| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
| |
features.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
Conflicts:
include/litmus/color.h
Signed-off-by: Christopher Kenna <cjk@cs.unc.edu>
|
| |
|
|
|
|
| |
This is not tested and probably won't work.
Signed-off-by: Christopher Kenna <cjk@cs.unc.edu>
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
Try and fix the cycle counter on ARMv7 CPUs (unsuccessful).
Use the multi-core timer on the ODROID-X to count cycles.
Add lockdown register access to proc.
Add proc options to disable prefetching for L1/L2 cache.
Add test to lockdown.c to print the time to read in various WSSs.
Signed-off-by: Christopher Kenna <cjk@cs.unc.edu>
|
| |
|
|
|
|
|
|
| |
Conflicts:
arch/arm/mach-omap2/board-omap4panda.c
Signed-off-by: Christopher Kenna <cjk@cs.unc.edu>
|
| |
|
|
| |
Signed-off-by: Christopher Kenna <cjk@cs.unc.edu>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While working on an ODROID-X (Samsung Exynos4412 ARM Cortex-A9), I
experienced non-determinism when reading and writing values to the
LITMUS^RT control page. Writes to the control page from user- (kernel-)
space where not always reflected in kernel- (user-) space.
Neither restricting the task to run on a single CPU nor inserting
general memory barriers (with mb()) fixed the problem. Mapping the
control page as uncachable in both kernel and user space did fix the
problem, which is what this patch does.
Also, since vunmap() cannot be called from an interrupt context, I had
to add a workqueue that unmaps and frees the control page when it is no
longer needed. (On my system, exit_litmus() was called from interrupt
context while the kernel reaped the task_struct.)
Conflicts:
include/litmus/rt_param.h
litmus/litmus.c
Does not make the color control page uncachable yet!
Signed-off-by: Christopher Kenna <cjk@cs.unc.edu>
|
| | |
|
| |\ |
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Added support for arbitrary deadlines.
Constraint: Relative deadline must be >= exec cost.
Use: Set relative deadline in rt_task::rdeadline. Set value to 0
to default to implicit deadlines.
Limitations: PFAIR not supported by this patch. PFAIR updated to
reject tasks that do not have implicit deadlines.
|
| | | |
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Conflicts:
include/litmus/budget.h
include/litmus/litmus.h
include/litmus/rt_param.h
include/litmus/sched_trace.h
include/litmus/trace.h
include/trace/events/litmus.h
litmus/Makefile
litmus/budget.c
litmus/ftdev.c
litmus/jobs.c
litmus/litmus.c
litmus/locking.c
litmus/preempt.c
litmus/rt_domain.c
litmus/sched_gsn_edf.c
litmus/trace.c
|
| | |/
| |
| |
| |
| | |
Track allocated pages and add a proc handler to reclaim free pages and
add control device for allocating colored memory with mmap.
|
| | |
| |
| |
| |
| |
| | |
User a 32-bit word for all non-preemptive section flags.
Set the "please yield soon" flag atomically when
accessing it on remotely-scheduled tasks.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Given a choice between several available CPUs (unlinked) on which
to schedule a task, let the scheduler select the CPU closest to
where that task was previously scheduled. Hopefully, this will
reduce cache migration penalties.
Notes: SCHED_CPU_AFFINITY is dependent upon x86 (only x86 is
supported at this time). Also PFair/PD^2 does not make use of
this feature.
Signed-off-by: Andrea Bastoni <bastoni@cs.unc.edu>
|
| | |
| |
| |
| |
| |
| | |
* Update prototypes for switched_to(), prio_changed(), select_task_rq().
* Fix missing pid field in printk output.
* Synchronize syscall numbers for arm and x86.
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| | |
Some things are working, but the kernel panics when you try and
deallocate an event group.
Committed so Jonathan can look at it.
|
| | | |
|
| | |
| |
| |
| |
| | |
Also started to fix typos in Jonathan's code, but there were too many
after I realized that event_group.o was not in the LITMUS Makefile.
|
| | | |
|
| |\ \
| | |
| | |
| | | |
ssh://cvs.cs.unc.edu/cvs/proj/litmus/repo/litmus2010 into wip-mc
|
| | | | |
|
| |/ / |
|
| | |
| |
| |
| |
| | |
This is not tested and missing everything for /proc. I just am checking
it in so that it the code is backed up in a git repository.
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add linux kernel configuration option CONFIG_LITMUS_MC.
Attempt to restore rt_param.h to its original state as much as possible.
Remove fields from rt_task and rt_job and move them into a new mc_data
struct. Added mc_data field to rt_param compiled in only if using MC
plugin.
Make a new MC plugin specific header that contains a mc_data struct,
which is a container for mc_task struct and a mc_job struct.
Update sched_mc.c to use the new data structures. Also, add some macros
that simplify the code, e.g., getting task criticality quickly.
Add system call to set MC plugin specific stuff. Check for the change in
liblitmus.
Add a few lines to exit_litmus to reclaim the MC plugin mc_data struct
in the task_struct on task exit.
|
| |
|
|
|
|
|
|
| |
It has always been LITMUS^RT policy that children of real-time tasks
may not skip the admissions test, etc. This used to be enforced, but
was apparently dropped during some port. This commit re-introduces
this policy. This fixes a kernel panic that occurred when "real-time
children" exited without proper initilization.
|
| |
|
|
|
|
|
|
|
|
|
| |
The CPU triggering the plugin switch should wait until all other CPUs
are in a proper state (synch_on_plugin_switch()) before performing the
actual switch.
Based on the original patch from Jeremy Erickson <jerickso@cs.unc.edu>.
This should solve (for most practical cases) the C-EDF-related
plugin-switch problem reported on the ML.
|
| |
|
|
|
|
|
|
|
|
| |
The TRACE() functionality doesn't need all of litmus.h. Currently,
it's impossible to use TRACE() in sched.h due to a circular
dependency. This patch moves TRACE() and friends to
litmus/sched_debug.h, which can be included in sched.h.
While at it, also fix some minor include ugliness that was revealed by
this change.
|
| | |
|
| |
|
|
|
|
| |
Change the Litmus proc layout so that loaded plugins are visible in
/proc/litmus/plugins/loaded and add Litmus functions make_plugin_proc_dir()
and remove_plugin_proc_dir() to add per-plugin proc directories.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Conflicts:
Makefile
arch/x86/include/asm/unistd_32.h
arch/x86/kernel/syscall_table_32.S
kernel/sched.c
kernel/time/tick-sched.c
Relevant API and functions changes (solved in this commit):
- (API) .enqueue_task() (enqueue_task_litmus),
dequeue_task() (dequeue_task_litmus),
[litmus/sched_litmus.c]
- (API) .select_task_rq() (select_task_rq_litmus)
[litmus/sched_litmus.c]
- (API) sysrq_dump_trace_buffer() and sysrq_handle_kill_rt_tasks()
[litmus/sched_trace.c]
- struct kfifo internal buffer name changed (buffer -> buf)
[litmus/sched_trace.c]
- add_wait_queue_exclusive_locked -> __add_wait_queue_tail_exclusive
[litmus/fmlp.c]
- syscall numbers for both x86_32 and x86_64
|
| |
|
|
|
| |
Allow all kinds of budget enforcement settings now that we have the
supporting infrastructure.
|
| |
|
|
|
|
|
|
| |
Introduces CONFIG_RELEASE_MASTER and makes release
master support dependent on the new symbol. This is
useful because dedicated interrupt handling only applies
to "large" multicore platforms. This will allow us to
not implement smp_send_pull_timers() for all platforms.
|
| |
|
|
|
| |
- With the "ALL" cluster size option the behavior of C-EDF is
equivalent to G-EDF (one single cluster)
|
| |
|
|
|
|
|
|
|
|
| |
NO_ENFORCEMENT - A job may execute beyond its declared execution time.
Jobs notify the kernel that they are complete via liblitmus's
sleep_next_period()
QUANTUM_ENFORCEMENT - The kernel terminates a job if its actual execution
time exceeds the declared execution time.
PRECISE_ENFORCEMENT - Hook declared, but not yet implemented. Plan to
support this policy through hrtimers. Error thrown if specified.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adapt to new schema for spinlock:
(tglx 20091217)
spinlock - the weakest one, which might sleep in RT
raw_spinlock - spinlock which always spins even on RT
arch_spinlock - the hardware level architecture dependent implementation
----
Most probably, all the spinlocks changed by this commit will be true
spinning lock (raw_spinlock) in PreemptRT (so hopefully we'll need few
changes when porting Litmmus to PreemptRT).
There are a couple of spinlock that the kernel still defines as
spinlock_t (therefore no changes reported in this commit) that might cause
us troubles:
- wait_queue_t lock is defined as spinlock_t; it is used in:
* fmlp.c -- sem->wait.lock
* sync.c -- ts_release.wait.lock
- rwlock_t used in fifo implementation in sched_trace.c
* this need probably to be changed to something always spinning in RT
at the expense of increased locking time.
----
This commit also fixes warnings and errors due to the need to include
slab.h when using kmalloc() and friends.
----
This commit does not compile.
|
| |
|
|
|
|
| |
Improved C-EDF plugin. C-EDF now supports different cluster sizes (based
on L2 and L3 cache sharing) and supports dynamic changes of cluster size
(this requires reloading the plugin).
|
| |
|
|
|
|
| |
Make sure the plugin is not used by any CPUs while switching.
The CPU performing the switch sends an IPI to all other CPUs forcing
them to synchronize on an atomic variable.
|
| |
|
|
|
|
|
|
| |
The od_table is strictly per-thread and should not be inherited across
a fork/clone. This caused memory corruption when a task exited, which
ultimately could lead to oopses in unrelated code.
Bug and testcase initially reported by Glenn.
|
| |
|
|
|
|
|
|
|
| |
This device only supports mmap()'ing a single page.
This page is shared RW between the kernel and userspace.
It is inteded to allow near-zero-overhead communication
between the kernel and userspace. It's first use will be a
proper implementation of user-signaled
non-preemptable section support.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a real-time task forks, then its LITMUS^RT-specific fields should be cleared,
because we don't want real-time tasks to spawn new real-time tasks that bypass
the plugin's admission control (if any).
This was broken in three ways:
1) kernel/fork.c did not erase all of tsk->rt_param, only the first few bytes due to
a wrong size argument to memset().
2) It should have been calling litmus_fork() instead anyway.
3) litmus_fork() was _also_ not clearing all of tsk->rt_param, due to another size
argument bug.
Interestingly, 1) and 2) can be traced back to the 2007->2008 port,
whereas 3) was added by Mitchell much later on (to dead code, no less).
I'm really surprised that this never blew up before.
|
| |
|
|
|
|
| |
- Binomial heap "heap" names conflicted with priority heap
of cgroup in kernel
- This patch change binomial heap "heap" names in "bheap"
|
| |
|
|
|
| |
Still to be merged:
- arm_release_timer() with no rq locking
|
