aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorAndrea Bastoni <bastoni@cs.unc.edu>2010-04-16 10:07:39 -0400
committerAndrea Bastoni <bastoni@cs.unc.edu>2010-04-16 10:07:39 -0400
commit5fd924b4177fff6904137fae17c8d79b9aeda570 (patch)
tree9901ab404f87f1bca2b1ea8bd3eab47b6e3f7ba3
parentc43679c57a5d89b703dd678d77f5f2186b9030cd (diff)
Add FIXME for crashes of C-EDF when reloading the cluster conf. on big SMP
- reloading the cluster configuration of C-EDF on big SMP machines is causing troubles as we do not check (ATM) that the plugin is used on some processor when we unload it
-rw-r--r--litmus/litmus.c6
1 files changed, 5 insertions, 1 deletions
diff --git a/litmus/litmus.c b/litmus/litmus.c
index a62ec1455cef..70a7f628455b 100644
--- a/litmus/litmus.c
+++ b/litmus/litmus.c
@@ -374,8 +374,12 @@ void litmus_exit_task(struct task_struct* tsk)
374 * (and that none is created in parallel) and that the plugin is not 374 * (and that none is created in parallel) and that the plugin is not
375 * currently in use on any processor (in theory). 375 * currently in use on any processor (in theory).
376 * 376 *
377 * For now, we don't enforce the second part since it is unlikely to cause 377 * FIXME we don't enforce the second part since it is unlikely to cause
378 * any trouble by itself as long as we don't unload modules. 378 * any trouble by itself as long as we don't unload modules.
379 * - This is causing problems with C-EDF when reloading clusters configuration
380 * on big SMP machines (e.g., 24 cores). The plugin is in use on some
381 * processor when we unload is, and this cause troubles as
382 * we free the clusters while these are used :(
379 */ 383 */
380int switch_sched_plugin(struct sched_plugin* plugin) 384int switch_sched_plugin(struct sched_plugin* plugin)
381{ 385{