Commit 852f0048f3 ("RAMBlock: make guest_memfd require uncoordinated
discard") highlighted that subsystems like VFIO may disable RAM block
discard. However, guest_memfd relies on discard operations for page
conversion between private and shared memory, potentially leading to
the stale IOMMU mapping issue when assigning hardware devices to
confidential VMs via shared memory. To address this and allow shared
device assignement, it is crucial to ensure the VFIO system refreshes
its IOMMU mappings.
RamDiscardManager is an existing interface (used by virtio-mem) to
adjust VFIO mappings in relation to VM page assignment. Effectively page
conversion is similar to hot-removing a page in one mode and adding it
back in the other. Therefore, similar actions are required for page
conversion events. Introduce the RamDiscardManager to guest_memfd to
facilitate this process.
Since guest_memfd is not an object, it cannot directly implement the
RamDiscardManager interface. Implementing it in HostMemoryBackend is
not appropriate because guest_memfd is per RAMBlock, and some RAMBlocks
have a memory backend while others do not. Notably, virtual BIOS
RAMBlocks using memory_region_init_ram_guest_memfd() do not have a
backend.
To manage RAMBlocks with guest_memfd, define a new object named
RamBlockAttributes to implement the RamDiscardManager interface. This
object can store the guest_memfd information such as the bitmap for
shared memory and the registered listeners for event notifications. A
new state_change() helper function is provided to notify listeners, such
as VFIO, allowing VFIO to do dynamically DMA map and unmap for the shared
memory according to conversion events. Note that in the current context
of RamDiscardManager for guest_memfd, the shared state is analogous to
being populated, while the private state can be considered discarded for
simplicity. In the future, it would be more complicated if considering
more states like private/shared/discarded at the same time.
In current implementation, memory state tracking is performed at the
host page size granularity, as the minimum conversion size can be one
page per request. Additionally, VFIO expected the DMA mapping for a
specific IOVA to be mapped and unmapped with the same granularity.
Confidential VMs may perform partial conversions, such as conversions on
small regions within a larger one. To prevent such invalid cases and
until support for DMA mapping cut operations is available, all
operations are performed with 4K granularity.
In addition, memory conversion failures cause QEMU to quit rather than
resuming the guest or retrying the operation at present. It would be
future work to add more error handling or rollback mechanisms once
conversion failures are allowed. For example, in-place conversion of
guest_memfd could retry the unmap operation during the conversion from
shared to private. For now, keep the complex error handling out of the
picture as it is not required.
Tested-by: Alexey Kardashevskiy <aik@amd.com>
Reviewed-by: Alexey Kardashevskiy <aik@amd.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@amd.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Link: https://lore.kernel.org/r/20250612082747.51539-5-chenyi.qiang@intel.com
[peterx: squash fixup from Chenyi to fix builds]
Signed-off-by: Peter Xu <peterx@redhat.com>
Update ReplayRamDiscard() function to return the result and unify the
ReplayRamPopulate() and ReplayRamDiscard() to ReplayRamDiscardState() at
the same time due to their identical definitions. This unification
simplifies related structures, such as VirtIOMEMReplayData, which makes
it cleaner.
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@amd.com>
Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Link: https://lore.kernel.org/r/20250612082747.51539-4-chenyi.qiang@intel.com
Signed-off-by: Peter Xu <peterx@redhat.com>
Modify memory_region_set_ram_discard_manager() to return -EBUSY if a
RamDiscardManager is already set in the MemoryRegion. The caller must
handle this failure, such as having virtio-mem undo its actions and fail
the realize() process. Opportunistically move the call earlier to avoid
complex error handling.
This change is beneficial when introducing a new RamDiscardManager
instance besides virtio-mem. After
ram_block_coordinated_discard_require(true) unlocks all
RamDiscardManager instances, only one instance is allowed to be set for
one MemoryRegion at present.
Suggested-by: David Hildenbrand <david@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@amd.com>
Tested-by: Alexey Kardashevskiy <aik@amd.com>
Reviewed-by: Alexey Kardashevskiy <aik@amd.com>
Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Link: https://lore.kernel.org/r/20250612082747.51539-3-chenyi.qiang@intel.com
Signed-off-by: Peter Xu <peterx@redhat.com>
Rename the helper to memory_region_section_intersect_range() to make it
more generic. Meanwhile, define the @end as Int128 and replace the
related operations with Int128_* format since the helper is exported as
a wider API.
Suggested-by: Alexey Kardashevskiy <aik@amd.com>
Reviewed-by: Alexey Kardashevskiy <aik@amd.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@amd.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Link: https://lore.kernel.org/r/20250612082747.51539-2-chenyi.qiang@intel.com
Signed-off-by: Peter Xu <peterx@redhat.com>
Syncing volatile memory provides no benefit, instead it can cause
performance issues in some cases. Only sync memory that is marked as
non-volatile after migration completes on destination.
Signed-off-by: Ben Chaney <bchaney@akamai.com>
Fixes: bd108a44bc (migration: ram: Switch to ram block writeback)
Link: https://lore.kernel.org/r/1CC43F59-336F-4A12-84AD-DB89E0A17A95@akamai.com
Signed-off-by: Peter Xu <peterx@redhat.com>
When the source VM attempts to connect to the destination VM's Unix
domain socket (cpr.sock) during a cpr-transfer test, race conditions can
occur if the socket file isn't ready. This can lead to connection
failures when running tests.
This patch creates and listens on the socket in advance, and passes the
pre-listened FD directly. This avoids timing issues and improves the
reliability of CPR tests.
Reviewed-by: Jason J. Herne <jjherne@linux.ibm.com>
Signed-off-by: Jaehoon Kim <jhkim@linux.ibm.com>
Reviewed-by: Steve Sistare <steven.sistare@oracle.com>
Link: https://lore.kernel.org/r/20250611205610.147008-2-jhkim@linux.ibm.com
[peterx: null-initialize opts_target, per Steve]
Signed-off-by: Peter Xu <peterx@redhat.com>
Extend cpr_transfer_input to handle SOCKET_ADDRESS_TYPE_FD alongside
SOCKET_ADDRESS_TYPE_UNIX. This change supports the use of pre-listened
socket file descriptors for cpr migration channels.
This change is particularly useful in qtest environments, where the
socket may be created externally and passed via fd.
Reviewed-by: Jason J. Herne <jjherne@linux.ibm.com>
Reviewed-by: Steve Sistare <steven.sistare@oracle.com>
Signed-off-by: Jaehoon Kim <jhkim@linux.ibm.com>
Link: https://lore.kernel.org/r/20250611205610.147008-3-jhkim@linux.ibm.com
Signed-off-by: Peter Xu <peterx@redhat.com>
If a virtual machine is paused for an extended period time, for example,
due to an incoming migration, there are also no changes on the screen.
VNC in such case increases the display update interval by
VNC_REFRESH_INTERVAL_INC (50 ms). The update interval can then grow up
to VNC_REFRESH_INTERVAL_MAX (3000 ms).
When the machine resumes, it can then take up to 3 seconds for the first
display update. Furthermore, the update interval is then halved with
each display update with changes on the screen. If there are moving
elements on the screen, such as a video, this can be perceived as
freezing and stuttering for few seconds before the movement is smooth
again.
This patch resolves this issue, by adding a listener to VM state changes
and changing the update interval when the VM state changes to RUNNING.
The update_displaychangelistener() function updates the internal timer,
and the display is refreshed immediately if the timer is expired.
Signed-off-by: Juraj Marcin <jmarcin@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Link: https://lore.kernel.org/r/20250521151616.3951178-1-jmarcin@redhat.com
Signed-off-by: Peter Xu <peterx@redhat.com>
Commit 407bc4bf90 ("qapi: Move include/qapi/qmp/ to include/qobject/")
brought the migration-helpers.c back by mistake. This file has been
replaced with migration/migration-qmp.c and
migration/migration-util.c.
Fixes: 407bc4bf90 ("qapi: Move include/qapi/qmp/ to include/qobject/")
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Message-id: 20200310152141.13959-1-peter.maydell@linaro.org
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Link: https://lore.kernel.org/r/20250523123023.19284-1-farosas@suse.de
Signed-off-by: Peter Xu <peterx@redhat.com>
There won't be any ram sync after the stage of save_complete, therefore
it's unnecessary to do manually protect for dirty pages being sent. Skip
to do this in last round can reduce noticeable downtime.
Signed-off-by: Yanfei Xu <yanfei.xu@bytedance.com>
Tested-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Link: https://lore.kernel.org/r/20250514115827.3216082-1-yanfei.xu@bytedance.com
[peterx: add comments]
Signed-off-by: Peter Xu <peterx@redhat.com>
Add property "quote-generation-socket" to tdx-guest, which is a property
of type SocketAddress to specify Quote Generation Service(QGS).
On request of GetQuote, it connects to the QGS socket, read request
data from shared guest memory, send the request data to the QGS,
and store the response into shared guest memory, at last notify
TD guest by interrupt.
command line example:
qemu-system-x86_64 \
-object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
-machine confidential-guest-support=tdx0
Note, above example uses the unix socket. It can be other types, like vsock,
which depends on the implementation of QGS.
To avoid no response from QGS server, setup a timer for the transaction.
If timeout, make it an error and interrupt guest. Define the threshold of
time to 30s at present, maybe change to other value if not appropriate.
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The error message is misleading - we successfully decoded the data,
the decoded data was simply with the wrong length.
Change the error message to show it is an length check failure with both
the received and expected values.
Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Link: https://lore.kernel.org/r/20250603050305.1704586-4-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Change sha348 to sha384.
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Link: https://lore.kernel.org/r/20250603050305.1704586-3-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The name of "enable_cpuid_0x1f" isn't right to its behavior because the
leaf 0x1f can be enabled even when "enable_cpuid_0x1f" is false.
Rename it to "force_cpuid_0x1f" to better reflect its behavior.
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Link: https://lore.kernel.org/r/20250603050305.1704586-2-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Currently, it gets below error when requesting any named cpu model with
"-cpu" to boot a TDX VM:
qemu-system-x86_64: KVM_TDX_INIT_VM failed: Invalid argument
It misleads people to think it's the bug of KVM or QEMU. It is just that
current QEMU doesn't support named cpu model for TDX.
To support named cpu models for TDX guest, there are opens to be
finalized and needs a mount of additional work.
For now, explicitly check the case when named cpu model is requested.
Error report a hint and exit.
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Link: https://lore.kernel.org/r/20250612133801.2238342-1-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When user requests PDCM explicitly via "+pdcm" without PMU enabled, emit
a warning to inform the user.
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Link: https://lore.kernel.org/r/20250304052450.465445-3-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
There is one entry relates to CPUID_EXT_PDCM in feature_dependencies[].
So it needs to get correct value of CPUID_EXT_PDCM before using
feature_dependencies[] to apply dependencies.
Besides, it also ensures CPUID_EXT_PDCM value is tracked in
env->features[FEAT_1_ECX].
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Link: https://lore.kernel.org/r/20250304052450.465445-2-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Nightly rustc complains that HPETAddrDecode has a lifetime but it is not
clearly noted that it comes from &self. Apply the compiler's suggestion
to shut it up.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
A log_mask_ln!() macro is provided which expects similar arguments as the
C version. However, the formatting works as one would expect from Rust.
To maximize code reuse the macro is just a thin wrapper around
qemu_log(). Also, just the bare minimum of logging masks is provided
which should suffice for the current use case of Rust in QEMU.
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250615112037.11992-2-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Once qemu-api is split in multiple crates, each of them will have
its own invocation of bindgen. There cannot be only one, because
there are occasional "impl" blocks for the bindgen-generated
structs (e.g. VMStateFlags or QOM classes) that have to
reside in the same crate as the bindgen-generated code.
For now, prepare for this new organization by invoking bindgen
within the qemu-api crate's build definitions; it's also a
much better place to list enums that need specific treatment
from bindgen.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When splitting the QEMU Rust bindings into multiple crates, the
bindgen-generated structs also have to be split so that it's
possible to add "impl" blocks (e.g. for Sync/Send or Default,
or even for utility methods in cases such as VMStateFlags).
Tweak various variable definitions in meson.build, to avoid naming
conflicts once there will be multiple bindgen invocations.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This removes undefined behavior associated to writing to uninitialized
fields, and makes it possible to remove "unsafe" from the instance_init
implementation.
However, the init function itself is still unsafe, because it must promise
(as a sort as MaybeUninit::assume_init) that all fields have been
initialized.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This is the trick that allows the parent-field initializer to be used
only for the object that it's meant to be initialized. This way,
the owner of a MemoryRegion must be the object that embeds it.
More information is in the comments; it's best explained with a simplified
example.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This is a smart pointer for MaybeUninit; it can be upcasted to the
already-initialized parent classes, or dereferenced to a MaybeUninit
for the class that is being initialized.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The array of BqlRefCell<HPETTimer> is not initialized yet at the
end of instance_init. In particular, the "state" field is NonNull
and therefore it is invalid to have it as zero bytes.
Note that MaybeUninit is necessary because assigning to self.timers[index]
would trigger Drop of the old value.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add a macro that makes it possible to convert a MaybeUninit<> into
another MaybeUninit<> for a single field within it. Furthermore, it is
possible to use the resulting MaybeUninitField<> in APIs that take the
parent object, such as memory_region_init_io().
This allows removing some of the undefined behavior from instance_init()
functions, though this may not be the definitive implementation.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Makes this custom_target() usage consistent with other ones in QEMU.
Fixes: 6e0dc9d2a8 ("meson: compile bundled device trees")
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250610204131.2862-3-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Commit fcb1ad456c ("system/datadir: Add new type constant for DTB files")
introduced a new type constant for DTB files and converted the boards with
bundled device trees to use it. Convert the other boards for consistency.
Fixes: fcb1ad456c ("system/datadir: Add new type constant for DTB files")
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250610204131.2862-2-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Before commit e54ef98c8a ("target/i386: do not trigger IRQ shadow for LSS"), any
write to SS in gen_movl_seg() would cause a TB exit. The changes introduced by
this commit were intended to restrict the DISAS_EOB_INHIBIT_IRQ exit to the case
where inhibit_irq is true, but missed that a DISAS_EOB_NEXT exit can still be
required when writing to SS and inhibit_irq is false.
Comparing the PE(s) && !VM86(s) section with the logic in x86_update_hflags(), we
can see that the DISAS_EOB_NEXT exit is still required for the !CODE32 case when
writing to SS in gen_movl_seg() because any change to the SS flags can affect
hflags. Similarly we can see that the existing CODE32 case is still correct since
a change to any of DS, ES and SS can affect hflags. Finally for the
gen_op_movl_seg_real() case an explicit TB exit is not needed because the segment
register selector does not affect hflags.
Update the logic in gen_movl_seg() so that a write to SS with inhibit_irq set to
false where PE(s) && !VM86(s) will generate a DISAS_EOB_NEXT exit along with the
inline comment. This has the effect of allowing Win98SE to boot in QEMU once
again.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Fixes: e54ef98c8a ("target/i386: do not trigger IRQ shadow for LSS")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2987
Link: https://lore.kernel.org/r/20250611130315.383151-1-mark.cave-ayland@ilande.co.uk
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
As pointed out by Akihiko Odaki, all Win32 libraries in MinGW have lowercase
names. This means that on (case-insensitive) Windows you can use the mixed-case
names suggested by Microsoft or all-lowercase names, while on Linux you need to
make them lowercase.
QEMU was already using lowercase names, so there is no need to test the
mixed-case name version of libSynchronization. Remove the unnecessary test
and while at it make all the tests use "required: true".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The build fails on Windows. Replace calls to Unix programs like ´cat´,
´sed´ and ´true´ with calls to ´python´ and wrap calls to
´os.path.relpath´ in try-except because it can fail when the two paths
are on different drives. Make sure to convert the Windows paths to Unix
paths to prevent warnings in generated files.
Signed-off-by: oltolm <oleg.tolmatcev@gmail.com>
Message-id: 20250612221521.1109-2-oleg.tolmatcev@gmail.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
* hw/arm/virt: Check bypass iommu is not set for iommu-map DT property
* tests/functional: Add a test for the realview-eb-mpcore machine
* qemu-options.hx: Fix reversed description of icount sleep behavior
* target/arm: Define raw write for PMU CLR registers
* docs/interop: convert qed_spec.txt to reStructuredText format
* hw/arm: make cpu targeted by arm_load_kernel the primary CPU.
* hw/intc/arm_gic: introduce a first-cpu-index property
* hw/arm/mps2: Configure the AN500 CPU with 16 MPU regions
* linux-user/arm: Fix return value of SYS_cacheflush
-----BEGIN PGP SIGNATURE-----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=9lWk
-----END PGP SIGNATURE-----
Merge tag 'pull-target-arm-20250616' of https://git.linaro.org/people/pmaydell/qemu-arm into staging
target-arm queue:
* hw/arm/virt: Check bypass iommu is not set for iommu-map DT property
* tests/functional: Add a test for the realview-eb-mpcore machine
* qemu-options.hx: Fix reversed description of icount sleep behavior
* target/arm: Define raw write for PMU CLR registers
* docs/interop: convert qed_spec.txt to reStructuredText format
* hw/arm: make cpu targeted by arm_load_kernel the primary CPU.
* hw/intc/arm_gic: introduce a first-cpu-index property
* hw/arm/mps2: Configure the AN500 CPU with 16 MPU regions
* linux-user/arm: Fix return value of SYS_cacheflush
# -----BEGIN PGP SIGNATURE-----
#
# iQJNBAABCAA3FiEE4aXFk81BneKOgxXPPCUl7RQ2DN4FAmhQJLgZHHBldGVyLm1h
# eWRlbGxAbGluYXJvLm9yZwAKCRA8JSXtFDYM3k+8EACPyIt6Tie/0kuTXG85r06P
# Iqlp59ENb4HNB43bOBlLqNNJydx1GplKnDeDVsz4MyjhtFTZ3xvY+52IbDOgb3YL
# 66uJfu1VgdnYoKOpfi3hEr9PM9QxNkENF/3+Fokn+X5tHzhNY4eo/RREQfz2tKJ5
# aKKvzmcZNYwqFLn8TKCNaAfjmqrdwVpuF+rt//4hGULZGwatt8qmlsXWLstozTAz
# SPx0utJfvblHj6a+49e0lXaL+xU5yho2J01OiNHrIGtED5Oq2Wmj/q0SV3qPiLDH
# hchuDXT71shE2JYoVeINaQpjRWzQNSPXvnEl9SyJ3Ghzop3sOZqE2iPVslvO0l5R
# rfrbR2rLnGcRPOwgntGnSMHiW+DH9ldhMUp7UvhlU3/5MJmWX2jpDfZJPsmmdyFH
# Evrf64yMO7JlLMQSMkTISZampz8aUgdlgrmQDJ4kGpdkp3GWqLE93eQGHCU16k2j
# PzjGxCIeOED4HEeGG3GrPTf0P1dy8WKla2z8ou2ANLqofrKpwFHTuGRl77ETn/bD
# 8c11hdS+uJ4Sx/efVPXWQhjCPNwLJL+DlCW0kKbGfTmiBdlcPPzPbrzmPNinVzJs
# luSGFzNit7Invfn0KZhy3l0iucMWGPdMDqN0B15RA5fFssdkuxZga5H2kOqCswOe
# 3hHo1s5PUbg+x1JAvQvi5g==
# =9lWk
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 16 Jun 2025 10:05:44 EDT
# gpg: using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg: issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [full]
# gpg: aka "Peter Maydell <pmaydell@gmail.com>" [full]
# gpg: aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [full]
# gpg: aka "Peter Maydell <peter@archaic.org.uk>" [unknown]
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83 15CF 3C25 25ED 1436 0CDE
* tag 'pull-target-arm-20250616' of https://git.linaro.org/people/pmaydell/qemu-arm:
linux-user/arm: Fix return value of SYS_cacheflush
hw/arm/mps2: Configure the AN500 CPU with 16 MPU regions
hw/intc/arm_gic: introduce a first-cpu-index property
hw/arm: make cpu targeted by arm_load_kernel the primary CPU.
docs/interop: convert qed_spec.txt to reStructuredText format
target/arm: Define raw write for PMU CLR registers
qemu-options.hx: Fix reversed description of icount sleep behavior
tests/functional: Add a test for the realview-eb-mpcore machine
hw/arm/virt: Check bypass iommu is not set for iommu-map DT property
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Although the emulated cacheflush syscall does nothing, it still needs to
return zero to indicate success.
Cc: qemu-stable@nongnu.org
Signed-off-by: J. Neuschäfer <j.neuschaefer@gmx.net>
Message-id: 20250613-cache-v1-1-ee9f4a9ba81b@gmx.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The AN500 application note documents that it configures the Cortex-M7
CPU to have 16 MPU regions. We weren't doing this in our emulation,
so the CPU had only the default 8 MPU regions. Set the mpu-ns-regions
property to 16 for this board.
This bug doesn't affect any of the other board types we model in
this source file, because they all use either the Cortex-M3 or
Cortex-M4. Those CPUs do not have an RTL configurable number of
MPU regions, and always provide 8 regions if the MPU is built in.
Cc: qemu-stable@nongnu.org
Reported-by: Corentin GENDRE <cocotroupe20@gmail.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20250605141801.1083266-1-peter.maydell@linaro.org
This introduces a first-cpu-index property to the arm-gic, as some SOCs
could have two separate GIC (ie: the zynqmp).
Signed-off-by: Clément Chigot <chigot@adacore.com>
Message-id: 20250526085523.809003-3-chigot@adacore.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: slightly expanded comment documenting GIC property]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Currently, arm booting processus assumes that the first_cpu is the CPU
that will boot: `arm_load_kernel` is powering off all but the `first_cpu`;
`do_cpu_reset` is setting the loader address only for this `first_cpu`.
For most of the boards, this isn't an issue as the kernel is loaded and
booted on the first CPU anyway. However, for zynqmp, the option
"boot-cpu" allows to choose any CPUs.
Create a new arm_boot_info entry `primary_cpu` recording which CPU will
be boot first. This one is set when `arm_boot_kernel` is called.
Signed-off-by: Clément Chigot <chigot@adacore.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20250526085523.809003-2-chigot@adacore.com
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Convert the qed_spec.txt file to reStructuredText and
include it in the manual.
buglink: https://gitlab.com/qemu-project/qemu/-/issues/527
Signed-off-by: Souleymane Conte <conte.souleymane@gmail.com>
Message-id: 20250609135124.45078-1-conte.souleymane@gmail.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: adjusted position of doc in the table of contents;
bulked up commit message; added file to MAINTAINERS section
for QED; made 'Consistency checking' a higher level section;
fixed one preexisting grammar nit (s/by from/from/)]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Raw writes to PMCNTENCLR and PMCNTENCLR_EL0 incorrectly used their
default write function, which clears written bits instead of writes the
raw value.
PMINTENCLR and PMINTENCLR_EL1 are similar registers, but they instead
had ARM_CP_NO_RAW. Commit 7a0e58fa64 ("target-arm: Split NO_MIGRATE
into ALIAS and NO_RAW") sugguests ARM_CP_ALIAS should be used instead of
ARM_CP_NO_RAW in such a case:
> We currently mark ARM coprocessor/system register definitions with
> the flag ARM_CP_NO_MIGRATE for two different reasons:
> 1) register is an alias on to state that's also visible via
> some other register, and that other register is the one
> responsible for migrating the state
> 2) register is not actually state at all (for instance the TLB
> or cache maintenance operation "registers") and it makes no
> sense to attempt to migrate it or otherwise access the raw state
>
> This works fine for identifying which registers should be ignored
> when performing migration, but we also use the same functions for
> synchronizing system register state between QEMU and the kernel
> when using KVM. In this case we don't want to try to sync state
> into registers in category 2, but we do want to sync into registers
> in category 1, because the kernel might have picked a different
> one of the aliases as its choice for which one to expose for
> migration.
These registers fall in category 1 (ARM_CP_ALIAS), not category 2
(ARM_CP_NO_RAW).
ARM_CP_NO_RAW also has another undesired side effect that hides
registers from GDB.
Properly set raw write functions and drop the ARM_CP_NO_RAW flag from
PMINTENCLR and PMINTENCLR_EL1; this fixes GDB/KVM state synchronization
of PMCNTENCLR and PMCNTENCLR_EL0, and exposes all these four registers
to GDB.
It is not necessary to add ARM_CP_ALIAS to these registers because the
flag is already set.
Signed-off-by: Akihiko Odaki <odaki@rsg.ci.i.u-tokyo.ac.jp>
Message-id: 20250531-clr-v3-1-377f9bf1746d@rsg.ci.i.u-tokyo.ac.jp
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The documentation for the -icount option incorrectly describes the behavior
of the sleep suboption. Based on the actual implementation and system
behavior, the effects of sleep=on and sleep=off were inadvertently reversed.
This commit updates the description to reflect their intended functionality.
Cc: qemu-stable@nongnu.org
Fixes: fa647905e6 ("qemu-options.hx: Fix minor issues in icount documentation")
Signed-off-by: Ethan Chen <ethan84@andestech.com>
Message-id: 20250606095728.3672832-1-ethan84@andestech.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The 'net_stream_send' method has a 68k byte array used for copying
data between guest and host. Skip the automatic zero-init of this
array to eliminate the performance overhead in the I/O hot path.
The 'buf1' array will be fully initialized when reading data off
the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-32-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The 'net_socket_send' method has a 68k byte array used for copying
data between guest and host. Skip the automatic zero-init of this
array to eliminate the performance overhead in the I/O hot path.
The 'buf1' array will be fully initialized when reading data off
the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-31-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The 'ufs_emulate_scsi_cmd' method has a 4k byte array used for
copying data from the device. Skip the automatic zero-init of
this array to eliminate the performance overhead in the I/O hot
path.
The 'outbuf' array will be fully initialized when data is copied
from the guest.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-30-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The 'megasas_dcmd_pd_get_list' and 'megasas_dcmd_get_properties'
methods have 4k structs used for copying data from the device.
Skip the automatic zero-init of this array to eliminate the
performance overhead in the I/O hot path.
The 'info' structs are manually initialized with memset(). The
compiler ought to be intelligent enough to turn the memset()
into a static initialization operation, and thus not duplicate
the automatic zero-init. Replacing memset() with '{}' makes it
unambiguous that the arrays are statically initialized.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-29-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
