Messages in this thread | | | From | Mike Frysinger <> | Date | Wed, 13 Apr 2011 02:14:39 -0400 | Subject | freezer: should barriers be smp ? |
| |
when we suspend/resume Blackfin SMP systems, we notice that the freezer code runs on multiple cores. this is of course what you want -- freeze processes in parallel. however, the code only uses non-smp based barriers which causes us problems ... our cores need software support to keep caches in sync, so our smp barriers do just that. but the non-smp barriers do not, and so the frozen/thawed processes randomly get stuck in the wrong task state.
thinking about it, shouldnt the freezer code be using smp barriers ? the point is to make sure that the updates are seen across all cores are not just the current one ? -mike
| |