netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: frank.blaschka@de.ibm.com
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, linux-s390@vger.kernel.org,
	Horst Hartmann <horsth@linux.vnet.ibm.com>
Subject: [patch 3/7] [PATCH] net,s390: provide architecture specific NET_SKB_PAD
Date: Tue, 01 Feb 2011 09:16:50 +0100	[thread overview]
Message-ID: <20110201081723.562745244@de.ibm.com> (raw)
In-Reply-To: 20110201081647.173934635@de.ibm.com

[-- Attachment #1: 606-net-skb-pad.diff --]
[-- Type: text/plain, Size: 1352 bytes --]

From: Horst Hartmann <horsth@linux.vnet.ibm.com>

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.
In order to fix this provide an architecture specific NET_SKB_PAD config symbol.

Signed-off-by: Horst Hartmann <horsth@linux.vnet.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
---


 arch/s390/Kconfig      |    4 ++++
 include/linux/skbuff.h |    4 +++-
 2 files changed, 7 insertions(+), 1 deletion(-)
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -49,6 +49,10 @@ config NO_DMA
 config ARCH_DMA_ADDR_T_64BIT
 	def_bool 64BIT
 
+config NET_SKB_PAD
+	int
+	default "32"
+
 config GENERIC_LOCKBREAK
 	def_bool y if SMP && PREEMPT
 
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -1431,8 +1431,10 @@ static inline int pskb_network_may_pull(
  * get_rps_cpus() for example only access one 64 bytes aligned block :
  * NET_IP_ALIGN(2) + ethernet_header(14) + IP_header(20/40) + ports(8)
  */
-#ifndef NET_SKB_PAD
+#ifndef CONFIG_NET_SKB_PAD
 #define NET_SKB_PAD	max(32, L1_CACHE_BYTES)
+#else
+#define NET_SKB_PAD	CONFIG_NET_SKB_PAD
 #endif
 
 extern int ___pskb_trim(struct sk_buff *skb, unsigned int len);


  parent reply	other threads:[~2011-02-01  8:18 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-02-01  8:16 [patch 0/7] s390: network patches for net-next frank.blaschka
2011-02-01  8:16 ` [patch 1/7] [PATCH] qeth: show new mac-address if its setting fails frank.blaschka
2011-02-01  8:16 ` [patch 2/7] [PATCH] qeth: add more strict MTU checking frank.blaschka
2011-02-01  8:16 ` frank.blaschka [this message]
2011-02-01 22:17   ` [patch 3/7] [PATCH] net,s390: provide architecture specific NET_SKB_PAD David Miller
2011-02-01  8:16 ` [patch 4/7] [PATCH] qeth: allow HiperSockets framesize change in suspend frank.blaschka
2011-02-01  8:16 ` [patch 5/7] [PATCH] qeth: allow OSA CHPARM change in suspend state frank.blaschka
2011-02-01  8:16 ` [patch 6/7] [PATCH] s390: Fix wrong size in memcmp (netiucv) frank.blaschka
2011-02-01  8:16 ` [patch 7/7] [PATCH] s390: Fix possibly wrong size in strncmp (smsgiucv) frank.blaschka

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20110201081723.562745244@de.ibm.com \
    --to=frank.blaschka@de.ibm.com \
    --cc=davem@davemloft.net \
    --cc=horsth@linux.vnet.ibm.com \
    --cc=linux-s390@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).