Use local variables for the captured timehand and timecounter in pps_event().
This fixes a potential issue in the nsec preparation for hardpps(). Here the
timecounter was accessed through the captured timehand after the generation was
checked.
Make a snapshot of the relevent timehand values early in pps_event(). Check
the timehand generation only once during the capture and event processing. Use
atomic_thread_fence_acq() similar to the other readers.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/604
Since 32c203577a5e by phk in 1999 (Make even more of the PPSAPI
implementations generic), the "nsec" parameter of hardpps() is a time
difference and no longer a time point. Change the name to "delta_nsec"
and adjust the comment.
Remove comment about a clock tick adjustment which is no longer in the code.
Pull Request: https://github.com/freebsd/freebsd-src/pull/640
Reviewed by: imp
The third argument to this function indicates whether the supplied
ticker is fixed or variable, i.e. requiring calibration. Give this
argument a type and name that better conveys this purpose.
Reviewed by: kib, markj
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35459
(which is used to calculate cputime from cpu ticks) has some imprecision and,
worse, huge timestep (about 20 minutes on 4GHz CPU) near 53.4 days of elapsed
time.
kern_time.c/cputick2timespec() (it is used for clock_gettime() for
querying process or thread consumed cpu time) Uses cputick2usec()
and then needlessly converting usec to nsec, obviously losing
precision even with fixed cputick2usec().
kern_time.c/kern_clock_getres() uses some weird (anyway wrong)
formula for getting cputick resolution.
PR: 262215
Reviewed by: gnn
Differential Revision: https://reviews.freebsd.org/D34558
Interrupts must be disabled during MMU remapping since the majority of
RTEMS including interrupts expects normal memory mapping semantics such
as unaligned accesses.
The whole cache invalidation and flushing functions only ended up
flusing the first N-1 levels of cache due to an off by one error. This
resovles that issue and makes consistent the usage of levels as they
relate to caching.
To ensure data consistency, the cache much be flushed before disabling
the MMU. When the MMU is disabled, all accesses are treated as
non-cachced and thus will bypass the cache.
The application configuration options are documented in
"cpukit/doxygen/appl-config.h". Since the application configuration
option defines are also present in multiple test program sources, the
"#OPTION" references cannot be mapped to a unique definition. Add an
anchor for each option and reference it to avoid the issues with the
multiple definitions.
Update #3994.
Rename the constaints section of application configuration options from
"Value Constraints" in "Constraints. Adjust the constraint wording
accordingly. This is in line with the RTEMS Classic API Guide.
Update #3994.
Each BSP may optionally provide default values for some application
configuration options. Remove the documentation of these items, since
the BSP provided defines are not application configuration options, they
are optional default values. Clarify CONFIGURE_DISABLE_BSP_SETTINGS
accordingly and move it into the "General System Configuration" group.
Update #3994.
The LEON3_Boot_Cpu global object is only used by start.S. Move the definition
of this object to start.S and use a local symbol .Lbootcpuindex for it.
Use a compare-and-swap instruction to assign the boot CPU. This allows a
concurrent initialization.
Close#4845.
This imports the full xil_exception.h instead of an empty stub. This is
required for some Xilinx drivers. The imported files adhere to the
current VERSION file.
This makes xil/sleep.h a stub to prevent multiple differing definitions
of sleep functions from toolchain and local headers. The non-standard
sleep definitions were not in use and can be added later if needed.
There are different cases for _ISR_lock_ISR_disable() and
_ISR_lock_ISR_enable() in the case RTEMS_SMP is defined or
RTEMS_PROFILING is defined, so remove the related code.
../../../cpukit/include/rtems/score/threaddispatch.h: In function '_Thread_Dispatch_disable':
../../../cpukit/include/rtems/score/threaddispatch.h:231:14: warning: 'lock_context' may be used uninitialized [-Wmaybe-uninitialized]
231 | cpu_self = _Thread_Dispatch_disable_critical( &lock_context );
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../../cpukit/include/rtems/score/threaddispatch.h:210:32: note: by argument 1 of type 'const ISR_lock_Context *' to '_Thread_Dispatch_disable_critical' declared here
210 | static inline Per_CPU_Control *_Thread_Dispatch_disable_critical(
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../../cpukit/include/rtems/score/threaddispatch.h:225:21: note: 'lock_context' declared here
225 | ISR_lock_Context lock_context;
| ^~~~~~~~~~~~
Update #4662.
The objxilinxsupport build object was accidentally included twice in
some of the ZynqMP BSPs by two different drivers that required it. This
commit manually deduplicates the inclusions by moving that inclusion to
the BSP. Duplication of object inclusions is considered a bug and can
cause race conditions in the build system.
The QSPI memory is initialized and used only when the BSP configure file
sets QSPI memory size to non-zero value. Currently QSPI is run in memory
mapped mode which allows future RTEMS binary linkage and upload into QSPI
memory.
Sponsored-By: Precidata
