[PATCH 2.6.24] b43: Reject new firmware early
We must reject new incompatible firmware early to avoid running into strange transmission failures. Signed-off-by: Michael Buesch <[EMAIL PROTECTED]> --- John, this must _only_ be applied to 2.6.24. 2.6.25 does have actual support for the new firmware. Index: wireless-2.6/drivers/net/wireless/b43/main.c === --- wireless-2.6.orig/drivers/net/wireless/b43/main.c 2008-01-16 11:33:55.0 +0100 +++ wireless-2.6/drivers/net/wireless/b43/main.c2008-01-16 11:44:06.0 +0100 @@ -1800,6 +1800,18 @@ static int b43_upload_microcode(struct b err = -EOPNOTSUPP; goto out; } + if (fwrev > 351) { + b43err(dev->wl, "YOUR FIRMWARE IS TOO NEW. Please downgrade your " + "firmware.\n"); + b43err(dev->wl, "Use this firmware tarball: " + "http://downloads.openwrt.org/sources/broadcom-wl-4.80.53.0.tar.bz2\n";); + b43err(dev->wl, "Use this b43-fwcutter tarball: " + "http://bu3sch.de/b43/fwcutter/b43-fwcutter-009.tar.bz2\n";); + b43err(dev->wl, "Read, understand and _do_ what this message says, please.\n"); + b43_write32(dev, B43_MMIO_MACCTL, 0); + err = -EOPNOTSUPP; + goto out; + } b43dbg(dev->wl, "Loading firmware version %u.%u " "(20%.2i-%.2i-%.2i %.2i:%.2i:%.2i)\n", fwrev, fwpatch, -- Greetings Michael. ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
Hi, > 1. Lights stay off on the card, even when it's working and associated. Weird. Especially when the 4318 works. Maybe the SPROM is programmed differently or something. > 2. There seems to be some timing issue with a Netgear WGR614 access point >(but no others I've tried). If I login with gnome, the network-manager >tries to connect automatically to the AP as it knows it from previous >sessions. This fails unless I have first removed the memory of this >ESSID from gconf. > rm -r .gconf/system/networking/wireless/networks// >I can then login, select the ESSID from the list, type in the passphrase >(it's WPA) and it connects fine. > > Attached is a dmesg from the current running system. I booted, logged in, > got the failure as in [2] above, logged out, reloaded the b43 module, > removed the gconf info, then logged in and successfully joined up to the > access point. There's no indication in your log that NM even tries to associate to the network before you reloaded the module so that looks more like a bug in NM or so. johannes signature.asc Description: This is a digitally signed message part ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
Gavin McCullagh wrote: > Hi, > > I'm the proud owner of the following Belkin f5d7011 cardbus card: > > 02:00.0 Network controller [0280]: Broadcom Corporation BCM4306 802.11b/g > Wireless LAN Controller [14e4:4320] (rev 03) > Subsystem: Belkin F5D7011 54g+ Wireless Network card [1799:7011] > Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- > Stepping- SERR- FastB2B- > Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- > SERR- Latency: 64 > Interrupt: pin A routed to IRQ 10 > Region 0: Memory at 2400 (32-bit, non-prefetchable) [size=8K] > > which broadly speaking, works reasonably well. I just updated to the > latest wireless-2.6 kernel and have a couple of issues to report. > > 1. Lights stay off on the card, even when it's working and associated. > > 2. There seems to be some timing issue with a Netgear WGR614 access point >(but no others I've tried). If I login with gnome, the network-manager >tries to connect automatically to the AP as it knows it from previous >sessions. This fails unless I have first removed the memory of this >ESSID from gconf. > rm -r .gconf/system/networking/wireless/networks// >I can then login, select the ESSID from the list, type in the passphrase >(it's WPA) and it connects fine. > > Attached is a dmesg from the current running system. I booted, logged in, > got the failure as in [2] above, logged out, reloaded the b43 module, > removed the gconf info, then logged in and successfully joined up to the > access point. > > I think both of the above have been mentioned on this list before, but I > thought I should report them. If someone would like me to debug this > further, I am open to instructions. > > I also own a bcm4318 Airforce One card which is slightly less stable > (though its lights do work!) so I might send a separate email regarding > that. > > Thanks for all the code and please let me know if I can be of help giving > up debug info for these cards. Please check that modules rfkill and rfkill-input are configured and loaded on your system. In your dmesg output, you are missing lines that look like Registered led device: b43-phy0:tx Registered led device: b43-phy0:rx Registered led device: b43-phy0:radio Are you running the panel applet? On my KDE system, I can right click on that icon in the lower-right hand corner, and select from any of the AP's in my vicinity. Larry ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
Hi, On Wed, 16 Jan 2008, Johannes Berg wrote: > > 1. Lights stay off on the card, even when it's working and associated. > > Weird. Especially when the 4318 works. Maybe the SPROM is programmed > differently or something. I'm not that familiar yet with the code, is there somewhere I should be looking to try to understand this. I've done some kernel coding before, though not on drivers. > There's no indication in your log that NM even tries to associate to the > network before you reloaded the module so that looks more like a bug in > NM or so. I'd go along with that except that it happens on this one AP and not on the number of others I regularly use. For that reason I guessed an issue between the AP and the card. Is there some more verbose info I can get my hands on? Gavin ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
> I'd go along with that except that it happens on this one AP and not on the > number of others I regularly use. For that reason I guessed an issue > between the AP and the card. > > Is there some more verbose info I can get my hands on? Hmm. Maybe run 'iwevent | tee /tmp/iweventlog' from a console while logging in. johannes signature.asc Description: This is a digitally signed message part ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
Hi, On Wed, 16 Jan 2008, Larry Finger wrote: > Please check that modules rfkill and rfkill-input are configured and loaded > on your system. In your dmesg output, you are missing lines that look like > > Registered led device: b43-phy0:tx > Registered led device: b43-phy0:rx > Registered led device: b43-phy0:radio I'll check this properly when I'm at the access point in question, though when I plug in the card here (different location), rfkill was already loaded (I have b43 in /etc/modules), rfkill-input loaded automatically and I got straight on. dmesg | grep " led " produces no output whatsoever. > Are you running the panel applet? On my KDE system, I can right click on > that icon in the lower-right hand corner, and select from any of the AP's > in my vicinity. I'm using gnome, but I am using the gnome-network-manager panel applet, yes. I also get a list of ESSIDs and pick from them. However, if NM has knowledge of a network it has previously been on it seems to try and join that one on its own initiative. When the failure happens (eg if I login and let it jump automatically to the AP in question), the NM applet swirls for a few seconds and then gives up, dropping to the little red exclamation mark symbol saying it has no network. If I then click to pick an ESSID, that ESSID is already selected, but trying again produces the same brief swirl before it gives up. Gavin ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
Hi, On Wed, 16 Jan 2008, Gavin McCullagh wrote: > > Registered led device: b43-phy0:tx > > Registered led device: b43-phy0:rx > > Registered led device: b43-phy0:radio > > I'll check this properly when I'm at the access point in question, Sorry, confused myself here as I rushed. The LEDs fail regardless of AP so the information below is the response. > when I plug in the card here (different location), rfkill was already > loaded (I have b43 in /etc/modules), rfkill-input loaded automatically and > I got straight on. > > dmesg | grep " led " > > produces no output whatsoever. If you need more info on this let me know. Gavin ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
On Wednesday 16 January 2008 17:58:26 Gavin McCullagh wrote: > Hi, > > On Wed, 16 Jan 2008, Larry Finger wrote: > > > Please check that modules rfkill and rfkill-input are configured and loaded > > on your system. In your dmesg output, you are missing lines that look like > > > > Registered led device: b43-phy0:tx > > Registered led device: b43-phy0:rx > > Registered led device: b43-phy0:radio > > I'll check this properly when I'm at the access point in question, though > when I plug in the card here (different location), rfkill was already > loaded (I have b43 in /etc/modules), rfkill-input loaded automatically and > I got straight on. > > dmesg | grep " led " > > produces no output whatsoever. Can you please post the SPROM data? $> cat $(find /sys -name ssb_sprom) -- Greetings Michael. ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
Hi, On Wed, 16 Jan 2008, Michael Buesch wrote: > > dmesg | grep " led " > > > > produces no output whatsoever. > > Can you please post the SPROM data? > $> cat $(find /sys -name ssb_sprom) [EMAIL PROTECTED]:~# cat $(find /sys -name ssb_sprom) 01401170991720430080020002100018811011009250B8FD4730531594FA85FE4C003E00490A01FF10FF023A Gavin ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: issues with bcm4306
On Wednesday 16 January 2008 23:00:09 Gavin McCullagh wrote: > Hi, > > On Wed, 16 Jan 2008, Michael Buesch wrote: > > > > dmesg | grep " led " > > > > > > produces no output whatsoever. > > > > > Can you please post the SPROM data? > > $> cat $(find /sys -name ssb_sprom) > > [EMAIL PROTECTED]:~# cat $(find /sys -name ssb_sprom) > 01401170991720430080020002100018811011009250B8FD4730531594FA85FE4C003E00490A01FF10FF023A So there's no special LEDs information in your SPROM. That means you must get the messages Larry quoted. I'm pretty sure that you didn't enable LEDs support in your kernel. Make sure CONFIG_LEDS_CLASS and CONFIG_MAC80211_LEDS are enabled. While you are at it also make sure CONFIG_RFKILL, CONFIG_RFKILL_INPUT and CONFIG_INPUT_POLLDEV are enabled. -- Greetings Michael. ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev