summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorHorst Hartmann <horsth@linux.vnet.ibm.com>2011-02-17 13:13:59 +0100
committerMartin Schwidefsky <sky@mschwide.boeblingen.de.ibm.com>2011-02-17 13:13:59 +0100
commitdd30ac327481f1663b6526fa9fa078569188b7cc (patch)
tree7e78bdbafc46bf0e70b89857c29c102676693d74
parent[S390] atomic: use inline asm (diff)
downloadlinux-dd30ac327481f1663b6526fa9fa078569188b7cc.tar.xz
linux-dd30ac327481f1663b6526fa9fa078569188b7cc.zip
[S390] net: provide architecture specific NET_SKB_PAD
NET_SKB_PAD has been increased from 32 to 64 and later to max(32, L1_CACHE_BYTES). This led to a 25% throughput decrease for streaming workloads accompanied by a 37% CPU cost increase on s390. Define a architecture specific NET_SKB_PAD with the old value of 32. Signed-off-by: Horst Hartmann <horsth@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
-rw-r--r--arch/s390/include/asm/cache.h1
1 files changed, 1 insertions, 0 deletions
diff --git a/arch/s390/include/asm/cache.h b/arch/s390/include/asm/cache.h
index 24aafa68b643..2a30d5ac0667 100644
--- a/arch/s390/include/asm/cache.h
+++ b/arch/s390/include/asm/cache.h
@@ -13,6 +13,7 @@
#define L1_CACHE_BYTES 256
#define L1_CACHE_SHIFT 8
+#define NET_SKB_PAD 32
#define __read_mostly __attribute__((__section__(".data..read_mostly")))