To be sorted in lexical order let's add a 50- prefix to the systemd
network units files.
Now these systemd network units will be processed before the other units
prefixed by 80-. And the first (in lexical order) of the network files
that matches a given device is applied, all later files are ignored,
even if they match as well.
Signed-off-by: Miguel Bernal Marin <miguel.bernal.marin@linux.intel.com>
Acked-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
We can not overwrite system files, there are others ways to overwrite
configuration.
Fixes: b7ad7052bd ("convert bridge creation to service")
Signed-off-by: Miguel Bernal Marin <miguel.bernal.marin@linux.intel.com>
Acked-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
According to MISRA C:2012, suffix 'U/UL' shall be for
unsigned const value, the member of enum variable should
not be used to compare with integer variable.
Add 'U/UL' for unsigned const value in the CPU module;
Use Macro insteading of enum feature_word since the member
of feature_word is used to compare with integer variable;
Use hex number insteading of Macro in the assembly code.
V1-->V2:
Update the suffix of some constant value as 'UL'
according to its'storage variable;
Split MACRO updates used in the assembly code
in other patch.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
The 23:16 bits in the IOAPIC version register is the highest entry in the I/O
redirection table, which is 1 smaller than the number of interrupt input
pins. But currently we use these bits directly as the number of pins.
This patch abstracts the pin count calculation code and adds 1 to the max rte
entry number to get the right pin count.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Some UOS need use FB device to display, while initial_modeset will disable fb,
so we need remove correlation between plane restriction and initial modeset
in kernel driver, also don't need set enable_initial_modeset in launch script.
Signed-off-by: Fei Jiang <fei.jiang@intel.com>
Reviewed-by: Min He <min.he@intel.com>
This patch is to fix the potential buffer overflow issues.
Signed-off-by: CHEN Gang <gang.c.chen@intel.com>
Reviewed-by: Zhi Jin <zhi.jin@intel.com>
Reviewed-by: xiaojin2 <xiaojing.liu@intel.com>
MISRA C requires that all unsigned constants should have
the suffix 'U/UL'(e.g. 0xffU), but the assembler may not
accept such C-style constants.
To work this around, all unsigned constants must be
explicitly spells out in assembly with a comment tracking
the original expression from which the magic number is
calculated.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
Here is how the recursion might happen:
when there is something wrong
|
sbuf_put -> memcpy_s -> pr_err -> do_logmsg
| |
-----------------------------------
Replace 'pr_err' with 'ASSERT' in 'memcpy_s' to break this kind of
recursion.
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
It will print error information inside memcpy_s if
the parameteter is invalid, the caller can not check
the return value for memcpy_s/strcpy_s/strncpy_s
code like this:
int a(void) {
return 0;
}
int b(void){
a();
}
fix as follow:
int a(void) {
return 0;
}
int b(void){
(void)a();
}
Signed-off-by: Mingqiang Chi <mingqiang.chi@intel.com>
The 'launch_uos.sh' script was updated recently and new lines
added to it. Adjust the 'emphasize-lines' directive accordingly.
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
If option -t is configured, a timer will be set with configured timeout when
launching acrntrace. When timer fired, acrntrace will exit.
Signed-off-by: Kaige Fu <kaige.fu@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Option -i is more suitable for interval. Meanwhile, we will use option -t for
timeout in later patch.
Signed-off-by: Kaige Fu <kaige.fu@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
acrntrace writes trace data to /tmp with unlimited size and will cause tmpfs
100% occupied. Consequently, some problem will raise up on SOS, like failing to
exec cmd.
This patch introduce an option -r to let user configre minimal space left on the
disk which ensures that acrntrace will exit when free storage space is less than
the reserved space. By default, we reserve 512M on the disk. Users can configure
reserved space through '-r'.
Signed-off-by: Kaige Fu <kaige.fu@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
To prevent the value is devided by zero, checks the denominator
before the calculation. Adding the if statement to check before use.
If the baud_rate is equal to zero, using default baud_rate.
Signed-off-by: Yang, Yu-chu <yu-chu.yang@intel.com>
There are some implicit type conversion in the CPU module,
this voilates MISRA C:2012 required rules.
We walk through each functions of CPU module, update
some variable type and the return value type of
the function, and add 'U/UL' for related const value.
V1-->V2:
Rebase the branch, resolve some conflict.
V2-->V3:
Update commit info for V2.
V3-->V4:
Few updates for fixing error instroduced during
resolving conflict.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
Adjusted the picture sizes a bit smaller, fixed some spelling errors,
and moved the boot-flow.dot image used by trusty into the common images
folder (and renamed the image to trusty-boot-flow.dot)
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
IOAPIC pins always fit in 8-bit and we already use uint8_t for virt_pins. This
patch converts pins in vioapic to uint8_t.
This is based on Arindam's previous patch ("was: hv: Cleanup and optimise
vioapic.c"), with SOS boot failure resolved, format string updated, complex
arithmetic expression with implicit type conversion decoupled. Also make some
local variables representing interrupt vectors uint32_t.
Signed-off-by: Arindam Roy <arindam.roy@intel.com>
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
A couple of diagram were written using text characters. This
commit changes that to use pictures instead.
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Move the existing Trusty document to the doc/ folder (where
it belongs) and convert the text to ReST.
The Documentation/ folder under hypervisor/ is removed as all
documents should be put under doc/.
All technical information has been preserved or was already
available in other documents.
Signed-off-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
No need to check the return value for memset
code like this:
int a(void) {
return 0;
}
int b(void){
a();
}
fix as follow:
int a(void) {
return 0;
}
int b(void){
(void)a();
}
Signed-off-by: Mingqiang Chi <mingqiang.chi@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Most of variables used when decoding instructions are unsigned bit-fields. This
patch adds the 'U' or 'UL' suffix to constants in unsigned context and changes
the type of arguments/local variables when necessary.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The SEG_DESC_* macros are mostly used as branch conditions though they evaluates
to signed int. This patch simplies their definitions and drop the unnecessary
casts accordingly.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch makes necessary integer narrowing and/or signedness conversion
explicit.
While some narrowing are expected behavior, the correctness of the others relies
on the specifications of some interfaces (e.g. the higher 32-bit of what
exec_vmread() returns is all 0s if the given field is 32-bit). Add a stub
for now to avoid missing them.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Translate the VIE_OP_* enum constants (which belong to an anonymous enum type)
to macros to ensure that they are always unsigned.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
CPL is represented by a plain int but calculated from shifts and bit-wise
operations. Convert it the uint8_t for consistency.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
get_vmcs_field() returns a VMCS field offset which is normally unsigned, but it
also returns negatives (-1 here) on invalid arguments. Following the convention
we use for vectors, pins, etc., use a special unsigned value to indicate such
errors.
v1 -> v2:
* Use a special value (VMX_INVALID_VMCS_FIELD) instead of a seperate output
parameter to indicate errors.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current register names in instr_emul are misleading since the register names
are not VM-specific. Rename VM_REG(_GUEST) to CPU_REG in both the hypervisor and
device model.
v1 -> v2:
* Introduced.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The vm_reg_name is a good example of a collection of discrete values. This patch
replaces signed integers with this type whenever applicable to avoid dependence
on the underlying value of such enumeration constants.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Op_sizes are mostly from vie->opsize which is a 4-bit field in struct vie. Use
uint8_t instead of int to represent them.
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
- rename 'cpu_set_logical_id()' to 'set_current_cpu_id()'
- rename 'cpu_find_logical_id()' to 'get_cpu_id_from_lapic_id()'
- some clean up in cpu.c & trampolines.s
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
An immediate reset or power down will cause a loss of write content.
The cause is the data write to disk is at cache within a short
time window before it's synced to storage media.
An explicit fsync() forces to sync the data to storage to prevent
the data loss of such immediate reset.
Signed-off-by: Huang Yang <yang.huang@intel.com>
Signed-off-by: duminx <minx.du@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
In the hypervisor, some strings are assigned to non const
object, this violates MISRA C:2012.
Update the type of the object as const type since it always
points to string.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
Once the specific interrupt is marked waiting for inject to
target vcpu, we don't need to mark it again if the same
interrupt is request to inject to same target vcpu.
One example is UP SOS + SMP UOS. It's possible that different
core of UOS try to notify SOS vcpu that there is ioreq pending.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
We need bitmap bit ops API for 32bit operators like vlapic irr
registers.
Signed-off-by: Yin Fengwei <fengwei.yin@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Some ACRN kernel components are using the API documentation methods of
the Linux kernel. While they use Sphinx for generating their
documentation, they don't use doxygen to collect the API information as
we do for the rest of the project. Instead, they use their own tools
called "kerneldoc". This PR incorporates those tools into our
documentation build process.
There is a prescribed directory structure for this to work: that the
acrn-hypervisor and acrn-kernel repos are cloned to sibling folders,
e.g.:
projectacrn
acrn-hypervisor
acrn-kernel
so that documentation references from acrn_hypervisor/doc can access the
source code in ../../acrn-kernel to do the kerneldoc processing. A full
display of the kerneldoc API material for a source file in the
acrn-kernel tree can be done using a sphinx extension directive:
.. kernel-doc:: /tools/virtio/linux/scatterlist.h
where the assumed root of these file references is ../../acrn-kernel.
The format for kerneldoc comments is documented in
https://www.kernel.org/doc/html/latest/doc-guide/kernel-doc.html and
references to kerneldoc API material in .rst files is documented in
https://www.kernel.org/doc/html/latest/doc-guide/kernel-doc.html#including-kernel-doc-comments
Without options, the kernel-doc directive includes all documentation
comments from the source file. With options, you can display subsets of
these comments.
The intention is to limit use of kerneldoc comments to the acrn-kernel
repo and not use them elsewhere within the ACRN project (where doxygen
comments are expected.)
While I'd prefer NOT to include the kerneldoc perl script here (it is
already in the acrn-kernel/sphinx folder), I don't want to create a
dependency on the acrn-kernel folder existing for documentation
generation, but this might be unavoidable once we have part of the API
material coming from there. We can update this in a later PR.
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
-Add acrntrace_format.py usage to README.rst
-Add acrntrace usage to README.rst
-Add acrnanalyze.py to README.rst
Signed-off-by: Kaige Fu <kaige.fu@intel.com>
Reviewed-by: Yan, Like <like.yan@intel.com>
Reviewed-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Acked-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
There are extra type conversion in the HV since logical_id
type is uint32_t and the input argument of bitmap operations
is uint16_t. BTW, the name of logical_id is not clear enough
to express its usage.
So the following updates are made in this patch:
Update logical_id type as unit_16 to reduce type casting;
Update related print argument;
Rename related logical_id as pcpu_id as needed.
Note: logical_id in the interrupt_init definition have been
updated in cpu_id cleanup patch.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
There are extra type conversion in the HV since cpu_id
type is uint32_t and the return value type of get_cpu_id
is uint16_t. BTW, the name of cpu_id is not clear enough
to express its usage.
So the following updates are made in this patch:
Update cpu_id type as unit_16 to reduce type casting;
Update related temporary variables type;
Update related print argument;
Change the input parameter name of interrupt_init as
cpu_id to keep align with function implement;
Rename cpu_id as pcpu_id as needed.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
In the hypervisor, virtual cpu id is defined as "int" or "uint32_t"
type in the hypervisor. So there are some sign conversion issues
about virtual cpu id (vcpu_id) reported by static analysis tool.
Sign conversion violates the rules of MISRA C:2012.
BTW, virtual cpu id has different names (vcpu_id, cpu_id, logical_id)
for different modules of HV, its type is defined as "int" or "uint32_t"
in the HV. cpu_id type and logical_id type clean up will be done in
other patchs.
V1-->V2:
More clean up the type of vcpu id;
"%hu" is for vcpu id in the print function.
Signed-off-by: Xiangyang Wu <xiangyang.wu@intel.com>
