[POWERPC] spufs: Don't call spu_run_init from spu_reacquire_runnable
authorAndre Detsch <adetsch@br.ibm.com>
Tue, 21 Aug 2007 02:06:22 +0000 (10:06 +0800)
committerPaul Mackerras <paulus@samba.org>
Thu, 30 Aug 2007 06:27:18 +0000 (16:27 +1000)
This fixes a major bug which was happening when a SPU thread advances
its execution right after being restored to a SPU.  A potentially
outdated NPC value was being (re)written to the SPU.

So, spu_run_init, in this case, was either not doing anything relevant,
or breaking the execution of the SPU thread.

This fixes a common problem of losing a mailbox write when it was done
to a saved context.

Signed-off-by: Andre Detsch <adetsch@br.ibm.com>
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
arch/powerpc/platforms/cell/spufs/run.c

index 6abdd8fe2098f29591a31ebd65a14952899067d4..958f10e90fddf5260ba7fef36fa160f97ba0de79 100644 (file)
@@ -193,11 +193,7 @@ static int spu_reacquire_runnable(struct spu_context *ctx, u32 *npc,
        if (ret)
                return ret;
 
-       ret = spu_run_init(ctx, npc);
-       if (ret) {
-               spu_release(ctx);
-               return ret;
-       }
+       spuctx_switch_state(ctx, SPU_UTIL_USER);
        return 0;
 }