It is necessary to notify the VHM and hypervisor on the completion of a VHM
request even when the UOS is in suspend or system reset mode because the VHM and
hypervisor rely on the notification to reset their own states on the
request.
Currently the VHM request state is checked against REQ_STATE_PROCESSING instead
of REQ_STATE_COMPLETE when handling system reset or suspend/resume, leading to a
completed request unnotified, and causing the HV to complain on an occupied VHM
request when it raises a new one.
This patch fixes this issue by properly notifying completed requests to the VHM
& hypervisor. Some concerns are raised during a discussion on the potential
races which does not hurt for now but may in the future. These considerations
and potential solutions are documented as comments for future reference.
Tracked-On: #895
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
This is the counterpart in DM to the VHM request state update in the
hypervisor. Major changes include:
* Remove accesses to the obsolete 'valid' member.
* Access the 'processed' member using atomic operations.
* Sync the documentation on vhm_request.
In addition, the new state transition also requires a VHM request to be always
handled properly, as there is no 'FAILED' state any more. Instead of crashing
the device model (and thus the UOS as well), the device model should return all
1s or ignore the request when it is to load from or store to an invalid address,
respectively.
Note: there is an issue in vm_system_reset() and vm_suspend_resume() where
completed VHM requests are not properly notified, causing the hypervisor to
complain as it sees uncompleted requests while trying to create a new one. This
issue will be resolved in a separate patch.
v1 -> v2:
* Use macro-defined constants for the default values for invalid PIO/MMIO
reads.
* Change the return type of vmexit_handler_t in DM to void as the return
values are no longer necessary.
* Remove VM_EXITCODE that are no longer used.
Tracked-On: #875
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
According to ACPI spec, when system back to working mode from
S3, WAK_STS bit (15) should be set.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Reviewed-by: Victor Sun <victor.sun@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
We do:
- pause target vm
- suspend all virtual devices
- wait for resume notification
- resume all virtual devices
- reset target vm
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
This function add high level reset_vdev function. Which is
implemented to call deinit/init pairing to emulate the virtual
device reset operation.
This patch also add the system reset which keep the UOS RAM
content functionality to DM.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Guest has erquirement to support system/full reboot and S3. Which could
trigger different reset path in guest
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
* introduce EXTRA_VERSION to replace RC_VERSION
* add daily tag into the version information
* unify the hypervisor and device model version
Signed-off-by: Jack Ren <jack.ren@intel.com>
- UOS will boot fail if 'strictio' is enabled ('-e' option), in this
case (with '-e'), device model will block all PIO accesses whose
handlers were not registered, after that, device model program will
exit, hence UOS boot fail.
actually, such kind of accesses exist, e.g. UOS would program
PIT registers (port address: 0x43) if hpet is disabled.
- For debug, we can trap unexpected PIO access in 'default_inout()'
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Current option of removing vGSI capability is global, which exposes
vIOAPIC link for all ptdev even only one need this. This patch makes
it as ptdev local option to lower the system level impact. To keep
vGSI for MSI capable ptdev, just explicitly append ",keep_gsi" in
option list, like "-s 14,passthru,0/e/0,keep_gsi"
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Update the vcpu id type as uint16_t for vm_create_vcpu, this
keeps alignment with the updates for the structure acrn_create_vcpu
used by hcall_create_vcpu in the hypervisor.
In the device model, the caller is responsible for vcpu id type
conversion; vcpu id type is uint16_t for external interface in the
current implement.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
when add/create vcpu failed, before it just exits,
and the resource will not released; now if failed,
let DM release the resource.
Signed-off-by: Minggui Cao <minggui.cao@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
pincpu and vmexit_pause both aligned to lower p, it is a bug, vmexit_pause should align with P(upper).
Missing enable_bvmcons usage so add -b in help prints.
remove redundant tab blank to keep same indentation.
Signed-off-by: Chris Ye <chris.ye@intel.com>
With '--ptdev_no_reset', DM doen not abort but warn when assign PCIe
dev without reset capability.
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
Current only 8 vioapic pins for pci irq (total 24 with 16 reserved),
which easily leads virtual GSI sharing with more and more passthrough
devices. This patch doulbes vioapic pin count and adds reboot hooks to
allocate from same pin after each reboot.
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Anthony Xu <anthony.xu@intel.com>
With curren code, DM will ignore the SIGHUP signal generated from
SOS reboot that causes DM will not release resource when SOS reboot
occurs.
Signed-off-by: Zheng, Gen <gen.zheng@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Move all virtual devices init/deinit to function to simplify the
failure path of main loop. In the future, new virtual device will
not touch main loop.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Acked-by: Eddie Dong <Eddie.dong@intel.com>
- make ioc_init/ioc_deinit take struct vmctx as argument
- ioc_init return int instead of pointer to struct ioc_dev
- add ioc_dev in vmctx to track ioc_dev
- remove the atkbdc.h included in vmmapi.h
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Reviewed-by: Liu, Yuan1 <yuan1.liu@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Reviewed-by: Wang Yu <yu1.wang@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Acked-by: Eddie Dong <Eddie.dong@intel.com>
- Move the variable local_time from main.c to rtc.c
- Change vrtc_init to return int instead of pointer to vrtc. We do
track vrtc in struct vmctx.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Following functions have never been used anywhere.
Let's remove them.
fbsdrun_muxed()
fbsdrun_vmexit_on_hlt()
fbsdrun_vmexit_on_pause()
fbsdrun_disable_x2apic()
Remove weird prefix "fbsdrun" from following functions' name.
fbsdrun_virtio_msix() --> virtio_uses_msix()
fbsdrun_start_thread() --> start_thread()
Signed-off-by: Jie Deng <jie.deng@intel.com>
Reviewed-by: Hao Li <hao.l.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Remove weird prefix "fbsdrun" from the function name.
Since "fbsdrun_addcpu" has never been called by external
functions, Let's remove its declaration from "dm.h" and
make it as a static function to keep consistency with
"fbsdrun_deletecpu".
Signed-off-by: Jie Deng <jie.deng@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Type '-T' in DM cmdline to enable hugetlb memory allocation.
It enabled 2 level of huge page:
Level 1: size 2M, path: /run/hugepage/acrn/huge_lv1/guid
Level 2: size 1G, path: /run/hugepage/acrn/huge_lv2/guid
NOTE:
before running, please make sure system already have enough hugepages
reserved under:
/sys/kernel/mm/hugepages/hugepages-xxxxkB/nr_hugepages
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
Move the bvmcons enable flag from main.c to consport.c.
So we don't need the flag in main.c for bvmcons.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Add "-i" in the DM boot command line for booting IOC mediator.
NOTE: currently only ioc_init will be called in the main entry, ioc_deinit
hasn't be called so far. The DM plans to add one virtual devices list inside
of struct vmctx in the near furture, and virtual devices data structure will
add one common deinit callbacks which will be called during VM exits.
Will support ioc_deinit once that patch merged.
Signed-off-by: Liu Yuan <yuan1.liu@intel.com>
Reviewed-by: Wang Yu <yu1.wang@intel.com>
Reviewed-by: Liu Shuo <shuo.a.liu@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Current, mevent cleanup path has issue. There are possible
following calling sequence happen when reboot/poweroff UOS:
1. mevent_dispatch() calls mevent_destroy
2. do_close_post calls some virtual device deinit to delete
mevent.
This patch introduce mevent init/deinit to make sure:
1. mevent init
2. mevent_add is called when init virtual device
3. mevent_del is called when deinit virtual device
4. mevent deinit
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Devices' de-init process might access some mapped memory space, such as
virtio virtqueues. Access after unmap will cause a fault.
Release the memory map after de-init processes can avoid it. Reading
more code, there are many error handling lost to unmap the memory.
Refined the code to do it.
Signed-off-by: Liu Shuo <shuo.a.liu@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
A monitor component will be added to acrn-dm, which crteats socket,
bind and listening at /run/acrn/vmname. Acrnctl & acrnd could conn
-ect to the socket for communication, using defined message, in
include/monitor_msg.h
For each defined message, a message handler callback could be
registered via monitor_add_msg_handler(). On received of a defined
message, a certain call back will be called. Each callback can only
see the message sender's socket-fd.
When acrn-dm want report something, not triggered by incoming message
it can send broadcast message, use monitor_broadcast().
Acked-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Yin, Fengwei <fengwei.yin@intel.com>
Signed-off-by: Tao, Yuhong <yuhong.tao@intel.com>
The patch includes:
1. vrtc_cleanup -> vrtc_deinit to align with other devices
2. delete timer created in vrtc_init
3. make call to vrtc_deinit in cleanup path
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <Eddie.dong@intel.com>
old code has no deinit functionality for virtual keyboard device.
Which will trigger resource leak when system is reboot.
deinit function is added to:
1. deinit low ps2 based keyboard and mouse
2. release memory/io resource of virtual keyboard device
NOTE: IRQ resource will be handed in pci irq module
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <Eddie.dong@intel.com>
The trusty enabled or not will be passed to HV
with create vm hypercall. It's passed to vSBL within
config page also.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Some guests needs guest partition info passed to support
A/B boot. DM needs to load guest partition info from file
and pass to vsbl.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
There are two things here:
1. Enable basic vsbl boot functionality.
2. add long option to DM to specific vsbl file. So DM could
choice to boot guest with vsbl.
It's hard to find a suitable short option. So only long option
is supported for vsbl file name option.
If long option "--vsbl='vsbl_file_name'" is given in DM
commandline, the "vsbl_file_name" will be loaded by DM
first and running. vsbl will be response to lead guest to
next boot stage.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
The common loader code is put to sw_load_common.c
The loader code for directly kernel loading is put to sw_load_bzimage.c
The further change for SBL will be put to sw_load_sbl.c
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
