From 93de2a6a4b91b72607136dd656edf03fb399d27f Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Wed, 23 Sep 2026 09:37:21 -0700 Subject: KVM: SEV: Do cache maintenance on the source VM during intra-host migration Manually perform cache maintenance on the source VM during intra-host migration to ensure no stale data is left in CPU caches after the VM is destroyed. Because the source VM is "converted" to a non-SEV VM, KVM's memory reclaim flows won't trigger cache maintenance, e.g. when all guest memory is reclaimed in response to detaching from the mmu_notifier. Note, relying on the destination VM to do cache maintenance isn't an option as KVM doesn't require identical guest memory configurations, i.e. the source VM may have access to memory that the destination VM does not. Enforcing equivalent memory configurations is infeasible, as it would require a *deep* comparison of memslots, e.g. to verify that not only are the memslot identical, but what the memslots point at is also identical. Fixes: b56639318bb2 ("KVM: SEV: Add support for SEV intra host migration") Cc: stable@vger.kernel.org Reported-by: Stefan Teodorescu Signed-off-by: Sean Christopherson Message-ID: <20260923163721.1584779-3-seanjc@google.com> Signed-off-by: Paolo Bonzini --- Documentation/gpu/vkms.rst | 287 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 287 insertions(+) create mode 100644 Documentation/gpu/vkms.rst (limited to 'Documentation/gpu/vkms.rst') diff --git a/Documentation/gpu/vkms.rst b/Documentation/gpu/vkms.rst new file mode 100644 index 000000000..1e79e62a6 --- /dev/null +++ b/Documentation/gpu/vkms.rst @@ -0,0 +1,287 @@ +.. _vkms: + +========================================== + drm/vkms Virtual Kernel Modesetting +========================================== + +.. kernel-doc:: drivers/gpu/drm/vkms/vkms_drv.c + :doc: vkms (Virtual Kernel Modesetting) + +Setup +===== + +The VKMS driver can be setup with the following steps: + +To check if VKMS is loaded, run:: + + lsmod | grep vkms + +This should list the VKMS driver. If no output is obtained, then +you need to enable and/or load the VKMS driver. +Ensure that the VKMS driver has been set as a loadable module in your +kernel config file. Do:: + + make nconfig + + Go to `Device Drivers> Graphics support` + + Enable `Virtual KMS (EXPERIMENTAL)` + +Compile and build the kernel for the changes to get reflected. +Now, to load the driver, use:: + + sudo modprobe vkms + +On running the lsmod command now, the VKMS driver will appear listed. +You can also observe the driver being loaded in the dmesg logs. + +The VKMS driver has optional features to simulate different kinds of hardware, +which are exposed as module options. You can use the `modinfo` command +to see the module options for vkms:: + + modinfo vkms + +Module options are helpful when testing, and enabling modules +can be done while loading vkms. For example, to load vkms with cursor enabled, +use:: + + sudo modprobe vkms enable_cursor=1 + +To disable the driver, use :: + + sudo modprobe -r vkms + +Configuring With Configfs +========================= + +It is possible to create and configure multiple VKMS instances via configfs. + +Start by mounting configfs and loading VKMS:: + + sudo mount -t configfs none /config + sudo modprobe vkms + +Once VKMS is loaded, ``/config/vkms`` is created automatically. Each directory +under ``/config/vkms`` represents a VKMS instance, create a new one:: + + sudo mkdir /config/vkms/my-vkms + +By default, the instance is disabled:: + + cat /config/vkms/my-vkms/enabled + 0 + +And directories are created for each configurable item of the display pipeline:: + + tree /config/vkms/my-vkms + ├── connectors + ├── crtcs + ├── enabled + ├── encoders + └── planes + +To add items to the display pipeline, create one or more directories under the +available paths. + +Start by creating one or more planes:: + + sudo mkdir /config/vkms/my-vkms/planes/plane0 + +Planes have 1 configurable attribute: + +- type: Plane type: 0 overlay, 1 primary, 2 cursor (same values as those + exposed by the "type" property of a plane) + +Continue by creating one or more CRTCs:: + + sudo mkdir /config/vkms/my-vkms/crtcs/crtc0 + +CRTCs have 1 configurable attribute: + +- writeback: Enable or disable writeback connector support by writing 1 or 0 + +Next, create one or more encoders:: + + sudo mkdir /config/vkms/my-vkms/encoders/encoder0 + +Last but not least, create one or more connectors:: + + sudo mkdir /config/vkms/my-vkms/connectors/connector0 + +Connectors have 1 configurable attribute: + +- status: Connection status: 1 connected, 2 disconnected, 3 unknown (same values + as those exposed by the "status" property of a connector) + +To finish the configuration, link the different pipeline items:: + + sudo ln -s /config/vkms/my-vkms/crtcs/crtc0 /config/vkms/my-vkms/planes/plane0/possible_crtcs + sudo ln -s /config/vkms/my-vkms/crtcs/crtc0 /config/vkms/my-vkms/encoders/encoder0/possible_crtcs + sudo ln -s /config/vkms/my-vkms/encoders/encoder0 /config/vkms/my-vkms/connectors/connector0/possible_encoders + +Since at least one primary plane is required, make sure to set the right type:: + + echo "1" | sudo tee /config/vkms/my-vkms/planes/plane0/type + +Once you are done configuring the VKMS instance, enable it:: + + echo "1" | sudo tee /config/vkms/my-vkms/enabled + +Finally, you can remove the VKMS instance disabling it:: + + echo "0" | sudo tee /config/vkms/my-vkms/enabled + +And removing the top level directory and its subdirectories:: + + sudo rm /config/vkms/my-vkms/planes/*/possible_crtcs/* + sudo rm /config/vkms/my-vkms/encoders/*/possible_crtcs/* + sudo rm /config/vkms/my-vkms/connectors/*/possible_encoders/* + sudo rmdir /config/vkms/my-vkms/planes/* + sudo rmdir /config/vkms/my-vkms/crtcs/* + sudo rmdir /config/vkms/my-vkms/encoders/* + sudo rmdir /config/vkms/my-vkms/connectors/* + sudo rmdir /config/vkms/my-vkms + +Testing With IGT +================ + +The IGT GPU Tools is a test suite used specifically for debugging and +development of the DRM drivers. +The IGT Tools can be installed from +`here `_ . + +The tests need to be run without a compositor, so you need to switch to text +only mode. You can do this by:: + + sudo systemctl isolate multi-user.target + +To return to graphical mode, do:: + + sudo systemctl isolate graphical.target + +Once you are in text only mode, you can run tests using the IGT_FORCE_DRIVER +variable to specify the device filter for the driver we want to test. +IGT_FORCE_DRIVER can also be used with the run-tests.sh script to run the +tests for a specific driver:: + + sudo IGT_FORCE_DRIVER="vkms" ./build/tests/ + sudo IGT_FORCE_DRIVER="vkms" ./scripts/run-tests.sh -t + +For example, to test the functionality of the writeback library, +we can run the kms_writeback test:: + + sudo IGT_FORCE_DRIVER="vkms" ./build/tests/kms_writeback + sudo IGT_FORCE_DRIVER="vkms" ./scripts/run-tests.sh -t kms_writeback + +You can also run subtests if you do not want to run the entire test:: + + sudo IGT_FORCE_DRIVER="vkms" ./build/tests/kms_flip --run-subtest basic-plain-flip + +Testing With KUnit +================== + +KUnit (Kernel unit testing framework) provides a common framework for unit tests +within the Linux kernel. +More information in ../dev-tools/kunit/index.rst . + +To run the VKMS KUnit tests:: + + tools/testing/kunit/kunit.py run --kunitconfig=drivers/gpu/drm/vkms/tests + +TODO +==== + +If you want to do any of the items listed below, please share your interest +with VKMS maintainers. + +IGT better support +------------------ + +Debugging: + +- kms_plane: some test cases are failing due to timeout on capturing CRC; + +Virtual hardware (vblank-less) mode: + +- VKMS already has support for vblanks simulated via hrtimers, which can be + tested with kms_flip test; in some way, we can say that VKMS already mimics + the real hardware vblank. However, we also have virtual hardware that does + not support vblank interrupt and completes page_flip events right away; in + this case, compositor developers may end up creating a busy loop on virtual + hardware. It would be useful to support Virtual Hardware behavior in VKMS + because this can help compositor developers to test their features in + multiple scenarios. + +Add Plane Features +------------------ + +There's lots of plane features we could add support for: + +- Add background color KMS property[Good to get started]. + +- Scaling. + +- Additional buffer formats. Low/high bpp RGB formats would be interesting + [Good to get started]. + +- Async updates (currently only possible on cursor plane using the legacy + cursor api). + +For all of these, we also want to review the igt test coverage and make sure +all relevant igt testcases work on vkms. They are good options for internship +project. + +Runtime Configuration +--------------------- + +We want to be able to reconfigure vkms instance without having to reload the +module through configfs. Use/Test-cases: + +- Hotplug/hotremove connectors on the fly (to be able to test DP MST handling + of compositors). + +- Change output configuration: Plug/unplug screens, change EDID, allow changing + the refresh rate. + +Writeback support +----------------- + +- The writeback and CRC capture operations share the use of composer_enabled + boolean to ensure vblanks. Probably, when these operations work together, + composer_enabled needs to refcounting the composer state to proper work. + [Good to get started] + +- Add support for cloned writeback outputs and related test cases using a + cloned output in the IGT kms_writeback. + +- As a v4l device. This is useful for debugging compositors on special vkms + configurations, so that developers see what's really going on. + +Output Features +--------------- + +- Variable refresh rate/freesync support. This probably needs prime buffer + sharing support, so that we can use vgem fences to simulate rendering in + testing. Also needs support to specify the EDID. + +- Add support for link status, so that compositors can validate their runtime + fallbacks when e.g. a Display Port link goes bad. + +CRC API Improvements +-------------------- + +- Optimize CRC computation ``compute_crc()`` and plane blending ``blend()`` + +Atomic Check using eBPF +----------------------- + +Atomic drivers have lots of restrictions which are not exposed to userspace in +any explicit form through e.g. possible property values. Userspace can only +inquiry about these limits through the atomic IOCTL, possibly using the +TEST_ONLY flag. Trying to add configurable code for all these limits, to allow +compositors to be tested against them, would be rather futile exercise. Instead +we could add support for eBPF to validate any kind of atomic state, and +implement a library of different restrictions. + +This needs a bunch of features (plane compositing, multiple outputs, ...) +enabled already to make sense. -- cgit v1.3.1