The vuart_read()/vuart_write() are coupled with PIO vuart type. Move
the non-type related code into vuart_read_reg()/vuart_write_reg(), so
that we can re-use them to handle MMIO request of pci-vuart type.
Tracked-On: #5394
Signed-off-by: Tao Yuhong <yuhong.tao@intel.com>
Reviewed-by: Wang, Yu1 <yu1.wang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Multiple devices could be nested under the same range. Skip remove if
the device is removed already
Tracked-On: #5437
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Changes:
1. assign 3 CPUs for WaaG on hybrid_rt scenario;
2. Passthrough NVME@9:0.0 for VM0 on hybrid_rt scenario;
3. Change rootfs from partition2 to partition3;
Tracked-On: #5390
Signed-off-by: Liang Yi <yi.liang@intel.com>
Acked-by: Victor Sun <victor.sun@intel.com>
Support enable ivshmem for SOS. Insert the ivshmem device information if
it is enabled.
1. get ivshmem vbar based:
- vbar[0] is size 0x100
- vbar[2] is specified MB size
2. get vbdf for ivshmem device
Tracked-On: #5426
Signed-off-by: Yang,Yu-chu <yu-chu,yang>
Add following to default output scenarios vm_configurations.c:
- pci_dev_num
- pci_devs = sos_pci_devs
Both was defineded in CONFIG_SOS_VM.
Tracked-On: #5426
Signed-off-by: Yang, Yu-chu <yu-chu.yang@intel.com>
- Refactor pci_dev_c.py to insert devices information per VMs
- Add function to get unused vbdf form bus:dev.func 00:00.0 to 00:1F.7
Add pci devices variables to vm_configurations.c
- To pass the pci vuart information form tool, add pci_dev_num and
pci_devs initialization by tool
- Change CONFIG_SOS_VM in hypervisor/include/arch/x86/vm_config.h to
compromise vm_configurations.c
Tracked-On: #5426
Signed-off-by: Yang, Yu-chu <yu-chu.yang@intel.com>
Allocate unused vbar for SOS and pre-launched VMs.
- For SOS, find unused vbar in the range which is assigned to pci host
bridge. The allocated vbar cannot have confilicts with any existing pci devices
- For pre-launched VMs, find unused vbar in the range 0x80000000 to
0xfffffff. The alloacted vbar cannot have confilicts with any
passthrough devices and mmio.
Tracked-On: #5426
Signed-off-by: Yang, Yu-chu <yu-chu.yang@intel.com>
automatically get PTCT table from native environment for the usage
of pre-launched VMs.
Tracked-On: #5418
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
fill physical PTCT table into pre-launched vACPI table when PSRAM
is configured to passthrough to pre-launched RTVM.
Tracked-On: #5418
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
add PSRAM configs in xmls, only enable PSRAM and passthrough to
pre-launched RTVM for hybrid_rt scenario on tgl-rvp board.
Tracked-On: #5418
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
Config tool UI will do mutually exclusive check the legacy vuart 0/1 and
PCI vuart 0/1 to make sure there is no legacy vuart and PCI vuart are
used at the same time for VMs.
Tracked-On: #5394
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
Add PCI VUART config for post-launched VMs in launch config UI. Users
can configure the console_vuart, configure or dynamically add or remove
communication_vuart based on the communication vuarts which are configured
from the scenario xml.
Tracked-On: #5394
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
Add PCI VUART dynamic config for VMs in scenario config UI, keep legacy
VUART config. PCI vuart base can be set to INVALID_PCI_BASE and PCI_VUART;
users will configure the target_vm_id and target_vuart_id when PCI vuart
base is set to PCI_VUART; users can dynamically add or delete PCI vuart
for VMs.
Tracked-On: #5394
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
- Explicitly reserve memory regions based on the input E820 map
- Revert "ovmf: reserve e820 table for PTCM"
Tracked-On: #5442
Signed-off-by: Peter Fang <peter.fang@intel.com>
We don't reserve PCI MMIO in e820 Table, it's included in DSDT ACPI Table.
About 0xA0000 - 0x100000 entry, we don't have any ACPI Table touch this region.
So we could remove it too.
After this change, we could only pass the reserved e820 table which we must
reserve to OVMF. In this case, the OVMF could trust ACRN-DM and pass the
reserved e820 table to guest instead of dropping it.
This patch needs the corresponding modify in OVMF. Otherwise, the guest could
not boot.
Tracked-On: #4550
Signed-off-by: Li Fei1 <fei1.li@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
The new (1.8.17) release of doxygen is complaining about errors in the
doxygen comments that were's reported by our current 1.8.13 release.
Let's fix these now. In a separate PR we'll also update some
configuration settings that will be obsolete, in preparation for moving
to this newer version.
[External_System_ID]ACRN-6774
Tracked-On: #5385
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
In pre-launched VM the GPA of vmsix BAR which is used for vmsix
over msi is calculated/allocated by acrn-config tool. The GPA
needs to be assigned to vdev when vdev is initialized. The
assignment is only needed for pre-launched VM. For SOS kernel
will reprogram the Bar base when startup. For post-launched VM
the Bar GPA will be assigned by device model via hypercall.
Tracked-On: #5316
Signed-off-by: Jian Jun Chen <jian.jun.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
When init_vmsix_on_msi is called during the initialization of a pt
device, the vmsix bar used for vmsix over msi is just created. No
mapping/unmapping is done and pci_vdev_write_vbar should be called
instead of vdev_pt_write_vbar at the time. Currently the Bar mapping
is delayed till OS sizing the Bar. Backup vbar base_gpa to mmio_gpa
is not required here becuase it will be done later when Bar mapping.
Tracked-On: #5316
Signed-off-by: Jian Jun Chen <jian.jun.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
changes:
1. Change SOS VM rootfs to nvme0;
2. Change hybrid_rt scenario VM0 mem size to 1GB;
Tracked-On: #5238
Signed-off-by: Victor Sun <victor.sun@intel.com>
The messages from the SOS socket can be safely read into the entire read
buffer.
Tracked-On: #5429
Signed-off-by: Peter Fang <peter.fang@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
pm_vuart stops listening and relays the message right away if it
encounters EAGAIN during read(). This causes the messages relayed to be
fragmented.
Only relay the message when it encounters a null character or a newline
character, or when the buffer is full.
Tracked-On: #5429
Signed-off-by: Peter Fang <peter.fang@intel.com>
Acked-by: Wang, Yu1 <yu1.wang@intel.com>
- Since de-privilege boot is removed, we no longer need to save boot
context in boot time.
- cpu_primary_start_64 is not an entry for ACRN hypervisor any more,
and can be removed.
Tracked-On: #5197
Signed-off-by: Zide Chen <zide.chen@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
update tgl-rvp.xml for tgl boards with IFWI of tsn version to enable
the embeded tsn device.
Tracked-On: #5427
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
Add functionality to get free vbar base for the vmsix devices.
- The devices size is 4k.
- The mmio range for non SOS VM is 2G to 4G
- The mmio range for SOS is depended on the range which is assigned to
PCI bus hostbridge
- The next vbar index is based on last device vbar index vbar_i
Tracked-On: #5422
Signed-off-by: Yang, Yu-chu <yu-chu.yang@intel.com>
Update some configuration settings that will be obsolete in the new
(1.8.17) doxygen release.
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
This patch enables doorbell feature for hv-land
ivshmem device to support interrupt notification
between VMs that use inter-VM(ivshmem) devices.
Tracked-On: #5407
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Reviewed-by: Li, Fei <fei1.li@intel.com>
Reviewed-by: Wang, Yu1 <yu1.wang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
This function can be used by other modules instead of hypercall
handling only, hence move it to vlapic.c
Tracked-On: #5407
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Reviewed-by: Li, Fei <fei1.li@intel.com>
Reviewed-by: Wang, Yu1 <yu1.wang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
- write_vmsix_cap_reg(): emulates vmsix cap writes.
write_pt_vmsix_cap_reg(): emulates msix cap write
for PT devices.
- rw_vmsix_table(): emulates vmsix table bar space access.
- vmsix_handle_table_mmio_access(): emulates the vmsix
bar space access only.
- pt_vmsix_handle_table_mmio_access(): emulates the vmsix
bar space access and remap msi entry for PT device if
write operation is executed.
- rename 'init_vmsix()' and 'deinit_vmsix()' to
'init_vmsix_pt()' and 'deinit_vmsix_pt()' respectively,
they're for PT devices only.
- remove below 2 functions,call
'pci_vdev_read_vcfg()' directly in cases they're used.
- 'read_vmsi_cap_reg()'
- 'read_vmsix_cap_reg()'
Tracked-On: #5407
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Reviewed-by: Li, Fei <fei1.li@intel.com>
Reviewed-by: Wang, Yu1 <yu1.wang@intel.com>
Acked-by: Eddie Done <eddie.dong@intel.com>
vmsix.c originally covers ptdev case but ACRN hypervisor
need to support pure virtual PCI mediator, such as ivshmem
device in this patch set.
For better understanding the code changes from patch
perspective, split the changes to several small patches.
This patch moves most original vmsix code to pci_pt.c
as they're mixed with ptdev specific operations.
The subsequent patches will start the detail abstraction change.
Tracked-On: #5407
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Reviewed-by: Li Fei <fei1.li@intel.com>
Reviewed-by: Wang, Yu1 <yu1.wang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Fix typo error "--apicdev_pt HID" to pass through a TPM device. Fix it to
"--acpidev_pt HID"
Tracked-On: #5401
Signed-off-by: Li Fei1 <fei1.li@intel.com>
Now ACRN supports direct boot mode, which could be SBL/ABL, or GRUB boot.
Thus the vboot wrapper layer can be removed and the direct boot functions
don't need to be wrapped in direct_boot.c:
- remove call to init_vboot(), and call e820_alloc_memory() directly at the
time when the trampoline buffer is actually needed.
- Similarly, call CPU_IRQ_ENABLE() instead of the wrapper init_vboot_irq().
- remove get_ap_trampoline_buf(), since the existing function
get_trampoline_start16_paddr() returns the exact same value.
- merge init_general_vm_boot_info() into init_vm_boot_info().
- remove vm_sw_loader pointer, and call direct_boot_sw_loader() directly.
- move get_rsdp_ptr() from vboot_wrapper.c to multiboot.c, and remove the
wrapper over two boot modes.
Tracked-On: #5197
Signed-off-by: Zide Chen <zide.chen@intel.com>
Since now we support direct boot only, we don't have to use FIRMWARE variable
to differentiate between sbl/GRUB and UEFI boot.
After this change:
- "FIRMWARE=sbl/uefi" should be removed from make commands.
- the firmware name is removed from the installed ACRN image. For example,
acrn.apl-up2.sbl.sdc.32.out will be changed to acrn.apl-up2.sdc.32.out.
Tracked-On: #5197
Signed-off-by: Zide Chen <zide.chen@intel.com>