Add a POSIX Option Group called XSI_REALTIME (with Kconfig
option CONFIG_XSI_REALTIME).
When XSI_REALTIME is selected (or when required POSIX Options
are enabled), define _XOPEN_REALTIME to be something other
than -1 (_XOPEN_VERSION seemed appropriate).
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Since XSI is composed of several distinct POSIX Option Groups
split Kconfig.xsi into separate files - one for each Option
Group.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
For the REALTIME_MINIMAL Application Environment Profile, uncomment
the POSIX_MEMLOCK, POSIX_MEMLOCK_RANGE, POSIX_MEMORY_PROTECTION,
POSIX_MAPPED_FILES, and POSIX_SHARED_MEMORY_OBJECTS options.
These should have been uncommented back when Kconfig options for
those features were added, so this can probably be called a
Kconfig bug.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Although it's quite unlikely that we will see shared memory
objects in practice that are greater than 2GiB, there is a
possibility of integer overflow when comparing intptr_t and
size_t, since the former is signed and the latter is unsigned.
Explicitly check to ensure that the addend is less than
INTPTR_MAX before subtraction.
This fixes CID 487734
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
The nsem_list_unlock() static inline function looks as though it
was designed to always ignore the return value of
k_mutex_unlock(), presumably for performance reasons.
A quick audit of the code looks as though the call should
always succeed.
Prepend (void) to the call to avoid the coverity defect in the
future.
CID 444386
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
It was possible that allocation was continuously attempting to find
space. It could happen if allocation interrupted consuming a packet
(which is marked as busy) and there is not enough space to allocate
requested packet but there would be if busy packet was freed.
Algorithm in that case was continuously going through the buffer
in search for packets that can be dropped to find space for the
new packet.
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
Utilize a code spell-checking tool to scan for and correct spelling errors
in various files within the `lib` directory.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
GCC versions before 4.5 do not have the -fno-lto flag and give an error if
it is used. The flag may also be unavailable on various other non-GCC
compilers.
Previously, the picolibc CMakeList.txt script assumed that the flag was
available, and hard-coded it as an addition to the compiler flags.
This patch improves compatibility by making use of the already existing
"prohibit_lto" CMake target compiler property, which is set to the
appropriate flag to disable LTO support, or unset if the compiler lacks
such an option.
Signed-off-by: Joel Holdsworth <jholdsworth@nvidia.com>
When using the SMF for a project discovered that events would sometimes
not propagate to parent states correctly. Could not create a minimum
reproducable test case for this, but it was found that these changes fixed
the bug. This commit creates a new function to reset internal state,
which is called on entry to smf_set_initial() and smf_set_state().
Closes#81300.
Signed-off-by: Geoffrey Hunter <gbmhunter@gmail.com>
To request heap statistics, a pointer to a heap structure is required.
This is straightforward for a user-defined heap. However, such a pointer
is not known for heaps created by other components or libraries, like
libc. Therefore, it is not possible to calculate the total heap memory.
The proposed solution is to use an array of pointers, which is filled in
on every sys_heap_init() call. One can then iterate through it to sum up
the total memory allocated for all heaps.
The array size is configurable. The default array size is zero,
which means the feature is disabled. Any other integer greater then zero
defines the array size and enables the feature.
A list of pointers instead of an array could be another approach,
but it requeres a heap, which is not always available.
Signed-off-by: Ivan Pankratov <ivan.pankratov@silabs.com>
Reverts the change made in d4d53010f00cadbb4f89c6d41391937646fc1740
(The changes was moved to another file in a restructuring)
The commit incorrectly assumed that no blocking would be allowed in
the syswq. This has caused issues observed among others
in #77241 and #80167
Signed-off-by: Håvard Reierstad <haavard.reierstad@nordicsemi.no>
Originally, when the POSIX_SHARED_MEMORY_OBJECTS option was
added, the O_TRUNC flag was not consistently available.
Now that it is consistently available, ensure it is used
within the shm_open() function.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Boot time initialization functions and data used there must be
available at boot. With demand paging, these may not exist in
memory when they are being used, resulting in page faults.
So pin these functions and data in linker sections to make
sure they are in memory at boot time.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Mostly a revert of commit b1def7145f ("arch: deprecate `_current`").
This commit was part of PR #80716 whose initial purpose was about providing
an architecture specific optimization for _current. The actual deprecation
was sneaked in later on without proper discussion.
The Zephyr core always used _current before and that was fine. It is quite
prevalent as well and the alternative is proving rather verbose.
Furthermore, as a concept, the "current thread" is not something that is
necessarily architecture specific. Therefore the primary abstraction
should not carry the arch_ prefix.
Hence this revert.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The Xtensa linker scripts seem to be injecting syntax errors when
MMU is enabled. Disable the implication in Kconfig.mem for Xtensa
until linker issues are resolved.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Declare _POSIX_C_SOURCE in a consistent way for both the
posix/options library as well as the tests/posix/fs
testsuite.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Check if the spinlock is held before attempting to wait by
taking the semaphore, as that would cause a context switch which
isn't allowed and will trigger an assertion error when
`CONFIG_SPIN_VALIDATE` is enabled.
Logging in spinlock-held context when the log buffer is full can lead
to an infinite assertion error loop, as the logging subsys attempts to
allocate buffer when there's none available, it will try to wait for
one and thus triggers the assertion error, the error message will be
printed through the logging sybsys but there's no buffer available,
so it will try to wait for one and triggers another assertion error..
This loop just goes on and on forever, and nothing gets printed to
the terminal.
Added a test to validate the fix.
Signed-off-by: Yong Cong Sin <ycsin@meta.com>
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
Signed-off-by: Maxim Adelman <imax@meta.com>
Previously, if a file object is was re-used, it could
inherit the offset field of the previously closed file object,
making reading from the beginning of the file impossible
until the offset was manually zero'ed.
The offset should *always* be zero when a file is ready to be
used.
The issue really only presents itself when implementing a
vtable backend.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
The `zvfs_finalize_typed_fd()` function notifies some backends
via `ioctl()` with `ZFD_IOCTL_SET_LOCK`. However, support for
this method and functionality is optional.
In backends that do not support locking, this benign failure can
set `errno` to 95 (`EOPNOTSUPP`) in many circumstances where a
change in `errno` (indicating some kind of failure) is not
appropriate.
Prevent errno poisoning by backing-up and restoring `errno`.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Previously, in libc-hooks.c, the signature of `_open()` was as
shown below.
```cpp
int _open(const char *name, int mode);
```
This conflicted with the signature of `_open()` from newlib,
which is
```cpp
int _open(const char *name, int flags, ...);
```
Moreover, the mode and flags field were reversed, but only for
the Xtensa architecture due to the `_open_r()` hook that is
present in `libc-hooks.c`.
This manifested itself via a call to `fopen(file, "w+")`, where
the expected flags should include `O_CREAT | O_TRUNC`, or
`0x200 | 0x400`. Instead, the unexpected flags passed to the
underlying `zvfs_open()` call were `0x1b6`.
This change corrects the function signature and order of the
arguments.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Do not set TC_PROVIDES_POSIX_C_LANG_SUPPORT_R as the default in
lib/posix/options/Kconfig.c_lang_r . TC_PROVIDES options are
only intended to be set by C libraries that implement parts of
the POSIX standard.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Oddly, even though CI passed when the file_system_r change was
merged, now CI has encountered a build error because MIN() was
not defined.
Include `<zephyr/sys/util.h>` to pull in the definition.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
A copy-paste error propogated this typo to a few different
Kconfig files.
Correct 'informnation' to 'information'.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Move the functionality of POSIX_FILE_SYSTEM_R to its own
compilation unit and remove the unnecessary dependency on
POSIX_FILE_SYSTEM.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
The compiler requires that much of the C library be built without using LTO
so that various symbols are available for use by generated code, including
things like memset and memcpy.
Add -fno-lto when building both picolibc itself as well as the Zephyr
interface bits.
Closes: #81674
Signed-off-by: Keith Packard <keithp@keithp.com>
Many out of the POSIX subsystem configs are enabled automatically
when merely CONFIG_POSIX_API is enabled, which is a prerequisite for
many networking samples. This causes a massive experimental warning
printout when building with warnings enabled.
Since the new POSIX Kconfig configuration options are already present
in Zephyr for 2 release cycles and seem settled, I suggest we move
them out of experimental phase.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
These functions can be used to join pthreads in a non-standard way.
The function pthread_tryjoin will not block and simply test whether the
thread has exited already. The function pthread_timed_join will only block
until the specified time. The functions are wrappers for calling the
k_thread_join with timeout K_NO_WAIT and with a specific timeout as opposed
to calling it with K_FOREVER.
Signed-off-by: Cla Galliard <clamattia@gmail.com>
Select TC_PROVIDES_POSIX_C_LANG_SUPPORT_R to keep Zephyr from
including the common libc implementation of the various _r APIs.
Signed-off-by: Fin Maaß <f.maass@vogl-electronic.com>
When the PRESTART thread state was removed, this changed the semantics
of k_thread_start() when thread was created with a K_FOREVER timeout,
suspended and then started with k_thread_start().
This sequence is used in p4wq to implement K_P4WQ_DELAYED_START
(which again is needed by K_P4WQ_USER_CPU_MASK).
With PRESTART removed, the following sequence:
z_mark_thread_as_not_suspended(thread);
k_thread_start(thread);
.. no longer starts the thread. As a result, p4wq users like SOF
multicore configurations, hit errors as p4wq threads never start.
Fix the implementation by removing the calls to change thread
suspended state explicitly, but rather rely on the new
k_thread_create() and k_thread_start() semantics.
Fixes: 7cdf40541b ("kernel/sched: Eliminate PRESTART thread state")
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Increase the number of locks to match what the ARC MWDT libc requires
now. The library wants to have 2 locks per each available FILE entry,
and then some more. Also do not include an internal libc header, as
all that is needed from that header is a simple typedef.
Signed-off-by: Ilya Tagunov <Ilya.Tagunov@synopsys.com>
If CONFIG_POSIX_API is enabled, but CONFIG_POSIX_DEVICE_IO_ALIAS_WRITE
is not, the _write replacement should be provided to make stdio work.
Signed-off-by: Ilya Tagunov <Ilya.Tagunov@synopsys.com>
As k_sleep returns int32_t, there is a possibility for integer overflow
during conversion from milliseconds to nanoseconds.
Signed-off-by: Ilya Tagunov <Ilya.Tagunov@synopsys.com>
Nanosecond time calculation overflows if the libc has 32-bit time_t.
One such libc is the classic ARC MWDT one, but there might be others.
Signed-off-by: Ilya Tagunov <Ilya.Tagunov@synopsys.com>
`_current` is now functionally equals to `arch_curr_thread()`, remove
its usage in-tree and deprecate it instead of removing it outright,
as it has been with us since forever.
Signed-off-by: Yong Cong Sin <ycsin@meta.com>
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
