summaryrefslogtreecommitdiffstats
path: root/cpukit/score/src (follow)
Commit message (Collapse)AuthorAgeFilesLines
...
* score: Allocate per-CPU data only if necessarySebastian Huber2018-09-191-13/+17
| | | | | | | | | The _Workspace_Allocate_aligned() would returns a non-NULL pointer for a zero size allocation request if there is enough memory available. This conflicts with the size estimate of zero in _Workspace_Space_for_per_CPU_data() if the per-CPU data set is empty. Update #3507.
* score: Add thread pin/unpin supportSebastian Huber2018-09-105-27/+266
| | | | | | | | | Add support to temporarily pin a thread to its current processor. This may be used to access per-processor data structures in critical sections with enabled thread dispatching, e.g. a pinned thread is allowed to block. Update #3508.
* score: Add flexible per-CPU dataSebastian Huber2018-09-101-21/+107
| | | | Update #3507.
* score: Fix EDF SMP schedulerSebastian Huber2018-09-031-1/+24
| | | | | Fix a special case: block a one-to-one scheduled thread while having a non-empty affine ready queue on the same processor.
* score: Generalize SMP scheduler block supportSebastian Huber2018-09-035-1/+6
| | | | | | Add extract from scheduled function to the _Scheduler_SMP_Block() operation. This allows a scheduler implementation to do extra work in case a scheduled node is blocked.
* Update FreeBSD kernel timespec supportSebastian Huber2018-08-231-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This change is based on the following FreeBSD commit: "Make timespecadd(3) and friends public The timespecadd(3) family of macros were imported from NetBSD back in r35029. However, they were initially guarded by #ifdef _KERNEL. In the meantime, we have grown at least 28 syscalls that use timespecs in some way, leading many programs both inside and outside of the base system to redefine those macros. It's better just to make the definitions public. Our kernel currently defines two-argument versions of timespecadd and timespecsub. NetBSD, OpenBSD, and FreeDesktop.org's libbsd, however, define three-argument versions. Solaris also defines a three-argument version, but only in its kernel. This revision changes our definition to match the common three-argument version. Bump _FreeBSD_version due to the breaking KPI change. Discussed with: cem, jilles, ian, bde Differential Revision: https://reviews.freebsd.org/D14725" To make the change public (outside #ifdef _KERNEL) it must be integrated in Newlib. Update #3472.
* score: Do not inline _Thread_Dispatch_enable()Sebastian Huber2018-08-231-0/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This function is slighly too complex for inlining with two if statements. The caller already needs a stack frame due to the potential call to _Thread_Do_dispatch(). In _Thread_Dispatch_enable() the call to _Thread_Do_dispatch() can be optimized to a tail call. A text size comparision (text size after patch - text size before patch) / text size before patch on sparc/erc32 with SMP enabled showed these results: Minimum -0.000697892 (fsdosfsname01.exe) Median -0.00745021 (psxtimes01.exe) Maximum -0.0233032 (spscheduler01.exe) A text size comparision text size after patch - text size before patch on sparc/erc32 with SMP enabled showed these results: Minimum -3312 (ada_sp09.exe) Median -1024 (tm15.exe) Maximum -592 (spglobalcon01.exe)
* score: RTEMS_PREDICT_TRUE(), RTEMS_PREDICT_FALSE()Sebastian Huber2018-07-254-15/+17
| | | | | | | Add RTEMS_PREDICT_TRUE() and RTEMS_PREDICT_FALSE() for static branch prediction hints. Close #3475.
* _SMP_Start_multitasking_on_secondary_processor()Sebastian Huber2018-07-251-9/+17
| | | | | Pass current processor control as first parameter to make dependency more explicit.
* score: Add a FALLTHROUGH comment to kvprintf()Sebastian Huber2018-07-181-1/+4
| | | | | | This change was added to the FreeBSD version of this function. This was spotted by Coverity Scan.
* score: Fix _Scheduler_EDF_SMP_Set_affinity()Sebastian Huber2018-07-181-7/+5
| | | | | | Commit 8744498752ad4f0eaf9fb5640c6a0e0f2dc92fda broke the _Scheduler_EDF_SMP_Set_affinity() implementation. We must test the overall affinity against the online processors.
* score: Add postorder tree iteration supportSebastian Huber2018-07-161-0/+81
| | | | Update #3465.
* riscv: Fix global constructionSebastian Huber2018-06-291-1/+1
| | | | Update #3433.
* Rework initialization and interrupt stack supportSebastian Huber2018-06-272-47/+40
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Statically initialize the interrupt stack area (_Configuration_Interrupt_stack_area_begin, _Configuration_Interrupt_stack_area_end, and _Configuration_Interrupt_stack_size) via <rtems/confdefs.h>. Place the interrupt stack area in a special section ".rtemsstack.interrupt". Let BSPs define the optimal placement of this section in their linker command files (e.g. in a fast on-chip memory). This change makes makes the CPU_HAS_SOFTWARE_INTERRUPT_STACK and CPU_HAS_HARDWARE_INTERRUPT_STACK CPU port defines superfluous, since the low level initialization code has all information available via global symbols. This change makes the CPU_ALLOCATE_INTERRUPT_STACK CPU port define superfluous, since the interrupt stacks are allocated by confdefs.h for all architectures. There is no need for BSP-specific linker command file magic (except the section placement), see previous ARM linker command file as a bad example. Remove _CPU_Install_interrupt_stack(). Initialize the hardware interrupt stack in _CPU_Initialize() if necessary (e.g. m68k_install_interrupt_stack()). The optional _CPU_Interrupt_stack_setup() is still useful to customize the registration of the interrupt stack area in the per-CPU information. The initialization stack can reuse the interrupt stack, since * interrupts are disabled during the sequential system initialization, and * the boot_card() function does not return. This stack resuse saves memory. Changes per architecture: arm: * Mostly replace the linker symbol based configuration of stacks with the standard <rtems/confdefs.h> configuration via CONFIGURE_INTERRUPT_STACK_SIZE. The size of the FIQ, ABT and UND mode stack is still defined via linker symbols. These modes are rarely used in applications and the default values provided by the BSP should be sufficient in most cases. * Remove the bsp_processor_count linker symbol hack used for the SMP support. This is possible since the interrupt stack area is now allocated by the linker and not allocated from the heap. This makes some configure.ac stuff obsolete. Remove the now superfluous BSP variants altcycv_devkit_smp and realview_pbx_a9_qemu_smp. bfin: * Remove unused magic linker command file allocation of initialization stack. Maybe a previous linker command file copy and paste problem? In the start.S the initialization stack is set to a hard coded value. lm32, m32c, mips, nios2, riscv, sh, v850: * Remove magic linker command file allocation of initialization stack. Reuse interrupt stack for initialization stack. m68k: * Remove magic linker command file allocation of initialization stack. Reuse interrupt stack for initialization stack. powerpc: * Remove magic linker command file allocation of initialization stack. Reuse interrupt stack for initialization stack. * Used dedicated memory region (REGION_RTEMSSTACK) for the interrupt stack on BSPs using the shared linkcmds.base (replacement for REGION_RWEXTRA). sparc: * Remove the hard coded initialization stack. Use the interrupt stack for the initialization stack on the boot processor. This saves 16KiB of RAM. Update #3459.
* score: Simplify _Objects_Name_to_string()Sebastian Huber2018-06-051-2/+12
| | | | | Do not use isprint() from <ctype.h> since it depends on the heavy weight C locale implementation in Newlib.
* mpci: Fix _MPCI_Enqueue_callout()Sebastian Huber2018-03-221-1/+2
| | | | | Update #3117. Update #3182.
* mpci: Update due to thread queue API changesSebastian Huber2018-03-221-6/+7
| | | | | Update #3117. Update #3182.
* Use _Thread_Dispatch_direct()Sebastian Huber2018-02-081-2/+2
| | | | | | | Use _Thread_Dispatch_direct() for operations that block the executing thread. This ensures that we get a fatal error (INTERNAL_ERROR_BAD_THREAD_DISPATCH_DISABLE_LEVEL) if we try to block in an invalid context, e.g. during system start or an interrupt handler.
* Add RTEMS thread APISebastian Huber2018-02-021-3/+96
| | | | Update #2843.
* score: Introduce new monotonic clockSebastian Huber2018-02-022-78/+66
| | | | | | | | Rename PER_CPU_WATCHDOG_MONOTONIC to PER_CPU_WATCHDOG_TICKS. Add new PER_CPU_WATCHDOG_MONOTONIC which is based on the system uptime (measured by timecounter). Close #3264.
* score: Rename _Watchdog_Realtime_from_*()Sebastian Huber2018-02-023-4/+4
| | | | | | Rename _Watchdog_Realtime_from_*() to _Watchdog_Ticks_from_*(). Update #3264.
* score: Optimize watchdog tickleSebastian Huber2018-02-022-34/+65
| | | | | | | Avoid unnecessary lock acquire/release operations. Get realtime via timecounter only if necessary. Update #3264.
* score: Use self-contained API mutexSebastian Huber2017-12-0415-143/+89
| | | | | | | | | | Use a self-contained recursive mutex for API_Mutex_Control. The API mutexes are protected against asynchronous thread cancellation. Add dedicated mutexes for libatomic and TOD. Close #2629. Close #2630.
* score: Simplify global constructionSebastian Huber2017-11-222-63/+46
| | | | Update #3243.
* score: Optimize scheduler priority updatesSebastian Huber2017-11-2016-628/+276
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Thread priority changes may append or prepend the thread to its priority group on the scheduler ready queue. Previously, a separate priority value and a prepend-it flag in the scheduler node were used to propagate a priority change to the scheduler. Now, use an append-it bit in the priority control and reduce the plain priority value to 63 bits. This change leads to a significant code size reduction (about 25%) of the SMP schedulers. The negligible increase of the standard priority scheduler is due to some additional shift operations (SCHEDULER_PRIORITY_MAP() and SCHEDULER_PRIORITY_UNMAP()). Before: text filename 136 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleblock.o 464 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimplechangepriority.o 24 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimple.o 108 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleschedule.o 292 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleunblock.o 264 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleyield.o text filename 280 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityblock.o 488 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerprioritychangepriority.o 200 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriority.o 164 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityschedule.o 328 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityunblock.o 200 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityyield.o text filename 24112 arm-rtems5/c/imx7/cpukit/score/src/libscore_a-scheduleredfsmp.o text filename 37204 sparc-rtems5/c/gr740/cpukit/score/src/libscore_a-scheduleredfsmp.o text filename 42236 powerpc-rtems5/c/qoriq_e6500_32/cpukit/score/src/libscore_a-scheduleredfsmp.o After: text filename 136 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleblock.o 272 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimplechangepriority.o 24 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimple.o 108 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleschedule.o 292 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleunblock.o 264 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulersimpleyield.o text filename 280 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityblock.o 488 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerprioritychangepriority.o 208 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriority.o 164 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityschedule.o 332 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityunblock.o 200 sparc-rtems5/c/erc32/cpukit/score/src/libscore_a-schedulerpriorityyield.o text filename 18860 arm-rtems5/c/imx7/cpukit/score/src/libscore_a-scheduleredfsmp.o text filename 28520 sparc-rtems5/c/gr740/cpukit/score/src/libscore_a-scheduleredfsmp.o text filename 32664 powerpc-rtems5/c/qoriq_e6500_32/cpukit/score/src/libscore_a-scheduleredfsmp.o
* score: Adjust _Scheduler_EDF_SMP_Insert_ready()Sebastian Huber2017-11-171-6/+4
| | | | | Use only one parameter to compute the next generation. Use index 0 for LIFO ordering, and index 1 for FIFO ordering.
* score: Fix priority ceiling updatesSebastian Huber2017-11-162-1/+8
| | | | | | | | | | | | We must not clear the priority updates in _Thread_queue_Extract_locked() since this function is used by the priority ceiling surrender operations after the ceiling priority handover from the previous owner to the new owner. This is especially important in SMP configurations. Move the _Thread_queue_Context_clear_priority_updates() invocation to the callers. Close #3237.
* score: Fix _Thread_queue_Flush_critical()Sebastian Huber2017-11-161-4/+14
| | | | | | | | | The thread queue extract operations performed by the _Thread_queue_Flush_critical() may result in a priority change of the thread queue owner. Carry out the scheduler priority update operation. This is especially important in SMP configurations. Close #3236.
* score: Change _Timecounter_Time_uptime to int32_tSebastian Huber2017-11-091-1/+2
| | | | | | | Move basic timecounter API shared with BSD network stack to <machine/_timecounter.h>. Update #3185.
* score: Use Processor_mask instead of cpu_set_tSebastian Huber2017-11-061-11/+5
|
* score: _Chain_Insert_ordered_unprotected()Sebastian Huber2017-11-064-19/+48
| | | | | | Change the chain order relation to use a directly specified left hand side value. This is similar to _RBTree_Insert_inline() and helps the compiler to better optimize the code.
* score: Add _IO_Printf() and _IO_Vprintf()Sebastian Huber2017-11-062-0/+396
| | | | | | | | | | | | | | The previous vprintk() implementation had a questionable licence header, lacks support for the 'z' and 'j' format specifiers, is not robust against invalid format specifiers, uses a global variable for output. Replace it with a stripped down version of the FreeBSD kernel kvprintf() function. The new implementation allows a low overhead rtems_snprintf() if necessary. Update #3199. Close #3216.
* score: Fix non-SMP buildSebastian Huber2017-10-291-1/+1
|
* score: Fix _Scheduler_Set_affinity()Sebastian Huber2017-10-282-5/+11
|
* score: Simplify SMP get lowest scheduledSebastian Huber2017-10-282-21/+14
| | | | | | There is no need to pass in the order relation since the scheduled threads reside on an already ordered chain. The caller will decide what to do with the lowest scheduled thread.
* score: Move thread queue timeout handlingSebastian Huber2017-10-2415-206/+227
| | | | | Update #3117. Update #3182.
* score: Rename function threadq support functionSebastian Huber2017-10-2410-10/+10
| | | | | | | | | Rename _Thread_queue_Context_set_do_nothing_enqueue_callout() into _Thread_queue_Context_set_enqueue_do_nothing_extra(). More _Thread_queue_Context_set_enqueue_*() functions will follow. Update #3117. Update #3182.
* score: Add _Thread_Continue()Sebastian Huber2017-10-241-8/+17
| | | | | Update #3117. Update #3182.
* score: Rename _Watchdog_Ticks_from_*()Sebastian Huber2017-10-242-2/+2
| | | | | | | | | | Rename _Watchdog_Ticks_from_*() to _Watchdog_Realtime_from_*(). This highlights that these routines are used for the CLOCK_REALTIME watchdogs (in contrast to CLOCK_MONOTONIC). Update #3117. Update #3182.
* score: Add _Thread_Add_timeout_ticks()Sebastian Huber2017-10-241-2/+1
| | | | | | | | Replace _Thread_Timer_insert_monotonic() with _Thread_Add_timeout_ticks(). Update #3117. Update #3182.
* score: Fix warningSebastian Huber2017-10-211-1/+1
|
* score: Rename watchdog variantsSebastian Huber2017-10-173-5/+5
| | | | | | | | | | | Rename PER_CPU_WATCHDOG_RELATIVE in PER_CPU_WATCHDOG_MONOTONIC to highlight the corresponding POSIX CLOCK_MONOTONIC. Rename PER_CPU_WATCHDOG_ABSOLUTE in PER_CPU_WATCHDOG_REALTIME to highlight the corresponding POSIX CLOCK_REALTIME. Update #3117. Update #3182.
* timecounter: Update FreeBSD identifiersSebastian Huber2017-10-121-1/+1
| | | | Update #3175.
* timecounter: Merge FreeBSD change r324528Konstantin Belousov2017-10-121-3/+1
| | | | | | | | | | | | The th_bintime, th_microtime and th_nanotime members of the timehand all cache the last system time (uptime + boottime). Only the format differs. Do not re-calculate the bintime and simply use the value used to calculate the microtime and nanotime. Group all the updates under the relevant comment. Remove obsoleted XXX part. Submitted by: Sebastian Huber <sebastian.huber@embedded-brains.de> MFC after: 1 week Update #3175.
* timecounter: Merge FreeBSD change r315287Eric van Gyzen2017-10-121-0/+10
| | | | | | | | | | | | | | | | | | Add missing pieces of r315280 I moved this branch from github to a private server, and pulled from the wrong one when committing r315280, so I failed to include two recent commits. Thankfully, they were only cosmetic and were included in the review. Specifically: Add documentation, polish comments, and improve style(9). Tested by: pho (r315280) MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9791 Update #3175.
* timecounter: Merge FreeBSD change r315280Eric van Gyzen2017-10-121-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When the RTC is adjusted, reevaluate absolute sleep times based on the RTC POSIX 2008 says this about clock_settime(2): If the value of the CLOCK_REALTIME clock is set via clock_settime(), the new value of the clock shall be used to determine the time of expiration for absolute time services based upon the CLOCK_REALTIME clock. This applies to the time at which armed absolute timers expire. If the absolute time requested at the invocation of such a time service is before the new value of the clock, the time service shall expire immediately as if the clock had reached the requested time normally. Setting the value of the CLOCK_REALTIME clock via clock_settime() shall have no effect on threads that are blocked waiting for a relative time service based upon this clock, including the nanosleep() function; nor on the expiration of relative timers based upon this clock. Consequently, these time services shall expire when the requested relative interval elapses, independently of the new or old value of the clock. When the real-time clock is adjusted, such as by clock_settime(3), wake any threads sleeping until an absolute real-clock time. Such a sleep is indicated by a non-zero td_rtcgen. The sleep functions will set that field to zero and return zero to tell the caller to reevaluate its sleep duration based on the new value of the clock. At present, this affects the following functions: pthread_cond_timedwait(3) pthread_mutex_timedlock(3) pthread_rwlock_timedrdlock(3) pthread_rwlock_timedwrlock(3) sem_timedwait(3) sem_clockwait_np(3) I'm working on adding clock_nanosleep(2), which will also be affected. Reported by: Sebastian Huber <sebastian.huber@embedded-brains.de> Reviewed by: jhb, kib MFC after: 2 weeks Relnotes: yes Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9791 Update #3175.
* timecounter: Merge FreeBSD change r310053Ed Schouten2017-10-121-2/+2
| | | | | | | | | | | | | | | | | | | | | | Add labels to sysctls related to clocks. Sysctls like kern.eventtimer.et.*.quality currently embed the name of the clock device. This is problematic for the Prometheus metrics exporter for two reasons: - Some of those clocks have dashes in their names, which Prometheus doesn't allow to be used in metric names. - It doesn't allow for extracting the same property of all clocks on the system from within a single query. Attach these nodes to have a label, so that the Prometheus metrics exporter gives these metric a uniform name with the name of the clock attached as a label. Reviewed by: cem Differential Revision: https://reviews.freebsd.org/D8775 Update #3175.
* timecounter: Merge FreeBSD change r304285Konstantin Belousov2017-10-121-5/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Implement userspace gettimeofday(2) with HPET timecounter. Right now, userspace (fast) gettimeofday(2) on x86 only works for RDTSC. For older machines, like Core2, where RDTSC is not C2/C3 invariant, and which fall to HPET hardware, this means that the call has both the penalty of the syscall and of the uncached hw behind the QPI or PCIe connection to the sought bridge. Nothing can me done against the access latency, but the syscall overhead can be removed. System already provides mappable /dev/hpetX devices, which gives straight access to the HPET registers page. Add yet another algorithm to the x86 'vdso' timehands. Libc is updated to handle both RDTSC and HPET. For HPET, the index of the hpet device to mmap is passed from kernel to userspace, index might be changed and libc invalidates its mapping as needed. Remove cpu_fill_vdso_timehands() KPI, instead require that timecounters which can be used from userspace, to provide tc_fill_vdso_timehands{,32}() methods. Merge i386 and amd64 libc/<arch>/sys/__vdso_gettc.c into one source file in the new libc/x86/sys location. __vdso_gettc() internal interface is changed to move timecounter algorithm detection into the MD code. Measurements show that RDTSC even with the syscall overhead is faster than userspace HPET access. But still, userspace HPET is three-four times faster than syscall HPET on several Core2 and SandyBridge machines. Tested by: Howard Su <howard0su@gmail.com> Sponsored by: The FreeBSD Foundation MFC after: 1 month Differential revision: https://reviews.freebsd.org/D7473 Update #3175.
* timecounter: Merge FreeBSD change r303548Konstantin Belousov2017-10-121-13/+11
| | | | | | | | | | Cache getbintime(9) answer in timehands, similarly to getnanotime(9) and getmicrotime(9). Suggested and reviewed by: bde (previous version) Sponsored by: The FreeBSD Foundation MFC after: 1 month Update #3175.
* timecounter: Merge FreeBSD change r303387Konstantin Belousov2017-10-121-46/+74
| | | | | | | | | | | | | | | | | | | Prevent parallel tc_windup() calls, both parallel top-level calls from setclock() and from simultaneous top-level and interrupt. For this, tc_windup() is protected with a tc_setclock_mtx spinlock, in the try mode when called from hardclock interrupt. If spinlock cannot be obtained without spinning from the interrupt context, this means that top-level executes tc_windup() on other core and our try may be avoided. The boottimebin and boottime variables should be adjusted from tc_windup(). To be correct, they must be part of the timehands and read using lockless protocol. Remove the globals and reimplement the getboottime(9)/getboottimebin(9) KPI using the timehands read protocol. Tested by: pho (as part of the whole patch) Reviewed by: jhb (same) Discussed wit: bde Sponsored by: The FreeBSD Foundation MFC after: 1 month X-Differential revision: https://reviews.freebsd.org/D7302 Update #3175.
generated by cgit v1.2.3 (git 2.25.1) at 2024-05-28 18:43:42 +0000