Re: Enabling CONFIG_DEBUG_SLAB breakes b43 module

2008-02-12 Thread Larry Finger 
Jörg Sommer wrote:
 Hi,
 
 I'm using the b43 driver to connect to a WPA2 encrypted WLAN. When I
 enable the kernel option CONFIG_DEBUG_SLAB my connection breaks.
 wpa_supplicant can establish the connection and I can make pings or DNS
 queries, but downloading a file kills the connection. The kernel log
 includes these lines:

I am not sure that this is a b43 problem. I tested transfers with my BCM94311/2 
by making a tar 
archive of one of my NFS shares. The NFS host is connected to the switch by a 
100TX connection. The 
802.11g wi-fi network had no other traffic, and the resulting tar file is 
nearly 1.8 GB. The b43 
host is x86_64 and has an AMD Turion 64 X2 CPU at 1.6 GHz.

With SLAB allocation, with or without debugging, the transfer crashed with a 
hung CPU. With SLUB, I
get no failures, with or without debugging enabled. I have no idea if this is a 
kernel regression, 
but the lesson is clear - use SLUB not SLAB.

Although the transfer works, there is a problem that leads to PHY transmission 
errors. I get 10-15
of them during each transfer, but they don't seem to cause any problems.

Larry


___
Bcm43xx-dev mailing list
Bcm43xx-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Re: Enabling CONFIG_DEBUG_SLAB breakes b43 module

2008-02-12 Thread Michael Buesch 
On Tuesday 12 February 2008 12:33:36 Jörg Sommer wrote:
  So ehm. Disable DEBUG_SLAB?
 
 Yes, but I thought this points to an error. Something like usage after
 free. No? I thought those DEBUG options are to find bugs.

Well, you said it breaks down on huge throughput only, right?
I'm not so surprised that it does that. We use a lot of GFP_ATOMIC
and stuff in networking and wireless. So if you enable heavy slab debugging
I'm not that surprised that it doesn't work as well as without it.

Besides this, what should I do? :) If I had a fix, I'd submit it.
If you think there's a bug nevertheless, please debug it. I can't do
anything about this.

-- 
Greetings Michael.
___
Bcm43xx-dev mailing list
Bcm43xx-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Re: Enabling CONFIG_DEBUG_SLAB breakes b43 module

2008-02-12 Thread Jörg Sommer 
Hallo Michael,

Michael Buesch [EMAIL PROTECTED] wrote:
 On Monday 11 February 2008 13:22:14 Jörg Sommer wrote:
 Hi,
 
 I'm using the b43 driver to connect to a WPA2 encrypted WLAN. When I
 enable the kernel option CONFIG_DEBUG_SLAB my connection breaks.
 wpa_supplicant can establish the connection and I can make pings or DNS
 queries, but downloading a file kills the connection.

 So ehm. Disable DEBUG_SLAB?

Yes, but I thought this points to an error. Something like usage after
free. No? I thought those DEBUG options are to find bugs.

Bye, Jörg.
-- 
“Hey, dad, you see how this man can twist his fingers? Amazing, isn't
it?” “No, son, not really. He's been using Emacs for ten years…”
___
Bcm43xx-dev mailing list
Bcm43xx-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Re: Enabling CONFIG_DEBUG_SLAB breakes b43 module

2008-02-11 Thread Michael Buesch 
On Monday 11 February 2008 13:22:14 Jörg Sommer wrote:
 Hi,
 
 I'm using the b43 driver to connect to a WPA2 encrypted WLAN. When I
 enable the kernel option CONFIG_DEBUG_SLAB my connection breaks.
 wpa_supplicant can establish the connection and I can make pings or DNS
 queries, but downloading a file kills the connection.

So ehm. Disable DEBUG_SLAB?

-- 
Greetings Michael.
___
Bcm43xx-dev mailing list
Bcm43xx-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Enabling CONFIG_DEBUG_SLAB breakes b43 module

2008-02-11 Thread Jörg Sommer 
Hi,

I'm using the b43 driver to connect to a WPA2 encrypted WLAN. When I
enable the kernel option CONFIG_DEBUG_SLAB my connection breaks.
wpa_supplicant can establish the connection and I can make pings or DNS
queries, but downloading a file kills the connection. The kernel log
includes these lines:

[  593.513610] eth1: RX protected frame, but have no key
[  593.574095] b43-phy0 debug: Removing Interface type 2
[  593.598874] b43-phy0 debug: Wireless interface stopped
[  593.599135] b43-phy0 debug: DMA-32 0x0200 (RX) max used slots: 4/64
[  593.601390] b43-phy0 debug: DMA-32 0x02A0 (TX) max used slots: 0/128
[  593.618607] b43-phy0 debug: DMA-32 0x0280 (TX) max used slots: 0/128
[  593.626237] b43-phy0 debug: DMA-32 0x0260 (TX) max used slots: 0/128
[  593.634301] b43-phy0 debug: DMA-32 0x0240 (TX) max used slots: 0/128
[  593.642281] b43-phy0 debug: DMA-32 0x0220 (TX) max used slots: 2/128
[  593.651068] b43-phy0 debug: DMA-32 0x0200 (TX) max used slots: 0/128
[  601.021777] b43-phy0 debug: Loading firmware version 351.126 (2006-07-29 
05:54:02)
[  602.751023] b43-phy0 debug: Chip initialized
[  602.760835] b43-phy0 debug: 30-bit DMA initialized
[  602.777127] b43-phy0 debug: Wireless interface started
[  602.823829] b43-phy0 debug: Adding Interface type 2
[  602.828855] ADDRCONF(NETDEV_UP): eth1: link is not ready
[  604.273055] eth1: Initial auth_alg=0
[  604.273082] eth1: authenticate with AP 00:14:1b:60:5d:00
[  604.276600] eth1: RX authentication from 00:14:1b:60:5d:00 (alg=0 
transaction=2 status=0)
[  604.276620] eth1: authenticated
[  604.276629] eth1: associate with AP 00:14:1b:60:5d:00
[  604.280697] eth1: RX AssocResp from 00:14:1b:60:5d:00 (capab=0x431 status=0 
aid=161)
[  604.280720] eth1: associated
[  604.280742] eth1: CTS protection enabled (BSSID=00:14:1b:60:5d:00)
[  604.280752] eth1: switched to short barker preamble (BSSID=00:14:1b:60:5d:00)
[  604.281061] eth1: WMM queue=2 aci=0 acm=0 aifs=3 cWmin=15 cWmax=1023 burst=0
[  604.281072] eth1: WMM queue=3 aci=1 acm=0 aifs=7 cWmin=15 cWmax=1023 burst=0
[  604.281081] eth1: WMM queue=1 aci=2 acm=0 aifs=2 cWmin=7 cWmax=15 burst=30
[  604.281091] eth1: WMM queue=0 aci=3 acm=0 aifs=2 cWmin=3 cWmax=7 burst=15
[  604.283358] ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
[  604.395160] eth1: RX protected frame, but have no key
[  604.396598] eth1: RX protected frame, but have no key
[  604.397609] eth1: RX protected frame, but have no key
[  607.038550] eth1: RX deauthentication from 00:14:1b:60:5d:00 (reason=14)
[  607.038567] eth1: deauthenticated
[  607.083301] ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
[  607.148534] eth1: Initial auth_alg=0
[  607.148560] eth1: authenticate with AP 00:14:1b:60:5d:00
[  607.151549] eth1: RX authentication from 00:14:1b:60:5d:00 (alg=0 
transaction=2 status=0)
[  607.151569] eth1: authenticated
[  607.151578] eth1: associate with AP 00:14:1b:60:5d:00
[  607.166211] eth1: RX ReassocResp from 00:14:1b:60:5d:00 (capab=0x0 status=13 
aid=0)
[  607.166229] eth1: AP denied association (code=13)
[  607.353166] eth1: associate with AP 00:14:1b:60:5d:00
[  607.355885] eth1: RX deauthentication from 00:14:1b:60:5d:00 (reason=2)
[  607.355904] eth1: deauthenticated
[  608.355322] eth1: authenticate with AP 00:14:1b:60:5d:00
[  608.357688] eth1: RX authentication from 00:14:1b:60:5d:00 (alg=0 
transaction=2 status=0)
[  608.357707] eth1: authenticated
[  608.357716] eth1: associate with AP 00:14:1b:60:5d:00
[  608.361568] eth1: RX AssocResp from 00:14:1b:60:5d:00 (capab=0x0 status=13 
aid=0)
[  608.361586] eth1: AP denied association (code=13)
[  608.561470] eth1: associate with AP 00:14:1b:60:5d:00
[  608.569470] eth1: RX deauthentication from 00:14:1b:60:5d:00 (reason=2)
[  608.569489] eth1: deauthenticated
[  609.568954] eth1: authenticate with AP 00:14:1b:60:5d:00
[  609.570574] eth1: RX authentication from 00:14:1b:60:5d:00 (alg=0 
transaction=2 status=0)
[  609.570589] eth1: authenticated
[  609.570598] eth1: associate with AP 00:14:1b:60:5d:00
[  609.572686] eth1: RX AssocResp from 00:14:1b:60:5d:00 (capab=0x0 status=13 
aid=0)
[  609.572698] eth1: AP denied association (code=13)

% uname -a
Linux ibook 2.6.24 #1 Thu Feb 7 10:22:52 CET 2008 ppc GNU/Linux

I've an Airport Extreme Card.
# lspci -s 10:12 -v
0001:10:12.0 Network controller: Broadcom Corporation BCM4306 802.11b/g 
Wireless LAN Controller (rev 03)
Subsystem: Apple Computer Inc. AirPort Extreme
Flags: bus master, fast devsel, latency 16, IRQ 52
Memory at 80084000 (32-bit, non-prefetchable) [size=8K]
Capabilities: [40] Power Management version 2
Kernel driver in use: b43-pci-bridge
Kernel modules: ssb

Bye, Jörg.
-- 
But in the case of git revert, it should be an ancestor (or the user
is just insane, in which case it doesn't matter - insane people can do
insane things)
Linus Torvalds [EMAIL PROTECTED]
___
Bcm43xx-dev mailing list