netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Johannes Berg <johannes@sipsolutions.net>
To: netdev@vger.kernel.org
Cc: Jouni Malinen <jkm@devicescape.com>,
	"John W. Linville" <linville@tuxdriver.com>,
	Jiri Benc <jbenc@suse.cz>,
	Johannes Berg <johannes@sipsolutions.net>
Subject: [PATCH 05/18] d80211:  get rid of WME bitfield
Date: Mon, 21 Aug 2006 09:41:12 +0200	[thread overview]
Message-ID: <20060821075159.740408530@sipsolutions.net> (raw)
In-Reply-To: 20060821074107.648561364@sipsolutions.net

[-- Attachment #1: d80211-wme-bitfield.patch --]
[-- Type: text/plain, Size: 3237 bytes --]

This patch gets rid of the endian-ness dependent bitfield
used for WME. It cleans up wme includes and adds some necessary
includes in other files that got them implicitly through wme.h

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>

--- wireless-dev.orig/net/d80211/wme.c	2006-08-20 14:56:17.418192788 +0200
+++ wireless-dev/net/d80211/wme.c	2006-08-20 14:56:19.058192788 +0200
@@ -12,7 +12,9 @@
 #include <linux/skbuff.h>
 #include <linux/module.h>
 #include <linux/if_arp.h>
+#include <linux/types.h>
 #include <net/ip.h>
+#include <net/pkt_sched.h>
 
 #include <net/d80211.h>
 #include "ieee80211_i.h"
@@ -246,15 +248,15 @@ static int wme_qdiscop_enqueue(struct sk
 	/* now we know the 1d priority, fill in the QoS header if there is one
 	 */
 	if (WLAN_FC_IS_QOS_DATA(fc)) {
-		struct qos_control *qc = (struct qos_control *)
-			(skb->data + ieee80211_get_hdrlen(fc) - 2);
-		u8 *p = (u8 *) qc;
-		*p++ = 0; /* do this due to gcc's lack of optimization on
-			   * bitfield ops */
-		*p = 0;
-		qc->tag1d = skb->priority;
+		u8 *p = skb->data + ieee80211_get_hdrlen(fc) - 2;
+		u8 qos_hdr = skb->priority & QOS_CONTROL_TAG1D_MASK;
 		if (local->wifi_wme_noack_test)
-			qc->ack_policy = 1;
+			qos_hdr |= QOS_CONTROL_ACK_POLICY_NOACK <<
+					QOS_CONTROL_ACK_POLICY_SHIFT;
+		/* qos header is 2 bytes, second reserved */
+		*p = qos_hdr;
+		p++;
+		*p = 0;
 	}
 
 	if (unlikely(queue >= local->hw->queues)) {
--- wireless-dev.orig/net/d80211/wme.h	2006-08-20 14:56:09.068192788 +0200
+++ wireless-dev/net/d80211/wme.h	2006-08-20 14:56:19.058192788 +0200
@@ -11,10 +11,8 @@
 #ifndef _WME_H
 #define _WME_H
 
-#include <asm/byteorder.h>
 #include <linux/netdevice.h>
-#include <linux/types.h>
-#include <net/pkt_sched.h>
+#include "ieee80211_i.h"
 
 #define QOS_CONTROL_LEN 2
 
@@ -24,26 +22,7 @@
 #define QOS_CONTROL_TID_MASK 0x0f
 #define QOS_CONTROL_ACK_POLICY_SHIFT 5
 
-/* This bit field structure should not be used; it can cause compiler to
- * generate unaligned accesses and inefficient code. */
-struct qos_control {
-#if defined(__LITTLE_ENDIAN_BITFIELD)
-	u8	tag1d:3, /* bits 0-2 */
-  		reserved1:1,
-		eosp:1,
-		ack_policy:2,
-  		reserved2:1;
-#elif defined (__BIG_ENDIAN_BITFIELD)
-	u8	reserved2:1,
-		ack_policy:2,
-		eosp:1,
-		reserved1:1,
-		tag1d:3; /* bits 0-2 */
-#else
-#error	"Please fix <asm/byteorder.h>"
-#endif
-	u8 reserved;
-} __attribute__ ((packed));
+#define QOS_CONTROL_TAG1D_MASK 0x07
 
 ieee80211_txrx_result
 ieee80211_rx_h_parse_qos(struct ieee80211_txrx_data *rx);
--- wireless-dev.orig/net/d80211/fifo_qdisc.c	2006-08-20 14:56:09.098192788 +0200
+++ wireless-dev/net/d80211/fifo_qdisc.c	2006-08-20 14:56:19.058192788 +0200
@@ -13,6 +13,7 @@
 #include <linux/config.h>
 #include <linux/version.h>
 #include <linux/netdevice.h>
+#include <net/pkt_sched.h>
 #include <net/d80211.h>
 #include "ieee80211_i.h"
 #include "wme.h"
--- wireless-dev.orig/net/d80211/ieee80211.c	2006-08-20 14:56:18.438192788 +0200
+++ wireless-dev/net/d80211/ieee80211.c	2006-08-20 14:56:19.058192788 +0200
@@ -18,6 +18,7 @@
 #include <linux/etherdevice.h>
 #include <linux/if_arp.h>
 #include <linux/wireless.h>
+#include <linux/rtnetlink.h>
 #include <net/iw_handler.h>
 #include <linux/compiler.h>
 

--

  parent reply	other threads:[~2006-08-21  8:02 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-08-21  7:41 [PATCH 00/18] d80211: various cleanups/fixes/changes Johannes Berg
2006-08-21  7:41 ` [PATCH 01/18] d80211: LED triggers Johannes Berg
2006-08-22  0:30   ` [PATCH 01/3] d80211: add support for SIOCSIWRATE, SIOCSIWTXPOW and SIOCSIWPOWER Mohamed Abbas
2006-08-22  0:36     ` [PATCH 02/3] d80211: iwlist scan Mohamed Abbas
2006-08-23 15:46       ` Jiri Benc
2006-08-28 20:37         ` [PATCH 0/7] d80211: support more wireless command mabbas
2006-08-22  0:38     ` [PATCH 03/3] d80211: adhoc Mohamed Abbas
2006-08-23 15:51       ` Jiri Benc
2006-08-23 15:19     ` [PATCH 01/3] d80211: add support for SIOCSIWRATE, SIOCSIWTXPOW and SIOCSIWPOWER Jiri Benc
2006-08-25 18:37     ` Jouni Malinen
2006-08-25 18:46       ` Mohamed Abbas
2006-08-22 16:54   ` [PATCH 01/18] d80211: LED triggers Jouni Malinen
2006-08-22 18:38     ` Jiri Benc
2006-08-23  7:02     ` Johannes Berg
2006-08-23 18:16   ` Jiri Benc
2006-08-24  7:03     ` Johannes Berg
2006-09-22 11:59   ` Jiri Benc
2006-08-21  7:41 ` [PATCH 02/18] d80211: master link Johannes Berg
2006-08-21  8:13   ` Johannes Berg
2006-08-21 19:08   ` Jiri Benc
2006-08-22  7:49     ` Johannes Berg
2006-08-21  7:41 ` [PATCH 03/18] d80211: pointers as extended booleans Johannes Berg
2006-08-22  6:43   ` Bill Fink
2006-08-22  8:39     ` Johannes Berg
2006-08-21  7:41 ` [PATCH 04/18] d80211: use kzalloc() Johannes Berg
2006-08-21  7:41 ` Johannes Berg [this message]
2006-08-21  7:41 ` [PATCH 06/18] d80211: rework rate control registration Johannes Berg
2006-08-21 19:19   ` Jiri Benc
2006-08-22  8:33     ` Johannes Berg
2006-08-21  7:41 ` [PATCH 07/18] d80211: get rid of sta_aid in favour of keeping track of TIM Johannes Berg
2006-08-22 18:36   ` Jiri Benc
2006-08-23  7:04     ` Johannes Berg
2006-08-23 10:04     ` [PATCH] " Johannes Berg
2006-08-23 10:05       ` Johannes Berg
2006-08-23 10:16     ` [PATCH ] " Johannes Berg
2006-08-21  7:41 ` [PATCH 08/18] d80211: clean up exports Johannes Berg
2006-08-22 16:44   ` Jouni Malinen
2006-08-23  7:01     ` Johannes Berg
2006-08-23 10:03     ` [PATCH] " Johannes Berg
2006-08-21  7:41 ` [PATCH 09/18] d80211: move out rate control registration code Johannes Berg
2006-08-21  7:41 ` [PATCH 10/18] d80211: clean up includes Johannes Berg
2006-08-21  7:41 ` [PATCH 11/18] d80211: clean up qdisc requeue Johannes Berg
2006-08-21 19:31   ` Jiri Benc
2006-08-22  7:48     ` Johannes Berg
2006-08-21  7:41 ` [PATCH 12/18] d80211: fix some sparse warnings Johannes Berg
2006-08-22 18:55   ` Jiri Benc
2006-08-21  7:41 ` [PATCH 13/18] d80211: clean up some coding style issues Johannes Berg
2006-08-21 19:35   ` Jiri Benc
2006-08-22  8:27     ` Johannes Berg
2006-08-21  7:41 ` [PATCH 14/18] d80211: make lowlevel TX framedump option visible Johannes Berg
2006-08-21  7:41 ` [PATCH 15/18] d80211: surface IBSS debug Johannes Berg
2006-08-21  7:41 ` [PATCH 16/18] d80211: get rid of MICHAEL_MIC_HWACCEL define Johannes Berg
2006-08-22 19:00   ` Jiri Benc
2006-08-23  7:05     ` Johannes Berg
2006-08-23  9:46       ` Jiri Benc
2006-08-21  7:41 ` [PATCH 17/18] d80211: surface powersave debug switch Johannes Berg
2006-08-21  7:41 ` [PATCH 18/18] d80211: fix some documentation Johannes Berg

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=20060821075159.740408530@sipsolutions.net \
    --to=johannes@sipsolutions.net \
    --cc=jbenc@suse.cz \
    --cc=jkm@devicescape.com \
    --cc=linville@tuxdriver.com \
    --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).