virtio devices should be able to specify capabilities
BUG=chromium:936567
TEST=boot vm
Change-Id: I049f9967eb59a7904528fff5aea844e30c636e28
Reviewed-on: https://chromium-review.googlesource.com/1493012
Commit-Ready: Daniel Prilik <prilik@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Stephen Barber <smbarber@chromium.org>
Avoiding Box<dyn Error> makes it less likely that we display errors with
insufficient context by accident.
Many of the errors touched in this CL already had helpful message
written! But those corresponding enum variants were never being
instantiated, and that bug was masked by Box<dyn Error>. For example see
the Error::LoadCmdline and Error::LoadKernel.
pub enum Error {
LoadCmdline(kernel_loader::Error),
...
}
Before this CL:
// Bug: boxes the underlying error without adding LoadCmdline
kernel_loader::load_cmdline(...)?;
After this CL:
kernel_loader::load_cmdline(...).map_err(Error::LoadCmdline)?;
TEST=cargo check
TEST=cargo check --all-features
TEST=cargo check --target aarch64-unknown-linux-gnu
Change-Id: I7c0cff843c2211565226b9dfb4142ad6b7fa15ac
Reviewed-on: https://chromium-review.googlesource.com/1502112
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
A bug has been filed to fix this differently. Until then, add a TODO and
a clippy disable so that clippy can be used to test for other issues
without stopping on this error.
BUG=928767
TEST=cargo clippy
Signed-off-by: Dylan Reid <dgreid@chromium.org>
Change-Id: Ic264bc9101653c30354415c913e9ee3752985706
Reviewed-on: https://chromium-review.googlesource.com/1506308
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Very similar to the trackpad device, it has the INPUT_PROP_DIRECT
property and does not support any buttons, just touch events.
Change-Id: I2c963013e402ff2aa1b4b529c6c494dd57f4add9
Reviewed-on: https://chromium-review.googlesource.com/1509697
Commit-Ready: Jorge Moreira Broche <jemoreira@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
This is convenient for types that embed Option<...>.
BUG=chromium:831850
TEST=cargo test
Change-Id: I65aa3515f49e4424eff54f17ae18dac8e85b0afa
Reviewed-on: https://chromium-review.googlesource.com/1510812
Commit-Ready: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: David Tolnay <dtolnay@chromium.org>
Reviewed-by: Jingkui Wang <jkwang@google.com>
underflow occurs when configuring a 64 bit register with a <33 bit
address.
BUG=chromium:924405
TEST=boot VM
Change-Id: I53a309b7bff3c91012bacb12d9fc9f8ceed68699
Reviewed-on: https://chromium-review.googlesource.com/1493011
Commit-Ready: Daniel Prilik <prilik@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-by: Zach Reizner <zachr@chromium.org>
In case crosvm starts with elevated capabilities (for example, we need
to start with CAP_SETGID to be able to map additional gids into plugin
jail), we should drop them before spawning VCPU threads.
BUG=b:117989168
TEST=Start plugin via concierge_client and verify the process does not
have any effective or permitted privileges.
tast run [] 'vm.*'
Change-Id: Ia1e80bfe19b296936d77fe9ffeda361211b41eed
Reviewed-on: https://chromium-review.googlesource.com/1506296
Commit-Ready: Dmitry Torokhov <dtor@chromium.org>
Tested-by: Dmitry Torokhov <dtor@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chirantan Ekbote <chirantan@chromium.org>
u64 register callback will only be invoked when the write is done.
BUG=chromium:831850
TEST=local build
CQ-DEPEND=CL:1509514
Change-Id: Id0be69535898fdcc4ba24d3151df7a5107a2725b
Reviewed-on: https://chromium-review.googlesource.com/1509515
Commit-Ready: Zach Reizner <zachr@chromium.org>
Tested-by: Jingkui Wang <jkwang@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Then we don't need to unwrap
BUG=chromium:831850
TEST=cargo test
CQ-DEPEND=CL:1506828
Change-Id: I4200ea6351d61df1974e5e4c8583e783b21ea0eb
Reviewed-on: https://chromium-review.googlesource.com/1509514
Commit-Ready: Zach Reizner <zachr@chromium.org>
Tested-by: Jingkui Wang <jkwang@google.com>
Reviewed-by: Jingkui Wang <jkwang@google.com>
Avoiding Box<dyn Error> makes it less likely that we display errors with
insufficient context by accident.
For example the following code which existed before this CL:
let dev_file = OpenOptions::new()
.read(true)
.write(true)
.open(dev_path)
.map_err(|e| Box::new(e))?;
This code converts io::Error directly to Box<dyn Error> without
providing enough context to debug what happened just from the io error
message.
The new code is forced to provide a dedicated Error enum variant which
carries a handwritten message and possibly further context:
.map_err(|e| Error::OpenVinput(dev_path.to_owned(), e))?;
TEST=cargo check
TEST=cargo check --all-features
TEST=cargo check --target aarch64-unknown-linux-gnu
Change-Id: I67d3f1f6f3f92a10d63462584e9546f8ad7074b2
Reviewed-on: https://chromium-review.googlesource.com/1501656
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
I made a separate function to create each distinct VirtioDeviceStub.
Some advantages:
- Makes it easier to see the top level structure of create_devices
(which used to be 463 lines all in one function) -- how it loops
over inputs, how it decides to conditionally create particular
device types.
- Makes it clearer to follow when resources are shared across multiple
devices, particularly resource_bridge_wl_socket. The uses of
resource_bridge_wl_socket used to be 87 lines apart before this CL.
Now it spans only 27 lines so fits on one screen.
- Reduces indentation to leave more space for rustfmt to format the
device creation in a more readable way.
- Improves the ability to use short variable names whose meanings are
scoped to one device. For example wayland and balloon device
creation used to have to manipulate wayland_device_socket and
balloon_device_socket. Now create_wayland_device and
create_balloon_device can each deal with just a thing called
'socket' whose meaning is specific to the appropriate device.
TEST=cargo check
TEST=cargo check --all-features
Change-Id: I50dc53051598668b4d3cabbae588add783b1fb79
Reviewed-on: https://chromium-review.googlesource.com/1501652
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Enough failure cases have been added to `add_pci_bar` and
`add_pci_capabilities` that they should return unique errors instead of
an `Option`.
BUG=none
TEST=cargo test in devices
Signed-off-by: Dylan Reid <dgreid@chromium.org>
Change-Id: Ice2a06d2944011f95707f113f9d709da15c90cfe
Reviewed-on: https://chromium-review.googlesource.com/1497740
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Builders of crosvm, such as the crosvm ebuild, should determine if they
want LTO enabled. By having lto removed from the release profile,
builders using `--release` will not be forced to use LTO.
BUG=None
TEST=cargo build --release should be heckin' fast
Change-Id: I4e231b6dac7670ab146d36c1c1660ab67935285a
Reviewed-on: https://chromium-review.googlesource.com/1497734
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Chirantan Ekbote <chirantan@chromium.org>
This changes the default Linux kernel command line from panic=1 (reboot
one second after panic) to panic=-1 (reboot immediately on panic). The
kernel should not normally panic; this is just to improve quality of
life for developer workflows, such as running bash as init and exiting
the shell to shut down the VM.
BUG=None
TEST=crosvm run -r vm_rootfs.img -p init=/bin/bash vm_kernel; exit shell
Change-Id: I7c9084ccf1786cd4455fd748512078e02fdb17fa
Signed-off-by: Daniel Verkamp <dverkamp@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1500872
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: David Tolnay <dtolnay@chromium.org>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
For the commands that verify the length of arguments, returning
Ok on an invalid length will confuse any external programs that
launch crosvm because they still get exit code of success from
crosvm even when the command failed.
Also add a missed return of 'create_qcow2' sub-command.
BUG=None
TEST=some basic tests:
() cargo check
() verify the return value from changed sub-commands is not 0 with
an invalid argument length.
() launch a VM with concierge_client from a root shell
Change-Id: I8278107a4d2fcf3cb6fafb65f30f431f97f7deb1
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-on: https://chromium-review.googlesource.com/1501552
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Check that the device can be created. This test would have caught the
bug with adding pci bars.
Change-Id: Ib0cc2edf0d8d1b2d95d9c3588ac325b5da886603
Signed-off-by: Dylan Reid <dgreid@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1497738
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chirantan Ekbote <chirantan@chromium.org>
On arm, if the device is not coherent, guest kernel will allocate dma memory as
no-cacheable. Crosvm on the host user space thinks the memory is
cacheable. Thus when guest kernel write to dma memory, it will bypass
the cache, crosvm won't see the change.
BUG=None
TEST=local build and test
Change-Id: If6cf2d28afec61d5beb136628116ff9e7e0483f4
Reviewed-on: https://chromium-review.googlesource.com/1497739
Commit-Ready: Jingkui Wang <jkwang@google.com>
Tested-by: Jingkui Wang <jkwang@google.com>
Reviewed-by: Sonny Rao <sonnyrao@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
The description method is deprecated and its signature forces less
helpful error messages than what Display can provide.
BUG=none
TEST=cargo check --all-features
TEST=cargo check --target aarch64-unknown-linux-gnu
Change-Id: I27fc99d59d0ef457c5273dc53e4c563ef439c2c0
Reviewed-on: https://chromium-review.googlesource.com/1497735
Commit-Ready: David Tolnay <dtolnay@chromium.org>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
When switching to PciBarConfiguration, the set_* functions were changed
to return self. The self for register index 1 was not being used.
TEST=boot a VM and check that there isn't a pci bus creation error.
Change-Id: I8d5162c70fcec1159a6283e26e744d0c3c76b804
Signed-off-by: Dylan Reid <dgreid@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1497737
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chirantan Ekbote <chirantan@chromium.org>
Due to repeated syscall rules in gpu_device and common_device policies,
minijail fails to compile the gpu_device.policy. This change unrefactors
that policy so that it may compile properly.
BUG=chromium:936633,chromium:837073
TEST=vmc start --enable-gpu termina
Change-Id: I09ab9296247279c3a9ba6e3a6852e2a7ae2612ed
Reviewed-on: https://chromium-review.googlesource.com/1493424
Commit-Ready: Dylan Reid <dgreid@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
When running in multiprocess mode, such as on a device, TPM state gets
placed in /run/vm/tpm.{pid} (e.g. /run/vm/tpm.22726) where pid is the
pid of the original crosvm process. The TPM simulator will write a
single file called NVChip of size 16384 bytes into this directory. The
directory and NVChip file will have uid and pid set to crosvm.
When running without multiprocess mode / without minijail / probably in
cros_sdk, TPM state is placed in /tmp/tpm-simulator as before. The
/run/vm directory is not present under cros_sdk.
Will follow up with a separate CL to remove the TPM state directory at
crosvm exit.
Tested by running the following on a grunt board (Barla) in dev mode:
sudo crosvm run \
--root rootfs.ext4 \
--socket crosvm.sock \
--seccomp-policy-dir seccomp \
--software-tpm \
-p init=/bin/bash \
-p panic=-1 \
vmlinux.bin
and confirming that /dev/tpm0 and /dev/tpmrm0 are present in the VM.
BUG=chromium:921841
TEST=manual testing on grunt
Change-Id: I1868896b9eb6f510d8b97022ba950b3604d9d40b
Reviewed-on: https://chromium-review.googlesource.com/1496910
Commit-Ready: David Tolnay <dtolnay@chromium.org>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
The same logic will be needed for the TPM device in a later CL.
BUG=none
TEST=cargo check
TEST=cargo check --features gpu
Change-Id: I1497a5eab51752db80b4b457834e45d669af69b7
Reviewed-on: https://chromium-review.googlesource.com/1497731
Commit-Ready: David Tolnay <dtolnay@chromium.org>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
Tested by running the following on a grunt board (Barla) in dev mode
together with CL:1496910:
sudo crosvm run \
--root rootfs.ext4 \
--socket crosvm.sock \
--seccomp-policy-dir seccomp \
--software-tpm \
-p init=/bin/bash \
-p panic=-1 \
vmlinux.bin
and confirming that /dev/tpm0 and /dev/tpmrm0 are present in the VM.
I needed to override the common device policy's `open` and `openat` and
was not able to get that working with the existing @include.
Note: untested on arm.
BUG=chromium:921841
TEST=manual testing on grunt
Change-Id: Ied7f18a631ce8c0ae280f8b6c01511ca20c3d1c8
Reviewed-on: https://chromium-review.googlesource.com/1496909
Commit-Ready: David Tolnay <dtolnay@chromium.org>
Tested-by: David Tolnay <dtolnay@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Rename functions and parameters that had 'virtio' in their names
because we also create non-virtio devices like audio.
BUG=none
TEST=emerge-eve crosvm and deploy it to the device, verify some of
devices are still created at /sys/bus/virtio/devices/
Change-Id: I3ea75159a865e5f00ecef349725b3c12f94afaca
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-on: https://chromium-review.googlesource.com/1480739
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: David Tolnay <dtolnay@chromium.org>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Plugin needs to access to various services such as cras, cups, and so
on, and therefore we need to make sure their respective groups are
mapped in the plugin jail.
BUG=b:117989168
TEST=Start plugin via concierge_client and verify that additional
groups besides crosvm are visible in the jail.
Change-Id: Ic1151fcfca0ca75c9ae6a22828853bbe6c9fe16c
Reviewed-on: https://chromium-review.googlesource.com/1495723
Commit-Ready: Dmitry Torokhov <dtor@chromium.org>
Tested-by: Dmitry Torokhov <dtor@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Not sure if adding the device addresses to the mmio bus
is the desired behavior, but it seems to work.
BUG=chromium:924405
TEST=boot VM
Change-Id: I7f6057b3e7d041a52b251af1203353ba7a0d3c22
Reviewed-on: https://chromium-review.googlesource.com/1480743
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
The idea is that virtio devices can specify additional memory
regions.
BUG=chromium:924405
TEST=run VM
Change-Id: I2a9f233ca8e2bc4fd9b05ee83101b11deb6e7b04
Reviewed-on: https://chromium-review.googlesource.com/1480742
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
This removes add_memory_region and add_io_region, and replaces
it with the add_pci_bar function.
BUG=chromium:924405
TEST=boot VM
Change-Id: Ifc637d174d3f8b1255cf13725a1a224b4cdf0a30
Reviewed-on: https://chromium-review.googlesource.com/1480741
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
We want to support 64-bit BARs and some additional functionality
is required.
BUG=chromium:924405
TEST=compile
Change-Id: I06aba41b6dfb9649437a417a32cb450d19d0d937
Reviewed-on: https://chromium-review.googlesource.com/1480740
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
Use a function to reduce the repetition.
BUG=None
TEST=some basic build and runtime tests on an eve device:
() run 'cargo check' and 'emerge-eve crosvm'
() launch default termina VM from GUI, install and open
eclipse app to test network, input and graphic features
() launch a test VM instance from shell and vsh into it,
verify virtio devices are created at
/sys/bus/virtio/devices/
Change-Id: If66defdd584cf9afa6f38531e41b85d327362dfa
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-on: https://chromium-review.googlesource.com/1480738
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: David Tolnay <dtolnay@chromium.org>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
Now that the connection oriented seqpacket sockets are used for vm
control messages, a response can be received by the requested. This
change prints out that response.
TEST=crosvm suspend|resume|balloon|stop <socket>
BUG=chromium:848187
Change-Id: I18ac23c26127332e2be498113cc0c3310fd09a7d
Reviewed-on: https://chromium-review.googlesource.com/1482370
Commit-Ready: Zach Reizner <zachr@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
The advantage of seqpacket is that they are connection oriented. A
listener can be created that accepts new connections, useful for the
path based VM control sockets. Previously, the only bidirectional
sockets in crosvm were either stream based or made using socketpair.
This change also whitelists sendmsg and recvmsg for the common device
policy.
TEST=cargo test
BUG=chromium:848187
Change-Id: I83fd46f54bce105a7730632cd013b5e7047db22b
Reviewed-on: https://chromium-review.googlesource.com/1470917
Commit-Ready: Zach Reizner <zachr@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
This change adds the ability to make seqpacket pairs, and set the
timeouts of the sockets. This also adds a TcpListener style api for
accepting UnixSeqpacket sockets.
TEST=cargo test -p sys_util
BUG=chromium:848187
Change-Id: I9f9bb5224cdfaf257d8e4a1bdaac8128be874951
Reviewed-on: https://chromium-review.googlesource.com/1482371
Commit-Ready: Zach Reizner <zachr@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Tested-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
These panic!()s might be user-triggerable, and in any event are not fatal errors.
BUG=chromium:908689
TEST=Unit tests in file.
Change-Id: I774bb633dc627247bd807727542589400b59ed07
Reviewed-on: https://chromium-review.googlesource.com/1487674
Commit-Ready: Miriam Zimmerman <mutexlox@chromium.org>
Tested-by: Miriam Zimmerman <mutexlox@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
This allows the default pivot root used by crosvm to be defined at
build time. If it is not set, /var/empty is used.
BUG=chromium:934513,chromium:933582
TEST=PreCQ passes
Change-Id: I2b461170e6b75324740d4ff709fd599085cefe18
Reviewed-on: https://chromium-review.googlesource.com/1483330
Commit-Ready: Allen Webb <allenwebb@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Daniel Verkamp <dverkamp@chromium.org>
