diff options
| author | Arnd Bergmann <arnd@arndb.de> | 2006-01-05 17:05:29 +0300 |
|---|---|---|
| committer | Paul Mackerras <paulus@samba.org> | 2006-01-09 07:44:57 +0300 |
| commit | 2fb9d2063626374dd8a2514b3a730facac8235d8 (patch) | |
| tree | b410dcdbc5aee656c37951be36951130450549e7 /arch/powerpc/platforms/cell/spu_base.c | |
| parent | aeb013772a2cc85a8d0baffd64977d2888bc781d (diff) | |
| download | linux-2fb9d2063626374dd8a2514b3a730facac8235d8.tar.xz | |
[PATCH] spufs: set irq affinity for running threads
For far, all SPU triggered interrupts always end up on
the first SMT thread, which is a bad solution.
This patch implements setting the affinity to the
CPU that was running last when entering execution on
an SPU. This should result in a significant reduction
in IPI calls and better cache locality for SPE thread
specific data.
Signed-off-by: Arnd Bergmann <arndb@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Diffstat (limited to 'arch/powerpc/platforms/cell/spu_base.c')
| -rw-r--r-- | arch/powerpc/platforms/cell/spu_base.c | 8 |
1 files changed, 8 insertions, 0 deletions
diff --git a/arch/powerpc/platforms/cell/spu_base.c b/arch/powerpc/platforms/cell/spu_base.c index 7fe3fa3da0e9..d75ae03df686 100644 --- a/arch/powerpc/platforms/cell/spu_base.c +++ b/arch/powerpc/platforms/cell/spu_base.c @@ -507,6 +507,14 @@ int spu_irq_class_1_bottom(struct spu *spu) return ret; } +void spu_irq_setaffinity(struct spu *spu, int cpu) +{ + u64 target = iic_get_target_id(cpu); + u64 route = target << 48 | target << 32 | target << 16; + spu_int_route_set(spu, route); +} +EXPORT_SYMBOL_GPL(spu_irq_setaffinity); + static void __iomem * __init map_spe_prop(struct device_node *n, const char *name) { |