Device model uses polling mode to fetch the virtio blk request in RTVM.
When RTVM brings up with io uring, the threads handling io uring and vq are
not same, which would cause competition. To fix this issue, device
model should handle vq and io uring in the same thread to avoid conflict.
Tracked-On: #8737
Signed-off-by: YuanXin-Intel <xin.yuan@intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
This patch is to change the vUART resource occupied by S5 function
between Service VM and guest VM to avoid the standard UART port
conflict when legacy UART passthrough to guest VM.
Tracked-On: #8622
Signed-off-by: YuanXin-Intel <xin.yuan@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
1. Enable Service VM to power off or restart the whole platform even when RTVM is running.
2. Allow Service VM stop the RTVM using acrnctl tool with option "stop -f".
3. Add 'Service VM supervisor role enabled' option in ACRN configurator
Tracked-On: #8618
Signed-off-by: YuanXin-Intel <xin.yuan@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Passthrough of lpss devices, such as sdio, spi, uart, is not supported for user
vm due to irq and acpi info missing.
Here provides new pci device passthrough options to pass irq and acpi dsdt info
by users. Considering spi dsdt info varies from HW, to add the flexibility of
configuration, it is designed to pass dsdt file of spi device by users rather
than hard code. Besides, remove the limit of the lpss device passthrough for rtvm.
Tracked-On: #8615
Signed-off-by: nacui <na.cui@intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
When multiple virtio-blk instances are created for one VM,
using the same `static struct virtio_ops virtio_blk_ops` for all instances
is buggy. It only works when all instances are created with the same number
of the virtqueues.
This patch fixes this issue by introducing a member in `struct virtio_blk`
to store the ops info for each virtio-blk instance.
Tracked-On: #8612
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
This patch updates the `iothread` option to specify the CPU affinity
of the iothread. Setting the iothread's CPU affinity could benefit the
Service VM's CPU utilization when Service VM owns limited dedicated CPUs.
It could be helpful to ensure the I/O mediator Quality of Service (QoS).
Once the performance tuning is done, the specific CPU affinity config could
pass to acrn-dm directly, letting the deployment more easily.
The format looks like below:
iothread=<num_iothread>@<cpu_affinity>
"@" is used to separate the following two settings:
- the number of iothread instances
- the CPU affinity settings for each iothread instance.
The format of `cpu_affinity` looks like below:
<cpu_affinity_0>/<cpu_affinity_1>/<cpu_affinity_2>/...
1. "/" is used to separate the CPU affinity setting for each iothread instance
(sequentially).
2. char '*' can be used to skip the setting for the specific iothread instance.
3. the number of cpu_affinity_x vs. the number of iothread instances
- If # of cpu_affinity_x is less than # of iothread instances,
no CPU affinity settings for the last few iothread instances.
- If # of cpu_affinity_x is more than # of iothread instances,
the extra cpu_affinity_x are discarded.
4. ":" is used to separate different CPU cores for each CPU affinity setting.
Examples to specify the CPU affinity of the iothread:
1. iothread=3@0:1:2/0:1
`add_virtual_device 9 virtio-blk iothread=3@0:1:2/0:1,mq=3,/dev/nvme1n1`
a) 3 iothread instances are created.
b) CPU affinity of iothread instances for this virtio-blk device:
- 1st iothread instance <-> pins to Service VM CPU 0,1,2
- 2nd iothread instance <-> pins to Service VM CPU 0,1
- 3rd iothread instance <-> No CPU affinity settings
2. iothread=3@0/*/1
`add_virtual_device 9 virtio-blk iothread=3@0/*/1,mq=3,/dev/nvme1n1`
a) 3 iothread instances are created.
b) CPU affinity of iothread instances for this virtio-blk device:
- 1st iothread instance <-> pins to Service VM CPU 0
- 2nd iothread instance <-> No CPU affinity settings
- 3rd iothread instance <-> pins to Service VM CPU 1
v1 -> v2:
* encapsulate one API in iothread.c to parse the iothread options, so that
other BE can also use it.
v2 -> v3:
* introduce one API iothread_free_options to free the elements that
are allocated dynamically in iothread_parse_options().
Tracked-On: #8612
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
This patch renames the iothread for better readability. For instance,
the new name of the iothread for virtio-blk device looks like `iothr-0-blk9:0`.
It could be helpful when tuning the performance and the CPU utilization.
v1 -> v2:
* add `const` qualifier for the input parameter of `iothread_create`
Tracked-On: #8612
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
Prior to this patch, one single iothread instance is created and initialized
in the `main` function. This single iothread monitors all the registered fds
and handles all the corresponding requests. It leads to the limited flexibility
of the iothread support.
To improve the flexibility of the iothread support, this patch does:
- add the support of multiple iothread instances.
`iothread_create` is introduced to create a certain number of iothread
instances. It shall be called at first by each virtual device owner (such as
virtio-blk BE) on initialization phase. Then, `iothread_add` can be called
to add the to be monitored fd to the specified iothread.
- update virtio-blk BE to let the acrn-dm option `iothread` accept a number
as the number of iothread instances to be created.
If `iothread` is contained in the parameters, but the number is not specified,
one iothread instance would be created by default.
Examples to specify the number of iothread instances:
1. Create 2 iothread instances
`add_virtual_device 9 virtio-blk iothread=2,mq=2,/dev/nvme1n1,writeback,aio=io_uring`
2. Create 1 iothread instances (by default)
`add_virtual_device 9 virtio-blk iothread,mq=2,/dev/nvme1n1,writeback,aio=io_uring`
- update virtio-blk BE to separate the request handling of different virtqueues
to different iothreads.
The request from one or more virtqueues can be handled in one iothread.
The mapping between virtqueues and iothreads is based on round robin.
v1 -> v2:
* add a mutex to protect the free ioctx slot allocation
Tracked-On: #8612
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
io_uring is a high-performance asynchronous I/O framework, primarily designed
to improve the efficiency of input and output (I/O) operations in user-space
applications.
This patch enables io_uring in block_if module. It utilizes the interfaces
provided by the user-space library `liburing` to interact with io_uring
in kernel-space.
To build the acrn-dm with io_uring support, `liburing-dev` package needs to be
installed. For example, it can be installed like below in Ubuntu 22.04.
sudo apt install liburing-dev
In order to support both the thread pool mechanism and the io_uring mechanism,
an acrn-dm option `aio` is introduced. By default, thread pool mechanism is
selected.
- Example to use io_uring:
`add_virtual_device 9 virtio-blk iothread,mq=2,/dev/nvme1n1,writeback,aio=io_uring`
- Example to use thread pool:
`add_virtual_device 9 virtio-blk iothread,mq=2,/dev/nvme1n1,writeback,aio=threads`
- Example to use thread pool (by default):
`add_virtual_device 9 virtio-blk iothread,mq=2,/dev/nvme1n1,writeback`
v2 -> v3:
* Update iothread_handler
- Use the unified eventfd interfaces to read the counter value of the
ioeventfd.
- Remove the while loop to read the ioeventfd. It is not necessary
because one read would reset the counter value to 0.
* Update iou_submit_sqe to return an error code
The caller of iou_submit_sqe shall check the return value.
If there is NO available submission queue entry in the submission queue,
need to break the while loop. Request can only be submitted when SQE is
available.
v1 -> v2:
* move the logic of reading out ioeventfd from iothread.c to virtio.c, because
it is specific to the virtqueue handling.
Tracked-On: #8612
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
block_if is the backend of ahci and virtio-blk. Only one queue is
supported by block_if now. Several worker threads are created as
the thread pool for the queue. One BIG mutex is used for the queue
and thread operation. With this patch block_if can support multiple
queues and each queue is backed by several worker threads. blockif_req
can be submited/enqueued into one specified queue. By spliting into
several queues contention from the BIG mutex can be relieved/eliminated.
This is used to support virtio-blk multiple queues feature.
Tracked-On: #8612
Signed-off-by: Jian Jun Chen <jian.jun.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
Virtio-blk can support multiple virtqueues (mq) which is negotiated
between FE and BE by the feature bit VIRTIO_BLK_F_MQ. The virtqueue
number of virtio-blk can be specified by "mq=x" in the parameter.
For example: "virtio-blk,iothread,mq=2,..."
Tracked-On: #8612
Signed-off-by: Jian Jun Chen <jian.jun.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
ACRN virtio devices are using a per device mutex to protect the
concurrent operations on the device's PIO/MMIO. This introduces
big contention in fast IO hence downgrades the IO performance,
for example virtio-blk with asyncio enabled. This patch introduces
per queue mutex to relieve such issues. Currently the per queue
mutex is only used in the asycio path when iothread is enabled.
Tracked-On: #8612
Signed-off-by: Jian Jun Chen <jian.jun.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
ACRN_IOEVENTFD_FLAG_ASYNCIO is not set when unregister ioeventfd
in the current implementation which will cause the old asyncio_desc
will be remained in hypervisor link list when switching from OVMF to
kernel.
Tracked-On: #8612
Signed-off-by: Jian Jun Chen <jian.jun.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
The meaning of lpc_uart_vdev::enabled was changed from runtime enablement
status of UARTs to user configuration in commit 852f10cc3. So it's
incorrect to clear it in lpc_deinit() which will cause UARTs disabled
after reboot.
Tracked-On: #8537
Fixes: 852f10cc3 ("dm: lpc: only emulate COM ports specified in command line")
Signed-off-by: Qiang Zhang <qiang4.zhang@intel.com>
1) region ID shall be configured by user via config tool.
2) region ID is programmed to "Subsystem ID" of PCI config space.
2) "Subsystem Vendor ID" is harded coded as 0x8086
Parameters to configure dm-land IVSHMEM device example generated
by config tool as below:
`add_virtual_device 8 ivshmem hv:/shm_region_0,256,2`
Tracked-On: #8566
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Currently lpc emulates all the supported COM ports no matter it is
configured or not in command line. Change the behavior to only emulate
those specified in command line.
Tracked-On: #8537
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Extend the devicemodel lpc uart emulation support to COM4. Since
COM1 is usually used for hv console and COM2 is taken by S5 feature,
only COM1 and COM2 is not enough.
Tracked-On: #8537
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Currently only devices on usb bus 0-4, port 0-19 can be passthrough to
the emulated XHCI controller. Remove this unnecessary limit.
Some unused definitions are also removed.
Tracked-On: #8506
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Host doorbell array write can be asynchronous, so add an async thread
which is used to deal doorbell write.
Tracked-On: #8504
Signed-off-by: Yuanyuan Zhao <yuanyuan.zhao@intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
doxygen will warn that documented return type is found for functions
that does not return anything in 1.9.4 or later versions. 'None' is
not a special keyword in doxyge, it will recognize it as description
to the return value that does not exist in void functions.
Tracked-On: #8425
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Currently the DSM (Data of Stolen Memory) size was hardcoded to 64M in
ACRN, meaning that users must set "DVMT Pre-Allocated" to 64M in order
to make IGD passthrough (GVT-d) to work. This patch reads the BIOS-
configured memory size from GGC (GMCH Graphics Control, 0x50) register
and passthrough corresponding area to guest.
Tracked-On: #8432
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
The register index and data format of BDSM (Base Data of Stolen Memory)
of Intel integrated graphics is changed in GPU Gen 11. Currently ACRN
uses a long device list for Gen11+ devices. This patch introduces a new
device allowlist for IGD passthrough in igd_pciids.h, covering IGD
device ids from Skylake, and handles passthrough by its generation.
If a device is not listed, it will be treated as a gen 11 device and a
warning will be printed, developers should add the new device to the
list if it is verified to work.
Tracked-On: #8432
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Since Intel Processor Graphics Gen11, Base Data of Stolen Memory (BDSM)
register is extended to 64 bit, allowing BIOS to assign an above 4G
address. This patch changes the type of hpa/gpa variables to uint64_t
to properly handle such case.
Tracked-On: #8432
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
In virtio_gpu_cmd_create_blob() and virtio_gpu_cmd_resource_attach_
backing(), entries may be accessed before initialization. Fix it by
using calloc() to allocate it instead of malloc().
Tracked-On: #8439
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
This patch checks the romfile path length in command line to avoid
possible buffer overflow, maximum path supported is 255 characters.
Tracked-On: #8439
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
This patch fix several issues that NULL pointers possibly be
dereferenced in display module.
Tracked-On: #8439
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Pointer 'vsock->vhost_vsock' returned from call to function
'vhost_vsock_init' may be NULL and will be dereferenced when
calling 'vhost_vsock_set_guest_cid()'.
Tracked-On: #8439
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
gvt_init_config() may perform out-of-range read on host_config, add
bound check before accessing it.
Tracked-On: #8382
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Reviewed-by: Jian Jun Chen <jian.jun.chen@intel.com>
Add the new parameter for register ioevent function, let the vhost
vq and viothread vq can share the register ioevent common API.
Tracked-On: #8323
Signed-off-by: Liu Long <long.liu@linux.intel.com>
Reviewed-by: Conghui <conghui.chen@intel.com>
To avoid assertions in devicemodel, remove all the assert() in vga.c,
use print error message to check errors.
All the behavior of registers follow this spec:
https://wiki.osdev.org/VGA_Hardware#Graphics_Mode
Tracked-On: #8125
Signed-off-by: Sun Peng <peng.p.sun@linux.intel.com>
Reviewed-by: Huang, Yonghua <yonghua.huang@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
Add a new flag in ioeventfd ioctl to support asyncio. After that, the IO
request will be processed in asyncio path by kernel and hypervisor.
Tracked-On: #8209
Signed-off-by: Conghui <conghui.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
Add the option of "romfile=file_location" to specify the rom file for rombar
Tracked-On: #8175
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Wang Yu <yu1.wang@intel.com>
The pci_reg 0x30 of PCI config_space is used to check whether the PCI rom
bar is supported. When the PCI rom is supported for the device in guest vm,
the 0x30 pci_reg is emulated and it can return the addr/enable bit.
Tracked-On: #8175
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
PCI ROM is the firmware specific to PCI device and it is provided by
the device vendor. The PCI rom resides in 0x30 offset of PCI config space.
This can be used to check whether the PCI rom exists. And when it exists,
it can load the firmware from the addr that is obtained from ROM bar addr.
For the user-vm, it will try to load the rom_file for the given PCI device and
enable the VM to access the firmware that is defined in rom_file.
BTW: The emulated rom_file is converted from efi image by using EfiRom. It has
no dependency on the ROM bar of physical PCI devices. Of course if the physical
PCI devices supports the ROM bar, the rom_file can also be dumped from the PCI
rom.
Now this is limited to PCI display device.
V2->V3: Add the function of pci_load_rombar/pci_release_rombar to handle the
rombar in course of passthrough_init/deinit.
Tracked-On: #8175
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Wang Yu <yu1.wang@intel.com>
Now the device model only supports the 0..5 PCI bar for PCI/PCIE devices.
This tries to allocate the PCI_MEM32 resource for PCI ROM bar.
V1->V2: Use the PCI_ROMBAR as bar index and PCIBAR_ROM bar type when calling
the pci_emul_alloc_bar to allocate the guest physical addr for PCI ROM bar.
And it will allocate the resource from PCIBAR_MEM32 region.
V2->V3: Add more comments that describes the parameter of pci_emul_alloc_bar.
Tracked-On: #8175
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Wang Yu <yu1.wang@intel.com>
Fix the bug in iothread handler, the event should be read out so that the
next epoll_wait not return directly as the fd can still readable.
Tracked-On: #8181
Signed-off-by: Conghui <conghui.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
Because of the iteration of i as is, the condition i <= XHCI_MAX_DEVS
always results as true and thus is unnecessary.
When compiling with cflag -Werror this condition will result in a
compilation failure.
Tracked-On: #8114
Signed-off-by: Marius Rodi <marius.rodi@linutronix.de>
Change virtio-i2c from legacy to modern device according to virtio spec v1.2.
Tracked-On: #8055
Signed-off-by: Liu Long <long.liu@linux.intel.com>
Signed-off-by: Conghui <conghui.chen@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
Return the parsed scanout_num to guest_vm so that the guest_vm can configure multi-display.
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
Sometimes the virtio_gpu in guest_vm doesn't pass the zero coord_x/y for the
displayed framebuffer. In such case the offset should be calibrated for pixman_image/dmabuf
so that it can display the image from the correct image.
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
The virtio-gpu display will be handled by the below cmd sequence:
1. Virtio_GPU_CMD_CREATE_RESOURCE
2. VIRTIO_GPU_CMD_SET_SCANOUT/SET_SCANOUT_BLOB
3. VIRTIO_GPU_CMD_FLUSH
And the VIRTIO_GPU_CMD_FLUSH will notify the vdisplay module to display the
framebuffer related with the scanout_id. But the virtio_gpu_cmd_flush doesn't
pass the scanout info.Instead it only passes the flushed_region and the
resource_id. So it needs to check whether the scanout_region is covered by the
flushed_region and then decide whether the scanout_win needs to be displayed.
v1->v2: Use the bpp instead of hardcode 4 for offset_calculation
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
The virtio-gpu in guest_vm also passes the scanout_id for cursor operation.
So this scanout_id should be used in course of handling cursor operation.
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
The scanout_id is included in the below virtio-gpu cmd request.
VIRTIO_GPU_GET_EDID
VIRTIO_GPU_SET_SCANOUT
VIRTIO_GPU_SET_SCANOUT_BLOB
VIRTIO_GPU_GET_DISPLAY_INFO
So the scanout_id and scanout_rect should be handled correctly. And it will
use the scanout_rect instead of resource width/height in course of handling
VIRTIO_GPU_SET_SCANOUT/SCANOUT_BLOB cmd.
v1->v2: change the position to zero when returning response for GET_DISPLAY_INFO
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
Now it only supports one scanout for virtio-gpu. So the scanout_id is ignored in
course of handling virtio-gpu cmd. In order to handle the virtio-gpu cmd correctly,
it adds the virtio_gpu_scanout structure so that it can record the scanout info.
v1->v2: Refine the field in virtio_gpu_scanout and error message for scanout_id
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
Vdisplay module will create the gui_window (vScreen) based on the virtio-gpu
parameter. When virtio-gpu device tries to establish the connection with
vdisplay module, it will return the supported number of vScreen.
Then the virtio-gpu device can initialize the correct info for the guest_vm.
BTW: The number is fixed to 1.
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
Currenly it can support only one display_win. When virtio-gpu wants to display
the framebuffer from Guest_vm, it will render it into the default win. In order
to support multi-window, the scanout_id is added for display API.
After display API is refined, the vdisplay and virtio-gpu can support the multi-win
independently.
BTW: Now the scanout_id is not implemented in vdisplay module and the virtio-gpu always
passes the zero. It will be implemeneted in the later patches.
Tracked-On: #7988
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Peng Sun <peng.p.sun@linux.intel.com>
There was some secure coding style violations of virtio net and tmp,
this patch add some NULL check to fix these violations.
Tracked-On: #6690
Signed-off-by: Chenli Wei <chenli.wei@linux.intel.com>
