aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
* Change default path to remove *2010 tag.2012.1wip-2011.2-bbbBjoern B. Brandenburg2011-11-24
| | | | | This is not the 2010 version anymore, so let's use a version-agnostic path.
* Switch to new np-section protocol.Bjoern B. Brandenburg2011-11-24
| | | | Everything is in one 32-bit word now.
* update headersBjoern B. Brandenburg2011-11-24
| | | | we need uint64_t in litmus.h
* tests: add valid budget_policy to rt_task parametersAndrea Bastoni2011-08-26
| | | | | | | Since set_rt_task_param() checks for wrong budget policies, make sure to specify a valid policy to avoid errors like the following. litmus: real-time task 3635 rejected because unsupported budget enforcement policy specified
* tests: allow 'LITMUS' as an alias of all non-Linux pluginsBjoern B. Brandenburg2011-02-03
| | | | This is useful for tests that apply to any plugin.
* add test cases for the FMLP and SRP under GSN- and PSN-EDFBjoern B. Brandenburg2011-02-03
|
* switch to generic locking system callsBjoern B. Brandenburg2011-02-03
| | | | | The new generic lock layer in LITMUS^RT does away with per-protocol system calls. Change accordingly.
* Export wctime() and cputime() from rtspin to library clientsBjoern B. Brandenburg2011-02-03
| | | | | The timing functions are quite handy when building benchmark tasks. Avoid copy&paste reuse by making them available via the library.
* bugfix: switch operangs in delay computationBjoern B. Brandenburg2011-02-03
| | | | | We want the positive loop length, not the negative, to prevent jobs from overruning their budget.
* add test that checks that RT class is not inherited across forkBjoern B. Brandenburg2011-02-01
|
* Remove old delay loop configuration cruft from rtspinBjoern B. Brandenburg2011-01-30
| | | | | Using cputime() is much more accurate than the old delay loop auto-configuration. There is no good reason to keep it around.
* Modify RTSpin so that it can read execution time data from a CSV file.Christopher Kenna2011-01-30
| | | | | Still need to figure out what to do with the period is less than the exec_cost (or budget).
* fix header detection on ARM2011.1Bjoern B. Brandenburg2011-01-25
| | | | | | The ARM port uses <litmus... instead of "litmus..., so let's use egrep with a proper regular expression to reliably find the line that includes the Litmus system calls.
* don't require LITMUS^RT headers for 'make help'Bjoern B. Brandenburg2011-01-25
| | | | It should be possible to see the help when the configuration is broken.
* add installation instructions to liblitmus (and 'make help')Bjoern B. Brandenburg2010-11-24
| | | | | Some initial documentation to aid new users get started with compiling the library.
* add cscope target to MakefileBjoern B. Brandenburg2010-11-23
| | | | | cscope is very nice for exploring the library and finding related code. Let's support it by pre-building an index of all files.
* add TAGS and tags targets to MakefileBjoern B. Brandenburg2010-11-23
| | | | | It's nice to have a TAGS file around when exploring the library. Let's generate one for vim and emacs.
* Patch so Make will die if it detects non-LITMUS^RT kernel headers.Christopher Kenna2010-11-22
|
* Provide standard Makefile rules for clients of liblitmusBjoern B. Brandenburg2010-11-22
| | | | | | We avoid doing the same thing over and over in all repositories using liblitmus if we just pull in common rules & configurations from liblitmus. This gives us the ability to cross-compile for free.
* rtspin: transition back to non-real-time modeBjoern B. Brandenburg2010-11-22
| | | | | rtspin produces prettier schedules if it transitions to non-real-time mode before terminating.
* remove asm/atomic.hBjoern B. Brandenburg2010-11-09
| | | | | | | | | | | For historic resons, we carry old atomic operations support in liblitmus. This is no longer useful: 1) There is actually no client for these calls in liblitmus. 2) There is now a standard gcc API for this purpose. http://gcc.gnu.org/onlinedocs/gcc/Atomic-Builtins.html Thus, we can reduce our maintenance burden.
* generate error message if build config seems fishyBjoern B. Brandenburg2010-11-09
| | | | | Failing loudly is a lot better than just producing a cryptic error message about some dependency file.
* output header to STDERR in measure_syscallBjoern B. Brandenburg2010-11-09
| | | | This hopefully helps a bit to clarify the output.
* add emergency exit to rtspinBjoern B. Brandenburg2010-11-09
| | | | | | This should only trigger if the execution time tracking is broken (as it was on District10). With the emergency exit, we at least prevent the system from becoming unresponsive.
* don't require /bin/bashBjoern B. Brandenburg2010-11-09
| | | | | We don't have BASH on District10 at the moment, and the scripts work just fine with busybox's /bin/sh replacement (ash).
* add ARM architecture-dependent filesBjoern B. Brandenburg2010-11-09
| | | | | The atomic.h file is currently only a dummy, but it is sufficient to get basic real-time tasks to run.
* refactor: remove all architecture-dependent code from include/Bjoern B. Brandenburg2010-11-09
| | | | Move the architecture-dependent code to the arch/ subtree.
* refactor: use architecture-specific includes for cycles.hBjoern B. Brandenburg2010-11-09
|
* avoid including header files directlyBjoern B. Brandenburg2010-11-09
| | | | | | | | | | | | | | | The kernel warns against including header files directly. For good reason: our previous approach (just -I$KERNEL/include) caused all kinds of files to be included that should have come from /usr/include instead. This patch rewrites the Makfile so that the (few) needed headers are copied into the liblitmus src tree before compiling the library. This avoids having to specify the kernel include directories with -I, and also makes it easier to link against liblitmus (external applications do not need to know where the kernel is). Finally, this allows us to enable -Werror.
* refactor: switch back from SCons to makeBjoern B. Brandenburg2010-11-09
| | | | | | | | | | | | | | | We originally switched from make to scons because 1) our makefiles were not very good; 2) SCons promised to make maintaining the build system simpler. Unfortunately, SCons has become more and more difficult to deal with as we moved to supporting several architecture and cross compilation, to the extend that we ended up re-creating make functionality in SCons. So let's switch back to make using a "clean" Makefile. Thanks a lot to Andrea Bastoni and Chris Kenna for feedback on previous iterations of these patches.
* Silence warning due to missing header file.Bjoern B. Brandenburg2010-11-09
|
* Don't show output from `which dialog` in setschedMac Mollison2010-11-09
| | | | | | | | Prevents `which dialog` from outputting (ugly) information to stderr in the case that dialog is not installed. Changes error message to make it more clear that dialog is strictly optional.
* Location change for the /proc loaded plugins file.Christopher Kenna2010-10-22
|
* rtspin: add -e option to enable PRECISE_ENFORCEMENT2010.2Bjoern B. Brandenburg2010-10-19
| | | | Useful to test how the kernel handles PRECISE_ENFORCEMENT.
* zero out rt_task parameters before setting themBjoern B. Brandenburg2010-09-22
| | | | | | | This avoids passing random stack contents to the kernel if the structure gained some fields due to plugin development. (Also, fix up some white space issues while we are at it.)
* Support budget enforcement policies. Allows tasks to specifyGlenn Elliott2010-05-20
| | | | | | | | how their execution budgets should be enforced: NO_ENFORCEMENT, QUANTUM_ENFORCEMENT, and PRECISE_ENFORCEMENT (unsupported). NOTE: Users of NO_ENFORCEMENT must call sleep_next_period() at the end of every job to signal to the kernel that its job is complete.
* Add atomic operations on x86_642010.1Andrea Bastoni2010-04-19
|
* Add cli() and sti() wrappers... use with care...Andrea Bastoni2010-04-12
|
* Add get_control_page() library functionAndrea Bastoni2010-04-12
|
* Improve build system.Bjoern B. Brandenburg2010-03-17
| | | | | | | | | | | | | | | Restructure SConstruct to reduce the likelihood of errors and to provide additional feedback. Features: - Better help. You can now run 'scons -h' to get a listing of all build options. - Local configuration. Build variables are picked up in a local .config file, if present. - Added --dump-config option to display the build configuration. - Use scons substitution system to avoid having to do parameter substitutions manually. - Some assorted fixes.
* Merge branch 'master' of ssh://cvs.cs.unc.edu/cvs/proj/litmus/repo/liblitmus2010Bjoern B. Brandenburg2010-03-10
|\
| * Build fix: support building on i686 systemsBjoern B. Brandenburg2010-03-10
| | | | | | | | | | | | | | 32bit Intel systems don't actually report 'x86' as their architecture, rather, they are i?86 systems. Provide a lookup table to map these codes to the 'x86' name that is used in the SConstruct file.
| * Fix error message.Bjoern B. Brandenburg2010-03-10
| | | | | | | | This is liblitmus, not ft_tools.
* | Add -v (verbose) switch to rtspin.Bjoern B. Brandenburg2010-03-10
|/ | | | | Can be used to dump the delay loop that rtspin is using.
* Update rt_launch usageAndrea Bastoni2010-03-02
|
* Add job control test case.Bjoern B. Brandenburg2010-02-26
| | | | Passes.
* Add test for set_rt_task_param parameter validation.Bjoern B. Brandenburg2010-02-26
| | | | All passed.
* Remove prototype for non-existing system call.Bjoern B. Brandenburg2010-02-26
| | | | | This syscall hasn't been around since Fall 2007... The RTCSA'08 paper details how we are doing it now.
* Add tests for invalid object types and od_open().Bjoern B. Brandenburg2010-02-25
|
* Provide additional information on test failures.Bjoern B. Brandenburg2010-02-25
|
generated by cgit v1.2.2 (git 2.25.0) at 2025-03-14 17:16:41 -0400