[Cbe-oss-dev] [PATCH] spusched: fix initial timeslice calculation

[Jeremy Kerr]

From: Christoph Hellwig <hch at lst.de>

Currently we calculate the first timeslice for every context
incorrectly - alloc_spu_context calls spu_set_timeslice before we set
ctx->prio so we always calculate the longest possible timeslice for the
lowest possible priority.

This patch makes sure to update the schedule-related fields before
calculating the timeslice and also makes sure we update the timeslice for
a non-running context when entering spu_run so a priority change affects
the context as soon as possible.

Signed-off-by: Christoph Hellwig <hch at lst.de>
Signed-off-by: Jeremy Kerr <jk at ozlabs.org>

---

Paul - another fix for 2.6.23. Tested on Cell.

---
 arch/powerpc/platforms/cell/spufs/context.c |    3 ++-
 arch/powerpc/platforms/cell/spufs/run.c     |    4 ++++
 2 files changed, 6 insertions(+), 1 deletions(-)

diff --git a/arch/powerpc/platforms/cell/spufs/context.c b/arch/powerpc/platforms/cell/spufs/context.c
index 6694f86..9cb081c 100644
--- a/arch/powerpc/platforms/cell/spufs/context.c
+++ b/arch/powerpc/platforms/cell/spufs/context.c
@@ -59,7 +59,8 @@ struct spu_context *alloc_spu_context(struct spu_gang *gang)
 	INIT_LIST_HEAD(&ctx->aff_list);
 	if (gang)
 		spu_gang_add_ctx(gang, ctx);
-	ctx->cpus_allowed = current->cpus_allowed;
+
+	__spu_update_sched_info(ctx);
 	spu_set_timeslice(ctx);
 	ctx->stats.util_state = SPU_UTIL_IDLE_LOADED;
 
diff --git a/arch/powerpc/platforms/cell/spufs/run.c b/arch/powerpc/platforms/cell/spufs/run.c
index 0b50fa5..6abdd8f 100644
--- a/arch/powerpc/platforms/cell/spufs/run.c
+++ b/arch/powerpc/platforms/cell/spufs/run.c
@@ -312,6 +312,7 @@ long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *event)
 	spu_acquire(ctx);
 	if (ctx->state == SPU_STATE_SAVED) {
 		__spu_update_sched_info(ctx);
+		spu_set_timeslice(ctx);
 
 		ret = spu_activate(ctx, 0);
 		if (ret) {
@@ -322,6 +323,9 @@ long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *event)
 		/*
 		 * We have to update the scheduling priority under active_mutex
 		 * to protect against find_victim().
+		 *
+		 * No need to update the timeslice ASAP, it will get updated
+		 * once the current one has expired.
 		 */
 		spu_update_sched_info(ctx);
 	}
-- 
1.5.0.rc4.g85b1