Messages in this thread | | | Date | Mon, 14 Oct 2013 18:16:50 +0200 | From | Juri Lelli <> | Subject | Re: [PATCH 03/14] sched: SCHED_DEADLINE structures & implementation. |
| |
On 10/14/2013 01:33 PM, Peter Zijlstra wrote: > On Mon, Oct 14, 2013 at 12:43:35PM +0200, Juri Lelli wrote: >> +static void replenish_dl_entity(struct sched_dl_entity *dl_se) >> +{ >> + struct dl_rq *dl_rq = dl_rq_of_se(dl_se); >> + struct rq *rq = rq_of_dl_rq(dl_rq); >> + >> + /* >> + * We keep moving the deadline away until we get some >> + * available runtime for the entity. This ensures correct >> + * handling of situations where the runtime overrun is >> + * arbitrary large. >> + */ >> + while (dl_se->runtime <= 0) { >> + dl_se->deadline += dl_se->dl_deadline; >> + dl_se->runtime += dl_se->dl_runtime; >> + } > > Are we sure GCC won't be 'smart' and bite us; that is do we need > something like: > > asm("" : "+rm" (dl_se->runtime)); > > in there? See: > > 0d98bb2656e9 sched: Prevent compiler from optimising the sched_avg_update() loop >
When disassembled everything seems fine, at least for x86 and ARM. Do I add the fake data hazard anyway?
Thanks,
- Juri
| |