aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
* Kludge: Add graph e-to-e period in sched_traceGlenn Elliott2014-01-29
| | | | | Kludge the expected graph end-to-end response time into record data to make data analysis easier.
* BUG: Trace IDs don't include ASSIGNEDGlenn Elliott2014-01-17
|
* Add combined PGM_SRC_SINK node type.Glenn Elliott2014-01-17
|
* Compilation Bug: Fix PGM trace macros.Glenn Elliott2014-01-17
|
* Don't unlink np tasks that block preemptions.Glenn Elliott2014-01-17
| | | | | | | This patch prevents non-preemptive tasks from being unlinked up a blocked preemption. Apparently was leading to lost tasks (they would suspend and never be heard from again).
* sched_trace: Trace PGM node type parametersGlenn Elliott2014-01-14
|
* Verbose admission test for CFL-splitGlenn Elliott2014-01-02
|
* Add PGM support to C-FLwip-pgmGlenn Elliott2013-12-06
| | | | This patch applies the same PGM changes made to C-EDF to C-FL.
* Add C-FL scheduler plugin.Glenn Elliott2013-11-26
| | | | | | This patch adds a C-FL scheduler plugin. Original work by Jeremy Erikson, port to latest Litmus by Namhoon Kim, and cleanup and commit by Glenn Elliott.
* PGM: Boost priority of producers, not consumers.Glenn Elliott2013-11-19
| | | | | | | This patch boosts the priority of PGM producers while they are sending tokens instead of boosting the priority of consumers while they are waiting for tokens. This improves schedulability analysis.
* Add PGM support to C-EDF.Glenn Elliott2013-09-27
| | | | Patch adds PGM priority boosting (and un-boosting) to C-EDF.
* Fix: Incorrectly expressed PGM adjustment thresh.Glenn Elliott2013-09-27
| | | | | | | PGM release/deadline adjustment is ignored if the difference between an adjusted time and current time falls below a given threshold. This threshold is supposed to be 200us, but was actually 200ns. This patch resolves this--threshold is now propertly 200us.
* Tracing for PGM release/deadline adjustment.Glenn Elliott2013-09-27
| | | | | | Patch adds tracing of job release/deadline adjustments of PGM tasks. Tracing is separate from regular job tracing so that we many observe the magnitude of adjustments/slippage.
* PGM support in GSN-EDF.Glenn Elliott2013-09-27
| | | | | | Patch enables PGM support by GSN-EDF. GSN-EDF boosts the priority of a job waiting for inbound tokens. Likewise, boosting is removed when inbound tokens have been received.
* PGM release/deadline adjustment.Glenn Elliott2013-09-27
| | | | | | Adds code that adjusts a jobs release and deadline according to when the job receives the necessary PGM tokens.
* Add CONFIG_SCHED_PGM to litmus/KconfigGlenn Elliott2013-09-27
| | | | Conditional compilation of PGM features. Off by default.
* Add control page variables for PGM.Glenn Elliott2013-09-27
| | | | | | | | | | | | | Currently, PGM token satisfaction is tracked in userspace. However, Litmus needs to be aware of when a PGM task is waiting for tokens in order to avoid unbounded priority inversions. The state of a PGM task is communicated to Litmus via the control page. We should be able to remove control page variables for PGM if/when token tracking is moved from userspace into the kernel.
* Fix: Set litmus timestamp in ftrace records.Glenn Elliott2013-09-16
| | | | | Patch updates ftrace.h to record a litmus_clock() time stamp in ftrace records.
* Fix litmus_admit_task() error pathBjoern Brandenburg2013-09-13
| | | | Don't try to release slab objects that were never allocated.
* litmus: Fix build error (implicit declaration of read_current_timer()) on ARMRoy Spliet2013-09-13
| | | | Signed-off-by: Roy Spliet <rspliet@mpi-sws.org>
* uncachedev: mmap memory that is not cached by CPUs2013.1archived-2013.1Glenn Elliott2013-08-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch creates a new character device, uncachedev. Pages of RAM allocated by this device are not cached by CPUs. Uses for such pages: 1) Determining *very* pessimistic emperical worst- case execution times. 2) Compare against performance with caches (quantify the avg. case benefit). 3) Deterministic memory accesses (access cannot cause a cache eviction.) 4) Theoretically, increased performance can be achieved by storing infrequently accessed data in uncache pages. uncachedev allocates pages with the pgprot_noncached() page attribute for user applications. Since pages allocated by uncachedev are not locked in memory by default, applications with any access level may mmap pages with uncachedev. Limitations: 1) Uncache pages must be MAP_PRIVATE. 2) Remapping not supported. Usage (user level): int size = NR_PAGES*PAGE_SIZE; int fd = open("/dev/litmus/uncache", O_RDWR); char *data = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0); <...do stuff...> munmap(data, size);
* Add PD^2 scheduler pluginFelipe Cerqueira2013-08-07
|
* Add C-EDF scheduler pluginFelipe Cerqueira2013-08-07
|
* Add P-FP scheduler pluginFelipe Cerqueira2013-08-07
|
* GSN-EDF: link tasks to local CPU if it is idleBjoern Brandenburg2013-08-07
| | | | | | | If a task is released on a CPU that is idle, then it is pointless to send an IPI to another CPU just to schedule the newly released task. This patch changes check_for_preemptions() to prefer the local CPU first if it is idle.
* Add GSN-EDF scheduler pluginFelipe Cerqueira2013-08-07
|
* Add PSN-EDF scheduler pluginFelipe Cerqueira2013-08-07
|
* Protect LITMUS^RT tasks from re-nicingBjoern Brandenburg2013-08-07
| | | | | Assigning a nice value to LITMUS^RT tasks is meaningless. Bail out early.
* Don't call set_tsk_need_resched() on remote LITMUS^RT taskBjoern Brandenburg2013-08-07
| | | | | This patch fixes a BUG_ON() in litmus/preempt.c:33 reported by Felipe Cerqueira & Manohar Vanga.
* Hook into rt scheduling class to protect LITMUS^RT tasksBjoern Brandenburg2013-08-07
| | | | | | The rt scheduling class thinks it's the highest-priority scheduling class around. It is not in LITMUS^RT. Don't go preempting remote cores that run SCHED_LITMUS tasks.
* Don't trigger load balancer in scheduler tick for LITMUS^RTBjoern Brandenburg2013-08-07
|
* Cache 'prev' in LITMUS^RT runqueueBjoern Brandenburg2013-08-07
| | | | | | LITMUS^RT plugins like to know who 'prev' is. pick_next_task() doesn't expose that info, so we just cache prev in the runqueue. Could robably be replaced by looking at 'current' instead.
* Hook into finish_switch()Bjoern Brandenburg2013-08-07
| | | | To keep track of stack usage and to notify plugin, if necessary.
* Reset SCHED_LITMUS scheduling class on forkBjoern Brandenburg2013-08-07
|
* Block sched_setaffinity() for SCHED_LITMUS tasksBjoern Brandenburg2013-08-07
|
* Integrate LITMUS^RT scheduling class with sched_setschedulerBjoern Brandenburg2013-08-07
|
* Disable cut-to-CFS optimization in Linux schedulerBjoern Brandenburg2013-08-07
| | | | | Global plugins require that the plugin be called even if there currently is no real-time task executing on the local core.
* Integrate LITMUS^RT with try_to_wake_up() pathBjoern Brandenburg2013-08-07
|
* Splice LITMUS^RT scheduling class into scheduler hierarchy above rt classBjoern Brandenburg2013-08-07
| | | | | Below stop_machine tasks, as such tasks never arise during normal operation.
* Call LITMUS^RT tick function from Linux tickBjoern Brandenburg2013-08-07
|
* Add LITMUS^RT scheduling class in kernel/sched/MakefileBjoern Brandenburg2013-08-07
|
* Introduce LITMUS^RT runqueue dummy into struct rqBjoern Brandenburg2013-08-07
|
* Integrate SRP ceiling blocking callback with Linux schedulerBjoern Brandenburg2013-08-07
| | | | Check whether a suspension is required at end of schedule().
* Hookup sched_trace_XXX() tracing in Linux schedulerBjoern Brandenburg2013-08-07
| | | | This patch adds context switch tracing to the main Linux scheduler.
* Augment rt_task() with is_realtime()Bjoern Brandenburg2013-08-07
| | | | | | | | | | | Whenever the kernel checks for rt_task() to avoid delaying real-time tasks, we want it to also not delay LITMUS^RT tasks. Hence, most calls to rt_task() should be matched by an equivalent call to is_realtime(). Notably, this affects the implementations of select() and nanosleep(), which use timer_slack_ns when setting up timers for non-real-time tasks.
* Hook into fork(), exec(), and exit()Felipe Cerqueira2013-08-07
| | | | | Allow LITMUS^RT to do some work when a process is created or terminated.
* Call sched_state_task_picked() from pick_next_task_stop()Bjoern Brandenburg2013-08-07
| | | | | Otherwise, the scheduler state machine becomes confused (and goes into a rescheduling loop) when stop-machine is triggered.
* Integrate preemption state machine with Linux schedulerBjoern Brandenburg2013-08-07
| | | | Track when a processor is going to schedule "soon".
* Move trace point definition to litmus/litmus.cBjoern Brandenburg2013-08-07
| | | | | | | | If !CONFIG_SCHED_TASK_TRACE, but CONFIG_SCHED_LITMUS_TRACEPOINT, then we still need to define the tracepoint structures. This patch should be integrated with the earlier sched_task_trace.c patches during one of the next major rebasing efforts.
* Add LITMUS^RT syscalls for x86Felipe Cerqueira2013-08-07
|
generated by cgit v1.2.2 (git 2.25.0) at 2024-11-23 13:12:16 -0500