Pointer arithmetic is currently used to calculate the address of a specific
Local Vector Table (LVT) register (except LVT_CMCI) in lapic, since the
registers are continuously placed with fixed padding in between. However each of
these registers are declared as a single uint32_t in struct lapic, resulting
pointer arithmetic on a non-array pointer which violates MISRA C requirements.
This patch refactors struct lapic by converting the LVT registers fields (again
except LVT_CMCI) to an array named lvt. The LVT indices are reordered to reflect
the order of the LVT registers on hardware, and reused to index this lvt array.
The code before and after the changes is semantically equivalent.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
filter out HW exception and NMI from idt_vectoring_info first:
- queue HW exception through vcpu_queue_exception
- make NMI request through vcpu_make_request
this is a complement patch for previous commit "exception: refine exception
injection path", here take care un-injected vectors for types HW exception &
NMI, the previous commit take care SW exception & external interrupt.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Tian, Kevin <kevin.tian@intel.com>
use func vcpu_queue_exception for vcpu_inject_gp and exception_vmexit_handler.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Tian, Kevin <kevin.tian@intel.com>
- inject exception based on exception_info
- inject exception according to priority
- for previous not injected vector - idt_vectoring_info, here only take
care about types for SW exception & external interrupt, the HW exception
& NMI will be taken care by vmexit_handler in the following commit "correct
idt_vectoring_info handling"
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Tian, Kevin <kevin.tian@intel.com>
add func vcpu_queue_exception to queue exception based on SDM Vol3 Table 6-5,
which may cause #DF or triple fault
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Tian, Kevin <kevin.tian@intel.com>
the pending_intr is not only serving for interrupt but also for different
request including TLB & TMR updating, so change the function & variants
name accordingly.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
It is an extension of GCC CPP to:
* allow omitting a variable macro argument entirely, and
* use ## __VA_ARGS__ to remove the the comma before ## __VA_ARGS__ when
__VA_ARGS__ is empty.
The only use of ## _VA_ARGS__ is to define the pr_xxx() macros, with the first
argument being the format string and the rest the to-be-formatted arguments. The
format string is explicitly spelled out because another macro pr_fmt() is used
to add to the format string a prefix which is customizable by defining what
pr_fmt() expands to.
For C99 compliance, this patch changes the pr_xxx() macros in the following
pattern.
- #define pr_fatal(fmt, ...) \
- do_logmsg(LOG_FATAL, pr_fmt(fmt), ## __VA_ARGS__); \
+ #define pr_fatal(...) \
+ do_logmsg(LOG_FATAL, pr_prefix __VA_ARGS__); \
Reference:
* https://gcc.gnu.org/onlinedocs/gcc/Variadic-Macros.html#Variadic-Macros
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
According to the syntax defined in C99, each struct/union field must have an
identifier. This patch adds names to the previously unnamed fields for C99
compatibility.
Here is a summary of the names (marked with a pair of *stars*) added.
struct trusty_mem:
union {
struct {
struct key_info key_info;
struct trusty_startup_param startup_param;
} *data*;
uint8_t page[CPU_PAGE_SIZE];
} first_page;
struct ptdev_remapping_info:
union {
struct ptdev_msi_info msi;
struct ptdev_intx_info intx;
} *ptdev_intr_info*;
union code_segment_descriptor:
uint64_t value;
struct {
union {
...
} low32;
union {
...
} high32;
} *fields*;
similar changes are made to the following structures.
* union data_segment_descriptor,
* union system_segment_descriptor,
* union tss_64_descriptor, and
* union idt_64_descriptor
struct trace_entry:
union {
struct {
uint32_t a, b, c, d;
} *fields_32*;
struct {
uint8_t a1, a2, a3, a4;
uint8_t b1, b2, b3, b4;
uint8_t c1, c2, c3, c4;
uint8_t d1, d2, d3, d4;
} *fields_8*;
struct {
uint64_t e;
uint64_t f;
} *fields_64*;
char str[16];
} *payload*;
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
According to the syntax defined in C99, each struct/union field must have an
identifier. This patch removes unnamed struct/union fields that can be easily
expressed in a C99-compatible way.
Here is a summary of structs/unions removed.
struct vhm_request:
union {
uint32_t type; uint32_t type;
int32_t reserved0[16]; => int32_t reserved0[15];
};
struct vhm_request_buffer:
struct vhm_request_buffer {
union { union vhm_request_buffer {
struct vhm_request ...; => struct vhm_request ...;
int8_t reserved[4096]; int8_t reserved[4096];
} }
}
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
The macro GET_DATA_SIZE's code style is not correct,
it is merged by accident,the Pull Request
<fix "obsolete use of designated initializer without '='">
have two commid id,the second commit changed the ucode.c
by mistake.
Now fixed.
Signed-off-by: huihuang.shi <huihuang.shi@intel.com>
According to the comments in hypervisor:
" This file includes config header file "bsp_cfg.h" and other
hypervisor used header files.
It should be included in all the source files."
this patch includes all common header files in hypervisor.h
then removes other redundant inclusions
Signed-off-by: Zide Chen <zide.chen@intel.com>
According to the C99 manual 6.5.15
syntax conditional-expression
logical-OR-expression
logical-OR-expression ? expression :conditional-expression
above bnf rules requires expression not be none
Signed-off-by: huihuang.shi <huihuang.shi@intel.com>
- move vmexit handling into vmexit_handler
- add error handling, failure will inject #GP
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
there are data transfer between guest(GPA) & hv(HPA), especially for
hypercall from guest.
guest should make sure these GPAs are address continous, but hv cannot
assure HPAs which mapped to these GPAs are address continous, for example,
after enable hugetlb, a contious GPA range could come from two different
2M pages.
this patch is handling such case by doing gpa page walking during
copy_from_vm & copy_to_vm.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
merge these two APIs to 'dump_intr_excp_frame'
Signed-off-by: Mingqiang Chi <mingqiang.chi@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
rename intr_ctx to intr_excp_ctx
Signed-off-by: Mingqiang Chi <mingqiang.chi@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
only trusty related hypercall will come from UOS, others should come from VM0
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Exception type is missing when inject a #GP.
GP should be injected even when guest irq disabled.
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <Eddie.dong@intel.com>
- MMIO access size is not initialized before
instruction emulation.
- rename 'analyze_instruction()' to 'decode_instruction()'
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
rename atomic_cmpxchg_int to atomic_cmpxchg
replace atomic_cmpset_long with atomic_cmpxchg64
rename atomic_readandclear_long to atomic_readandclear64
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@Intel.com>
1. vpic_wire_mode no need to atomically set its value.
2. uart open conut want to atomically set its value not atomically set
its bits.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Check ept rwx misconfigurations when config memory attribute,
if misconfig it will assert.
Signed-off-by: Mingqiang Chi <mingqiang.chi@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Lock should be on target vcpu in function pause_vcpu(), not current vcpu.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Add set_memmaps hypercall to support multi regions memmap.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
allocated all pcpus to vm0 to handle possible AP wakeup flow for all cpus,
as we pass org ACPI table to VM0 - that means VM0 can see all CPUs.
SOS(VM0) start expected CPUs through "maxcpus=" kernel cmdline option.
During first hypercall from SOS, calling vm_fixup to free un-expect-enabled
vcpus from VM0.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
remove sipi_from_efi_boot_service_exit & efi_deferred_wakeup_pcpu workaround
for uefi boot flow
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
wakeup AP need INIT-SIPI-SIPI sequence in old time, now we only need
INIT-SIPI.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
- device driver should register valid i/o handlers
in any cases, avoid referencing to default handler
- remove i/o handler test code as they shall
never be NULL.
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
If the optimization option is enabled, it is possible that one variable is
not initialized before using in the get_vioapic_info. (In fact the warning is
bogus)
This is only to reduce the compiling warning.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Now one uint64_t type is used to obtain the corresponding descriptor_table
for GDT/IDT. This will cause the stack protect corruption under -O2.
So the descriptor_table struct is added to configure the GDT/IDT of VMCS.
V1->V2: Move the descriptor_table into vmx.h header file
And its type is renamed from dt_addr_t to descriptor_table.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Zheng Gen <gen.zheng@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
The RFLAGS will be touched in some inline assembly.(exec_vmxon/
RFLAGS_RESTORE). The "cc" constraint should be added. Otherwise
it won't be handled under -O2 option.
And "%%XXX" register should also be added into constraints.
Otherwise it will be optimized incorrectly.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Zheng Gen <gen.zheng@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Add vlapic_create_timer/vlapic_reset_timer to setup/reset a timer.
Add vlapic_update_lvtt to disarm timer when mode changes.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Using __func__ instead of function string name.
Using tab instead of more whitespace.
Using macro instead of numeric constants.
Remove unnecessary function declaration.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
These code is useless since it just pseudocode. Current it doesn't
support vlapic one-shot/periodic timer.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
make the timer list be ordered to speed up expried timer
process and next timer event finding.
Add timer would not schedule timer unless it's the next
timer event.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
V2->V3: Updated variable name: trampoline_code_paddr
V1->V2: changed variable name: init_ap_code_addr
These page tablea are sitting right after the trampoline code, so adjust it according to
the actual loaded address for trampoline code
Signed-off-by: Zheng, Gen <gen.zheng@intel.com>
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Signed-off-by: Zide Chen <zide.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
V3->V4: Updated function/variable names for accurancy
V2->V3: Changed a few function/variable names to make it less confusing
V1->V2: removed the unneccesary cache flushing
- For UEFI boot, allocate memory for trampoline code in ACRN EFI,
and pass the pointer to HV through efi_ctx
- For other boot, scan E820 to allocate memory in HV run time
- update_trampoline_code_refs() updates all the references that need the
absolute PA with the actual load address
Signed-off-by: Zheng, Gen <gen.zheng@intel.com>
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Signed-off-by: Zide Chen <zide.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
V2->V3: Fixed the booting issue on MRB board and removed the restriction
of allocate memory from address 0
1) Fix the booting from MRB issue
-#define CONFIG_LOW_RAM_SIZE 0x000CF000
+#define CONFIG_LOW_RAM_SIZE 0x00010000
2) changed e820_alloc_low_memory() to handle corner case of unaligned e820 entries
and enable it to allocate memory at address 0
+ a length = end > start ? (end - start) : 0;
- /* We don't want the first page */
- if ((length == size) && (start == 0))
- continue;
3) changed emalloc_for_low_mem() to enable to allocate memory at address 0
- /* We don't want the first page */
- if (start == 0)
- start = EFI_PAGE_SIZE;
V1->V2: moved e820_alloc_low_memory() to guest.c and added the logic to
handle unaligned E820 entries
emalloc_for_low_mem() is used if CONFIG_EFI_STUB is defined.
e820_alloc_low_memory() is used for other cases
In either case, the allocated memory will be marked with E820_TYPE_RESERVED
Signed-off-by: Zheng, Gen <gen.zheng@intel.com>
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Signed-off-by: Zide Chen <zide.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
V1->V2: removed CONFIG_LOW_RAM_START and added ".org 0" to
cpu_secondary.S
The assumption is trampoline code is relocated while HV is not, so:
trampoline code is built at address 0, and CS register is updated
by SIPI to reflect the correct vector
in real mode part, added extra pointers for page tables and long jump buffer
so it's possible for HV code to patch the relocation offset
in long mode part, use absolute addressing when referring HV symbols,
and use relative addressing for symbols within trampoline code
Signed-off-by: Zheng, Gen <gen.zheng@intel.com>
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Signed-off-by: Zide Chen <zide.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
Currently config_page_table_attr() treats MMU_MEM_ATTR_READ exactly as
MMU_MEM_ATTR_BIT_READ_WRITE for PTT_HOST, so even when MMU_MEM_ATTR_WRITE
is not used, the R/W bit in PTE is still being set
Signed-off-by: Zide Chen <zide.chen@intel.com>
- According to Intel SDM 24.9.2,Vol3, should check the
validity of "VM-exit interruption information" before
extracting the vector of interrupt.
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
- rename it to 'io_shared_page' to keep consistent
with ACRN HDL foils.
- update related code that reference this data structure.
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
- Add "DBG" or "REL" to indicate the DBG build or REL build explicityly;
- Change the build time format to "%F %T".
Example:
HV version 0.1-rc4-2018-04-28 14:20:32-b2d7282-dirty DBG build by like
Change-Id: Ib410064b0a6603e3c90f30dffa722237c07fc069
Signed-off-by: Yan, Like <like.yan@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Make assert on max px cnt of boot cpu data, since it shouldn't happen if
px data is properly initialized in boot process.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
The initial of iobitmap pointer should be moved out of loop since address
is sequentially incremented.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
with this patch guest could access idle io port and enter idle normally.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
Each VM would have its own Cx data, for now we copy it from boot_cpu_info.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
The cx data is hardcoded within HV, load it to boot_cpu_data when HV boot.
The patch provide a3960 soc cx data for example.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
currently, pass-thru devices are managed by per-vm's remapping entries
which is virtual based:
- MSI entry is identified by virt_bdf+msix_index
- INTx entry is identified by virt_pin+vpin_src
it works but it's not a good design for physical resource management, for
example a physical IOAPIC pin could belong to different vm's INTx entries,
the Device Model then must make sure there is no resource conflict from
application's level.
This patch change the design from virtual to physical based:
- MSI entry is identified by phys_bdf+msix_index
- INTx entry is identified by phys_pin
The physical resource is directly managed in hypervisor, a miss adding
entry will be found by hypervisor and return error message with failure.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
add initialize_timer to initialize or reset a timer;
add_timer add timer to corresponding physical cpu timer list.
del_timer delete timer from corresponding physical cpu timer list.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Get tsc hz by cpuid 0x15 if we supported, otherwise
calibrate tsc by pit timer.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
If guest reboot is issued before trusty init hypercall is issued,
we shouldn't destroy ept fo trusty memory because the ept is not
created yet.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Using eax will truncate the high 32bit part of 64bit virtual address.
And the type of sync is unsigned long, so using rbx instead of ebx.
Signed-off-by: Zheng, Gen <gen.zheng@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
This is to do the clean-up of IOAPIC mmio-access. Use the same API to
access the IOAPIC register. At the same time it also helps to avoid the
optimization in direct access mode.(The volatile is already added in
mmio_read_long/mmio_write_long)
V1->V2: Follow Fengwei's suggestion to use the mmio_read/write_long
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Now just add some basic feature/capability detect (not all). Vapic
didn't add here for if we must support vapic then the code which
for vapic not supported must remove, like mmio apic r/w.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Split pm.c from cpu_state_tbl.c to put guest power management related
functions, keep cpu_state_tbl.c to store host cpu state table and
related functions.
Signed-off-by: Victor Sun <victor.sun@intel.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
for timer list is operated by per-cpu; and no interrupt
service operates it too. So it's unnecessary for spinlock.
Signed-off-by: Minggui Cao <minggui.cao@intel.com>
Reviewed-by: Yin Fengwei <fengwei.yin@intel.com>
Reviewed-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
adding "hugepagesz=1G" and "hugepages=X" into SOS cmdline, for X, current
strategy is making it equal
e820_mem.total_mem_size -CONFIG_REMAIN_1G_PAGES
if CONFIG_REMAIN_1G_PAGES is not set, it will use 3 by default.
CONFIG_CMA is added to indicate using cma cmdline option for SOS kernel,
by default system will use hugetlb cmdline option if no CONFIG_CMA defined.
Signed-off-by: Jason Chen CJ <jason.cj.chen@intel.com>
Acked-by: Xu, Anthony <anthony.xu@intel.com>
