Update the instructions on how to increase the size of a UOS disk image so that
the entire operation can be performed on the development host. This has the
advantage that it can be done completely offline (the UOS must be powered off
in fact) and no extra tools need to be installed in the UOS to do this.
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
For open source, just need to refer FuSa standard instead
of listing concrete fulfillment matrix.
Add related FuSa standard reference and remove related
fulfillment matrix.
V1-->V2:
Make reference more clear according to David'comments
Signed-off-by: Xiangyang Wu <xiangyang.wu@linux.intel.com>
Currently MSR IA32_MISC_ENABLE is passthrough to guest.
However, guest may change the value of this MSR, which will cause issue in hypervisor.
This patch uses VMX MSR store area to isolate the MSR IA32_MISC_ENABLE between guest and host.
TODO:
Some bits of the MSR IA32_MISC_ENABLE is not just per core, but per package.
So need to check if need to prevent guest from setting or clearing these bits that may affect other cores.
Tracked-On: #2834
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Rename the field msr_num to msr_index, which is more accurate,
in struct msr_store_entry.
Tracked-On: #2834
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
-- move this api from misc.c to timer.c to avoid
reverse dependency, and remove misc.c
Tracked-On: #1842
Signed-off-by: Mingqiang Chi <mingqiang.chi@intel.com>
- for all cases of referring guest bootargs size, replace MEM_2K with
CONFIG_MAX_BOOTARGS_SIZE for better readability.
- remove duplicated MAX_BOOTARGS_SIZE definition from vm_config.h.
Also fix one minor issue in general_sw_loader() which uses copy_to_gpa()
to copy a string. Since copy_to_gpa() makes use of memncpy_s() to do the
job, the size parameter should include the string null ternimator.
Tracked-On: #2806
Signed-off-by: Zide Chen <zide.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
print each IRQ descriptor interrupts number and all of IRQ descriptors
interrupts when UOS requests or releases a GPIO IRQ.
Tracked-On: #2512
Signed-off-by: Yuan Liu <yuan1.liu@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
Support reading GPIO value when the GPIO switches to IRQ mode.
Tracked-On: #2512
Signed-off-by: Yuan Liu <yuan1.liu@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
GPIO IRQ controller emulation is used to handle level trigger and
edge trigger interrupts. Use GPIO IRQ virtqueue to handle IRQ chip
operations and GPIO event virtqueue to indicate IRQ source to UOS.
Tracked-On: #2512
Signed-off-by: Yuan Liu <yuan1.liu@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
add the GPIO IRQ definitions, and implement the GPIO IRQ
initialization and deinitialization.
Tracked-On: #2512
Signed-off-by: Yuan Liu <yuan1.liu@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
There are two virtqueues for irq, one for handling the operations of
front-end irq controller and the other for triggering the interrupt.
Tracked-On: #2512
Signed-off-by: Yuan Liu <yuan1.liu@intel.com>
Acked-by: Yu Wang <yu1.wang@intel.com>
Remove the unused seed parsing source files under
hypervisor/boot/sbl and related header files.
Tracked-On: #2724
Signed-off-by: Qi Yadong <yadong.qi@intel.com>
Added a struct acrn_vm in firmware.h to remove
a compiler warning.
No change in logic.
Tracked-On: #2830
Signed-off-by: Arindam Roy <arindam.roy@intel.com>
OVMF requires a more descriptive mechanism than RTC CMOS to retrieve
ACRN's memory layout, so we now pass the E820 map to it, starting at
0xEF000 (ROM area).
ACRN currently uses [4GB, 5GB) as its 64-bit PCI host aperture. This is
inconsistent with OVMF's assumption of its platform's memory layout,
because it derives the size of high memory from RTC CMOS, which is
incapable of describing the 64-bit PCI hole.
By default, OVMF uses RTC CMOS 0x5b/0x5c/0x5d to determine the size of
high memory. This value only tells OVMF how much memory is above 4GB,
but not the platform's memory layout above 4GB.
Using RTC CMOS works for QEMU, because QEMU places its 64-bit PCI host
aperture above its highmem. Therefore, OVMF can always assume highmem is
located at [4GB, 4GB + highmem), which is not where ACRN's highmem is
located. For example, if we have 1GB of usable memory above 4GB, ACRN
will place it at [5GB, 6GB).
This change allows OVMF to correctly identify the guest's memory layout.
It will consider any reserved region above 4GB as 64-bit PCI host
aperture.
MP table, SMBIOS and ACPI tables are all located above 0xF0000 so it is
guaranteed that there is no overlap. There can only be a maximum of 128
E820 entries.
v1 -> v2:
- provide more explanation to this commit
- add signature before E820 map for OVMF backward compatibility
Tracked-On: #2792
Signed-off-by: Peter Fang <peter.fang@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
While building the E820 map for UOS, [lowmem, lowmem_limit) and [5GB,
highmem) can be empty. Remove the empty entries if they appear.
Tracked-On: #2792
Signed-off-by: Peter Fang <peter.fang@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
- fix comments
- update the first RAM region to [0, 0xA0000) because [0xA0000, 1MB) is
designated as video memory and ROM area
- use 2GB as lowmem_limit
Tracked-On: #2792
Signed-off-by: Peter Fang <peter.fang@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
Add the PCI MMIO fallback handler to the 64-bit PCI host aperture, so
that the guest won't inadvertently crash acrn-dm due to unhandled MMIO.
Tracked-On: #2792
Signed-off-by: Peter Fang <peter.fang@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
- use strncmp() instead of comparing string pointers to make no
assumptions about the toolchain's literal pool
- re-shuffle the functions so they're consistent with mem.h
- make non-public functions static
- increase code re-use
Tracked-On: #2792
Signed-off-by: Peter Fang <peter.fang@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
Per commit dbd9ab07e1, GUEST_CFG_OFFSET is
no longer needed.
Tracked-On: #2792
Signed-off-by: Peter Fang <peter.fang@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Acked-by: Yin Fengwei <fengwei.yin@intel.com>
This patch add tutorial about using zephyr as guest.
Tracked-On: #2713
Signed-off-by: Kaige Fu <kaige.fu@intel.com>
---
v3 -> v4:
- Addressed two minor comments from David
v2 -> v3:
- Detail section `Build Zephyr`
- Add missing sudo
- Automatically boot zephyr
- Other grammar, phrasing fix
v1 -> v2:
- NUC7i5DNH -> NUC7i5DNHE.
- Add note about doc to build acrn supported zephyr image.
- Refine section `Boot Zephyr as User OS`.
- Some minor fix.
Enhance the ACRN shell interactive help. It is close to a 1-1 mapping with
the online documentation but cut a little shorter in various places to make it
more user-friendly when using it from the ACRN console.
Tracked-On: #2829
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
To merge the multiboot bootargs within sbl_init_vm_boot_info(), buffer
overflow could happen when it doesn't provide correct 'dmax' argument
to strncpy_s().
Also, currently it doesn't check the availability of the dest buffer before
overwriting '\0' with a whitespace, which theoretically the dest string
could end up with no null terminator within it's array boundary.
This patch also creates a separate function to merge the cmdline strings,
because after the above fixes some lines in sbl_init_vm_boot_info()
function could have up to 7 tabs in front of the first character, which
looks messy and sbl_init_vm_boot_info() is getting too complicated.
Tracked-On: #2806
Signed-off-by: Zide Chen <zide.chen@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Linux access TSC_ADJUST to verify it has not tampered every time when enter idle.
So for RTVM running rt-linux, the access will cause vm exit which affect real-time performance.
This commit pass through TSC_ADJUST to VM with lapic_pt, to avoid TSC_ADJUST caused vm_exit.
For other VMs, TSC_ADJUST msr access is still trapped and emulated.
Tracked-On: #2813
Signed-off-by: Yan, Like <like.yan@intel.com>
This patch hide Memory Protection Extention (MPX) capability from guest.
- vCPUID change:
Clear cpuid.07H.0.ebx[14]
Clear cpuid.0DH.0.eax[4:3]
- vMSR change:
Add MSR_IA32_BNDCFGS to un-supported MSR array.
- XCR0[4:3] is not allowed to set by guest.
Tracked-On: #2821
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Both sharing mode and partition mode should follow the same coding logic/style
for similar functions:
vdev cfgread/cfgwrite: should all return -ENODEV if the pci reg access is not handled by
it, but previously the partition mode code is not following this logic
vpci cfgread/cfgwrite: if the vdev cfgread/cfgwrite does not handle this reg,
pass on to next vdev cfgread/cfgwrite, if no vdev handles that req, passthru to
physical pci device
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Do the pci cfg read/write sanity checking before the request is dispatched to
submodules, so that the checking is centralized rather than scattered across multiple
files/places
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
In preparation for vpci ops function removal, so that these functions can be
called directly instead by vpci code
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Remove vdev ops for partition mode, change related code to directly call the corresponding
functions instead
Remove struct pci_vdev_ops from vpci.h
Add @pre for pci_find_vdev_by_pbdf and pci_find_vdev_by_vbdf/partition_mode_vpci_init
Change the return value from int32_t to void to comply with misra c and
add ASSERT/panic in the functions (if necessary):
vdev_hostbridge_init
vdev_hostbridge_deinit
vdev_pt_init
vdev_pt_deinit
Still use pr_err in partition_mode_cfgread and partition_mode_cfgwrite to check if vdev cfgread/cfgwrite
access is aligned on 1/2/4 bytes, which is the only case that vdev cfgread/cfgwrite will return
nonzero, pr_err will be removed in subsequent patch titled "unify the sharing
mode and partition mode coding style for similar functions"
Remove @pre for local variables
Add ASSERT in partition_mode_pdev_init to check if pdev is NULL (user config
error)
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
ACRN uses global invalidation for all DMAR translation caches. Whenever
a UOS is shutdown or rebooted, it ends up clearing entries in translation
caches belonging to other VMs/domains. This patch adds support for
domain/device level invalidation for DMA translation caches and index
based invalidation for Interrupt Remapping Cache.
Tracked-On: #2738
Signed-off-by: Sainath Grandhi sainath.grandhi@intel.com
Acked-by: Eddie Dong eddie.dong@intel.com
strncpy_s(d, dmax, s, slen): the 'dmax' includes the null terminator, while
slen doesn't. Thus if (dmax == slen == strlen(s)), strncpy_s() chooses to
discard the last character from s and instead write '\0' to d[dmax - 1].
strnlen_s(s, maxsize): if there is no terminating null character in the
first maxsize characters pointed to by s, strnlen_s() returns maxsize.
So in the following example or similar cases, we need to increase the size
of d[] by 1 to accommodate the null terminator, and add '1' to the dmax
argument to strncpy_s().
uint8_t d[MAX_LEN];
size = strnlen_s(s, MAX_LEN);
strncpy_s(d, MAX_LEN, s, size);
Tracked-On: #861
Signed-off-by: Zide Chen <zide.chen@intel.com>
Add a tutorial on how to increase the disk size of a User OS (UOS)
based on Clear Linux. The tutorial is largely inspired from Clear Linux
tutorial and refers to it. It mostly explains how to resize the .img
file using the 'qemu-img' tool since ACRN does not have any dedicated
tool for manipulating virtual disk image files.
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Co-Authored-By: David B. Kinder <david.b.kinder@intel.com>
Add 'rdmsr' and 'wrmsr' to the "ACRN Shell Commands" documentation. Both are
already described via the interactive help command but not (yet) included in our
on-line documentation.
Tracked-On: #2684
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Currently board names are used inconsistently across the project.
* Name of defconfigs for various boards use lowercase.
* Directory of config files for various boards use lowercase.
* CONFIG_BOARD uses uppercase.
This confuses the configuration scripts and leads to unintended overwriting of
.config, as well as missing of board-specific headers during compilation because
the include paths are case-sensitive.
This patch converts the default board names to lowercase to resolve such
issues. Users are still free to define their own boards in either uppercase or
lowercase as long as they keep the cases consistent.
Tracked-On: #2794
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Preparing for hybrid mode:
- create vE820 for pre-launched VMs and do other init code when the
vm_config->type is PRE_LAUNCHED_VM.
- create ve820.c for each board because without wrapping by
CONFIG_PARTITION_MODE, ve820_entry[] needs to be visible even when
compiling target boards that haven't enabled pre-launched VMs.
- remove create_prelaunched_vm_e820() from vm.c and implement board
specific function for each $(CONFIG_BOARD)/ve820.c. The reasons being:
- don't need to define ve820_entry[32] for those boards that don't
support pre-launched VMs.
- more importantly, this makes it much easier to create different per-VM
vE820 when it's needed.
Tracked-On: #2291
Signed-off-by: Zide Chen <zide.chen@intel.com>
Previously vrtc is for partition mode only, now enable it for sharing mode;
Tracked-On: #2291
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Suppose run_ctx.cr0/cr4 are correct when do world switching, so call
vcpu_set_cr0/cr4() to update cr0/cr4 directly before resume to guest.
This design is only for trusty world switching.
Tracked-On: #2773
Signed-off-by: Qi Yadong <yadong.qi@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Remove vdev ops for sharing mode, directly call the corresponding functions
instead of calling the ops callbacks (indirectly)
Remove alloc_pci_vdev() and merge its code into init_vdev_for_pdev() to simplify code
Remove @pre for local variables
Change the return value from int32_t to void to comply with misra c and
add ASSERT in the functions (if necessary) to verify the assumptions for debug build:
vmsi_init
vmsix_init
vmsi_deinit
vmsix_deinit
Add @pre for vmsix_init_helper and make it a void function, use ASSERT to verify
the assumption for debug build.
Add ASSERT in get_sos_vm
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
This is to fix the following misra c violation:
Pointer param should be declared pointer to const. : vdev
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
No need to use these 2 variables as global (per pci_misx), can simply use local
variables in code instead.
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Define has_msi_cap and has_msix_cap inline functions to do sanity checking for
msi and msix ops, the corresponding code block in existing code is replaced with
a call to these new functions.
A few minor coding style fix.
Tracked-On: #2534
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Now we only configure "hide MTRR" explicitly to false for SOS. For other VMs,
we don't configure it which means hide_mtrr is false by default.
And remove global config MTRR_ENABLED
Tracked-On: #1842
Signed-off-by: Li, Fei1 <fei1.li@intel.com>