[POWERPC] spufs: fix scheduler starvation by idle contexts

2.6.25 has a regression where we can starve the scheduler by creating
(N_SPES+1) contexts, then running them one at a time.

The final context will never be run, as the other contexts are loaded on
the SPEs, none of which are repoted as free (ie, spu->alloc_state !=
SPU_FREE), so spu_get_idle() doesn't give us a spu to run on. Because
all of the contexts are stopped, none are descheduled by the scheduler
tick, as spusched_tick returns if spu_stopped(ctx).

This change replaces the spu_stopped() check with checking for SCHED_IDLE
in ctx->policy. We set a context's policy to SCHED_IDLE when we're not
in spu_run(). We also favour SCHED_IDLE contexts when looking for contexts
to unbind, but leave their timeslice intact for later resumption.

This patch fixes the following test in the spufs-testsuite:
  tests/20-scheduler/02-yield-starvation

Signed-off-by: Jeremy Kerr <jk@ozlabs.org>

diff --git a/arch/powerpc/platforms/cell/spufs/run.c b/arch/powerpc/platforms/cell/spufs/run.c
index fca22e1..6221968 100644 (file)
--- a/arch/powerpc/platforms/cell/spufs/run.c
+++ b/arch/powerpc/platforms/cell/spufs/run.c
@@ -234,6 +234,7 @@ static int spu_run_fini(struct spu_context *ctx, u32 *npc,
        *npc = ctx->ops->npc_read(ctx);
 
        spuctx_switch_state(ctx, SPU_UTIL_IDLE_LOADED);
+       ctx->policy = SCHED_IDLE;
        spu_release(ctx);
 
        if (signal_pending(current))

diff --git a/arch/powerpc/platforms/cell/spufs/sched.c b/arch/powerpc/platforms/cell/spufs/sched.c
index 5915343..3a59721 100644 (file)
--- a/arch/powerpc/platforms/cell/spufs/sched.c
+++ b/arch/powerpc/platforms/cell/spufs/sched.c
@@ -856,21 +856,18 @@ static noinline void spusched_tick(struct spu_context *ctx)
 {
        struct spu_context *new = NULL;
        struct spu *spu = NULL;
-       u32 status;
 
        if (spu_acquire(ctx))
                BUG();  /* a kernel thread never has signals pending */
 
        if (ctx->state != SPU_STATE_RUNNABLE)
                goto out;
-       if (spu_stopped(ctx, &status))
-               goto out;
        if (ctx->flags & SPU_CREATE_NOSCHED)
                goto out;
        if (ctx->policy == SCHED_FIFO)
                goto out;
 
-       if (--ctx->time_slice)
+       if (--ctx->time_slice && ctx->policy != SCHED_IDLE)
                goto out;
 
        spu = ctx->spu;
@@ -880,7 +877,8 @@ static noinline void spusched_tick(struct spu_context *ctx)
        new = grab_runnable_context(ctx->prio + 1, spu->node);
        if (new) {
                spu_unschedule(spu, ctx);
-               spu_add_to_rq(ctx);
+               if (ctx->policy != SCHED_IDLE)
+                       spu_add_to_rq(ctx);
        } else {
                spu_context_nospu_trace(spusched_tick__newslice, ctx);
                ctx->time_slice++;