aboutsummaryrefslogtreecommitdiffstats
path: root/litmus
Commit message (Collapse)AuthorAge
* fixes for RTSSbudgetingztong2021-06-12
|
* Changed syscall name from exit_fz to cancel_watchdogztong2021-03-10
|
* Merged in job release race condition fixes from MC^2, added Tracing ↵ztong2021-03-07
| | | | functionality
* Backport logging, timer, and race condition bugfixes from MC^2Joshua Bakita2021-03-07
| | | | | | | | | | | | | | | Fixes are: 1. Fix a potential race condition with budget replenishment in GEDF reservations 2. Fix bug where release_jobs_before_now() could race with itself and corrupt the release heap state (the conditional that checked for a non-empty queue was prior to lock aquisition, and another thread could potentially clear the queue while we spun). 3. Fix a problem with incorrect HRTIMER_MODE in rt_domain. With a preemptible kernel, these need to explicitly expire in a hard interrupt context (new with LITMUS-RT on 5.4). 4. Fix a logging bug in rt_domain where non-existant variable `rh->timer` was being printed rather than `rt->timer`.
* Syscall for exiting forbidden zone addedztong2021-03-02
|
* Readded sending signal code in fz watchdogztong2021-03-01
|
* Took out unnecessary assertsztong2021-02-26
|
* Cleaned up FZ wakeup logicztong2021-02-26
|
* Added support for early releasing tasks in clear_priority_inheritanceztong2021-02-23
|
* WIP complete?ztong2021-02-23
|
* WIPztong2021-02-21
|
* WIPztong2021-02-21
|
* WIP5ztong2021-02-21
|
* WIP4ztong2021-02-21
|
* Made wait_queue_entry_t in fz_police timeout persistant after timeoutztong2021-02-21
|
* Test sleeping with litmus_reschedulztong2021-02-21
|
* WIP 3ztong2021-02-19
|
* WIP2ztong2021-02-18
|
* WIP checkpointztong2021-02-18
|
* omlp_fz_police bug fixupdate_litmus_2019ztong2021-02-17
|
* Fixed access_fz_check bug and changed fz enforcement conventionztong2021-02-16
|
* UNTESTED code for forbidden zone policingztong2021-02-16
|
* Corrected logic for pinned CPU task assignmentztong2021-02-09
|
* Added capability of pinning a task to a cpu in gedf_reservationztong2021-02-09
| | | | Untested
* Added KUTrace Supportztong2021-02-03
| | | | Moved sched_litmus back to the top scheduling class
* EXT-RES Ported to kernel version 5.4.0-rc7ztong2020-12-23
|
* Switch to HRTIMER_MODE_ABS_PINNED_HARD5.4-EXT-RESleochanj1052020-11-09
| | | Scheduling timers need to expire in a hard IRQ context which must be explicitly requested with Linux 5.4.
* style problemsleochanj052020-09-26
|
* handled prev==NULL in pfp and psnedfleochanj052020-09-23
|
* initial fixesleochanj052020-09-23
|
* Handle prev=NULL in schedule() for P-RESJoshua Bakita2020-05-08
| | | | | | P-RES doesn't appear to use `prev` for anything besides sanity checks, so fix should not and does not appear to break any of the logical correctness.
* Handle prev=NULL in C-EDF and PFAIR's schedule()Nathan O2020-02-13
| | | | | | | | - Updated the C-EDF and PFAIR scheduler plugins so that they should be able to handle the case when their schedule functions are passed a NULL pointer in place of prev. They can handle tasks without crashing, but I have not gone to any length to ensure their behavior is completely correct.
* Fix balancing issue, GEDF pluginNathan O2020-02-13
| | | | | | | | | | | | | | | - Fixed a bug where some new balancing code wasn't commented out when it needed to be. (It wasn't in the version of Linux that the previous version of LITMUS was based on, so I missed it earlier.) - The GSN-EDF plugin can now be activated, and runs tasks without crashing (though I haven't verified that it schedules stuff in the correct order yet). - *All remaining plugins probably need to be modified in the same was as the GSN-EDF plugin*. Specifically, the "prev" argument to the "<plugin>_schedule(...)" function may be NULL, and the plugin must be able to handle such a case!
* Fix remaining compilation problemsNathan O2019-12-10
| | | | | | | | | | | - This change addresses additional changes that are needed to LITMUS code so that the kernel is able to compile. - The kernel is now able to compile. - I will probably squash this commit with the previous commit sometime later, so that all of the LITMUS-updating changes are in one commit and easy to see, *if* this kernel works.
* Re-add LITMUS files, but don't "connect the wires"Nathan O2019-11-15
- Added the LITMUS^RT source files from the most recent version of LITMUS (based on the old kernel). - This change does *not* actually make use of any of the new files, because I wanted to make sure that any changes for getting LITMUS to actually work are clearly visible in future commits. If I made any such changes before committing the files themselves, then it wouldn't be as easy to see what needed to change in the LITMUS source files, and the large number of changes would make it more difficult to see what needed to change in the base Linux sources, too.
generated by cgit v1.2.2 (git 2.25.0) at 2025-04-16 05:08:30 -0400