[PATCH 0/2] iwlwifi fixes for 2.6.33

[PATCH 0/2] iwlwifi fixes for 2.6.33

[Reinette Chatre]

Two fixes are included. The first problem being fixed is to stop data
transmission when mac80211 indicates that station left IBSS. The second fix
is a companion fix to the one already included as "iwl3945: fix panic in
iwl3945 driver" that also addresses the panic. With this patch the panic is
addressed and we still need to figure out why bad data arrives from device.
This patch also targets stable.

Abhijeet Kolekar (1):
  iwlwifi: disable tx on beacon update notification

Zhu Yi (1):
  iwlwifi: fix iwl_queue_used bug when read_ptr == write_ptr

 drivers/net/wireless/iwlwifi/iwl-core.c |   45 +++++++++++++++++-------------
 drivers/net/wireless/iwlwifi/iwl-dev.h  |    2 +-
 2 files changed, 26 insertions(+), 21 deletions(-)

These patches are also available from wireless-2.6 branch on
git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-2.6.git

I had some problems compiling wireless-2.6 as I ran into the problem
described in http://patchwork.kernel.org/patch/68463/ . I assume that the
"CONFIG_KPROBES=n" relies on patch "[PATCH] x86: compile insn.c and inat.c
only for KPROBES", which is not present in wireless-2.6. I am using
gawk. For testing I manually applied this patch and disabled kprobes to be
able to build.

Thank you

Reinette

[PATCH 1/2] iwlwifi: fix iwl_queue_used bug when read_ptr == write_ptr

[Reinette Chatre]

From: Zhu Yi <yi.zhu@intel.com>

When txq read_ptr equals to write_ptr, iwl_queue_used should
always return false. Because there is no used TFD in this case.

This is a complementary fix to the fix already included in commit "iwl3945:
fix panic in iwl3945 driver". Both fixes are needed to address the panic
below.

This problem was discussed on linux-wireless in
http://thread.gmane.org/gmane.linux.kernel.wireless.general/43568

<1>[ 7290.414172] IP: [<ffffffffa0dd53a1>] iwl3945_rx_reply_tx+0xc1/0x450 [iwl3945]
<4>[ 7290.414205] PGD 0
<1>[ 7290.414214] Thread overran stack, or stack corrupted
<0>[ 7290.414229] Oops: 0002 [#1] PREEMPT SMP
<0>[ 7290.414246] last sysfs file: /sys/devices/platform/coretemp.1/temp1_input
<4>[ 7290.414265] CPU 0
<4>[ 7290.414274] Modules linked in: af_packet nfsd usb_storage usb_libusual cpufreq_powersave exportfs cpufreq_conservative iwl3945 nfs cpufreq_userspace snd_hda_codec_realtek acpi_cpufreq uvcvideo lockd iwlcore snd_hda_intel joydev coretemp nfs_acl videodev snd_hda_codec mac80211 v4l1_compat snd_hwdep sbp2 v4l2_compat_ioctl32 uhci_hcd psmouse auth_rpcgss ohci1394 cfg80211 ehci_hcd video ieee1394 snd_pcm serio_raw battery ac nvidia(P) usbcore output sunrpc evdev lirc_ene0100 snd_page_alloc rfkill tg3 libphy fuse lzo lzo_decompress lzo_compress
<6>[ 7290.414486] Pid: 0, comm: swapper Tainted: P           2.6.32-rc8-wl #213 Aspire 5720
<6>[ 7290.414507] RIP: 0010:[<ffffffffa0dd53a1>]  [<ffffffffa0dd53a1>] iwl3945_rx_reply_tx+0xc1/0x450 [iwl3945]
<6>[ 7290.414541] RSP: 0018:ffff880002203d60  EFLAGS: 00010246
<6>[ 7290.414557] RAX: 000000000000004f RBX: ffff880064c11600 RCX: 0000000000000013
<6>[ 7290.414576] RDX: ffffffffa0ddcf20 RSI: ffff8800512b7008 RDI: 0000000000000038
<6>[ 7290.414596] RBP: ffff880002203dd0 R08: 0000000000000000 R09: 0000000000000100
<6>[ 7290.414616] R10: 0000000000000001 R11: 0000000000000000 R12: 00000000000000a0
<6>[ 7290.414635] R13: 0000000000000002 R14: 0000000000000013 R15: 0000000000020201
<6>[ 7290.414655] FS:  0000000000000000(0000) GS:ffff880002200000(0000) knlGS:0000000000000000
<6>[ 7290.414677] CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
<6>[ 7290.414693] CR2: 0000000000000041 CR3: 0000000001001000 CR4: 00000000000006f0
<6>[ 7290.414712] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
<6>[ 7290.414732] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
<4>[ 7290.414752] Process swapper (pid: 0, threadinfo ffffffff81524000, task ffffffff81528b60)
<0>[ 7290.414772] Stack:
<4>[ 7290.414780]  ffff880002203da0 0000000000000046 0000000000000000 0000000000000046
<4>[ 7290.414804] <0> 0000000000000282 0000000000000282 0000000000000282 ffff880064c12010
<4>[ 7290.414830] <0> ffff880002203db0 ffff880064c11600 ffff880064c12e50 ffff8800512b7000
<0>[ 7290.414858] Call Trace:
<0>[ 7290.414867]  <IRQ>
<4>[ 7290.414884]  [<ffffffffa0dc8c47>] iwl3945_irq_tasklet+0x657/0x1740 [iwl3945]
<4>[ 7290.414910]  [<ffffffff8138fc60>] ? _spin_unlock+0x30/0x60
<4>[ 7290.414931]  [<ffffffff81049a21>] tasklet_action+0x101/0x110
<4>[ 7290.414950]  [<ffffffff8104a3d0>] __do_softirq+0xc0/0x160
<4>[ 7290.414968]  [<ffffffff8100d01c>] call_softirq+0x1c/0x30
<4>[ 7290.414986]  [<ffffffff8100eff5>] do_softirq+0x75/0xb0
<4>[ 7290.415003]  [<ffffffff81049ee5>] irq_exit+0x95/0xa0
<4>[ 7290.415020]  [<ffffffff8100e547>] do_IRQ+0x77/0xf0
<4>[ 7290.415038]  [<ffffffff8100c7d3>] ret_from_intr+0x0/0xf
<0>[ 7290.415052]  <EOI>
<4>[ 7290.415067]  [<ffffffff81234efa>] ? acpi_idle_enter_bm+0x270/0x2a5
<4>[ 7290.415087]  [<ffffffff81234f04>] ? acpi_idle_enter_bm+0x27a/0x2a5
<4>[ 7290.415107]  [<ffffffff81234efa>] ? acpi_idle_enter_bm+0x270/0x2a5
<4>[ 7290.415130]  [<ffffffff812c11f3>] ? cpuidle_idle_call+0x93/0xf0
<4>[ 7290.415149]  [<ffffffff8100b0d7>] ? cpu_idle+0xa7/0x110
<4>[ 7290.415168]  [<ffffffff8137b3d5>] ? rest_init+0x75/0x80
<4>[ 7290.415187]  [<ffffffff8158cd0a>] ? start_kernel+0x3a7/0x3b3
<4>[ 7290.415206]  [<ffffffff8158c315>] ? x86_64_start_reservations+0x125/0x129
<4>[ 7290.415227]  [<ffffffff8158c3fd>] ? x86_64_start_kernel+0xe4/0xeb
<0>[ 7290.415243] Code: 00 41 39 ce 0f 8d e8 01 00 00 48 8b 47 40 48 63 d2 48 69 d2 98 00 00 00 4c 8b 04 02 48 c7 c2 20 cf dd a0 49 8d 78 38 49 8d 40 4f <c6> 47 09 00 c6 47 0c 00 c6 47 0f 00 c6 47 12 00 c6 47 15 00 49
<1>[ 7290.415382] RIP  [<ffffffffa0dd53a1>] iwl3945_rx_reply_tx+0xc1/0x450 [iwl3945]
<4>[ 7290.415410]  RSP <ffff880002203d60>
<0>[ 7290.415421] CR2: 0000000000000041
<4>[ 7290.415436] ---[ end trace ec46807277caa515 ]---
<0>[ 7290.415450] Kernel panic - not syncing: Fatal exception in interrupt
<4>[ 7290.415468] Pid: 0, comm: swapper Tainted: P      D    2.6.32-rc8-wl #213
<4>[ 7290.415486] Call Trace:
<4>[ 7290.415495]  <IRQ>  [<ffffffff8138c040>] panic+0x7d/0x13a
<4>[ 7290.415519]  [<ffffffff8101071a>] oops_end+0xda/0xe0
<4>[ 7290.415538]  [<ffffffff8102e1ea>] no_context+0xea/0x250
<4>[ 7290.415557]  [<ffffffff81038991>] ? select_task_rq_fair+0x511/0x780
<4>[ 7290.415578]  [<ffffffff8102e475>] __bad_area_nosemaphore+0x125/0x1e0
<4>[ 7290.415597]  [<ffffffff81038d0c>] ? __enqueue_entity+0x7c/0x80
<4>[ 7290.415616]  [<ffffffff81039201>] ? enqueue_task_fair+0x111/0x150
<4>[ 7290.415636]  [<ffffffff8102e53e>] bad_area_nosemaphore+0xe/0x10
<4>[ 7290.415656]  [<ffffffff8102e8fa>] do_page_fault+0x26a/0x320
<4>[ 7290.415674]  [<ffffffff813905df>] page_fault+0x1f/0x30
<4>[ 7290.415697]  [<ffffffffa0dd53a1>] ? iwl3945_rx_reply_tx+0xc1/0x450 [iwl3945]
<4>[ 7290.415723]  [<ffffffffa0dc8c47>] iwl3945_irq_tasklet+0x657/0x1740 [iwl3945]
<4>[ 7290.415746]  [<ffffffff8138fc60>] ? _spin_unlock+0x30/0x60
<4>[ 7290.415764]  [<ffffffff81049a21>] tasklet_action+0x101/0x110
<4>[ 7290.415783]  [<ffffffff8104a3d0>] __do_softirq+0xc0/0x160
<4>[ 7290.415801]  [<ffffffff8100d01c>] call_softirq+0x1c/0x30
<4>[ 7290.415818]  [<ffffffff8100eff5>] do_softirq+0x75/0xb0
<4>[ 7290.415835]  [<ffffffff81049ee5>] irq_exit+0x95/0xa0
<4>[ 7290.415852]  [<ffffffff8100e547>] do_IRQ+0x77/0xf0
<4>[ 7290.415869]  [<ffffffff8100c7d3>] ret_from_intr+0x0/0xf
<4>[ 7290.415883]  <EOI>  [<ffffffff81234efa>] ? acpi_idle_enter_bm+0x270/0x2a5
<4>[ 7290.415911]  [<ffffffff81234f04>] ? acpi_idle_enter_bm+0x27a/0x2a5
<4>[ 7290.415931]  [<ffffffff81234efa>] ? acpi_idle_enter_bm+0x270/0x2a5
<4>[ 7290.415952]  [<ffffffff812c11f3>] ? cpuidle_idle_call+0x93/0xf0
<4>[ 7290.415971]  [<ffffffff8100b0d7>] ? cpu_idle+0xa7/0x110
<4>[ 7290.415989]  [<ffffffff8137b3d5>] ? rest_init+0x75/0x80
<4>[ 7290.416007]  [<ffffffff8158cd0a>] ? start_kernel+0x3a7/0x3b3
<4>[ 7290.416026]  [<ffffffff8158c315>] ? x86_64_start_reservations+0x125/0x129
<4>[ 7290.416047]  [<ffffffff8158c3fd>] ? x86_64_start_kernel+0xe4/0xeb

Reported-by: Maxim Levitsky <maximlevitsky@gmail.com>
Tested-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
CC: stable@kernel.org
---
 drivers/net/wireless/iwlwifi/iwl-dev.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/wireless/iwlwifi/iwl-dev.h b/drivers/net/wireless/iwlwifi/iwl-dev.h
index 165d1f6..3822cf5 100644
--- a/drivers/net/wireless/iwlwifi/iwl-dev.h
+++ b/drivers/net/wireless/iwlwifi/iwl-dev.h
@@ -711,7 +711,7 @@ extern void iwl_txq_ctx_stop(struct iwl_priv *priv);
 extern int iwl_queue_space(const struct iwl_queue *q);
 static inline int iwl_queue_used(const struct iwl_queue *q, int i)
 {
-	return q->write_ptr > q->read_ptr ?
+	return q->write_ptr >= q->read_ptr ?
 		(i >= q->read_ptr && i < q->write_ptr) :
 		!(i < q->read_ptr && i >= q->write_ptr);
 }
-- 
1.5.6.3

[PATCH 2/2] iwlwifi: disable tx on beacon update notification

[Reinette Chatre]

From: Abhijeet Kolekar <abhijeet.kolekar@intel.com>

On beacon change update notification from mac we are not disabling
the tx in adhoc mode. Mac sends BSS_CHANGED_BEACON_ENABLED when
station leaves IBSS. Driver should indicate uCode to not to send
anything on receiving this notification.

Functionality to indicate uCode is duplicated across
two notifications so created a common function called iwl_set_no_assoc.

Fix the issue at
http://bugzilla.intellinuxwireless.org/show_bug.cgi?id=2133.

Signed-off-by: Abhijeet Kolekar <abhijeet.kolekar@intel.com>
Tested-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
---
 drivers/net/wireless/iwlwifi/iwl-core.c |   45 +++++++++++++++++-------------
 1 files changed, 25 insertions(+), 20 deletions(-)

diff --git a/drivers/net/wireless/iwlwifi/iwl-core.c b/drivers/net/wireless/iwlwifi/iwl-core.c
index 574d366..5461f10 100644
--- a/drivers/net/wireless/iwlwifi/iwl-core.c
+++ b/drivers/net/wireless/iwlwifi/iwl-core.c
@@ -2344,6 +2344,21 @@ static void iwl_ht_conf(struct iwl_priv *priv,
 	IWL_DEBUG_MAC80211(priv, "leave\n");
 }
 
+static inline void iwl_set_no_assoc(struct iwl_priv *priv)
+{
+	priv->assoc_id = 0;
+	iwl_led_disassociate(priv);
+	/*
+	 * inform the ucode that there is no longer an
+	 * association and that no more packets should be
+	 * sent
+	 */
+	priv->staging_rxon.filter_flags &=
+		~RXON_FILTER_ASSOC_MSK;
+	priv->staging_rxon.assoc_id = 0;
+	iwlcore_commit_rxon(priv);
+}
+
 #define IWL_DELAY_NEXT_SCAN_AFTER_ASSOC (HZ*6)
 void iwl_bss_info_changed(struct ieee80211_hw *hw,
 			  struct ieee80211_vif *vif,
@@ -2475,20 +2490,8 @@ void iwl_bss_info_changed(struct ieee80211_hw *hw,
 					IWL_DELAY_NEXT_SCAN_AFTER_ASSOC;
 			if (!iwl_is_rfkill(priv))
 				priv->cfg->ops->lib->post_associate(priv);
-		} else {
-			priv->assoc_id = 0;
-			iwl_led_disassociate(priv);
-
-			/*
-			 * inform the ucode that there is no longer an
-			 * association and that no more packets should be
-			 * send
-			 */
-			priv->staging_rxon.filter_flags &=
-				~RXON_FILTER_ASSOC_MSK;
-			priv->staging_rxon.assoc_id = 0;
-			iwlcore_commit_rxon(priv);
-		}
+		} else
+			iwl_set_no_assoc(priv);
 	}
 
 	if (changes && iwl_is_associated(priv) && priv->assoc_id) {
@@ -2503,12 +2506,14 @@ void iwl_bss_info_changed(struct ieee80211_hw *hw,
 		}
 	}
 
-	if ((changes & BSS_CHANGED_BEACON_ENABLED) &&
-	    vif->bss_conf.enable_beacon) {
-		memcpy(priv->staging_rxon.bssid_addr,
-		       bss_conf->bssid, ETH_ALEN);
-		memcpy(priv->bssid, bss_conf->bssid, ETH_ALEN);
-		iwlcore_config_ap(priv);
+	if (changes & BSS_CHANGED_BEACON_ENABLED) {
+		if (vif->bss_conf.enable_beacon) {
+			memcpy(priv->staging_rxon.bssid_addr,
+			       bss_conf->bssid, ETH_ALEN);
+			memcpy(priv->bssid, bss_conf->bssid, ETH_ALEN);
+			iwlcore_config_ap(priv);
+		} else
+			iwl_set_no_assoc(priv);
 	}
 
 	mutex_unlock(&priv->mutex);
-- 
1.5.6.3

Re: [PATCH 0/2] iwlwifi fixes for 2.6.33

[reinette chatre]

On Fri, 2010-01-08 at 10:04 -0800, Chatre, Reinette wrote:
> I had some problems compiling wireless-2.6 as I ran into the problem
> described in http://patchwork.kernel.org/patch/68463/ . I assume that the
> "CONFIG_KPROBES=n" relies on patch "[PATCH] x86: compile insn.c and inat.c
> only for KPROBES", which is not present in wireless-2.6. I am using
> gawk. For testing I manually applied this patch and disabled kprobes to be
> able to build.

This problem went away after I upgraded gawk from 3.1.6 to 3.1.7.

Reinette

[PATCH 0/2] iwlwifi fixes for 2.6.33

[Reinette Chatre]

The license was omitted from some new files, fix it here. We also fix a
data transfer stall issue on 5000 series hardware. An equivalent fix has
already been submitted for 1000 and 6000 series hardware.

These patches are also available from wireless-2.6 branch on
git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-2.6.git

Reinette Chatre (1):
  iwlwifi: add license to tracing files

Wey-Yi Guy (1):
  iwlwifi: Fix throughput stall issue in HT mode for 5000

 drivers/net/wireless/iwlwifi/iwl-5000.c     |    5 +++++
 drivers/net/wireless/iwlwifi/iwl-devtrace.c |   26 ++++++++++++++++++++++++++
 drivers/net/wireless/iwlwifi/iwl-devtrace.h |   26 ++++++++++++++++++++++++++
 3 files changed, 57 insertions(+), 0 deletions(-)