* tag 'for-upstream-8.0' of https://gitlab.com/bonzini/qemu:
i386: Add new CPU model SapphireRapids
target/i386: KVM: allow fast string operations if host supports them
target/i386: add FZRM, FSRS, FSRC
target/i386: add FSRM to TCG
MAINTAINERS: Cover RCU documentation
ci, docker: update CentOS and OpenSUSE Python to non-EOL versions
docs/devel: update and clarify lcitool instructions
lcitool: update submodule
configure: Look for auxiliary Python installations
configure: protect against escaping venv when running Meson
meson: stop looking for 'sphinx-build-3'
meson: Avoid duplicates in generated config-poison.h again
target/i386: Fix BZHI instruction
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Wang, Lei [Fri, 12 Aug 2022 05:57:51 +0000 (22:57 -0700)]
i386: Add new CPU model SapphireRapids
The new CPU model mostly inherits features from Icelake-Server, while
adding new features:
- AMX (Advance Matrix eXtensions)
- Bus Lock Debug Exception
and new instructions:
- AVX VNNI (Vector Neural Network Instruction):
- VPDPBUS: Multiply and Add Unsigned and Signed Bytes
- VPDPBUSDS: Multiply and Add Unsigned and Signed Bytes with Saturation
- VPDPWSSD: Multiply and Add Signed Word Integers
- VPDPWSSDS: Multiply and Add Signed Integers with Saturation
- FP16: Replicates existing AVX512 computational SP (FP32) instructions
using FP16 instead of FP32 for ~2X performance gain
- SERIALIZE: Provide software with a simple way to force the processor to
complete all modifications, faster, allowed in all privilege levels and
not causing an unconditional VM exit
- TSX Suspend Load Address Tracking: Allows programmers to choose which
memory accesses do not need to be tracked in the TSX read set
- AVX512_BF16: Vector Neural Network Instructions supporting BFLOAT16
inputs and conversion instructions from IEEE single precision
- fast zero-length MOVSB (KVM doesn't support yet)
- fast short STOSB (KVM doesn't support yet)
- fast short CMPSB, SCASB (KVM doesn't support yet)
Features that may be added in future versions:
- CET (virtualization support hasn't been merged)
Signed-off-by: Wang, Lei <lei4.wang@intel.com> Reviewed-by: Robert Hoo <robert.hu@linux.intel.com>
Message-Id: <20220812055751.14553-1-lei4.wang@intel.com> Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Mon, 27 Feb 2023 09:41:46 +0000 (10:41 +0100)]
target/i386: KVM: allow fast string operations if host supports them
These are just a flag that documents the performance characteristic of
an instruction; it needs no hypervisor support. So include them even
if KVM does not show them. In particular, FZRM/FSRS/FSRC have only
been added very recently, but they are available on Sapphire Rapids
processors.
Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Mon, 27 Feb 2023 09:55:46 +0000 (10:55 +0100)]
target/i386: add FZRM, FSRS, FSRC
These are three more markers for string operation optimizations.
They can all be added to TCG, whose string operations are more or
less as fast as they can be for short lengths.
Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
OBJECT_DECLARE_SIMPLE_TYPE() macro provides the OrIRQState
declaration for free. Besides, the QOM code style is to use
the structure name as typedef, and QEMU style is to use Camel
Case, so rename qemu_or_irq as OrIRQState.
Mechanical change using:
$ sed -i -e 's/qemu_or_irq/OrIRQState/g' $(git grep -l qemu_or_irq)
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20230113200138.52869-5-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
hw/or-irq: Declare QOM macros using OBJECT_DECLARE_SIMPLE_TYPE()
Missed during automatic conversion from commit 8063396bf3
("Use OBJECT_DECLARE_SIMPLE_TYPE when possible").
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20230113200138.52869-4-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
hw/irq: Declare QOM macros using OBJECT_DECLARE_SIMPLE_TYPE()
QOM *DECLARE* macros expect a typedef as first argument,
not a structure. Replace 'struct IRQState' by 'IRQState'
to avoid when modifying the macros:
../hw/core/irq.c:29:1: error: declaration of anonymous struct must be a definition
DECLARE_INSTANCE_CHECKER(struct IRQState, IRQ,
^
Use OBJECT_DECLARE_SIMPLE_TYPE instead of DECLARE_INSTANCE_CHECKER.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20230113200138.52869-3-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Since commit be8d853766 ("iothread: add I/O thread object") we
never used IOThreadClass / IOTHREAD_CLASS() / IOTHREAD_GET_CLASS(),
remove these definitions.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20230113200138.52869-2-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20230220115114.25237-9-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20230220115114.25237-8-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
cmsdk_apb_uart_create() is only used twice in the same
file. Open-code it.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20230220115114.25237-7-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Open-code the single use of xilinx_uartlite_create().
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20230220115114.25237-6-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
hw/char/xilinx_uartlite: Expose XILINX_UARTLITE QOM type
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20230220115114.25237-5-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
pl011_luminary_create() is only used for the Stellaris board,
open-code it.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20230220115114.25237-4-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
pl011_create() is only used in DeviceRealize handlers,
not a hot-path. Inlining is not justified.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20230220115114.25237-3-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Since &I2C_SLAVE(dev)->qdev == dev, no need to go back and
forth with QOM type casting. Directly use 'dev'.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20230220115114.25237-2-philmd@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Fabiano Rosas [Fri, 17 Feb 2023 20:11:36 +0000 (17:11 -0300)]
cpu-defs.h: Expose CPUTLBEntryFull to non-TCG code
This struct has no dependencies on TCG code and it is being used in
target/arm/ptw.c to simplify the passing around of page table walk
results. Those routines can be reached by KVM code via the gdbstub
breakpoint code, so take the structure out of CONFIG_TCG to make it
visible when building with --disable-tcg.
Signed-off-by: Fabiano Rosas <farosas@suse.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Fabiano Rosas [Fri, 17 Feb 2023 20:11:34 +0000 (17:11 -0300)]
target/arm: Move regime_using_lpae_format into internal.h
This function is needed by common code (ptw.c), so move it along with
the other regime_* functions in internal.h. When we enable the build
without TCG, the tlb_helper.c file will not be present.
Signed-off-by: Fabiano Rosas <farosas@suse.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Fabiano Rosas [Fri, 17 Feb 2023 20:11:28 +0000 (17:11 -0300)]
target/arm: Wrap TCG-only code in debug_helper.c
The next few patches will move helpers under CONFIG_TCG. We'd prefer
to keep the debug helpers and debug registers close together, so
rearrange the file a bit to be able to wrap the helpers with a TCG
ifdef.
Signed-off-by: Fabiano Rosas <farosas@suse.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Thomas Huth [Tue, 21 Feb 2023 13:37:14 +0000 (13:37 +0000)]
include/hw/arm/allwinner-a10.h: Remove superfluous includes from the header
pci_device.h is not needed at all in allwinner-a10.h, and serial.h
is only needed by the corresponding .c file.
Signed-off-by: Thomas Huth <thuth@redhat.com> Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20230215152233.210024-1-thuth@redhat.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Mon, 27 Feb 2023 11:58:43 +0000 (11:58 +0000)]
Merge tag 'pull-request-2023-02-27' of https://gitlab.com/thuth/qemu into staging
* Simplify device casting in w/vfio/ccw.c
* Fix memory corruption in the s390x dump code
* Various s390x TCG clean-ups
* s390x PV support for asynchronous teardown for reboot
* qemu-keymap related fixes
* Improvements for the duration of the gitlab-CI
* Deprecate the "-no-acpi" command line switch
* tag 'pull-request-2023-02-27' of https://gitlab.com/thuth/qemu: (33 commits)
Deprecate the "-no-acpi" command line switch
gitlab-ci.d/base: Mark jobs as interruptible by default
gitlab-ci.d: Build with --enable-fdt=system by default
gitlab-ci.d/buildtest-template: Simplify the configure step
gitlab-ci.d/buildtest: Disintegrate the build-coroutine-sigaltstack job
gitlab-ci.d/buildtest: Remove aarch64-softmmu from the build-system-ubuntu job
Updated the FSF address to <https://www.gnu.org/licenses/>
meson: fix dependency on qemu-keymap
qemu-keymap: Silence memory leak warning from Clang's sanitizer
configure: Add 'mkdir build' check
tests/tcg/s390x: Add sam.S
tests/tcg/s390x: Add bal.S
target/s390x: Use tcg_constant_* in translate_vx.c.inc
target/s390x: Use tcg_constant_i32 for fpinst_extract_m34
target/s390x: Use tcg_constant_* for DisasCompare
target/s390x: Use tcg_constant_* in local contexts
s390x/pv: Add support for asynchronous teardown for reboot
target/s390x: Hoist some computation in access_memmove
target/s390x: Inline do_access_{get,set}_byte
target/s390x: Remove TLB_NOTDIRTY workarounds
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Paolo Bonzini [Wed, 12 Oct 2022 16:20:44 +0000 (18:20 +0200)]
ci, docker: update CentOS and OpenSUSE Python to non-EOL versions
Python 3.6 is at end-of-life. Update the libvirt-ci module to a
version that supports overrides for targets and package mappings;
this way, QEMU can use the newer versions provided by CentOS 8 (Python
3.8) and OpenSUSE 15.3 (Python 3.9).
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Thu, 9 Feb 2023 11:24:26 +0000 (12:24 +0100)]
docs/devel: update and clarify lcitool instructions
Shorten a bit the description of what libvirt-ci does, the name of the
data files is not relevant at that point. However, the procedures to add
new build prerequisites are lacking some information, particularly with
respect to regenerating the output test files for lcitool's unit tests.
While at it, also update the paths in the libvirt-ci repository.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
John Snow [Fri, 10 Feb 2023 00:31:43 +0000 (01:31 +0100)]
configure: Look for auxiliary Python installations
At the moment, we look for just "python3" and "python", which is good
enough almost all of the time. But ... if you are on a platform that
uses an older Python by default and only offers a newer Python as an
option, you'll have to specify --python=/usr/bin/foo every time.
We can be kind and instead make a cursory attempt to locate a suitable
Python binary ourselves, looking for the remaining well-known binaries.
This configure loop will prefer, in order:
1. Whatever is specified in $PYTHON
2. python3
3. python
4. python3.11 down through python3.6
Notes:
- Python virtual environment provides binaries for "python3", "python",
and whichever version you used to create the venv,
e.g. "python3.8". If configure is invoked from inside of a venv, this
configure loop will not "break out" of that venv unless that venv is
created using an explicitly non-suitable version of Python that we
cannot use.
- In the event that no suitable python is found, the first python found
is the version used to generate the human-readable error message.
- The error message isn't printed right away to allow later
configuration code to pick up an explicitly configured python.
Signed-off-by: John Snow <jsnow@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Fri, 10 Feb 2023 16:18:54 +0000 (17:18 +0100)]
configure: protect against escaping venv when running Meson
If neither --python nor --meson are specified, Meson's generated
build.ninja will invoke Python script using the interpreter *that Meson
itself is running under*; not the one identified by configure.
This is only an issue if Meson's Python interpreter is not "the first
one in the path", which is the one that is used if --python is not
specified. A common case where this happen is when the "python3" binary
comes from a virtual environment but Meson is not installed (with pip)
in the virtual environment. In this case (presumably) whoever set up
the venv wanted to use the venv's Python interpreter to build QEMU,
while Meson might use a different one, for example an enterprise
distro's older runtime.
So, detect whether a virtual environment is setup, and if the virtual
environment does not have Meson, use the meson submodule. Meson will
then run under the virtual environment's Python interpreter.
Reported-by: John Snow <jsnow@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
John Snow [Tue, 21 Feb 2023 01:24:55 +0000 (20:24 -0500)]
meson: stop looking for 'sphinx-build-3'
Once upon a time, "sphinx-build" on certain RPM platforms invoked
specifically a Python 2.x version, while "sphinx-build-3" was a distro
shim for the Python 3.x version.
These days, none of our supported platforms utilize a 2.x version, and
those that still have 'sphinx-build-3' make it a symbolic link to
'sphinx-build'. Not searching for 'sphinx-build-3' will prefer
pip/venv installed versions of sphinx if they're available.
This adds an extremely convenient ability to test document building
ability in QEMU across multiple versions of Sphinx for the purposes of
compatibility testing.
Signed-off-by: John Snow <jsnow@redhat.com>
Message-Id: <20230221012456.2607692-6-jsnow@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Thomas Huth [Fri, 24 Feb 2023 09:05:43 +0000 (10:05 +0100)]
Deprecate the "-no-acpi" command line switch
Similar to "-no-hpet", the "-no-acpi" switch is a legacy command
line option that should be replaced with the "acpi" machine parameter
nowadays.
Message-Id: <20230224090543.1129677-1-thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Sunil V L <sunilvl@ventanamicro.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Thu, 23 Feb 2023 19:13:43 +0000 (20:13 +0100)]
gitlab-ci.d/base: Mark jobs as interruptible by default
When handling pull requests in the staging branch, it often happens
that one of the job fails due to a problem, so that the pull request
can't be merged. Peter/Richard/Stefan then informs the sender of the
pull request and continues by pushing the next pending pull request
from another subsystem maintainer. Now the problem is that there might
still be lots of other running jobs in the pipeline of the first pull
request, eating up precious CI minutes though the pipeline is not
needed anymore. We can avoid this by marking the jobs as "interruptible".
With this setting, the jobs from previous pipelines are automatically
terminated when pushing a new one. If someone does not like this auto-
matic termination, it can still be disabled in the settings of the
repository. See this URL for details:
meson: Avoid duplicates in generated config-poison.h again
Commit eed56e9a89f "configure, meson: move config-poison.h to meson"
lost a "| sort -u". Restore it. config-poison shrinks from ~4500 to
~700 lines when all targets are enabled.
Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Tested-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230206102037.3621709-1-armbru@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Thomas Huth [Tue, 7 Feb 2023 20:14:47 +0000 (21:14 +0100)]
gitlab-ci.d: Build with --enable-fdt=system by default
By using --enable-fdt=system we can make sure that the configure
script does not try to check out the "dtc" submodule. This should
help to safe some precious CI minutes in the long run.
While we're at it, also drop some now-redundant --enable-slirp
and --enable-capstone statements. These used to have the "=system"
suffix in the past, too, which has been dropped when the their
corresponding submodules had been removed. Since these features
are auto-enabled anyway now (since the containers have the right
libraries installed), we do not need the explicit --enable-...
statements anymore.
Message-Id: <20230207201447.566661-6-thuth@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Tue, 7 Feb 2023 20:14:46 +0000 (21:14 +0100)]
gitlab-ci.d/buildtest-template: Simplify the configure step
It's easier to use ${TARGETS:+--target-list="$TARGETS"} to add
a --target-list parameter depending on whether the TARGETS variable
is set or not.
Message-Id: <20230207201447.566661-5-thuth@redhat.com> Reviewed-by: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Tue, 7 Feb 2023 20:14:45 +0000 (21:14 +0100)]
gitlab-ci.d/buildtest: Disintegrate the build-coroutine-sigaltstack job
We can get rid of the build-coroutine-sigaltstack job by moving
the configure flags that should be tested here to other jobs:
Move --with-coroutine=sigaltstack to the build-system-debian job
(where the coroutines should get some more test coverage with
"make check-block", too) and --enable-trace-backends=ftrace to
the cross-s390x-kvm-only job.
Message-Id: <20230207201447.566661-4-thuth@redhat.com> Reviewed-by: David Woodhouse <dwmw2@infradead.org> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Tue, 7 Feb 2023 20:14:44 +0000 (21:14 +0100)]
gitlab-ci.d/buildtest: Remove aarch64-softmmu from the build-system-ubuntu job
aarch64-softmmu is also checked on the same version of Ubuntu in the
gcov job, so it is redundant to check again in the normal ubuntu job.
Message-Id: <20230207201447.566661-3-thuth@redhat.com> Reviewed-by: David Woodhouse <dwmw2@infradead.org> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Khadija Kamran [Mon, 20 Feb 2023 07:01:09 +0000 (07:01 +0000)]
Updated the FSF address to <https://www.gnu.org/licenses/>
The Free Software Foundation moved to a new address and some
sources in QEMU referred to their old location.
The address should be updated and replaced by a pointer to
<https://www.gnu.org/licenses/>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/379 Signed-off-by: Khadija Kamran <kkamran.bese16seecs@seecs.edu.pk>
Message-Id: <576ee9203fdac99d7251a98faa66b9ce1e7febc5.1675941486.git.kkamran.bese16seecs@seecs.edu.pk> Reviewed-by: Stefan Weil <sw@weilnetz.de> Signed-off-by: Thomas Huth <thuth@redhat.com>
Steve Sistare [Mon, 6 Feb 2023 18:34:02 +0000 (10:34 -0800)]
meson: fix dependency on qemu-keymap
When qemu-keymap is not available on the host, and enable-xkbcommon
is specified, parallel make fails with:
% make clean
...
% make -j 32
...
FAILED: pc-bios/keymaps/is
./qemu-keymap -f pc-bios/keymaps/is -l is
/bin/sh: ./qemu-keymap: No such file or directory
... many similar messages ...
The code always runs find_program, rather than waiting to build
qemu-keymap, because it looks for CONFIG_XKBCOMMON in config_host
rather than config_host_data. Making serially succeeds, by soft
linking files from pc-bios/keymaps, but that is not the desired
result for enable-xkbcommon.
Examining all occurrences of 'in config_host' for similar bugs shows one
instance in the docs, which is also fixed here.
Fixes: 4113f4cfee ("meson: move xkbcommon to meson") Signed-off-by: Steve Sistare <steven.sistare@oracle.com> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <1675708442-74966-1-git-send-email-steven.sistare@oracle.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Direct leak of 136 byte(s) in 1 object(s) allocated from:
#0 0x5642830d0820 in __interceptor_calloc.part.11 asan_malloc_linux.cpp.o
#1 0x7f31873b8d2b in xkb_state_new (/lib64/libxkbcommon.so.0+0x1dd2b) (BuildId: dd32581e2248833243f3f646324ae9b98469f025)
SUMMARY: AddressSanitizer: 136 byte(s) leaked in 1 allocation(s).
It can be silenced by properly releasing the "state" again
after it has been used.
Message-Id: <20230221122440.612281-1-thuth@redhat.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
target/s390x: Use tcg_constant_* in translate_vx.c.inc
In most cases, this is a simple local allocate and free
replaced by tcg_constant_*. In three cases, a variable
temp was initialized with a constant value -- reorg to
localize the constant. In gen_acc, this fixes a leak.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-Id: <20230220184052.163465-5-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
The a and b fields are not modified by the consumer,
and while we need not free a constant, tcg will quietly
ignore such frees, so free_compare need not be changed.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-Id: <20230220184052.163465-3-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Claudio Imbrenda [Tue, 14 Feb 2023 16:30:35 +0000 (17:30 +0100)]
s390x/pv: Add support for asynchronous teardown for reboot
This patch adds support for the asynchronous teardown for reboot for
protected VMs.
When attempting to tear down a protected VM, try to use the new
asynchronous interface first. If that fails, fall back to the classic
synchronous one.
The asynchronous interface involves invoking the new
KVM_PV_ASYNC_DISABLE_PREPARE command for the KVM_S390_PV_COMMAND ioctl.
This will prepare the current protected VM for asynchronous teardown.
Once the protected VM is prepared for teardown, execution can continue
immediately.
Once the protected VM has been prepared, a new thread is started to
actually perform the teardown. The new thread uses the new
KVM_PV_ASYNC_DISABLE command for the KVM_S390_PV_COMMAND ioctl. The
previously prepared protected VM is torn down in the new thread.
Once KVM_PV_ASYNC_DISABLE is invoked, it is possible to use
KVM_PV_ASYNC_DISABLE_PREPARE again. If a protected VM has already been
prepared and its cleanup has not started, it will not be possible to
prepare a new VM. In that case the classic synchronous teardown has to
be performed.
The synchronous teardown will now also clean up any prepared VMs whose
asynchronous teardown has not been initiated yet.
This considerably speeds up the reboot of a protected VM; for large VMs
especially, it could take a long time to perform a reboot with the
traditional synchronous teardown, while with this patch it is almost
immediate.
Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com> Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230214163035.44104-3-imbrenda@linux.ibm.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
target/s390x: Hoist some computation in access_memmove
Ensure that the total length is in a local variable
across the byte loop. Compute size1 difference once.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-8-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Inline into the parent functions with a simple test
to select the page, and a new define to remove ifdefs.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-7-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
When this code was written, it was using tlb_vaddr_to_host,
which does not handle TLB_DIRTY. Since then, it has been
converted to probe_access_flags, which does.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Acked-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-6-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Assign to access struct immediately, rather than waiting
until the end of the function. This means we can pass
address of haddr struct members instead of allocating
extra space on the local stack.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-5-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
The interface from probe_access_flags is void*, and matching
that will be helpful. We already rely on the gcc extension
for byte arithmetic on void*.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-4-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
target/s390x: Pass S390Access pointer into access_prepare
Passing a pointer from the caller down to access_prepare_nf
eliminates a structure copy.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-3-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
In db9aab5783a2 we broke the contract of s390_probe_access, in that it
no longer returned an exception code, nor set __excp_addr. Fix both.
Reported-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20230109201856.3916639-2-richard.henderson@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Wed, 15 Feb 2023 08:57:03 +0000 (09:57 +0100)]
target/s390x/arch_dump: Simplify memory allocation in s390x_write_elf64_notes()
We are not on a hot path here, so there is no real need for the logic
here with the split heap and stack space allocation. Simplify it by
always allocating memory from the heap.
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230215085703.746788-1-thuth@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Tue, 14 Feb 2023 14:10:56 +0000 (15:10 +0100)]
target/s390x/arch_dump: Fix memory corruption in s390x_write_elf64_notes()
"note_size" can be smaller than sizeof(note), so unconditionally calling
memset(notep, 0, sizeof(note)) could cause a memory corruption here in
case notep has been allocated dynamically, thus let's use note_size as
length argument for memset() instead.
Reported-by: Sebastian Mitterle <smitterl@redhat.com> Fixes: 113d8f4e95 ("s390x: pv: Add dump support")
Message-Id: <20230214141056.680969-1-thuth@redhat.com> Reviewed-by: Janosch Frank <frankja@linux.ibm.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
hw/vfio/ccw: Replace DO_UPCAST(VFIOCCWDevice) by VFIO_CCW()
Use the VFIO_CCW() QOM type-checking macro to avoid DO_UPCAST().
Reviewed-by: Eric Farman <farman@linux.ibm.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230213170145.45666-7-philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
No need to double-cast, call CCW_DEVICE() on VFIOCCWDevice.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Eric Farman <farman@linux.ibm.com>
Message-Id: <20230213170145.45666-6-philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
hw/vfio/ccw: Replace DO_UPCAST(S390CCWDevice) by S390_CCW_DEVICE()
Use the S390_CCW_DEVICE() QOM type-checking macro to avoid DO_UPCAST().
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230213170145.45666-5-philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
hw/vfio/ccw: Use intermediate S390CCWDevice variable
'cdev' is VFIOCCWDevice's private parent object.
Access it using the S390_CCW_DEVICE() QOM macro.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Eric Farman <farman@linux.ibm.com>
Message-Id: <20230213170145.45666-4-philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
We can directly use the QOM DEVICE() macro to get the parent object.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Eric Farman <farman@linux.ibm.com>
Message-Id: <20230213170145.45666-3-philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Thu, 16 Feb 2023 14:29:15 +0000 (15:29 +0100)]
Do not include hw/hw.h if it is not necessary
hw.h only contains the protoype of one function nowadays, hw_error(),
so all files that do not use this function anymore also do not need
to include this header anymore.
Message-Id: <20230216142915.304481-1-thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Thomas Huth <thuth@redhat.com>
Thomas Huth [Wed, 15 Feb 2023 12:41:22 +0000 (13:41 +0100)]
tests/qtest/rtl8139-test: Make the test less verbose by default
We are facing the issues that some test logs in the gitlab CI are
too big (and thus cut off). The rtl8139-test is one of the few qtests
that prints many lines of output by default when running with V=1, so
it contributes to this problem. Almost all other qtests are silent
with V=1 and only print debug messages with V=2 and higher. Thus let's
change the rtl8139-test to behave more like the other tests and only
print the debug messages with V=2 (or higher).
Message-Id: <20230215124122.72037-1-thuth@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
Peter Maydell [Sun, 26 Feb 2023 20:14:46 +0000 (20:14 +0000)]
Merge tag 'pull-riscv-to-apply-20230224' of github.com:palmer-dabbelt/qemu into staging
Fourth RISC-V PR for QEMU 8.0, Attempt 2
* A triplet of cleanups to the kernel/initrd loader that avoids
duplication between the various boards.
* Weiwei Li, Daniel Henrique Barboza, and Liu Zhiwei have been added as
reviewers. Thanks for the help!
* A fix for PMP matching to avoid incorrectly appling the default
permissions on PMP permission violations.
* A cleanup to avoid an unnecessary avoid env_archcpu() in
cpu_get_tb_cpu_state().
* Fixes for the vector slide instructions to avoid truncating 64-bit
values (such as doubles) on 32-bit targets.
# -----BEGIN PGP SIGNATURE-----
#
# iQJHBAABCAAxFiEEKzw3R0RoQ7JKlDp6LhMZ81+7GIkFAmP5Br8THHBhbG1lckBk
# YWJiZWx0LmNvbQAKCRAuExnzX7sYiT4RD/9hdSlQlR1g/2h4fbCJ3U0GvyNH0T7N
# mt3AX8hFvmfR1O63qqVVebJSHM1dTm6WsA19vKE5tdtbjV5V8UZuBTSqYeRBSrLd
# LK9IHhwv3k9OQ/EG8CgRo7HEMxAurpC26zTf3chnfwa1Wyl5XxCXNx5hPbhu18G9
# oxw0sBi51T0Tb+N6lOVVSfmiEZWLXRq+lDCZdV0j864brsSjo4x8VEGrLaFTOJLf
# X4MW6vBI4Pcb7EGnHjj5WvRKsf8gdahdx8bSTjORIm8oGri9Iyw6Vrg2khuhjnuH
# 99sD1O06cvrylp+sCOVei8H3S6/xCepQXUXnCBCd1/cetgV+olo+ZR78Z8ZjXPED
# jhZ23lsDcge+4W141lsCiwLgzI0YO3Ac+84zQLIvcx16c8zow3G9FO9sTlBSsgnW
# 0XJrsUF7AZB6quUSMytG7WK+OBizzCRwj7ItC+Mty68wLrei5lDVj8b0t8hAQEdr
# dOb7jku+Dz8OspGZx1aDKKifGDO+Ppv4PjAM2G44OmkM824SvvFg8+FEr9NgbKbp
# VgTZDCeVC6IEpzthKsK8WeompLo7Sc33KITqwMbGiyGs+gsnmgKP2bcTLF8YTlFk
# dqFBWjo3tjH5oukgTLCSYY4xPaHR9q418vGAfRox15GtUVliQ9iL5oH47PVXg4U7
# YsNZ74nD1pUueg==
# =Umli
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 24 Feb 2023 18:49:35 GMT
# gpg: using RSA key 2B3C3747446843B24A943A7A2E1319F35FBB1889
# gpg: issuer "palmer@dabbelt.com"
# gpg: Good signature from "Palmer Dabbelt <palmer@dabbelt.com>" [unknown]
# gpg: aka "Palmer Dabbelt <palmerdabbelt@google.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg: There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 00CE 76D1 8349 60DF CE88 6DF8 EF4C A150 2CCB AB41
# Subkey fingerprint: 2B3C 3747 4468 43B2 4A94 3A7A 2E13 19F3 5FBB 1889
* tag 'pull-riscv-to-apply-20230224' of github.com:palmer-dabbelt/qemu:
target/riscv: Fix vslide1up.vf and vslide1down.vf
target/riscv: avoid env_archcpu() in cpu_get_tb_cpu_state()
target/riscv: Smepmp: Skip applying default rules when address matches
MAINTAINERS: Add some RISC-V reviewers
target/riscv: Remove privileged spec version restriction for RVV
hw/riscv/boot.c: make riscv_load_initrd() static
hw/riscv/boot.c: consolidate all kernel init in riscv_load_kernel()
hw/riscv: handle 32 bit CPUs kernel_entry in riscv_load_kernel()
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
* tag 'for-upstream' of https://repo.or.cz/qemu/kevin: (29 commits)
block/rbd: Add support for layered encryption
block/rbd: Add luks-any encryption opening option
block/rbd: Remove redundant stack variable passphrase_len
virtio-scsi: reset SCSI devices from main loop thread
dma-helpers: prevent dma_blk_cb() vs dma_aio_cancel() race
scsi: protect req->aiocb with AioContext lock
block: Mark bdrv_co_refresh_total_sectors() and callers GRAPH_RDLOCK
block: Mark bdrv_*_dirty_bitmap() and callers GRAPH_RDLOCK
block: Mark bdrv_co_delete_file() and callers GRAPH_RDLOCK
block: Mark bdrv_(un)register_buf() GRAPH_RDLOCK
block: Mark bdrv_co_eject/lock_medium() and callers GRAPH_RDLOCK
block: Mark bdrv_co_is_inserted() and callers GRAPH_RDLOCK
block: Mark bdrv_co_io_(un)plug() and callers GRAPH_RDLOCK
block: Mark bdrv_co_create() and callers GRAPH_RDLOCK
block: Mark preadv_snapshot/snapshot_block_status GRAPH_RDLOCK
block: Mark bdrv_co_copy_range() GRAPH_RDLOCK
block: Mark bdrv_co_do_pwrite_zeroes() GRAPH_RDLOCK
block: Mark bdrv_co_pwrite_sync() and callers GRAPH_RDLOCK
block: Mark public read/write functions GRAPH_RDLOCK
block: Mark read/write in block/io.c GRAPH_RDLOCK
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
* tag 'for-upstream' of https://gitlab.com/bonzini/qemu:
target/i386/gdbstub: Fix a bug about order of FPU stack in 'g' packets.
docs: build-platforms: refine requirements on Python build dependencies
thread-posix: add support for setting threads name on OpenBSD
target/i386: Fix 32-bit AD[CO]X insns in 64-bit mode
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Fri, 24 Feb 2023 10:05:43 +0000 (10:05 +0000)]
Merge tag 'pull-tcg-20230221' of https://gitlab.com/rth7680/qemu into staging
tcg: Allow first half of insn in ram, and second half in mmio
linux-user/sparc: SIGILL for unknown trap vectors
linux-user/microblaze: SIGILL for privileged insns
linux-user: Fix deadlock while exiting due to signal
target/microblaze: Add gdbstub xml
util: Adjust cacheflush for windows-arm64
include/sysemu/os-win32: Adjust setjmp/longjmp for windows-arm64
* tag 'pull-tcg-20230221' of https://gitlab.com/rth7680/qemu:
sysemu/os-win32: fix setjmp/longjmp on windows-arm64
util/cacheflush: fix cache on windows-arm64
target/microblaze: Add gdbstub xml
linux-user/microblaze: Handle privileged exception
cpus: Make {start,end}_exclusive() recursive
linux-user: Always exit from exclusive state in fork_end()
linux-user/sparc: Raise SIGILL for all unhandled software traps
accel/tcg: Allow the second page of an instruction to be MMIO
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
LIU Zhiwei [Mon, 13 Feb 2023 09:45:50 +0000 (17:45 +0800)]
target/riscv: Fix vslide1up.vf and vslide1down.vf
vslide1up_##BITWIDTH is used by the vslide1up.vx and vslide1up.vf. So its
scalar input should be uint64_t to hold the 64 bits float register.And the
same for vslide1down_##BITWIDTH.
This bug is caught when run these instructions on qemu-riscv32.
Signed-off-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com> Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn> Reviewed-by: Frank Chang <frank.chang@sifive.com>
Message-ID: <20230213094550.29621-1-zhiwei_liu@linux.alibaba.com> Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
target/riscv: Smepmp: Skip applying default rules when address matches
When MSECCFG.MML is set, after checking the address range in PMP if the
asked permissions are not same as programmed in PMP, the default
permissions are applied. This should only be the case when there
is no matching address is found.
This patch skips applying default rules when matching address range
is found. It returns the index of the match PMP entry.
Fixes: 824cac681c3 (target/riscv: Fix PMP propagation for tlb) Signed-off-by: Himanshu Chauhan <hchauhan@ventanamicro.com> Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20230209055206.229392-1-hchauhan@ventanamicro.com> Signed-off-by: Alistair Francis <alistair.francis@wdc.com> Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
This patch adds some active RISC-V members as reviewers to the
MAINTAINERS file.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com> Acked-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com> Acked-by: Weiwei Li <liweiwei@iscas.ac.cn> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com> Reviewed-by: Frank Chang <frank.chang@sifive.com>
Message-Id: <20230209003308.738237-1-alistair.francis@opensource.wdc.com> Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Frank Chang [Wed, 8 Feb 2023 06:32:08 +0000 (14:32 +0800)]
target/riscv: Remove privileged spec version restriction for RVV
The RVV specification does not require that the core needs to support
the privileged specification v1.12.0 to support RVV, and there is no
dependency from ISA level.
This commit removes the restriction from both RVV CSRs and extension CPU
ISA string.
Signed-off-by: Frank Chang <frank.chang@sifive.com> Reviewed-by: Bin Meng <bmeng@tinylab.org> Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com> Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20230208063209.27279-1-frank.chang@sifive.com> Signed-off-by: Alistair Francis <alistair.francis@wdc.com> Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Or Ozeri [Sun, 29 Jan 2023 11:31:20 +0000 (05:31 -0600)]
block/rbd: Add support for layered encryption
Starting from ceph Reef, RBD has built-in support for layered encryption,
where each ancestor image (in a cloned image setting) can be possibly
encrypted using a unique passphrase.
A new function, rbd_encryption_load2, was added to librbd API.
This new function supports an array of passphrases (via "spec" structs).
This commit extends the qemu rbd driver API to use this new librbd API,
in order to support this new layered encryption feature.
Signed-off-by: Or Ozeri <oro@il.ibm.com>
Message-Id: <20230129113120.722708-4-oro@oro.sl.cloud9.ibm.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Or Ozeri [Sun, 29 Jan 2023 11:31:19 +0000 (05:31 -0600)]
block/rbd: Add luks-any encryption opening option
Ceph RBD encryption API required specifying the encryption format
for loading encryption. The supported formats were LUKS (v1) and LUKS2.
Starting from Reef release, RBD also supports loading with "luks-any" format,
which works for both versions of LUKS.
This commit extends the qemu rbd driver API to enable qemu users to use
this luks-any wildcard format.
Signed-off-by: Or Ozeri <oro@il.ibm.com>
Message-Id: <20230129113120.722708-3-oro@oro.sl.cloud9.ibm.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Or Ozeri <oro@il.ibm.com>
Message-Id: <20230129113120.722708-2-oro@oro.sl.cloud9.ibm.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Tue, 21 Feb 2023 21:22:18 +0000 (16:22 -0500)]
virtio-scsi: reset SCSI devices from main loop thread
When an IOThread is configured, the ctrl virtqueue is processed in the
IOThread. TMFs that reset SCSI devices are currently called directly
from the IOThread and trigger an assertion failure in blk_drain() from
the following call stack:
The blk_drain() function is not designed to be called from an IOThread
because it needs the Big QEMU Lock (BQL).
This patch defers TMFs that reset SCSI devices to a Bottom Half (BH)
that runs in the main loop thread under the BQL. This way it's safe to
call blk_drain() and the assertion failure is avoided.
Introduce s->tmf_bh_list for tracking TMF requests that have been
deferred to the BH. When the BH runs it will grab the entire list and
process all requests. Care must be taken to clear the list when the
virtio-scsi device is reset or unrealized. Otherwise deferred TMF
requests could execute later and lead to use-after-free or other
undefined behavior.
The s->resetting counter that's used by TMFs that reset SCSI devices is
accessed from multiple threads. This patch makes that explicit by using
atomic accessor functions. With this patch applied the counter is only
modified by the main loop thread under the BQL but can be read by any
thread.
Reported-by: Qing Wang <qinwang@redhat.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230221212218.1378734-4-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Tue, 21 Feb 2023 21:22:17 +0000 (16:22 -0500)]
dma-helpers: prevent dma_blk_cb() vs dma_aio_cancel() race
dma_blk_cb() only takes the AioContext lock around ->io_func(). That
means the rest of dma_blk_cb() is not protected. In particular, the
DMAAIOCB field accesses happen outside the lock.
There is a race when the main loop thread holds the AioContext lock and
invokes scsi_device_purge_requests() -> bdrv_aio_cancel() ->
dma_aio_cancel() while an IOThread executes dma_blk_cb(). The dbs->acb
field determines how cancellation proceeds. If dma_aio_cancel() sees
dbs->acb == NULL while dma_blk_cb() is still running, the request can be
completed twice (-ECANCELED and the actual return value).
The following assertion can occur with virtio-scsi when an IOThread is
used:
Fix the race by holding the AioContext across dma_blk_cb(). Now
dma_aio_cancel() under the AioContext lock will not see
inconsistent/intermediate states.
Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230221212218.1378734-3-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Tue, 21 Feb 2023 21:22:16 +0000 (16:22 -0500)]
scsi: protect req->aiocb with AioContext lock
If requests are being processed in the IOThread when a SCSIDevice is
unplugged, scsi_device_purge_requests() -> scsi_req_cancel_async() races
with I/O completion callbacks. Both threads load and store req->aiocb.
This can lead to assert(r->req.aiocb == NULL) failures and undefined
behavior.
Protect r->req.aiocb with the AioContext lock to prevent the race.
Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230221212218.1378734-2-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:22:02 +0000 (16:22 +0100)]
block: Mark bdrv_co_refresh_total_sectors() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_refresh_total_sectors() need to hold a reader lock for the
graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-24-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:22:01 +0000 (16:22 +0100)]
block: Mark bdrv_*_dirty_bitmap() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_*_dirty_bitmap() need to hold a reader lock for the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-23-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:22:00 +0000 (16:22 +0100)]
block: Mark bdrv_co_delete_file() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_delete_file() need to hold a reader lock for the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-22-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:21:59 +0000 (16:21 +0100)]
block: Mark bdrv_(un)register_buf() GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_register_buf() and bdrv_unregister_buf() need to hold a reader lock
for the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-21-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:21:58 +0000 (16:21 +0100)]
block: Mark bdrv_co_eject/lock_medium() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_eject() and bdrv_co_lock_medium() need to hold a reader lock for
the graph.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-20-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
block: Mark bdrv_co_is_inserted() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_is_inserted() need to hold a reader lock for the graph.
blk_is_inserted() is done as a co_wrapper_mixed_bdrv_rdlock (unlike most
other blk_* functions) because it is called a lot from other blk_co_*()
functions that already hold the lock. These calls go through
blk_is_available(), which becomes a co_wrapper_mixed_bdrv_rdlock, too,
for the same reason.
Functions that run in a coroutine and can call bdrv_co_is_available()
directly are changed to do so, which results in better TSA coverage.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-19-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:21:56 +0000 (16:21 +0100)]
block: Mark bdrv_co_io_(un)plug() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_io_plug() and bdrv_co_io_unplug() need to hold a reader lock for
the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-18-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 3 Feb 2023 15:21:55 +0000 (16:21 +0100)]
block: Mark bdrv_co_create() and callers GRAPH_RDLOCK
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_create() need to hold a reader lock for the graph.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-17-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
projects / mirror_qemu.git / log