch defines the terms for "Automated tests", "Unit
testing", "Functional testing", "System testing", "Flaky tests",
"Gating", and "Continuous Integration".
Signed-off-by: Willian Rampazzo <willianr@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210831152939.97570-2-willianr@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Peter Maydell [Sun, 5 Sep 2021 14:48:42 +0000 (15:48 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
pc,pci,virtio: fixes, cleanups
Fixes, cleanups all over the place.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Sat 04 Sep 2021 22:34:10 BST
# gpg: using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469
# gpg: issuer "mst@redhat.com"
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [full]
# gpg: aka "Michael S. Tsirkin <mst@redhat.com>" [full]
# Primary key fingerprint: 0270 606B 6F3C DF3D 0B17 0970 C350 3912 AFBE 8E67
# Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA 8A0D 281F 0DB8 D28D 5469
* remotes/mst/tags/for_upstream: (35 commits)
vhost-vdpa: remove the unncessary queue_index assignment
vhost-vdpa: fix the wrong assertion in vhost_vdpa_init()
vhost-vdpa: tweak the error label in vhost_vdpa_add()
vhost-vdpa: fix leaking of vhost_net in vhost_vdpa_add()
vhost-vdpa: don't cleanup twice in vhost_vdpa_add()
vhost-vdpa: remove the unnecessary check in vhost_vdpa_add()
vhost_net: do not assume nvqs is always 2
vhost: use unsigned int for nvqs
vhost_net: remove the meaningless assignment in vhost_net_start_one()
vhost-vdpa: correctly return err in vhost_vdpa_set_backend_cap()
vhost-vdpa: remove unused variable "acked_features"
tests/vhost-user-bridge.c: Fix typo in help message
tests/vhost-user-bridge.c: Sanity check socket path length
hw/virtio: Add flatview update in vhost_user_cleanup()
hw/virtio: Remove NULL check in virtio_free_region_cache()
hw/virtio: Document virtio_queue_packed_empty_rcu is called within RCU
MAINTAINERS: Added myself as a reviewer for acpi/smbios subsystem
hw/acpi: use existing references to pci device struct within functions
hw/pci: remove all references to find_i440fx function
hw/i386/acpi-build: Get NUMA information from struct NumaState
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Jason Wang [Fri, 3 Sep 2021 09:10:21 +0000 (17:10 +0800)]
vhost-vdpa: remove the unncessary queue_index assignment
The queue_index of NetClientState should be assigned in set_netdev()
afterwards, so trying to net_vhost_vdpa_init() is meaningless. This
patch removes this.
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-12-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:20 +0000 (17:10 +0800)]
vhost-vdpa: fix the wrong assertion in vhost_vdpa_init()
Vhost_vdpa_add() can fail for various reasons, so the assertion of the
succeed is wrong. Instead, we should free the NetClientState and
propagate the error to the caller
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-11-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:19 +0000 (17:10 +0800)]
vhost-vdpa: tweak the error label in vhost_vdpa_add()
Introduce new error label to avoid the unnecessary checking of net
pointer.
Fixes: 1e0a84ea49b68 ("vhost-vdpa: introduce vhost-vdpa net client") Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-10-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:18 +0000 (17:10 +0800)]
vhost-vdpa: fix leaking of vhost_net in vhost_vdpa_add()
Fixes: 1e0a84ea49b68 ("vhost-vdpa: introduce vhost-vdpa net client") Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-9-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:17 +0000 (17:10 +0800)]
vhost-vdpa: don't cleanup twice in vhost_vdpa_add()
The previous vhost_net_cleanup is sufficient for freeing, calling
vhost_vdpa_del() in this case will lead an extra round of free. Note
that this kind of "double free" is safe since vhost_dev_cleanup() zero
the whole structure.
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-8-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:16 +0000 (17:10 +0800)]
vhost-vdpa: remove the unnecessary check in vhost_vdpa_add()
The VhostVDPAState is just allocated by qemu_new_net_client() via
g_malloc0() in net_vhost_vdpa_init(). So s->vhost_net is NULL for
sure, let's remove this unnecessary check in vhost_vdpa_add().
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-7-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:15 +0000 (17:10 +0800)]
vhost_net: do not assume nvqs is always 2
This patch switches to initialize dev.nvqs from the VhostNetOptions
instead of assuming it was 2. This is useful for implementing control
virtqueue support which will be a single vhost_net structure with a
single cvq.
Note that nvqs is still set to 2 for all users and this patch does not
change functionality.
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Eli Cohen <elic@nvidia.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-6-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:14 +0000 (17:10 +0800)]
vhost: use unsigned int for nvqs
Switch to use unsigned int for nvqs since it's not expected to be
negative.
Reviewed-by: Eli Cohen <elic@nvidia.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-5-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:13 +0000 (17:10 +0800)]
vhost_net: remove the meaningless assignment in vhost_net_start_one()
The nvqs and vqs have been initialized during vhost_net_init() and are
not expected to change during the life cycle of vhost_net
structure. So this patch removes the meaningless assignment.
Reviewed-by: Eli Cohen <elic@nvidia.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-4-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Fri, 3 Sep 2021 09:10:12 +0000 (17:10 +0800)]
vhost-vdpa: correctly return err in vhost_vdpa_set_backend_cap()
We should return error code instead of zero, otherwise there's no way
for the caller to detect the failure.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-3-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210903091031.47303-2-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Peter Maydell [Wed, 1 Sep 2021 15:27:13 +0000 (16:27 +0100)]
tests/vhost-user-bridge.c: Fix typo in help message
Fix a typo in the help message printed by vhost-user-bridge.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20210901152713.25701-1-peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The vhost-user-bridge binary accepts a UNIX socket path on
the command line. Sanity check that this is short enough to
fit into a sockaddr_un before copying it in.
Fixes: Coverity CID 1432866 Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20210901152632.25511-1-peter.maydell@linaro.org> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Yuwei Zhang [Mon, 30 Aug 2021 12:34:33 +0000 (20:34 +0800)]
hw/virtio: Add flatview update in vhost_user_cleanup()
Qemu will crash on vhost backend unexpected exit and re-connect │
in some case due to access released memory.
Signed-off-by: Yuwei Zhang <zhangyuwei.9149@bytedance.com>
Message-Id: <20210830123433.45727-1-zhangyuwei.9149@bytedance.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
hw/virtio: Remove NULL check in virtio_free_region_cache()
virtio_free_region_cache() is called within call_rcu(),
always with a non-NULL argument. Ensure new code keep it
that way by replacing the NULL check by an assertion.
Add a comment this function is called within call_rcu().
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210826172658.2116840-3-philmd@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
hw/virtio: Document virtio_queue_packed_empty_rcu is called within RCU
While virtio_queue_packed_empty_rcu() uses the '_rcu' suffix,
it is not obvious it is called within rcu_read_lock(). All other
functions from this file called with the RCU locked have a comment
describing it. Document this one similarly for consistency.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210826172658.2116840-2-philmd@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Ani Sinha [Wed, 25 Aug 2021 03:19:47 +0000 (08:49 +0530)]
MAINTAINERS: Added myself as a reviewer for acpi/smbios subsystem
I have developed an interest in this space and hopefully can lend some
helping hand to Igor and Michael in reviewing simpler patches.
Signed-off-by: Ani Sinha <ani@anisinha.ca> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Acked-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20210825031949.919376-4-ani@anisinha.ca> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Ani Sinha [Wed, 25 Aug 2021 03:19:46 +0000 (08:49 +0530)]
hw/acpi: use existing references to pci device struct within functions
There is no need to use fresh typecasts to get references to pci device structs
when there is an existing reference to pci device struct. Use existing reference.
Minor cleanup.
Signed-off-by: Ani Sinha <ani@anisinha.ca> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210825031949.919376-3-ani@anisinha.ca> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Ani Sinha [Wed, 25 Aug 2021 03:19:45 +0000 (08:49 +0530)]
hw/pci: remove all references to find_i440fx function
commit c0e427d6eb5fefc538 ("hw/acpi/ich9: Enable ACPI PCI hot-plug") removed all
uses of find_i440fx() function. This has been replaced by the more generic call
acpi_get_i386_pci_host() which maybe able to find the root bus both for i440fx
machine type as well as for the q35 machine type. There seems to be no more any
need to maintain a i440fx specific version of the api call. Remove it.
Tested by building from a clean tree successfully.
Signed-off-by: Ani Sinha <ani@anisinha.ca> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210825031949.919376-2-ani@anisinha.ca> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jingqi Liu [Mon, 23 Aug 2021 01:12:54 +0000 (09:12 +0800)]
hw/i386/acpi-build: Get NUMA information from struct NumaState
Since commits aa57020774b ("numa: move numa global variable
nb_numa_nodes into MachineState") and 7e721e7b10e ("numa: move
numa global variable numa_info into MachineState"), we can get
NUMA information completely from MachineState::numa_state.
Remove PCMachineState::numa_nodes and PCMachineState::node_mem,
since they are just copied from MachineState::numa_state.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Signed-off-by: Jingqi Liu <jingqi.liu@intel.com>
Message-Id: <20210823011254.28506-1-jingqi.liu@intel.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Wed, 4 Aug 2021 03:48:03 +0000 (11:48 +0800)]
vhost: correctly detect the enabling IOMMU
Vhost used to compare the dma_as against the address_space_memory to
detect whether the IOMMU is enabled or not. This might not work well
since the virito-bus may call get_dma_as if VIRTIO_F_IOMMU_PLATFORM is
set without an actual IOMMU enabled when device is plugged. In the
case of PCI where pci_get_address_space() is used, the bus master as
is returned. So vhost actually tries to enable device IOTLB even if
the IOMMU is not enabled. This will lead a lots of unnecessary
transactions between vhost and Qemu and will introduce a huge drop of
the performance.
For PCI, an ideal approach is to use pci_device_iommu_address_space()
just for get_dma_as. But Qemu may choose to initialize the IOMMU after
the virtio-pci which lead a wrong address space is returned during
device plugged. So this patch switch to use transport specific way via
iommu_enabled() to detect the IOMMU during vhost start. In this case,
we are fine since we know the IOMMU is initialized correctly.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210804034803.1644-4-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Wed, 4 Aug 2021 03:48:02 +0000 (11:48 +0800)]
virtio-pci: implement iommu_enabled()
This patch implements the PCI transport version of iommu_enabled. This
is done by comparing the address space returned by
pci_device_iommu_address_space() against address_space_memory.
Note that an ideal approach is to use pci_device_iommu_address_space()
in get_dma_as(), but it might not work well since the IOMMU could be
initialized after the virtio-pci device is initialized.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210804034803.1644-3-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Jason Wang [Wed, 4 Aug 2021 03:48:01 +0000 (11:48 +0800)]
virtio-bus: introduce iommu_enabled()
This patch introduce a new method for the virtio-bus for the transport
to report whether or not the IOMMU is enabled for the device.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210804034803.1644-2-jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Let's compress the code a bit to improve readability. We can drop the
vm_running check in virtio_balloon_free_page_start() as it's already
properly checked in the single caller.
Cc: Wei Wang <wei.w.wang@intel.com> Cc: Michael S. Tsirkin <mst@redhat.com> Cc: Philippe Mathieu-Daudé <philmd@redhat.com> Cc: Alexander Duyck <alexander.duyck@gmail.com> Cc: Juan Quintela <quintela@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Peter Xu <peterx@redhat.com> Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210708095339.20274-3-david@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
virtio-balloon: don't start free page hinting if postcopy is possible
Postcopy never worked properly with 'free-page-hint=on', as there are
at least two issues:
1) With postcopy, the guest will never receive a VIRTIO_BALLOON_CMD_ID_DONE
and consequently won't release free pages back to the OS once
migration finishes.
The issue is that for postcopy, we won't do a final bitmap sync while
the guest is stopped on the source and
virtio_balloon_free_page_hint_notify() will only call
virtio_balloon_free_page_done() on the source during
PRECOPY_NOTIFY_CLEANUP, after the VM state was already migrated to
the destination.
2) Once the VM touches a page on the destination that has been excluded
from migration on the source via qemu_guest_free_page_hint() while
postcopy is active, that thread will stall until postcopy finishes
and all threads are woken up. (with older Linux kernels that won't
retry faults when woken up via userfaultfd, we might actually get a
SEGFAULT)
The issue is that the source will refuse to migrate any pages that
are not marked as dirty in the dirty bmap -- for example, because the
page might just have been sent. Consequently, the faulting thread will
stall, waiting for the page to be migrated -- which could take quite
a while and result in guest OS issues.
While we could fix 1) comparatively easily, 2) is harder to get right and
might require more involved RAM migration changes on source and destination
[1].
As it never worked properly, let's not start free page hinting in the
precopy notifier if the postcopy migration capability was enabled to fix
it easily. Capabilities cannot be enabled once migration is already
running.
Note 1: in the future we might either adjust migration code on the source
to track pages that have actually been sent or adjust
migration code on source and destination to eventually send
pages multiple times from the source and and deal with pages
that are sent multiple times on the destination.
Note 2: virtio-mem has similar issues, however, access to "unplugged"
memory by the guest is very rare and we would have to be very
lucky for it to happen during migration. The spec states
"The driver SHOULD NOT read from unplugged memory blocks ..."
and "The driver MUST NOT write to unplugged memory blocks".
virtio-mem will move away from virtio_balloon_free_page_done()
soon and handle this case explicitly on the destination.
Fixes: c13c4153f76d ("virtio-balloon: VIRTIO_BALLOON_F_FREE_PAGE_HINT") Cc: qemu-stable@nongnu.org Cc: Wei Wang <wei.w.wang@intel.com> Cc: Michael S. Tsirkin <mst@redhat.com> Cc: Philippe Mathieu-Daudé <philmd@redhat.com> Cc: Alexander Duyck <alexander.duyck@gmail.com> Cc: Juan Quintela <quintela@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Peter Xu <peterx@redhat.com> Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210708095339.20274-2-david@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com>
* remotes/armbru/tags/pull-qapi-2021-09-03:
qapi: Tweak error messages for unknown / conflicting 'if' keys
qapi: Tweak error messages for missing / conflicting meta-type
tests/qapi-schema: Hide OrderedDict in test output
qapi: Use re.fullmatch() where appropriate
qapi: Use "not COND" instead of "!COND" for generated documentation
qapi: Avoid redundant parens in code generated for conditionals
qapi: Factor common recursion out of cgen_ifcond(), docgen_ifcond()
qapi: Fix C code generation for 'if'
tests/qapi-schema: Demonstrate broken C code for 'if'
tests/qapi-schema: Correct two 'if' conditionals
qapi: Simplify how QAPISchemaIfCond represents "no condition"
qapi: Simplify QAPISchemaIfCond's interface for generating C
qapi: Set boolean value correctly in examples
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Eduardo Habkost [Thu, 5 Aug 2021 19:34:31 +0000 (15:34 -0400)]
Use PCI_HOST_BRIDGE macro
OBJECT_CHECK(PciHostState, ..., TYPE_PCI_HOST_BRIDGE) is exactly
what the PCI_HOST_BRIDGE macro does. We can just use the macro
instead of using OBJECT_CHECK manually.
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20210805193431.307761-7-ehabkost@redhat.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Eduardo Habkost [Thu, 5 Aug 2021 19:34:26 +0000 (15:34 -0400)]
acpi: Delete broken ACPI_GED_X86 macro
The macro never worked and never will, because the
AcpiGedX86State type never existed.
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20210805193431.307761-2-ehabkost@redhat.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Gerd Hoffmann <kraxel@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Alyssa Ross [Fri, 6 Aug 2021 14:39:26 +0000 (14:39 +0000)]
vhost-user: add missing space in error message
This would previously give error messages like
> Received unexpected msg type.Expected 0 received 1
Signed-off-by: Alyssa Ross <hi@alyssa.is>
Message-Id: <20210806143926.315725-1-hi@alyssa.is> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
hw/virtio: move vhost_set_backend_type() to vhost.c
Just a small refactor patch.
vhost_set_backend_type() gets called only in vhost.c, so we can move the
function there and make it static. We can then extern the visibility of
kernel_ops, to match the other VhostOps in vhost-backend.h.
The VhostOps constants now make more sense in vhost.h
Suggested-by: Raphael Norwitz <raphael.norwitz@nutanix.com> Signed-off-by: Tiberiu Georgescu <tiberiu.georgescu@nutanix.com>
Message-Id: <20210809134015.67941-1-tiberiu.georgescu@nutanix.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Ani Sinha [Thu, 12 Aug 2021 07:14:09 +0000 (12:44 +0530)]
hw/acpi: refactor acpi hp modules so that targets can just use what they need
Currently various acpi hotplug modules like cpu hotplug, memory hotplug, pci
hotplug, nvdimm hotplug are all pulled in when CONFIG_ACPI_X86 is turned on.
This brings in support for whole lot of subsystems that some targets like
mips does not need. They are added just to satisfy symbol dependencies. This
is ugly and should be avoided. Targets should be able to pull in just what they
need and no more. For example, mips only needs support for PIIX4 and does not
need acpi pci hotplug support or cpu hotplug support or memory hotplug support
etc. This change is an effort to clean this up.
In this change, new config variables are added for various acpi hotplug
subsystems. Targets like mips can only enable PIIX4 support and not the rest
of all the other modules which were being previously pulled in as a part of
CONFIG_ACPI_X86. Function stubs make sure that symbols which piix4 needs but
are not required by mips (for example, symbols specific to pci hotplug etc)
are available to satisfy the dependencies.
Currently, this change only addresses issues with mips malta targets. In future
we might be able to clean up other targets which are similarly pulling in lot
of unnecessary hotplug modules by enabling ACPI_X86.
This change should also address issues such as the following:
https://gitlab.com/qemu-project/qemu/-/issues/221
https://gitlab.com/qemu-project/qemu/-/issues/193
Signed-off-by: Ani Sinha <ani@anisinha.ca>
Message-Id: <20210812071409.492299-1-ani@anisinha.ca> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Gerd Hoffmann [Thu, 12 Aug 2021 10:23:41 +0000 (12:23 +0200)]
q35: catch invalid cpu hotplug configuration
Related: https://bugzilla.redhat.com//show_bug.cgi?id=1985924 Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-Id: <20210812102341.3316254-1-kraxel@redhat.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Ani Sinha [Mon, 16 Aug 2021 08:32:14 +0000 (14:02 +0530)]
hw/acpi: define PIIX4 acpi pci hotplug property strings at a single place
Now that we have "acpi-pci-hotplug-with-bridge-support" PIIX4 PM property being
used for both q35 and i440fx machine types, it is better that we defined this
property string at a single place within a header file like other PIIX4
properties. We can then use this single definition at all the places that needs
it instead of duplicating the string everywhere. While at it, this change also
adds a definition for "acpi-root-pci-hotplug" PIIX4 PM property and uses
this definition at all places that were formally using the string value.
Signed-off-by: Ani Sinha <ani@anisinha.ca>
Message-Id: <20210816083214.105740-1-ani@anisinha.ca> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Denis Plotnikov [Mon, 9 Aug 2021 10:48:24 +0000 (13:48 +0300)]
vhost: make SET_VRING_ADDR, SET_FEATURES send replies
On vhost-user-blk migration, qemu normally sends a number of commands
to enable logging if VHOST_USER_PROTOCOL_F_LOG_SHMFD is negotiated.
Qemu sends VHOST_USER_SET_FEATURES to enable buffers logging and
VHOST_USER_SET_VRING_ADDR per each started ring to enable "used ring"
data logging.
The issue is that qemu doesn't wait for reply from the vhost daemon
for these commands which may result in races between qemu expectation
of logging starting and actual login starting in vhost daemon.
The race can appear as follows: on migration setup, qemu enables dirty page
logging by sending VHOST_USER_SET_FEATURES. The command doesn't arrive to a
vhost-user-blk daemon immediately and the daemon needs some time to turn the
logging on internally. If qemu doesn't wait for reply, after sending the
command, qemu may start migrateing memory pages to a destination. At this time,
the logging may not be actually turned on in the daemon but some guest pages,
which the daemon is about to write to, may have already been transferred
without logging to the destination. Since the logging wasn't turned on,
those pages won't be transferred again as dirty. So we may end up with
corrupted data on the destination.
The same scenario is applicable for "used ring" data logging, which is
turned on with VHOST_USER_SET_VRING_ADDR command.
To resolve this issue, this patch makes qemu wait for the command result
explicitly if VHOST_USER_PROTOCOL_F_REPLY_ACK is negotiated and logging enabled.
Signed-off-by: Denis Plotnikov <den-plotnikov@yandex-team.ru>
Message-Id: <20210809104824.78830-1-den-plotnikov@yandex-team.ru> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Yajun Wu [Mon, 16 Aug 2021 01:32:35 +0000 (04:32 +0300)]
hw/virtio: Fix leak of host-notifier memory-region
If call virtio_queue_set_host_notifier_mr fails, should free
host-notifier memory-region.
Fixes: 44866521bd ("vhost-user: support registering external host notifiers") Signed-off-by: Yajun Wu <yajunw@nvidia.com>
Message-Id: <1629077555-19907-1-git-send-email-yajunw@nvidia.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Eugenio Pérez [Thu, 12 Aug 2021 14:09:33 +0000 (16:09 +0200)]
vhost-vdpa: Do not send empty IOTLB update batches
With the introduction of the batch hinting, meaningless batches can be
created with no IOTLB updates if the memory region was skipped by
vhost_vdpa_listener_skipped_section. This is the case of host notifiers
memory regions, device un/realize, and others. This causes the vdpa
device to receive dma mapping settings with no changes, a possibly
expensive operation for nothing.
To avoid that, VHOST_IOTLB_BATCH_BEGIN hint is delayed until we have a
meaningful (not skipped section) mapping or unmapping operation, and
VHOST_IOTLB_BATCH_END is not written unless at least one of _UPDATE /
_INVALIDATE has been issued.
v3:
* Use a bool instead of a counter avoiding potential number wrapping
* Fix bad check on _commit
* Move VHOST_BACKEND_F_IOTLB_BATCH check to
vhost_vdpa_iotlb_batch_begin_once
v2 (from RFC):
* Rename misleading name
* Abstract start batching function for listener_add/del
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20210812140933.226288-1-eperezma@redhat.com> Acked-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
tests/qapi-schema: Hide OrderedDict in test output
Since commit 5d83b9a130 "qapi: replace if condition list with dict
{'all': [...]}", we represent if conditionals as trees consisting of
OrderedDict, list and str. This results in less than legible test
output. For instance:
if OrderedDict([('not', OrderedDict([('any', [OrderedDict([('not', 'TEST_IF_EVT')]), OrderedDict([('not', 'TEST_IF_STRUCT')])])]))])
We intend to replace OrderedDict by dict when we get Python 3.7, which
will result in more legible output:
if {'not': {'any': [{'not': 'TEST_IF_EVT'}, {'not': 'TEST_IF_STRUCT'}]}}
Can't wait: put in a hack to get that now, with a comment to revert it
when we replace OrderedDict.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20210831123809.1107782-11-armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
qapi: Avoid redundant parens in code generated for conditionals
Commit 6cc2e4817f "qapi: introduce QAPISchemaIfCond.cgen()" caused a
minor regression: redundant parenthesis. Subsequent commits
eliminated of many of them, but not all. Get rid of the rest now.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20210831123809.1107782-8-armbru@redhat.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
When commit 5d83b9a130 "qapi: replace if condition list with dict
{'all': [...]}" made cgen_ifcond() and docgen_ifcond() recursive, it
messed up parenthesises in the former, and got them right in the
latter, as the previous commit demonstrates.
To fix, adopt the latter's working code for the former. This
generates the correct code from the previous commit's commit message.
projects / mirror_qemu.git / log