From: "W. van den Akker" <listsrv@wilsoft.nl>
To: linux-wireless@vger.kernel.org
Subject: Re: Still problems to get ath9k in accessmode. Kernel hangs
Date: Tue, 6 Jan 2009 15:52:45 +0100
References: <200901011444.20333.listsrv@wilsoft.nl> <20090105094102.GA27836@jm.kir.nu> <200901060857.41357.listsrv@wilsoft.nl>
In-Reply-To: <200901060857.41357.listsrv@wilsoft.nl>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Message-Id: <200901061552.45729.listsrv@wilsoft.nl> (sfid-20090106_155311_941857_A18BFEC1)
Sender: linux-wireless-owner@vger.kernel.org

On Tuesday 06 January 2009, you wrote:
> On Monday 05 January 2009, you wrote:
> > On Thu, Jan 01, 2009 at 02:44:20PM +0100, W. van den Akker wrote:
> > > I have tested on the server  with the latest wireless-testing the past
> > > days (2.6.28-rc9-wl). The wireless-testing looks somewhat more stable
> > > then the compat-wireless.
> > >
> > > I have also installed the netconsole module to watch any kernel prints.
> > > After running hostapd (v0.6.6) from command line the AP is
> > > up-and-running. Then if I connect with the laptop (tested it with
> > > kernel .26 and .28 on the laptop) I get a connection and can start a
> > > browser en explore the internet. If I start a second session (Kmail or
> > > second browser) the
> > > server hangs within a minute. Leaving again no trace on the console,
> > > hostapd-log or netconsole.
> >
> > Just to make sure I understood this correctly.. Does this happen with
> > just a single association station (your laptop)? What are you referring
> > to with a "session"? Just another program using network (i.e., no new
> > wireless associations, etc.)? As far as the ath9k driver is concerned,
> > new applications (e.g., new TCP connection) should not really matter
> > much.. If you have IEEE 802.11n enabled, could you please run a test
> > with it disabled to make sure this is not related to frame aggregation?
> >
> > As far as getting any output from the system is concerned, the next step
> > could be to see if NMI watchdog would be able to provide any help.
> > Please take a look at Documentation/nmi_watchdog.txt in the kernel
> > source code and see if you could use that on the AP to get some output
> > when the system hangs (running without X on console to make sure you see
> > the output).
>
> Hi Jouni,
>
> Yes you understood correctly. Only a laptop, start a browser then start
> Kmail over the same wireless connection. Then it hangs. This is always
> reproducable. There are also other sorts or hangups (for example when
> starting hostapd, but that is not always reproducable when it happens).
>
> I will try today the watchdog-suggestion. Hope that it will provide some
> help.

I have enabled the NMI_watchdog (see below)
          CPU0       CPU1
  0:     322372          0   IO-APIC-edge      timer
  1:         12          0   IO-APIC-edge      i8042
  4:        101          0   IO-APIC-edge      serial
  6:          3          0   IO-APIC-edge      floppy
  7:          0          0   IO-APIC-edge      parport0
  8:          2          0   IO-APIC-edge      rtc0
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb1
 17:       6282          0   IO-APIC-fasteoi   eth1
 18:     364152          0   IO-APIC-fasteoi   ata_piix, ata_piix, eth0
 19:          0          0   IO-APIC-fasteoi   uhci_hcd:usb2
 21:       1727          0   IO-APIC-fasteoi   ath
 23:          0          0   IO-APIC-fasteoi   ehci_hcd:usb3
 27:         89          0   IO-APIC-fasteoi   aic79xx
NMI:     322364     322307   Non-maskable interrupts
LOC:          0     322302   Local timer interrupts
RES:      10695      24691   Rescheduling interrupts
CAL:         36        112   Function call interrupts
TLB:       2371       3165   TLB shootdowns
TRM:          0          0   Thermal event interrupts
SPU:          0          0   Spurious interrupts
ERR:          0
MIS:          0

However, after starting a internet session and Kmail session the server hangs 
again without any output. Does the NMI-handler not process per CPU? So if one 
CPU hangs, the other would still proceed?

I have no 802.11n enabled, nor X.

This means I think the system halted hard 'hlt-instruction'

Any further suggestion? I am little bit hopeless now. I have placed the PCI 
card into another slot, no result.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.