fuzz/generic-fuzz: add a limit on DMA bytes written
As we have repplaced fork-based fuzzing, with reboots - we can no longer
use a timeout+exit() to avoid slow inputs. Libfuzzer has its own timer
that it uses to catch slow inputs, however these timeouts are usually
seconds-minutes long: more than enough to bog-down the fuzzing process.
However, I found that slow inputs often attempt to fill overly large DMA
requests. Thus, we can mitigate most timeouts by setting a cap on the
total number of DMA bytes written by an input.
Signed-off-by: Alexander Bulekov <alxndr@bu.edu> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
We use sparse-mem for fuzzing. For long-running fuzzing processes, we
eventually end up with many allocated sparse-mem pages. To avoid this,
clear the allocated pages on system-reset.
Signed-off-by: Alexander Bulekov <alxndr@bu.edu> Reviewed-by: Darren Kenny <darren.kenny@oracle.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Peter Maydell [Thu, 16 Feb 2023 13:09:51 +0000 (13:09 +0000)]
Merge tag 'migration-20230215-pull-request' of https://gitlab.com/juan.quintela/qemu into staging
Migration Pull request
This pull request contains:
* Add qemu_file_get_to_fd() a.k.a. make vfio happy(Avihai)
* migration/block is now DPRINTF() free zone (Philippe)
* remove res_compat and improve docs (me)
* tag 'migration-20230215-pull-request' of https://gitlab.com/juan.quintela/qemu:
migration: Rename res_{postcopy,precopy}_only
migration: Remove unused res_compatible
migration: In case of postcopy, the memory ends in res_postcopy_only
migration/block: Convert remaining DPRINTF() debug macro to trace events
migration/qemu-file: Add qemu_file_get_to_fd()
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Thu, 16 Feb 2023 11:16:39 +0000 (11:16 +0000)]
Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging
* block/iscsi: fix double-free on BUSY or similar statuses
* catch [accel] entry without accelerator
* target/i386: various fixes for BMI and ADX instructions
* make the contents of meson-buildoptions.sh stable
* tag 'for-upstream' of https://gitlab.com/bonzini/qemu:
libqtest: ensure waitpid() is only called once
libqtest: split qtest_spawn_qemu function
target/i386: fix ADOX followed by ADCX
target/i386: Fix C flag for BLSI, BLSMSK, BLSR
target/i386: Fix BEXTR instruction
tests/tcg/i386: Introduce and use reg_t consistently
vl: catch [accel] entry without accelerator
block/iscsi: fix double-free on BUSY or similar statuses
remove unnecessary extern "C" blocks
build: make meson-buildoptions.sh stable
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Juan Quintela [Wed, 8 Feb 2023 13:48:02 +0000 (14:48 +0100)]
migration: Rename res_{postcopy,precopy}_only
Once that res_compatible is removed, they don't make sense anymore.
We remove the _only preffix. And to make things clearer we rename
them to must_precopy and can_postcopy.
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
Peter Maydell [Wed, 15 Feb 2023 18:37:27 +0000 (18:37 +0000)]
Merge tag 'misc-next-pull-request' of https://gitlab.com/berrange/qemu into staging
* Document 'password-secret' option for -iscsi
* Deprecate iSCSI 'password' in favour of 'password-secret'
* Remove deprecated 'password' option for SPICE
* Fix handling of cached read buffers with TLS
migration/block: Convert remaining DPRINTF() debug macro to trace events
Finish the conversion from commit fe80c0241d
("migration: using trace_ to replace DPRINTF").
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
This has been replaced by the 'password-secret' option,
which references a 'secret' object instance.
Reviewed-by: Fabiano Rosas <farosas@suse.de> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
iscsi: add support for getting CHAP password via QCryptoSecret API
but was not mentioned in the command line docs
Reviewed-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Fabiano Rosas <farosas@suse.de> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Antoine Damhet [Tue, 15 Nov 2022 14:23:29 +0000 (15:23 +0100)]
io/channel-tls: fix handling of bigger read buffers
Since the TLS backend can read more data from the underlying QIOChannel
we introduce a minimal child GSource to notify if we still have more
data available to be read.
Signed-off-by: Antoine Damhet <antoine.damhet@shadow.tech> Signed-off-by: Charles Frey <charles.frey@shadow.tech> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Antoine Damhet [Tue, 15 Nov 2022 14:23:28 +0000 (15:23 +0100)]
crypto: TLS: introduce `check_pending`
The new `qcrypto_tls_session_check_pending` function allows the caller
to know if data have already been consumed from the backend and is
already available.
Signed-off-by: Antoine Damhet <antoine.damhet@shadow.tech> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Peter Maydell [Tue, 14 Feb 2023 14:46:10 +0000 (14:46 +0000)]
Merge tag 'pull-request-2023-02-14' of https://gitlab.com/thuth/qemu into staging
* Bump minimum Clang version to 10.0
* Improve the handling of the libdw library
* Deprecate --enable-gprof builds and remove them from CI
* Remove the deprecated "sga" device
* Some header #include clean-ups
* Make qtests more flexible with regards to missing devices
* Some small s390x-related fixes/improvements
* tag 'pull-request-2023-02-14' of https://gitlab.com/thuth/qemu: (22 commits)
hw/s390x/event-facility: Replace DO_UPCAST(SCLPEvent) by SCLP_EVENT()
tests/tcg/s390x: Use -nostdlib for softmmu tests
tests/qtest: Don't build virtio-serial-test.c if device not present
tests/qtest: bios-tables-test: Skip if missing configs
tests/qemu-iotests: Require virtio-scsi-pci
tests/qtest: Do not include hexloader-test if loader device is not present
tests/qtest: Check for devices in bios-tables-test
tests/qtest: drive_del-test: Skip tests that require missing devices
tests/qtest: Skip unplug tests that use missing devices
test/qtest: Fix coding style in device-plug-test.c
tests/qtest: hd-geo-test: Check for missing devices
tests/qtest: Add dependence on PCIE_PORT for virtio-net-failover.c
tests/qtest: Do not run lsi53c895a test if device is not present
tests/qtest: Skip PXE tests for missing devices
Do not include "qemu/error-report.h" in headers that do not need it
include/hw: Do not include "hw/registerfields.h" in headers that don't need it
hw/misc/sga: Remove the deprecated "sga" device
tests/qtest/npcm7xx_pwm-test: Be less verbose unless V=2
build: deprecate --enable-gprof builds and remove from CI
meson: Disable libdw for static builds by default
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
hw/s390x/event-facility: Replace DO_UPCAST(SCLPEvent) by SCLP_EVENT()
Use the SCLP_EVENT() QOM type-checking macro to avoid DO_UPCAST().
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230212225144.58660-16-philmd@linaro.org> Reviewed-by: Thomas Huth <thuth@redhat.com> Reviewed-by: Eric Farman <farman@linux.ibm.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Ilya Leoshkevich [Tue, 31 Jan 2023 18:20:57 +0000 (19:20 +0100)]
tests/tcg/s390x: Use -nostdlib for softmmu tests
The code currently uses -nostartfiles, but this does not prevent
linking with libc. On Fedora there is no cross-libc, so the linking
step fails.
Fix by using the more comprehensive -nostdlib (that's also what
probe_target_compiler() checks for as well).
Fixes: 503e549e441e ("tests/tcg/s390x: Test unaligned accesses to lowcore") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-Id: <20230131182057.2261614-1-iii@linux.ibm.com> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Fabiano Rosas [Wed, 8 Feb 2023 19:47:00 +0000 (16:47 -0300)]
tests/qtest: bios-tables-test: Skip if missing configs
If we build with --without-default-devices, CONFIG_HPET and
CONFIG_PARALLEL are set to N, which makes the respective devices go
missing from acpi tables.
Signed-off-by: Fabiano Rosas <farosas@suse.de> Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230208194700.11035-13-farosas@suse.de> Signed-off-by: Thomas Huth <thuth@redhat.com>
Fabiano Rosas [Wed, 8 Feb 2023 19:46:57 +0000 (16:46 -0300)]
tests/qtest: Check for devices in bios-tables-test
Do not include tests that require devices that are not available in
the QEMU build.
Signed-off-by: Fabiano Rosas <farosas@suse.de> Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20230208194700.11035-10-farosas@suse.de> Signed-off-by: Thomas Huth <thuth@redhat.com>
Fabiano Rosas [Wed, 8 Feb 2023 19:46:50 +0000 (16:46 -0300)]
tests/qtest: Do not run lsi53c895a test if device is not present
The tests are built once for all the targets, so as long as one QEMU
binary is built with CONFIG_LSI_SCSI_PCI=y, this test will
run. However some binaries might not include the device. So check this
again in runtime.
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Message-Id: <20230208194700.11035-3-farosas@suse.de> Reviewed-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Fabiano Rosas [Wed, 8 Feb 2023 19:46:49 +0000 (16:46 -0300)]
tests/qtest: Skip PXE tests for missing devices
Check if the devices we're trying to add are present in the QEMU
binary. They could have been removed from the build via Kconfig or the
--without-default-devices option.
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Message-Id: <20230208194700.11035-2-farosas@suse.de> Reviewed-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Fri, 10 Feb 2023 11:19:31 +0000 (12:19 +0100)]
Do not include "qemu/error-report.h" in headers that do not need it
Include it in the .c files instead that use the error reporting
functions.
Message-Id: <20230210111931.1115489-1-thuth@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Fri, 10 Feb 2023 11:23:15 +0000 (12:23 +0100)]
include/hw: Do not include "hw/registerfields.h" in headers that don't need it
Include "hw/registerfields.h" in the .c files instead (if needed).
Message-Id: <20230210112315.1116966-1-thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Peter Maydell [Thu, 9 Feb 2023 13:50:47 +0000 (13:50 +0000)]
tests/qtest/npcm7xx_pwm-test: Be less verbose unless V=2
The npcm7xx_pwm-test produces a lot of output at V=1, which
means that on our CI tests the log files exceed the gitlab
500KB limit. Suppress the messages about exactly what is
being tested unless at V=2 and above.
This follows the pattern we use with qom-test.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20230209135047.1753081-1-peter.maydell@linaro.org> Reviewed-by: Thomas Huth <thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Alex Bennée [Tue, 31 Jan 2023 09:42:23 +0000 (09:42 +0000)]
build: deprecate --enable-gprof builds and remove from CI
As gprof relies on instrumentation you rarely get useful data compared
to a real optimised build. Lets deprecate the build option and
simplify the CI configuration as a result.
Buglink: https://gitlab.com/qemu-project/qemu/-/issues/1338 Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230131094224.861621-1-alex.bennee@linaro.org> Reviewed-by: Thomas Huth <thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Ilya Leoshkevich [Fri, 10 Feb 2023 00:52:08 +0000 (01:52 +0100)]
meson: Disable libdw for static builds by default
Static QEMU build fails on Debian Bullseye:
/usr/bin/ld: /usr/lib/x86_64-linux-gnu/libdw.a(debuginfod-client.o): in function `__libdwfl_debuginfod_init':
(.text.startup+0x17): undefined reference to `dlopen'
The reason is that pkg-config does not suggest -ldl for libdw, and
adding --extra-ldflags="-ldl" resolves the issue. However, static
linking with libdw is an unclear topic:
* Linux perf does it.
* Debian's libdw-dev description says:
Only link to the static version for special cases and when you
don't need anything from the ebl backends.
* As the error message above indicates, -ldl is also needed for
debuginfod support.
The functionality provided by libdw is needed for analyzing performance
of JITed code, which is mostly useful to developers and researchers.
Therefore, in order to avoid unpleasant surprises for people who don't
need this, simply disable libdw for static builds by default. It can
still be enabled explicitly if needed.
Reported-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-Id: <20230210005208.438142-2-iii@linux.ibm.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Tue, 31 Jan 2023 18:02:39 +0000 (19:02 +0100)]
configure: Bump minimum Clang version to 10.0
Anthony Perard recently reported some problems with Clang v6.0 from
Ubuntu Bionic (with regards to the -Wmissing-braces configure test).
Since we're not officially supporting that version of Ubuntu anymore,
we should better bump our minimum version check in the configure script
instead of using our time to fix problems of unsupported compilers.
According to repology.org, our supported distros ship these versions
of Clang (looking at the highest version only):
While it seems like we could update to v12.0.0 from that point of view,
the default version on Ubuntu 20.04 is still v10.0, and we use that for
our CI tests based via the tests/docker/dockerfiles/ubuntu2004.docker
file.
Thus let's make v10.0 our minimum version now (which corresponds to
Apple Clang version v12.0). The -Wmissing-braces check can then be
removed, too, since both our minimum GCC and our minimum Clang version
now handle this correctly.
Message-Id: <20230131180239.1582302-1-thuth@redhat.com> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Peter Maydell [Mon, 13 Feb 2023 11:54:05 +0000 (11:54 +0000)]
Merge tag 'migration-20230213-pull-request' of https://gitlab.com/juan.quintela/qemu into staging
Migration Pull request (take3)
Hi
In this PULL request:
- Added to leonardo fixes: Fixes: b5eea99ec2 ("migration: Add yank feature") Reported-by: Li Xiaohui <xiaohli@redhat.com>
Please apply.
[take 2]
- rebase to latest upstream
- fix compilation of linux-user (if have_system was missing) (me)
- cleanup multifd_load_cleanup(leonardo)
- Document RAM flags (me)
Please apply.
[take 1]
This are all the reviewed patches for migration:
- AVX512 support for xbzrle (Ling Xu)
- /dev/userfaultd support (Peter Xu)
- Improve ordering of channels (Peter Xu)
- multifd cleanups (Li Zhang)
- Remove spurious files from last merge (me)
Rebase makes that to you
- Fix mixup between state_pending_{exact,estimate} (me)
- Cache RAM size during migration (me)
- cleanup several functions (me)
* tag 'migration-20230213-pull-request' of https://gitlab.com/juan.quintela/qemu: (22 commits)
ram: Document migration ram flags
migration/multifd: Move load_cleanup inside incoming_state_destroy
migration/multifd: Join all multifd threads in order to avoid leaks
migration/multifd: Remove unnecessary assignment on multifd_load_cleanup()
migration/multifd: Change multifd_load_cleanup() signature and usage
migration: Postpone postcopy preempt channel to be after main
migration: Add a semaphore to count PONGs
migration: Cleanup postcopy_preempt_setup()
migration: Rework multi-channel checks on URI
Update bench-code for addressing CI problem
AVX512 support for xbzrle_encode_buffer
migration: I messed state_pending_exact/estimate
migration: Make ram_save_target_page() a pointer
migration: Calculate ram size once
migration: Split ram_bytes_total_common() in two functions
migration: Make find_dirty_block() return a single parameter
migration: Simplify ram_find_and_save_block()
util/userfaultfd: Support /dev/userfaultfd
linux-headers: Update to v6.1
multifd: Remove some redundant code
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The above error happens in the target host, when multifd is being used
for precopy, and then postcopy is triggered and the migration finishes.
This will crash the VM in the target host.
To avoid that, move multifd_load_cleanup() inside
migration_incoming_state_destroy(), so that the load cleanup becomes part
of the incoming state destroying process.
Running multifd_load_cleanup() twice can become an issue, though, but the
only scenario it could be ran twice is on process_incoming_migration_bh().
So removing this extra call is necessary.
On the other hand, this multifd_load_cleanup() call happens way before the
migration_incoming_state_destroy() and having this happening before
dirty_bitmap_mig_before_vm_start() and vm_start() may be a need.
So introduce a new function multifd_load_shutdown() that will mainly stop
all multifd threads and close their QIOChannels. Then use this function
instead of multifd_load_cleanup() to make sure nothing else is received
before dirty_bitmap_mig_before_vm_start().
Fixes: b5eea99ec2 ("migration: Add yank feature") Reported-by: Li Xiaohui <xiaohli@redhat.com> Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Leonardo Bras [Fri, 10 Feb 2023 06:36:30 +0000 (03:36 -0300)]
migration/multifd: Join all multifd threads in order to avoid leaks
Current approach will only join threads that are still running.
For the threads not joined, resources or private memory are always kept in
the process space and never reclaimed before process end, and this risks
serious memory leaks.
This should usually not represent a big problem, since multifd migration
is usually just ran at most a few times, and after it succeeds there is
not much to be done before exiting the process.
Yet still, it should not hurt performance to join all of them.
Fixes: b5eea99ec2 ("migration: Add yank feature") Reported-by: Li Xiaohui <xiaohli@redhat.com> Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Leonardo Bras [Fri, 10 Feb 2023 06:36:29 +0000 (03:36 -0300)]
migration/multifd: Remove unnecessary assignment on multifd_load_cleanup()
Before assigning "p->quit = true" for every multifd channel,
multifd_load_cleanup() will call multifd_recv_terminate_threads() which
already does the same assignment, while protected by a mutex.
So there is no point doing the same assignment again.
Fixes: b5eea99ec2 ("migration: Add yank feature") Reported-by: Li Xiaohui <xiaohli@redhat.com> Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Leonardo Bras [Fri, 10 Feb 2023 06:36:28 +0000 (03:36 -0300)]
migration/multifd: Change multifd_load_cleanup() signature and usage
Since it's introduction in commit f986c3d256 ("migration: Create multifd
migration threads"), multifd_load_cleanup() never returned any value
different than 0, neither set up any error on errp.
Even though, on process_incoming_migration_bh() an if clause uses it's
return value to decide on setting autostart = false, which will never
happen.
In order to simplify the codebase, change multifd_load_cleanup() signature
to 'void multifd_load_cleanup(void)', and for every usage remove error
handling or decision made based on return value != 0.
Fixes: b5eea99ec2 ("migration: Add yank feature") Reported-by: Li Xiaohui <xiaohli@redhat.com> Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Peter Xu [Wed, 8 Feb 2023 20:28:13 +0000 (15:28 -0500)]
migration: Postpone postcopy preempt channel to be after main
Postcopy with preempt-mode enabled needs two channels to communicate. The
order of channel establishment is not guaranteed. It can happen that the
dest QEMU got the preempt channel connection request before the main
channel is established, then the migration may make no progress even during
precopy due to the wrong order.
To fix it, create the preempt channel only if we know the main channel is
established.
For a general postcopy migration, we delay it until postcopy_start(),
that's where we already went through some part of precopy on the main
channel. To make sure dest QEMU has already established the channel, we
wait until we got the first PONG received. That's something we do at the
start of precopy when postcopy enabled so it's guaranteed to happen sooner
or later.
For a postcopy recovery, we delay it to qemu_savevm_state_resume_prepare()
where we'll have round trips of data on bitmap synchronizations, which
means the main channel must have been established.
Signed-off-by: Peter Xu <peterx@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Peter Xu [Wed, 8 Feb 2023 20:28:11 +0000 (15:28 -0500)]
migration: Cleanup postcopy_preempt_setup()
Since we just dropped the only case where postcopy_preempt_setup() can
return an error, it doesn't need a retval anymore because it never fails.
Move the preempt check to the caller, preparing it to be used elsewhere to
do nothing but as simple as kicking the async connection.
Signed-off-by: Peter Xu <peterx@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Peter Xu [Wed, 8 Feb 2023 20:28:10 +0000 (15:28 -0500)]
migration: Rework multi-channel checks on URI
The whole idea of multi-channel checks was not properly done, IMHO.
Currently we check multi-channel in a lot of places, but actually that's
not needed because we only need to check it right after we get the URI and
that should be it.
If the URI check succeeded, we should never need to check it again because
we must have it. If it check fails, we should fail immediately on either
the qmp_migrate or qmp_migrate_incoming, instead of failingg it later after
the connection established.
Neither should we fail any set capabiliities like what we used to do here:
5ad15e8614 ("migration: allow enabling mutilfd for specific protocol only", 2021-10-19)
Because logically the URI will only be set later after the capability is
set, so it doesn't make a lot of sense to check the URI type when setting
the capability, because we're checking the cap with an old URI passed in,
and that may not even be the URI we're going to use later.
This patch mostly reverted all such checks for before, dropping the
variable migrate_allow_multi_channels and helpers. Instead, add a common
helper to check URI for multi-channels for either qmp_migrate and
qmp_migrate_incoming and that should do all the proper checks. The failure
will only trigger with the "migrate" or "migrate_incoming" command, or when
user specified "-incoming xxx" where "xxx" is not "defer".
Signed-off-by: Peter Xu <peterx@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
ling xu [Wed, 16 Nov 2022 15:29:23 +0000 (23:29 +0800)]
Update bench-code for addressing CI problem
Unit test code is in test-xbzrle.c, and benchmark code is in xbzrle-bench.c
for performance benchmarking. we have modified xbzrle-bench.c to address
CI problem.
Signed-off-by: ling xu <ling1.xu@intel.com> Co-authored-by: Zhou Zhao <zhou.zhao@intel.com> Co-authored-by: Jun Jin <jun.i.jin@intel.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
ling xu [Wed, 16 Nov 2022 15:29:22 +0000 (23:29 +0800)]
AVX512 support for xbzrle_encode_buffer
This commit is the same with [PATCH v6 1/2], and provides avx512 support for xbzrle_encode_buffer
function to accelerate xbzrle encoding speed. Runtime check of avx512
support and benchmark for this feature are added. Compared with C
version of xbzrle_encode_buffer function, avx512 version can achieve
50%-70% performance improvement on benchmarking. In addition, if dirty
data is randomly located in 4K page, the avx512 version can achieve
almost 140% performance gain.
Signed-off-by: ling xu <ling1.xu@intel.com> Co-authored-by: Zhou Zhao <zhou.zhao@intel.com> Co-authored-by: Jun Jin <jun.i.jin@intel.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Juan Quintela [Tue, 21 Jun 2022 11:20:35 +0000 (13:20 +0200)]
migration: Simplify ram_find_and_save_block()
We will need later that find_dirty_block() return errors, so
simplify the loop.
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Peter Xu [Tue, 7 Feb 2023 20:57:11 +0000 (15:57 -0500)]
util/userfaultfd: Support /dev/userfaultfd
Teach QEMU to use /dev/userfaultfd when it existed and fallback to the
system call if either it's not there or doesn't have enough permission.
Firstly, as long as the app has permission to access /dev/userfaultfd, it
always have the ability to trap kernel faults which QEMU mostly wants.
Meanwhile, in some context (e.g. containers) the userfaultfd syscall can be
forbidden, so it can be the major way to use postcopy in a restricted
environment with strict seccomp setup.
Signed-off-by: Peter Xu <peterx@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
Make IO channel flush call after the inflight request has been drained
in multifd thread, or else we may missed to flush the inflight request.
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
To make things worse, it appears like Zhenzhong is the one to blame.
for(int i=0; i < 1000000; i++) {
printf("I will not do rebases when I am tired\n");
}
Sorry, Juan.
Reviewed-by: Cédric Le Goater <clg@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Juan Quintela <quintela@redhat.com>
Paolo Bonzini [Fri, 13 Jan 2023 11:01:20 +0000 (12:01 +0100)]
libqtest: ensure waitpid() is only called once
If a test aborts after qtest_wait_qemu() is called, the SIGABRT hooks are
still in place and waitpid() is called again. The second time it is called,
the process does not exist anymore and the system call fails.
Move the s->qemu_pid = -1 assignment to qtest_wait_qemu() to make it
idempotent, and anyway remove the SIGABRT hook as well to avoid that
qtest_check_status() is called twice. Because of the extra call,
qtest_remove_abrt_handler() now has to be made idempotent as well.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Fri, 13 Jan 2023 10:27:14 +0000 (11:27 +0100)]
libqtest: split qtest_spawn_qemu function
In order to create a function that allows testing of invalid command
lines, extract the parts of qtest_init_without_qmp_handshake that do
not require any successful set up of sockets.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Tue, 31 Jan 2023 08:48:03 +0000 (09:48 +0100)]
target/i386: fix ADOX followed by ADCX
When ADCX is followed by ADOX or vice versa, the second instruction's
carry comes from EFLAGS and the condition codes use the CC_OP_ADCOX
operation. Retrieving the carry from EFLAGS is handled by this bit
of gen_ADCOX:
Unfortunately, in this case cc_op has been overwritten by the previous
"if" statement to CC_OP_ADCOX. This works by chance when the first
instruction is ADCX; however, if the first instruction is ADOX,
ADCX will incorrectly take its carry from OF instead of CF.
Fix by moving the computation of the new cc_op at the end of the function.
The included exhaustive test case fails without this patch and passes
afterwards.
Because ADCX/ADOX need not be invoked through the VEX prefix, this
regression bisects to commit 16fc5726a6e2 ("target/i386: reimplement
0x0f 0x38, add AVX", 2022-10-18). However, the mistake happened a
little earlier, when BMI instructions were rewritten using the new
decoder framework.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1471 Reported-by: Paul Jolly <https://gitlab.com/myitcv> Fixes: 1d0b926150e5 ("target/i386: move scalar 0F 38 and 0F 3A instruction to new decoder", 2022-10-18) Cc: qemu-stable@nongnu.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
tests/tcg/i386: Introduce and use reg_t consistently
Define reg_t based on the actual register width.
Define the inlines using that type. This will allow
input registers to 32-bit insns to be set to 64-bit
values on x86-64, which allows testing various edge cases.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230114230542.3116013-2-richard.henderson@linaro.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Fri, 13 Jan 2023 10:35:40 +0000 (11:35 +0100)]
vl: catch [accel] entry without accelerator
Avoid a SIGSEGV and return an error instead.
Reported-by: Thomas Huth <thuth@redhat.com>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1439 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Tue, 10 Jan 2023 16:36:33 +0000 (17:36 +0100)]
block/iscsi: fix double-free on BUSY or similar statuses
Commit 8c460269aa77 ("iscsi: base all handling of check condition on
scsi_sense_to_errno", 2019-07-15) removed a "goto out" so that the
same coroutine is re-entered twice; once from iscsi_co_generic_cb,
once from the timer callback iscsi_retry_timer_expired. This can
cause a crash.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1378 Reported-by: Grzegorz Zdanowski <https://gitlab.com/kiler129> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Tue, 10 Jan 2023 08:19:31 +0000 (09:19 +0100)]
remove unnecessary extern "C" blocks
A handful of header files in QEMU are wrapped with extern "C" blocks.
These are not necessary: there are C++ source files anymore in QEMU,
and even where there were some, they did not include most of these
files anyway.
Remove them for consistency.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Mon, 6 Feb 2023 12:32:32 +0000 (13:32 +0100)]
build: make meson-buildoptions.sh stable
The library directory can change depending on the multilib setup of the host.
It would be even better to detect it in configure with the same algorithm
that Meson uses, but the important thing to avoid confusing developers is
to have identical contents of scripts/meson-buildoptions.sh, independent
of the distro and architecture on which it was created.
So, for now just give a custom default value to libdir.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
* tag 'block-pull-request' of https://gitlab.com/stefanha/qemu:
iotests/detect-zeroes-registered-buf: add new test
qemu-io: add -r option to register I/O buffer
qemu-io: use BdrvRequestFlags instead of int
block: fix detect-zeroes= with BDRV_REQ_REGISTERED_BUF
virtio-blk: add missing AioContext lock
vhost-user-fs: Back up vqs before cleaning up vhost_dev
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Stefan Hajnoczi [Tue, 7 Feb 2023 20:37:19 +0000 (15:37 -0500)]
iotests/detect-zeroes-registered-buf: add new test
This regression test demonstrates that detect-zeroes works with
registered buffers. Bug details:
https://gitlab.com/qemu-project/qemu/-/issues/1404
Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230207203719.242926-5-stefanha@redhat.com>
Stefan Hajnoczi [Tue, 7 Feb 2023 20:37:18 +0000 (15:37 -0500)]
qemu-io: add -r option to register I/O buffer
The blk_register_buf() API is an optimization hint that allows some
block drivers to avoid I/O buffer housekeeping or bounce buffers.
Add an -r option to register the I/O buffer so that qemu-io can be used
to test the blk_register_buf() API. The next commit will add a test that
uses the new option.
Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230207203719.242926-4-stefanha@redhat.com>
Stefan Hajnoczi [Tue, 7 Feb 2023 20:37:17 +0000 (15:37 -0500)]
qemu-io: use BdrvRequestFlags instead of int
The block layer APIs use BdrvRequestFlags while qemu-io code uses int.
Although the code compiles and runs fine, BdrvRequestFlags is clearer
because it differentiates between other types of flags like bdrv_open()
flags.
This is purely refactoring.
Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230207203719.242926-3-stefanha@redhat.com>
Stefan Hajnoczi [Tue, 7 Feb 2023 20:37:16 +0000 (15:37 -0500)]
block: fix detect-zeroes= with BDRV_REQ_REGISTERED_BUF
When a write request is converted into a write zeroes request by the
detect-zeroes= feature, it is no longer associated with an I/O buffer.
The BDRV_REQ_REGISTERED_BUF flag doesn't make sense without an I/O
buffer and must be cleared because bdrv_co_do_pwrite_zeroes() fails with
-EINVAL when it's set.
Fiona Ebner <f.ebner@proxmox.com> bisected and diagnosed this QEMU 7.2
regression where writes containing zeroes to a blockdev with
discard=unmap,detect-zeroes=unmap fail.
Buglink: https://gitlab.com/qemu-project/qemu/-/issues/1404 Fixes: e8b6535533be ("block: add BDRV_REQ_REGISTERED_BUF request flag") Tested-by: Fiona Ebner <f.ebner@proxmox.com> Cc: qemu-stable@nongnu.org Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230207203719.242926-2-stefanha@redhat.com>
virtio_blk_update_config() calls blk_get_geometry and blk_getlength,
and both functions eventually end up calling bdrv_poll_co when not
running in a coroutine:
- blk_getlength is a co_wrapper_mixed function
- blk_get_geometry calls bdrv_get_geometry -> bdrv_nb_sectors, a
co_wrapper_mixed function too
Since we are not running in a coroutine, we need to take s->blk
AioContext lock, otherwise bdrv_poll_co will inevitably call
AIO_WAIT_WHILE and therefore try to un unlock() an AioContext lock
that was never acquired.
Steps to reproduce the issue: simply boot a VM with
-object '{"qom-type":"iothread","id":"iothread1"}' \
-blockdev '{"driver":"file","filename":"$QCOW2","aio":"native","node-name":"libvirt-1-storage","cache":{"direct":true,"no-flush":false},"auto-read-only":true,"discard":"unmap"}' \
-blockdev '{"node-name":"libvirt-1-format","read-only":false,"cache":{"direct":true,"no-flush":false},"driver":"qcow2","file":"libvirt-1-storage"}' \
-device virtio-blk-pci,iothread=iothread1,drive=libvirt-1-format,id=virtio-disk0,bootindex=1,write-cache=on
and observe that it will fail not manage to boot with "qemu_mutex_unlock_impl: Operation not permitted"
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Tested-by: Lukáš Doktor <ldoktor@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230208111148.1040083-1-eesposit@redhat.com>
This may just be because when the machine is under heavy load
running the CI tests it hits the timeout before the QEMU
under test has started to the point of being able to respond
to HMP queries.
Bump the timeout to 60 seconds to see if the intermittent
goes away.
Acked-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20230207165119.1479132-1-peter.maydell@linaro.org
* tag 'pull-include-2023-02-06-v2' of https://repo.or.cz/qemu/armbru:
Drop duplicate #include
Don't include headers already included by qemu/osdep.h
Fix non-first inclusions of qemu/osdep.h
accel: Clean up includes
block: Clean up includes
riscv: Clean up includes
target/hexagon: Clean up includes
net: Clean up includes
migration: Clean up includes
qga: Clean up includes
hw/tricore: Clean up includes
hw/input: Clean up includes
hw/cxl: Clean up includes
crypto: Clean up includes
bsd-user: Clean up includes
scripts/clean-includes: Improve --git commit message
scripts/clean-includes: Skip symbolic links
scripts/clean-includes: Don't claim duplicate headers found when not
scripts/clean-includes: Fully skip / ignore files
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
if cpu_gpr_d[r3] == 0 then we were shifting the lower register to the
right by 32 which is undefined behaviour. In this case the TriCore would
do nothing an just return the higher register cpu_reg_d[r1]. We fixed
that by detecting whether cpu_gpr_d[r3] was zero and cleared the lower
register.
tests/tcg/tricore: Add test for OPC2_32_RCRW_INSERT
DREG_RS2 and DREG_CALC_RESULT were mapped to the same register which
would not trigger https://gitlab.com/qemu-project/qemu/-/issues/653. So
let's make each register unique.
we were mixing up the "c" and "d" registers. We used "d" as a
destination register und "c" as the source. According to the TriCore ISA
manual 1.6 vol 2 it is the other way round.
we were mixing up the "c" and "d" registers. We used "d" as a
destination register und "c" as the source. According to the TriCore ISA
manual 1.6 vol 2 it is the other way round.
Tracked down with the help of scripts/clean-includes.
Signed-off-by: Markus Armbruster <armbru@redhat.com> Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com> Reviewed-by: Greg Kurz <groug@kaod.org> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230202133830.2152150-21-armbru@redhat.com>
projects / qemu-xen.git / log