Re: 2.6.19-rc5-mm2 -- bcm43xx busted (backing out the bcm43xx patches fixes it)

Re: 2.6.19-rc5-mm2 -- bcm43xx busted (backing out the bcm43xx patches fixes it)

[Larry Finger]

Rafael J. Wysocki wrote:
> On Wednesday, 15 November 2006 00:21, Miles Lane wrote:
>> Hello,
>>
>> The last three MM kernels have fail to give me a working bcm43xx driver.
>> The odd thing is that dmesg output seems to indicate that the driver
>> is working okay.  NetworkManager doesn't see the driver, though.
>> "iwlist scan" fails to find any access points, too.  iwconfig shows
>> "Access Point: invalid".
> 
> I can confirm the symptoms, I see them too on my test boxes.
> 
>> I tried backing out the following patches, and it fixes the problem:
>>
>> drivers/net/wireless/bcm43xx/bcm43xx.h
>> drivers/net/wireless/bcm43xx/bcm43xx_main.c
>> drivers/net/wireless/bcm43xx/bcm43xx_power.c
>> drivers/net/wireless/bcm43xx/bcm43xx_wx.c
>>> drivers/net/wireless/bcm43xx/bcm43xx_xmit.c 

The missing patch is shown below. This patch was entitled "[PATCH] bcm43xx: Readd dropped
assignment" and submitted to wireless-2.6 by Daniel Drake on 10/17/06, but it seems to have fallen
through the cracks. It is a fix to a patch entitled "[PATCH] ieee80211: Move IV/ICV stripping into
ieee80211_rx" also submitted by Daniel Drake on 9/26/2006.

NOTE to maintainers: This problem affects BOTH wireless-2.6 and 2.6.19-rcX-mmY. At present, the
"Move IV/ICV" patch has not been incorporated into 2.6.19-rcX and it is OK.

Larry


In the patch sent by Daniel Drake under the title "[PATCH] ieee80211: Move
IV/ICV stripping into ieee80211_rx", a needed line was accidentally removed.
As my current copy of wireless-2.6.git does not contain this line, I am
(re)submitting a patch to restore that line.

Signed-off-by: Larry Finger <Larry.Finger-tQ5ms3gMjBLk1uMJSBkQmQ@public.gmane.org>
---

Index: wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
===================================================================
--- wireless-2.6.orig/drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
+++ wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
@@ -543,6 +543,7 @@ int bcm43xx_rx(struct bcm43xx_private *b
  		break;
  	}

+	frame_ctl = le16_to_cpu(wlhdr->frame_ctl);
  	switch (WLAN_FC_GET_TYPE(frame_ctl)) {
  	case IEEE80211_FTYPE_MGMT:
  		ieee80211_rx_mgt(bcm->ieee, wlhdr, &stats);

Re: 2.6.19-rc5-mm2 -- bcm43xx busted (backing out the bcm43xx patches fixes it)

[John W. Linville]

On Wed, Nov 15, 2006 at 08:16:37PM -0600, Larry Finger wrote:

> NOTE to maintainers: This problem affects BOTH wireless-2.6 and 
> 2.6.19-rcX-mmY. At present, the
> "Move IV/ICV" patch has not been incorporated into 2.6.19-rcX and it is OK.

This patch is available in the 'upstream' branch pull request I posted
yesterday evening:

	http://marc.theaimsgroup.com/?l=linux-netdev&m=116355611923558&w=2

Hth!

John
-- 
John W. Linville
linville@tuxdriver.com

Re: 2.6.19-rc5-mm2 -- bcm43xx busted (backing out the bcm43xx patches fixes it)

[Miles Lane]

On 11/15/06, Larry Finger <Larry.Finger@lwfinger.net> wrote:
> Rafael J. Wysocki wrote:
> > On Wednesday, 15 November 2006 00:21, Miles Lane wrote:
> >> Hello,
> >>
> >> The last three MM kernels have fail to give me a working bcm43xx driver.
> >> The odd thing is that dmesg output seems to indicate that the driver
> >> is working okay.  NetworkManager doesn't see the driver, though.
> >> "iwlist scan" fails to find any access points, too.  iwconfig shows
> >> "Access Point: invalid".
> >
> > I can confirm the symptoms, I see them too on my test boxes.
> >
> >> I tried backing out the following patches, and it fixes the problem:
> >>
> >> drivers/net/wireless/bcm43xx/bcm43xx.h
> >> drivers/net/wireless/bcm43xx/bcm43xx_main.c
> >> drivers/net/wireless/bcm43xx/bcm43xx_power.c
> >> drivers/net/wireless/bcm43xx/bcm43xx_wx.c
> >>> drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
>
> The missing patch is shown below. This patch was entitled "[PATCH] bcm43xx: Readd dropped
> assignment" and submitted to wireless-2.6 by Daniel Drake on 10/17/06, but it seems to have fallen
> through the cracks. It is a fix to a patch entitled "[PATCH] ieee80211: Move IV/ICV stripping into
> ieee80211_rx" also submitted by Daniel Drake on 9/26/2006.
>
> NOTE to maintainers: This problem affects BOTH wireless-2.6 and 2.6.19-rcX-mmY. At present, the
> "Move IV/ICV" patch has not been incorporated into 2.6.19-rcX and it is OK.
>
> Larry
>
>
> In the patch sent by Daniel Drake under the title "[PATCH] ieee80211: Move
> IV/ICV stripping into ieee80211_rx", a needed line was accidentally removed.
> As my current copy of wireless-2.6.git does not contain this line, I am
> (re)submitting a patch to restore that line.
>
> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
> ---
>
> Index: wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
> ===================================================================
> --- wireless-2.6.orig/drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
> +++ wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
> @@ -543,6 +543,7 @@ int bcm43xx_rx(struct bcm43xx_private *b
>                 break;
>         }
>
> +       frame_ctl = le16_to_cpu(wlhdr->frame_ctl);
>         switch (WLAN_FC_GET_TYPE(frame_ctl)) {
>         case IEEE80211_FTYPE_MGMT:
>                 ieee80211_rx_mgt(bcm->ieee, wlhdr, &stats);
>
>

Thanks!  I have verified that this patch solves the problem for me.

      Miles