Just like with _Swap(), we need two variants of these utilities which
can atomically release a lock and context switch. The naming shifts
(for byte count reasons) to _reschedule/_pend_curr, and both have an
_irqlock variant which takes the traditional locking.
Just refactoring. No logic changes.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Implements osThreadJoin and osThreadDetach.
This implementation uses a semaphore to signal when a thread is
exiting so any join operations are signalled to continue. It supports
multiple join operations on a single thread, and ensures joins are
aborted if a thread is detached.
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
This was never a long-term solution, more of a gross hack
to get test cases working until we could figure out a good
end-to-end solution for memory domains that generated
appropriate linker sections. Now that we have this with
the app shared memory feature, and have converted all tests
to remove it, delete this feature.
To date all userspace APIs have been tagged as 'experimental'
which sidesteps deprecation policies.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
* Newlib now defines a special z_newlib_partition containing
all globals relevant to newlib. Most of these are in libc.a
with a heap tracking variable in newlib's hooks.
* Both C libraries now expose a k_mem_partition containing the
bounds of the malloc heap arena. Threads that want to use
libc malloc() will need to add this to their memory domain.
* z_newlib_get_heap_bounds has been removed, in favor of the
memory partition for the heap arena
* ztest now includes the C library partitions in its memory
domain.
* The mem_alloc test now runs in user mode to prove that this
all works for both C libraries.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
If an unitialized/zeroed optional attribute was passed to osThreadNew
the priority would be osThreadNone i.e. uninitialized. This causes an
ASSERT to be hit as the priority isn't valid (it is not between
osPriorityIdle and osPriorityISR).
The fix checks the passed in priority is not osPriorityNone and assigns
osPriorityNormal. This is the correct CMSIS behaviour.
The ASSERT will still be hit if the priority is invalid (<0).
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
Fixed an issue whereby if an attribute structure was passed into a CMSIS
RTOS v2 'new' function with an invalid address i.e. NULL assigned to the
name (char*) member the memcpy at the end of each new function
would cause a segmentation fault i.e. read from an invalid
address.
This has been fixed by checking if the name is NULL and using the
default name from the init struct if it is. This is the same name
that would be used if not passing in the optional attr function
argument.
Changed the memcpy to strncpy to ensure that the copy does not read
beyond the end of the source string and changed the length from 16 to 15
(by means of a `sizeof(...)-1`) of the destination buffer to ensure that
it will always be nul-terminated.
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
Implemented dynamic thread stacks for CMSIS threads by declaring an
array of default sized thread stacks. Allocation cannot be done on the
heap as some architectures require strict alignment for stacks so the
macro must be used to define the stack to ensure most compatibility.
Added a Kconfig variable to limit the number of dynamic threads on the
system (they also count towards total CMSIS thread count). This is so a
developer can have fine grained control over how many dynamic threads
can be allocated because all their stacks must be allocated up front so
could use a lot of memory needlessly if oversubscribed. The default
value is 0 which effectively disabled dynamic threads but also reduces
the memory impact to almost none.
Fixed an assert bug where thread_num was being tested against the
maximum allowed CMSIS threads - it previous checked for less than or
equal which actually (due to when the increment happens) allowed there
to be one more thread. The check now correctly uses less than and only
allowed up to the defined maximum.
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
Implemented dynamic allocation of memory pools in a similar to manner to
what was already implemented for message queues. Added all the same
checks on size vs. maximum allowed and current heap.
Added an additional Kconfig variable to define the maximum size of a
dynamically allocated memory pool.
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
Added some additional checks when creating a message queue to ensure the
size of the queue does not exceed the size of the buffer passed in via
the optional attributes.
Added a new Kconfig option to limit the maximum size of a message queue
dynamically allocated on the heap.
Added a check to ensure the heap is at least large enough to hold a
maximum size dynamically allocated queue.
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
Added Kconfig dependency that NUM_PREEMPT_PRIORITIES must be at least
osPriorityISR (56). This was enforced by a build assert message but not
decribed in the Kconfig.
Signed-off-by: Carlos Stuart <carlosstuart1970@gmail.com>
After #12732, 6904501173
asserts call k_panic.
Before this, the POSIX arch had its own hack in the
__ASSERT_POST implementation to terminate the process instead
of spining forever.
But the POSIX arch does implement k_panic properly, so there
is no need anymore for this hack.
=> Remove the special treatment for POSIX ARCH
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Replaced forever loop in assert with call to a function.
In post_assert_action() function, k_panic is called.
Forever loop was preventing logs to be printed and had behavior
ependent on the context (low prioriy thread - system continue to
ork, irq - system is blocked).
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
lib/ was starting to get messy and inconsitent. Files being either
dumped in the root or in sub-directories without a clear plan.
Move all library components into one single folder and call it 'os'.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
mq_maxmsg and mq_msgsize are defined to be of
type long in POSIX standard. So use long for
variables that hold its value in mq_open().
Signed-off-by: Niranjhana N <niranjhana.n@intel.com>
This patch adds a x86_64 architecture and qemu_x86_64 board to Zephyr.
Only the basic architecture support needed to run 64 bit code is
added; no drivers are added, though a low-level console exists and is
wired to printk().
The support is built on top of a "X86 underkernel" layer, which can be
built in isolation as a unit test on a Linux host.
Limitations:
+ Right now the SDK lacks an x86_64 toolchain. The build will fall
back to a host toolchain if it finds no cross compiler defined,
which is tested to work on gcc 8.2.1 right now.
+ No x87/SSE/AVX usage is allowed. This is a stronger limitation than
other architectures where the instructions work from one thread even
if the context switch code doesn't support it. We are passing
-no-sse to prevent gcc from automatically generating SSE
instructions for non-floating-point purposes, which has the side
effect of changing the ABI. Future work to handle the FPU registers
will need to be combined with an "application" ABI distinct from the
kernel one (or just to require USERSPACE).
+ Paging is enabled (it has to be in long mode), but is a 1:1 mapping
of all memory. No MMU/USERSPACE support yet.
+ We are building with -mno-red-zone for stack size reasons, but this
is a valuable optimization. Enabling it requires automatic stack
switching, which requires a TSS, which means it has to happen after
MMU support.
+ The OS runs in 64 bit mode, but for compatibility reasons is
compiled to the 32 bit "X32" ABI. So while the full 64 bit
registers and instruction set are available, C pointers are 32 bits
long and Zephyr is constrained to run in the bottom 4G of memory.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Added glue logic to interface Zephyr with LittlevGL GUI library
This includes:
* KConfig options for all lvgl options
* Kernel & user space memory management
* Zephyr to lvgl FS call mapping
* Color space conversion function
Signed-off-by: Jan Van Winkel <jan.van_winkel@dxplore.eu>
Following the standard GCC RISC-V convetion use __riscv for the RISC-V
specific define:
41d6b10e96/gcc/config/riscv/riscv-c.c (L37)
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Used as a checksum on command messages when talking with MMC cards.
Implemented using the unwound bytewise implementation from
https://en.wikipedia.org/wiki/Computation_of_cyclic_redundancy_checks
which is a good mix of size and speed.
The API and naming matches lib/crc7.c in Linux.
Signed-off-by: Michael Hope <mlhx@google.com>
If any of the Zephyr version numbers went beyond 99, the "%2d" printf
specifiers would expand to fit and the string would run over the
memory on the stack used for os_str[].
Recent GCC versions (remember native_posix and x86_64 use the host
compiler) were actually detecting this and correctly issuing a warning
(but only if the 3-digit char value would overflow the actual array
size!), which was breaking sanitycheck for me on Fedora 28 and Ubuntu
18.04 build hosts. Pretty impresive warning.
As it happens this was wasteful anyway; we were spending bytes on the
stack (and in rodata to store the constant which, and the cycles
needed to copy it into place on the stack where it would be
overwritten immediately) when we could just snprintf() directly into
the buffer the user gave us.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Thread Flags are used to trigger execution states between threads.
These APIs provide functionalities like set, clear and wait.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Events are used to trigger execution states between threads.
These APIs provide functionalities like event set, clear and
wait.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
These APIs allow creating, allocating and freeing of mempools.
Note: "Mempool" in CMSIS actually means memslabs in Zephyr.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
These APIs provide the support of virtual timers. All timers
can be started, restarted, or stopped. Timers can be configured
as one-shot or periodic.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
APIs to introduce wait i.e osDelay and osDelayUntil are defined
here. They are analogous to k_sleep in the kernel.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Implement support for Kernel management APIs like
osKernelInitialize, osKernelGetTickCount, osKernelGetSysTimerCount
etc.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
According with MISRA-C the value returned by a non-void function has
to be used. As memcpy return is almost useless, we are explicitly
ignoring it.
MISRA-C rule 17.7
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
ioctl() just dispatches to the corresponding vmethod of an fd.
fcntl() handles fdtable-level operations (so far doesn't handle
actually, returning "not implemented" error), and forwards
fd-specific operations to ioctl vmethod just the same (i.e.
ioctl and fcntl operations share the same namespace, but otherwise
disjoint).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
As extend fdtable usage to more cases, there regularly arises a need
to forward ioctl/fcntl arguments to another ioctl vmethod, which is
complicated because it defined as taking variadic arguments. The only
portable solution is to convert variadic arguments to va_list at the
first point of entry from client code, and then pass va_list around.
To facilitate calling ioctl with variadic arguments from system code,
z_fdtable_call_ioctl() helper function is added.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
The function atomic_set return the previous value of the
target. Sometimes this value is irrelevant, e.g when initializing a
variable.
As MISRA-C rule 17.7 requires that the value returned by a non-void
function must be used, we have to explicitly ignore some cases.
MISRA-C rule 17.7
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This allows for workqueues to be started in user mode.
No additional kernel objects or system calls are defined
other than starting the workqueue in user mode; for
permission purposes the embedded queue and thread objects
are sufficient.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Similar to the last patch, there was a spot in block recombination
where the lock would be released while the combined block was being
held allocated. That means that when recombining a single top-level
block, it was possible for the entire heap to look allocated.
Make the combination and re-addition of the larger block atomic.
Requires a little surgery to the structure of the code, so this is a
little more involved than the earlier fix.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Andy Ross