Rather than checking if the trace_marker file in sysfs exists and then
immediately proceeding to open it, we can just try to open it and handle
the resulting error if it doesn't exist.
Additionally, log the filename if the file can't be opened to make the
debugging process easier.
BUG=b:256052372
TEST=cargo build --features=trace_marker
Change-Id: I117324b0903afdbd351d03a18755321cacf1f42c
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4262323
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Keiichi Watanabe <keiichiw@chromium.org>
Many tests for virtio-blk is using TempDir just to create a temporary
file. These usage can be simplified with TempFile. So, this change does
that simplification.
BUG=None
Test=`cargo test -p devices`
Change-Id: I2b3bbd66894cc4a8f88cbf8faeb6e06dda600f30
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4272805
Reviewed-by: Keiichi Watanabe <keiichiw@chromium.org>
Commit-Queue: Takaya Saeki <takayas@chromium.org>
This reverts commit e5aec1e53d.
Reason for revert: The android fstab which is capped limited up to `X86_64_FDT_MAX_SIZE` exceeds the `SETUP_DATA_END` at least on an octopus DUT.
Original change's description:
> x86_64: place setup_data immediately after cmdline
>
> This moves the setup_data out of the way of the bzImage decompression
> stub, so it won't get overwritten.
>
> Fixes a regression with some kernel configurations introduced in commit
> 5bedfa3408 ("x86_64: add SETUP_RNG_SEED setup_data entry"), which
> unconditionally adds a setup_data entry.
>
> BUG=b:268397895
> TEST=boot x86-64 bzImage in crosvm that previously failed
> TEST=cargo test -p x86_64
>
> Change-Id: Ie5efc19d607aad16c1ca80bcdd2fb62933dd73e1
> Fixes: 5bedfa3408 ("x86_64: add SETUP_RNG_SEED setup_data entry")
> Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4233967
> Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
> Reviewed-by: Dmitrii Osipenko <dmitry.osipenko@collabora.corp-partner.google.com>
> Reviewed-by: Zihan Chen <zihanchen@google.com>
Bug: b:268397895
Change-Id: Ie430caa38b745cb1d5e69ac644a2c76813069221
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4272810
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Reviewed-by: Keiichi Watanabe <keiichiw@chromium.org>
Commit-Queue: Keiichi Watanabe <keiichiw@chromium.org>
create_plugin_jail() and create_sandbox_minijail() do the same settings
except (1) `RLIMIT_NOFILE` and (2) plugin jail does not support embeded
bpf seccomp filter. But both difference won't be a problem.
(1) Given that Parallels plugin is closed source we will not be able t
ocame up with a reasonable limit, so it might need to go the way GPU
device does it and set the limit to 32K.
(2) We validate seccomp_policy_dir in jail_config is not none.
The fallback logic for the pivot root path for plugin is a bit changed,
but it should not be a problem.
* before: cfg.plugin_root > DEFAULT_PIVOT_ROOT env > "/var/empty"
* after: cfg.plugin_root > jail_config.pivot_root > DEFAULT_PIVOT_ROOT
env > "/var/empty"
This migration also introduce minor updates to jail helpers.
* create_base_minijail() validates root path to be absolute path.
* otherwise, chroot fails and minijail fails.
* The source for mounting proc is changed ("/proc" -> "proc").
* "proc" looks the proper source to mount according to proc(5) man
page.
* "/proc" as the source was introduced by https://crrev.com/c/1717739,
but it looks they didn't have any concrete reason for it.
BUG=b:268281422
TEST=./tools/run_tests2
TEST=tast run $DUT arc.PlayStore.vm
Change-Id: Ie014cadf3cec4a049ef783546782209e26d9902a
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4262049
Reviewed-by: Dmitry Torokhov <dtor@chromium.org>
Commit-Queue: Shin Kawamura <kawasin@google.com>
WorkerThread provides a mechanism to start a thread that can be stopped
by sending an event from another thread. It also handles automatically
stopping the thread if the WorkerThread object is dropped.
BUG=b:269187468
TEST=tools/presubmit --all
Change-Id: Ifeef3f34f75fcd0496e87e038b60fa9e5e60a681
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4237530
Reviewed-by: Noah Gold <nkgold@google.com>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Frederick Mayle <fmayle@google.com>
This CL addresses lack of clear output queue command on VDA. It omits queueing
output buffer to backend if a PictureReady response is pending. This
eliminates double enqueueing of buffers and double ownership for buffers.
However in the event of clearing queue with output buffers while VDA owns
output buffers, the double ownership still occurs until a specific
buffer is enqueued again.
BUG=b:252950991
BUG=b:269093605
TEST=CtsMediaV2TestCases android.mediav2.cts.CodecDecoderSurfaceTest
TEST=CtsMediaTestCases android.media.cts.DecoderTest
Change-Id: I2f2560dc9d23077e6b6c523a1e6efa837d2d3184
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4254766
Reviewed-by: Alexandre Courbot <acourbot@chromium.org>
Commit-Queue: Bartłomiej Grzesik <bgrzesik@google.com>
Opeing a file on the monitor process is blocking the upcoming minijail
migration. Open the swap file on the main process and pass the file
descriptor to the monitor process.
Using a single swap file and mmap different ranges of it for different
regions is efficient.
BUG=b:269372016
TEST=cargo test -p swap
Change-Id: I4be150419d263ebe76ad00dae2cafce6e8660cbd
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4253294
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Shin Kawamura <kawasin@google.com>
Reviewed-by: David Stevens <stevensd@chromium.org>
The GuestMemory type is cloned to create the instance used in the virtio
device worker threads, so it is not necessary to replace the one stored
inside VirtioPciDevice when sleep() is called; it will already be
another clone of the same GuestMemory.
BUG=b:232437513
TEST=tools/presubmit --all
Change-Id: I6c047f9ab42922c45194c4940fb0f7261c3e5fc6
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4257208
Reviewed-by: Keiichi Watanabe <keiichiw@chromium.org>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Elie Kheirallah <khei@google.com>
This is admittedly an odd feature with a narrow use case. It will lower
the risk of crash analysis tools accidentally trying to read guest
memory that has been unshared from the host (i.e. in a protected VM).
Hoperfully we'll eventually have a better solution, where we only mmap
the memory regions that are shared with the host. There is no hypervisor
support for that yet.
BUG=b:238324526
Change-Id: Iac7ed38aa52778a43ac5ba22bab5a3df7a554ee6
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4251719
Commit-Queue: Frederick Mayle <fmayle@google.com>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
The crosvm-upstream repository in the ChromiumOS git host was removed,
so use crosvm instead.
BUG=None
TEST=mdbook build # and click through links
Change-Id: Ib35d930505c1a3381ce4f4fc53f54d626c810d1e
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4233968
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Zihan Chen <zihanchen@google.com>
This moves the setup_data out of the way of the bzImage decompression
stub, so it won't get overwritten.
Fixes a regression with some kernel configurations introduced in commit
5bedfa3408 ("x86_64: add SETUP_RNG_SEED setup_data entry"), which
unconditionally adds a setup_data entry.
BUG=b:268397895
TEST=boot x86-64 bzImage in crosvm that previously failed
TEST=cargo test -p x86_64
Change-Id: Ie5efc19d607aad16c1ca80bcdd2fb62933dd73e1
Fixes: 5bedfa3408 ("x86_64: add SETUP_RNG_SEED setup_data entry")
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4233967
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Dmitrii Osipenko <dmitry.osipenko@collabora.corp-partner.google.com>
Reviewed-by: Zihan Chen <zihanchen@google.com>
This is the limit used by Linux kernels since the 2.6 era, as reported
in the bzImage header as well as the COMMAND_LINE_SIZE definition in
setup.h.
The previous allocation of almost 2 MB was excessive and would have
also overlapped the memory region used for ACPI tables had a long
enough command line been provided.
We will use the area freed up by this change for the setup_data in the
next change.
BUG=b:268397895
TEST=boot x86-64 bzImage in crosvm
Change-Id: I3b5a3771549c3352e609a9685f591be38ae9fa63
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4233966
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Zihan Chen <zihanchen@google.com>
The shm_vios backend for virtio-snd has a worker thread that needs to be
stopped when the device is shutting down. Before this change, it kept a
separate `running` value in a shared mutex-protected field, using an
Event to signal the thread to observe the new value of `running`.
However, it is sufficient to just exit the worker when the Event is
signaled, so we can remove the extra `running` field. The previous
behavior of handling all events returned from the latest WaitContext
wait() call is preserved (rather than breaking out of the loop directly)
to ensure behavior is not changed due to this patch.
BUG=b:269187468
TEST=tools/presubmit --all
Change-Id: I7c053d2d3e05c05d4e71dacf1b45836a2dc6d618
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4252080
Reviewed-by: Chih-Yang Hsia <paulhsia@chromium.org>
Reviewed-by: Pattara Teerapong <pteerapong@chromium.org>
Reviewed-by: Jorge Moreira Broche <jemoreira@google.com>
Commit-Queue: Dennis Kempin <denniskempin@google.com>
They are flaky and that is expected since the feature is incomplete.
BUG=b:269174979
Change-Id: I929e822b7f81742a5dba0fef637d6143f18302b1
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4256510
Reviewed-by: Dennis Kempin <denniskempin@google.com>
Reviewed-by: Elie Kheirallah <khei@google.com>
Commit-Queue: Frederick Mayle <fmayle@google.com>
This makes the terminal behave appropriately when using a serial console
connected to stdin/stdout. In particular, there is no extra buffering of
input, and echo is disabled so input is not printed twice.
Attempting to set the terminal mode when running without an attached
console (e.g. within a GUI app where the standard handles are not
connected) will fail silently, which is fine.
BUG=b:264428954
TEST=Use bash running in crosvm from Windows command prompt
Change-Id: I2a267748401ea7d792196f5722c817ba22e1707c
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4200972
Reviewed-by: Vikram Auradkar <auradkar@google.com>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
This roughly matches the unix::terminal interface and can be used
to configure the terminal for use with a serial console (no line
buffering, interpret ANSI escapes, and so on).
BUG=b:264428954
TEST=tools/presubmit --all
Change-Id: I7efa8c405b44dadec55e08993b5f7a4c0236cf69
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4200971
Reviewed-by: Vikram Auradkar <auradkar@google.com>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
page fault logger is only for debugging purpose. It has a security
concern about exposing the trigger to enable the logger. The logger will
make the page fault handling slower and consumes disk resources.
Instead, the page fault logger is only enabled by compiler feature flag
like:
```
cargo build --features=swap,swap/log_page_fault
```
Opening a file on the monitor process is blocking the upcoming minijail
migration. Open the log file on the main process and pass the file
descriptor to the monitor process.
BUG=b:269372016
TEST=manual test
Change-Id: I5df2a32da7fb149680fa24617ffd905b2483108b
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4253287
Reviewed-by: David Stevens <stevensd@chromium.org>
Commit-Queue: Shin Kawamura <kawasin@google.com>
This is a public facing API that concierge depends upon, we should add
some test.
BUG=b:269184617
TEST=unit test
Change-Id: Ia340b6c2e89689d783a4a1e69cbfa9702a26ad83
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4248037
Reviewed-by: Keiichi Watanabe <keiichiw@chromium.org>
Reviewed-by: Alexandre Courbot <acourbot@chromium.org>
Commit-Queue: Junichi Uekawa <uekawa@chromium.org>
On some Linux hpage_pmd_size returns the size with a line break which
cause failure on parse.
BUG=b:258351526
TEST=manual test
Change-Id: If8b9a43022bd0dee5035069b9aaa68114ebaaa86
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4252438
Reviewed-by: David Stevens <stevensd@chromium.org>
Commit-Queue: Shin Kawamura <kawasin@google.com>
Cargo by default does not strip symbols for release profile builds.
Specify strip symbols for binary size measurements make it more
meaningful (chromeOS builds are already doing so). We don't want to
make this the default for all release builds as that will hinder
user's ability to report bugs and crashes.
TEST=./tools/dev_container ./tools/build_release --strip
BUG=b:257303497
Change-Id: Ic46393fbb4176bd7631dd838d72ddbead9626774
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4251718
Auto-Submit: Zihan Chen <zihanchen@google.com>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Zihan Chen <zihanchen@google.com>
Make vhost-user sound Windows compatible following the other devices.
BUG=b:269174100
TEST=downstream
Change-Id: I87ebf9a34abcac3c1d2dbbc3eecf1312c01e4c48
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4246703
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Idan Raiter <idanr@google.com>
Creates the jail create and move all policy files and helper methods to
the crate to make jail helpers available to outside of the main crate
(i.e. swap crate).
This also move devices::Minijail and JailConfig to jail crate.
BUG=b:258351526
TEST=cargo build
Change-Id: If9a148bdb3b18f8b746875d47d1077fb17707c18
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4230456
Commit-Queue: Shin Kawamura <kawasin@google.com>
Reviewed-by: Dennis Kempin <denniskempin@google.com>
Reviewed-by: Noah Gold <nkgold@google.com>
Reviewed-by: David Stevens <stevensd@chromium.org>
Reviewed-by: Alexandre Courbot <acourbot@chromium.org>
Prior to this CL, the doctest for write_obj_volatile() was failing on
a Windows device, and the doctest for read_obj_volatile() called
read_obj() rather than read_obj_volatile(), not testing the correct
function. After this CL, both tests pass on Linux and Windows, and the
doctest for read_obj_volatile() correctly tests the volatile version of
the function.
TEST=cargo test --doc in base/src, CQ
Change-Id: I4d3fd77a7ededed12231c980899187d747b61234
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4247139
Reviewed-by: Vikram Auradkar <auradkar@google.com>
Commit-Queue: Clarissa Garvey <clarissagarvey@chromium.org>
This CL replaces the old test runner at run_tests with the new one
from run_tests2. For compatibility, a symlink is maintained until
builders have updated their recipes.
The presubmit script is updated to use the new CLI interface while
running the same selection of tests.
Note: An alias --build-only is added for --no-run to ensure
compatibility with the windows builder.
BUG=b:269169405
TEST=CQ
Change-Id: Id1e664af250f53d797c2262e9ac54aa1d7a09e4b
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4246704
Commit-Queue: Zihan Chen <zihanchen@google.com>
Auto-Submit: Dennis Kempin <denniskempin@google.com>
Reviewed-by: Zihan Chen <zihanchen@google.com>
The crosvm kernel loader failed to set the high 32 bits of the command
line address, so it would use whatever value was originally loaded from
the bzImage header. Set the ext_cmd_line_ptr to the high 32 bits of the
address where crosvm loads the command line (usually 0) so the kernel
can correctly locate the cmdline in guest memory.
This fixes booting kernels with CONFIG_X86_5LEVEL=y enabled. The 5-level
paging code checks for the "no5lvl" option in early boot using the
cmdline_find_option_bool() function, which eventually calls the
get_cmd_line_ptr() function and builds a command line pointer by
combining cmd_line_ptr and ext_cmd_line_ptr. Without this patch, this
will potentially access data outside of guest memory and crash.
BUG=b:269162598
TEST=boot x86-64 Linux with CONFIG_X86_5LEVEL=y in crosvm
Change-Id: Ic59a3143529333e21ba2c37c45967314cbb0f1f5
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4246699
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Dmitry Torokhov <dtor@chromium.org>
With the previous CLs migrating many functions and types to use
zerocopy, finally we can do a big clean up of`unsafe impl DataInit`.
We are down to 93 instances now from 230 at 5th CL in this series.
Also fixes a bug introduced previously in this series where I set
the trait bound of a few write_at_addr unsafe functions to
`FromBytes` when `AsBytes` should also be required.
TEST=CQ
BUG=b:204409584
Change-Id: I6658dd246a8932493ef6a652054a23ecfde76198
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4237765
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Zihan Chen <zihanchen@google.com>
Merges have been running smoothly enough that we no longer need these.
BUG=None
TEST=None
Change-Id: If4100966eba25038dda15dded158438a5a3824d2
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4240538
Auto-Submit: Dennis Kempin <denniskempin@google.com>
Reviewed-by: Zihan Chen <zihanchen@google.com>
Commit-Queue: Zihan Chen <zihanchen@google.com>
The llvm tools provide access to commonly used llvm tools through
cargo - specifically we need llvm-strip for stripping debug data
before packaging tests for remote execution.
This is already installed by install-deps. But having it part of the
toolchain file will automatically install it on toolchain updates
as well.
BUG=None
TEST=None
Change-Id: I433ebccb5d7a2c419c5cab7476146a142967e439
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4240537
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Dennis Kempin <denniskempin@google.com>
virtio-input has a queue processing thread that handles multiple queues.
Unfortunately the thread collects all of the `needs_interrupt` booleans,
together and then only sends an interrupt for one of the queues (the
eventq). This means that the statusq will never receive its interrupt.
This means we never tell the driver that we have more buffers for it,
and so the device starves and stops sending us guest -> host input
events.
This CL splits up the interrupt tracking so that we interrupt the right
queues as needed.
BUG=b:268572174
TEST=ran emulator with a workload that uses the queue & verified that it
doesn't stall out anymore.
Change-Id: I8a5eb09caff5186a111450da29fcdf0a67087161
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4240542
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Commit-Queue: Noah Gold <nkgold@google.com>
This method was taking ownership of two objects (the handler and the
listener) without one being more eligible to being called `self` than
the other, so it makes sense to make it a regular function of the
listener, as moving it there allows us to make it private.
BUG=b:217480043
TEST=send data between the host and a VM connected to a vhost-user vsock device using socat.
TEST=send data between a VVU vsock device inside a device VM and another guest using socat.
Change-Id: I0e212f91bc488faf88730717dd67450fda420c9d
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4222842
Reviewed-by: Keiichi Watanabe <keiichiw@chromium.org>
Commit-Queue: Alexandre Courbot <acourbot@chromium.org>
Reviewed-by: Ryuichiro Chiba <chibar@chromium.org>
This object temporarily owns the list of Bus objects to sleep/wake and
allows the snapshot/restore code to be simplified by removing the manual
cleanup on error exits.
BUG=None
TEST=crosvm run ... -s /tmp/crosvm bzImage
TEST=crosvm snapshot /tmp/snapshot/tmp/crosvm
Change-Id: I7147d5145f49fa149eefcc518dfe03dec3499d97
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4226504
Reviewed-by: Noah Gold <nkgold@google.com>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Frederick Mayle <fmayle@google.com>
Reviewed-by: Elie Kheirallah <khei@google.com>
This CL removed many uses of DataInit in devices. Some paddings
are manually added/fixed to allow AsBytes to derive without ABI
changes.
TESTED=CQ
BUG=b:204409584
Change-Id: I1f8c2d5304fc8e685cc3e5166c73481f6a3f78f7
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4235224
Commit-Queue: Dennis Kempin <denniskempin@google.com>
Auto-Submit: Zihan Chen <zihanchen@google.com>
Reviewed-by: Dennis Kempin <denniskempin@google.com>
libvda is only used on ChromeOS, which compiles with
--no-default-features.
Even though libvda may not be enabled by default, it is still
part of the workspace and will be compiled (and fail without
libvda-stub).
This enables crosvm builds with the default feature set on plain
linux.
BUG=b:266129169
TEST=cargo nextest run --workspace
Change-Id: I521bb5e715852aa6be69484db04e2cf4dbc3684e
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4235223
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Auto-Submit: Dennis Kempin <denniskempin@google.com>
Reviewed-by: Alexandre Courbot <acourbot@chromium.org>
Commit-Queue: Alexandre Courbot <acourbot@chromium.org>
The repeat_tests/coverage/retry_tests options have been replaced
with profiles for pre/post-submit in https://crrev.com/c/4234898
BUG=b:261600801
TEST=lucicfg validate main.star
Change-Id: Ie55793b722a0d374996d63053d7be59a31a5ced2
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4234905
Reviewed-by: Zihan Chen <zihanchen@google.com>
Commit-Queue: Dennis Kempin <denniskempin@google.com>
Add sleep_virtio_devices which let's use get back the queues and mem
from the virtio device, to preserve state.
Add implementation for Wl.
BUG=b:232437513
Test=./tools/presubmit --all
Change-Id: Id7f652f5ae6409c65cff87e0f9458a84f0e8b370
Reviewed-on: https://chromium-review.googlesource.com/c/crosvm/crosvm/+/4184261
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Noah Gold <nkgold@google.com>
Commit-Queue: Elie Kheirallah <khei@google.com>
