linux-wireless.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [RFC/T V2] b43: Fix rfkill radio LED
@ 2007-12-10 17:23 Larry Finger
  2007-12-10 18:16 ` Michael Buesch
  0 siblings, 1 reply; 3+ messages in thread
From: Larry Finger @ 2007-12-10 17:23 UTC (permalink / raw)
  To: Michael Buesch; +Cc: Bcm43xx-dev, linux-wireless

Since addition of the rfkill callback, the LED associated with the off
switch on the radio has not worked for several reasons:

(1) Essential data in the rfkill structure were missing.
(2) The rfkill structure was initialized after the LED initialization.
(3) There was a minor memory leak if the radio LED structure was inited.

Once the above problems were fixed, additional difficulties were noted:

(4) The radio LED was in the wrong state at startup.
(5) The radio switch had to be manipulated twice for each state change.
(6) A circular mutex locking situation existed.

This patch fixes all of the above.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
---

Michael,

This version of the rfkill switch patch is pretty straight forward. Please
comment on the dropping of wl->mutex before rfkill initialization. This is
the only way I could avoid the circular locking without a much earlier
rfkill initialization.

One "gotcha" with this patch is that module rfkill-input is not automatically
loaded - a separate modprobe command must be used. I have sent a message to Ivo
and Dmitry regarding this issue, and hope to resolve it before the final patch
is submitted.

You will also note that this version calls b43_led_turn_on to initialize the
LED associated with the rfkill switch. I think this version avoids the
criticism that you had for an earlier version - this one only does the call
if a led_radio device has been setup. I need to do this as I could find no
reliable way to get the LED on at driver load time.

Larry
---

 leds.c   |    5 +++--
 leds.h   |    2 ++
 main.c   |   24 +++++++++++++++---------
 rfkill.c |   14 ++++++++++++--
 4 files changed, 32 insertions(+), 13 deletions(-)

Index: wireless-2.6/drivers/net/wireless/b43/rfkill.c
===================================================================
--- wireless-2.6.orig/drivers/net/wireless/b43/rfkill.c
+++ wireless-2.6/drivers/net/wireless/b43/rfkill.c
@@ -60,8 +60,12 @@ static void b43_rfkill_poll(struct input
 	}
 	mutex_unlock(&wl->mutex);
 
-	if (unlikely(report_change))
-		input_report_key(poll_dev->input, KEY_WLAN, enabled);
+	/* send the radio switch event to the system - note both a key press
+	 * and a release are required */
+	if (unlikely(report_change)) {
+		input_report_key(poll_dev->input, KEY_WLAN, 1);
+		input_report_key(poll_dev->input, KEY_WLAN, 0);
+	}
 }
 
 /* Called when the RFKILL toggled in software. */
@@ -133,6 +137,12 @@ void b43_rfkill_init(struct b43_wldev *d
 	rfk->poll_dev->poll = b43_rfkill_poll;
 	rfk->poll_dev->poll_interval = 1000; /* msecs */
 
+	rfk->poll_dev->input->name = rfk->name;
+	rfk->poll_dev->input->id.bustype = BUS_HOST;
+	rfk->poll_dev->input->id.vendor = dev->dev->bus->boardinfo.vendor;
+	rfk->poll_dev->input->evbit[0] = BIT(EV_KEY);
+	set_bit(KEY_WLAN, rfk->poll_dev->input->keybit);
+
 	err = rfkill_register(rfk->rfkill);
 	if (err)
 		goto err_free_polldev;
Index: wireless-2.6/drivers/net/wireless/b43/main.c
===================================================================
--- wireless-2.6.orig/drivers/net/wireless/b43/main.c
+++ wireless-2.6/drivers/net/wireless/b43/main.c
@@ -2165,7 +2165,6 @@ static void b43_mgmtframe_txantenna(stru
 static void b43_chip_exit(struct b43_wldev *dev)
 {
 	b43_radio_turn_off(dev, 1);
-	b43_leds_exit(dev);
 	b43_gpio_cleanup(dev);
 	/* firmware is released later */
 }
@@ -2193,11 +2192,10 @@ static int b43_chip_init(struct b43_wlde
 	err = b43_gpio_init(dev);
 	if (err)
 		goto out;	/* firmware is released later */
-	b43_leds_init(dev);
 
 	err = b43_upload_initvals(dev);
 	if (err)
-		goto err_leds_exit;
+		goto err_gpio_clean;
 	b43_radio_turn_on(dev);
 
 	b43_write16(dev, 0x03E6, 0x0000);
@@ -2279,8 +2277,7 @@ out:
 
 err_radio_off:
 	b43_radio_turn_off(dev, 1);
-err_leds_exit:
-	b43_leds_exit(dev);
+err_gpio_clean:
 	b43_gpio_cleanup(dev);
 	return err;
 }
@@ -2718,7 +2715,7 @@ static int b43_antenna_from_ieee80211(u8
 static int b43_op_config(struct ieee80211_hw *hw, struct ieee80211_conf *conf)
 {
 	struct b43_wl *wl = hw_to_b43_wl(hw);
-	struct b43_wldev *dev;
+	struct b43_wldev *uninitialized_var(dev);
 	struct b43_phy *phy;
 	unsigned long flags;
 	unsigned int new_phymode = 0xFFFF;
@@ -3305,6 +3302,7 @@ static void b43_wireless_core_exit(struc
 	b43_rfkill_exit(dev);
 	mutex_lock(&dev->wl->mutex);
 
+	b43_leds_exit(dev);
 	b43_rng_exit(dev->wl);
 	b43_pio_free(dev);
 	b43_dma_free(dev);
@@ -3426,12 +3424,20 @@ static int b43_wireless_core_init(struct
 	memset(wl->mac_addr, 0, ETH_ALEN);
 	b43_upload_card_macaddress(dev);
 	b43_security_init(dev);
-	b43_rfkill_init(dev);
 	b43_rng_init(wl);
-
 	b43_set_status(dev, B43_STAT_INITIALIZED);
 
-      out:
+	mutex_unlock(&dev->wl->mutex);
+	b43_rfkill_init(dev);
+	mutex_lock(&dev->wl->mutex);
+	b43_leds_init(dev);
+	/* if an LED is associated with the rfkill switch,
+	 * and the hardware radio enable switch is on, now is the
+	 * time to turn that LED on */
+	if (dev->led_radio.dev && dev->radio_hw_enable)
+		b43_led_turn_on(dev, dev->led_radio.index,
+				dev->led_radio.activelow);
+out:
 	return err;
 
       err_chip_exit:
Index: wireless-2.6/drivers/net/wireless/b43/leds.c
===================================================================
--- wireless-2.6.orig/drivers/net/wireless/b43/leds.c
+++ wireless-2.6/drivers/net/wireless/b43/leds.c
@@ -30,7 +30,7 @@
 #include "leds.h"
 
 
-static void b43_led_turn_on(struct b43_wldev *dev, u8 led_index,
+void b43_led_turn_on(struct b43_wldev *dev, u8 led_index,
 			    bool activelow)
 {
 	struct b43_wl *wl = dev->wl;
@@ -232,4 +232,5 @@ void b43_leds_exit(struct b43_wldev *dev
 	b43_unregister_led(&dev->led_tx);
 	b43_unregister_led(&dev->led_rx);
 	b43_unregister_led(&dev->led_assoc);
+	b43_unregister_led(&dev->led_radio);
 }
Index: wireless-2.6/drivers/net/wireless/b43/leds.h
===================================================================
--- wireless-2.6.orig/drivers/net/wireless/b43/leds.h
+++ wireless-2.6/drivers/net/wireless/b43/leds.h
@@ -44,6 +44,8 @@ enum b43_led_behaviour {
 
 void b43_leds_init(struct b43_wldev *dev);
 void b43_leds_exit(struct b43_wldev *dev);
+void b43_led_turn_on(struct b43_wldev *dev, u8 led_index,
+			    bool activelow);
 
 
 #else /* CONFIG_B43_LEDS */

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [RFC/T V2] b43: Fix rfkill radio LED
  2007-12-10 17:23 [RFC/T V2] b43: Fix rfkill radio LED Larry Finger
@ 2007-12-10 18:16 ` Michael Buesch
  2007-12-10 19:22   ` Larry Finger
  0 siblings, 1 reply; 3+ messages in thread
From: Michael Buesch @ 2007-12-10 18:16 UTC (permalink / raw)
  To: Larry Finger; +Cc: Bcm43xx-dev, linux-wireless

On Monday 10 December 2007 18:23:21 Larry Finger wrote:
> This version of the rfkill switch patch is pretty straight forward. Please
> comment on the dropping of wl->mutex before rfkill initialization. This is
> the only way I could avoid the circular locking without a much earlier
> rfkill initialization.

I think that's not acceptable, as it introduces a nasty (although unlikely)
race condition with the band switch. I will think about it and will fix
it myself then. If there's no way to properly fix it, I think it may also
be OK to live with this damn unlikely race.

Please also add the call to request_module() into the rfkill init and test
if it works properly. After that, please send the complete patch back to me
and I will try to fix the locking issue.

Anyway, this patch looks nice. Good work.

-- 
Greetings Michael.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [RFC/T V2] b43: Fix rfkill radio LED
  2007-12-10 18:16 ` Michael Buesch
@ 2007-12-10 19:22   ` Larry Finger
  0 siblings, 0 replies; 3+ messages in thread
From: Larry Finger @ 2007-12-10 19:22 UTC (permalink / raw)
  To: Michael Buesch; +Cc: Bcm43xx-dev, linux-wireless

Michael Buesch wrote:
> 
> I think that's not acceptable, as it introduces a nasty (although unlikely)
> race condition with the band switch. I will think about it and will fix
> it myself then. If there's no way to properly fix it, I think it may also
> be OK to live with this damn unlikely race.
> 
> Please also add the call to request_module() into the rfkill init and test
> if it works properly. After that, please send the complete patch back to me
> and I will try to fix the locking issue.

The request_module() call is in the version I just sent you. If rfkill-input is a module, it loads
correctly. If the code is built-in or if the code was not built at al, it generates an error;
therefore, my error message is a little ambiguous. If I use the CONFIG variables, I could generate
really proper messages, but I decided to skip that part.

Larry


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2007-12-10 19:22 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-12-10 17:23 [RFC/T V2] b43: Fix rfkill radio LED Larry Finger
2007-12-10 18:16 ` Michael Buesch
2007-12-10 19:22   ` Larry Finger

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).