2.6.24-rc1: NULL pointer dereference using netconsole

2.6.24-rc1: NULL pointer dereference using netconsole

[Andrew Nelless]

Hi,

I booted up 2.6.24-rc1 this morning [Real early over a brew ;-)] and was
having a problems with multiple ~5 second hangs on SATA/drive init
(Something to do with "EH" something-or-other and resets but I'll email
in separately about it later unless its fixed by the time I get the chance).

Anyway, I went to fire up netconsole to get a decent log dump and hit
across the following nasty. Netconsole works fine in 2.6.23.1 with a
similar config and the same kernel parameters.

A shot of the screen is the only method I could come up with to capture
the log, I hope that is OK, it is pretty readable.


The nasty:
http://andotnet.nfshost.com/linux/2.6.24-rc1-netconsole-nullderef.jpg

The config:
http://andotnet.nfshost.com/linux/config-2.6.24-rc1.txt

The 'old' 2.6.23.1 config:
http://andotnet.nfshost.com/linux/config-2.6.23.1.txt

Netconsole log from successfully booting 2.6.23.1:
http://andotnet.nfshost.com/linux/successful-nclog-2.6.23.1.txt

The hardware, etc:
eth0          : Netgear FA311 10/100Mbps ethernet adapter (netsemi driver)
                connected directly to another PC using a crossover cable.
cpu           : athlon64 3500+
compiler      : gcc 4.2.2
kernel params : root=/dev/sda3 vga=0x31B nmi_watchdog=0
                netconsole=@192.168.2.1/eth0,6666@192.168.2.2/00:11:d8:c0:5e:96


I hope this report is helpful and it's not too early in the release cycle,
I'm still pretty new and not subscribed, so please CC :-)

Btw, is it preferable for non-subscribers to use the kernel bugzilla? Or
post to the list?


Andrew (On his 3rd brew of the day)

[patch] natsemi: fix oops, link back netdevice from private-struct

[Ingo Molnar]

* Andrew Nelless <andrew@nelless.net> wrote:

> Hi,
> 
> I booted up 2.6.24-rc1 this morning [Real early over a brew ;-)] and 
> was having a problems with multiple ~5 second hangs on SATA/drive init 
> (Something to do with "EH" something-or-other and resets but I'll 
> email in separately about it later unless its fixed by the time I get 
> the chance).
> 
> Anyway, I went to fire up netconsole to get a decent log dump and hit 
> across the following nasty. Netconsole works fine in 2.6.23.1 with a 
> similar config and the same kernel parameters.
> 
> A shot of the screen is the only method I could come up with to 
> capture the log, I hope that is OK, it is pretty readable.
> 
> 
> The nasty: 
> http://andotnet.nfshost.com/linux/2.6.24-rc1-netconsole-nullderef.jpg


the NULL dereference is here:

 (gdb) list *0xffffffff804a9504
 0xffffffff804a9504 is in natsemi_poll (drivers/net/natsemi.c:717).
 712             return count;
 713     }
 714
 715     static inline void __iomem *ns_ioaddr(struct net_device *dev)
 716     {
 717             return (void __iomem *) dev->base_addr;
 718     }
 719

which is this code from natsemi.c:

 2227            struct net_device *dev = np->dev;
 2228            void __iomem * ioaddr = ns_ioaddr(dev);
 2229            int work_done = 0;

seems like the NAPI changes in -rc1 added an np->dev field but forgot to 
initialize it ...

does the patch below fix the oops for you?

	Ingo

-------------------->
Subject: natsemi: fix oops, link back netdevice from private-struct
From: Ingo Molnar <mingo@elte.hu>

this commit:

  commit bea3348eef27e6044b6161fd04c3152215f96411
  Author: Stephen Hemminger <shemminger@linux-foundation.org>
  Date:   Wed Oct 3 16:41:36 2007 -0700

      [NET]: Make NAPI polling independent of struct net_device objects.

added np->dev to drivers/net/natsemi.c's struct netdev_private, but
forgot to initialize that new field upon driver init. The result was
a predictable NULL dereference oops the first time the hardware
generated an interrupt.

Reported-by: Andrew Nelless <andrew@nelless.net>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 drivers/net/natsemi.c |    1 +
 1 file changed, 1 insertion(+)

Index: linux/drivers/net/natsemi.c
===================================================================
--- linux.orig/drivers/net/natsemi.c
+++ linux/drivers/net/natsemi.c
@@ -864,6 +864,7 @@ static int __devinit natsemi_probe1 (str
 
 	np = netdev_priv(dev);
 	netif_napi_add(dev, &np->napi, natsemi_poll, 64);
+	np->dev = dev;
 
 	np->pci_dev = pdev;
 	pci_set_drvdata(pdev, dev);

Re: [patch] natsemi: fix oops, link back netdevice from private-struct

[Andrew]

On Wed, October 24, 2007 12:58, Ingo Molnar wrote:
>
> the NULL dereference is here:
>
> (gdb) list *0xffffffff804a9504
> 0xffffffff804a9504 is in natsemi_poll (drivers/net/natsemi.c:717).
> 712             return count;
> 713     }
> 714
> 715     static inline void __iomem *ns_ioaddr(struct net_device *dev)
> 716     {
> 717             return (void __iomem *) dev->base_addr;
> 718     }
> 719
>
>
> which is this code from natsemi.c:
>
> 2227            struct net_device *dev = np->dev;
> 2228            void __iomem * ioaddr = ns_ioaddr(dev);
> 2229            int work_done = 0;
>
>
> seems like the NAPI changes in -rc1 added an np->dev field but forgot to initialize it ...
>
> does the patch below fix the oops for you?
>
> Ingo
>
>

Yep, that got it, thanks.

Re: [patch] natsemi: fix oops, link back netdevice from private-struct

[Jeff Garzik]

Ingo Molnar wrote:
> * Andrew Nelless <andrew@nelless.net> wrote:
> 
>> Hi,
>>
>> I booted up 2.6.24-rc1 this morning [Real early over a brew ;-)] and 
>> was having a problems with multiple ~5 second hangs on SATA/drive init 
>> (Something to do with "EH" something-or-other and resets but I'll 
>> email in separately about it later unless its fixed by the time I get 
>> the chance).
>>
>> Anyway, I went to fire up netconsole to get a decent log dump and hit 
>> across the following nasty. Netconsole works fine in 2.6.23.1 with a 
>> similar config and the same kernel parameters.
>>
>> A shot of the screen is the only method I could come up with to 
>> capture the log, I hope that is OK, it is pretty readable.
>>
>>
>> The nasty: 
>> http://andotnet.nfshost.com/linux/2.6.24-rc1-netconsole-nullderef.jpg
> 
> 
> the NULL dereference is here:
> 
>  (gdb) list *0xffffffff804a9504
>  0xffffffff804a9504 is in natsemi_poll (drivers/net/natsemi.c:717).
>  712             return count;
>  713     }
>  714
>  715     static inline void __iomem *ns_ioaddr(struct net_device *dev)
>  716     {
>  717             return (void __iomem *) dev->base_addr;
>  718     }
>  719
> 
> which is this code from natsemi.c:
> 
>  2227            struct net_device *dev = np->dev;
>  2228            void __iomem * ioaddr = ns_ioaddr(dev);
>  2229            int work_done = 0;
> 
> seems like the NAPI changes in -rc1 added an np->dev field but forgot to 
> initialize it ...
> 
> does the patch below fix the oops for you?
> 
> 	Ingo
> 
> -------------------->
> Subject: natsemi: fix oops, link back netdevice from private-struct
> From: Ingo Molnar <mingo@elte.hu>
> 
> this commit:
> 
>   commit bea3348eef27e6044b6161fd04c3152215f96411
>   Author: Stephen Hemminger <shemminger@linux-foundation.org>
>   Date:   Wed Oct 3 16:41:36 2007 -0700
> 
>       [NET]: Make NAPI polling independent of struct net_device objects.
> 
> added np->dev to drivers/net/natsemi.c's struct netdev_private, but
> forgot to initialize that new field upon driver init. The result was
> a predictable NULL dereference oops the first time the hardware
> generated an interrupt.
> 
> Reported-by: Andrew Nelless <andrew@nelless.net>
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> ---
>  drivers/net/natsemi.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> Index: linux/drivers/net/natsemi.c
> ===================================================================
> --- linux.orig/drivers/net/natsemi.c
> +++ linux/drivers/net/natsemi.c
> @@ -864,6 +864,7 @@ static int __devinit natsemi_probe1 (str
>  
>  	np = netdev_priv(dev);
>  	netif_napi_add(dev, &np->napi, natsemi_poll, 64);
> +	np->dev = dev;
>  
>  	np->pci_dev = pdev;
>  	pci_set_drvdata(pdev, dev);

applied