Panic With Large Network Copy

[Scott Willson]

On Jun 4, 2007, at 10:49 AM, Scott Willson wrote:

>
> On May 29, 2007, at 4:26 PM, Kris Kennaway wrote:
>
>> On Tue, May 29, 2007 at 03:36:49PM -0700, Scott Willson wrote:
>>> I am seeing hard (often no core dump) crashes on a new AMD64 box
>>> running 6.2 RELEASE. When I try to rsync 10+ GB of backup files to
>>> the new box, I can reliably crash it after about 20 minutes; often
>>> quicker if I do something else intensive at the same time, like
>>> compile MySQL. Here are the box specs:
>>> ASUS M2NPV-VM motherboard
>>> AMD A64 3800+ 2.4G CPU
>>> ...
>
>>> Most times, I don't even get a core dump. Here's one I did get:
>>> panic: double fault
>>> ...
>
>>> #9  0xffffffff804371f0 in m_freem (mb=0x0) at uma.h:303
>>> #10 0xffffffff80634125 in nve_ospackettx (ctx=0xffffff00798aac00,
>>> id=0xffffffffb19ea6d0, success=0) at /usr/src/sys/dev/nve/if_nve.c: 
>>> 1551
>>
>> This looks like a nve driver bug to me.  You may wish to try the  
>> nfe driver.
>>
>> Kris
>
> OK, my box is running nicely now. The nfe driver was indeed a good  
> idea, thanks! Here are the details if anyone else has similar  
> problems.
>
> 10baseT hub + nve = kernal panics under high load
> This is the default FreeBSD 6.2 RELEASE configuration.
>
> 10baseT hub + nfe + e100phy patch = errors under high load (tx v2  
> error 0x6204<UNDERFLOW>, watchdog timeout)
> http://www.se.hiroshima-u.ac.jp/~shigeaki/software/freebsd-nfe.html
> This is a replacement driver + recommended path for my hardware. No  
> panics, but many errors.
>
> 10baseT hub + nfe with no patches = errors under high load (tx v2  
> error 0x6204<UNDERFLOW>, watchdog timeout)
>
> 10/100/1000baseT switch + nfe + e100phy patch = errors under high  
> load (tx v2 error 0x6204<UNDERFLOW>, watchdog timeout)
>
> 10/100/1000baseT switch + nfe = No errors!
> This is a new switch and the nfe driver with no patch. In dmesg, I  
> see 'ukphy0' when I boot.
>
> So, as you may have surmised, my motherboard + an old 10baseT hub  
> doesn't work right with any driver. I replaced my very old hub with  
> a new switch, and I am now running the nfe driver with ukphy0. This  
> combination works great.

Well, turns out after all that, the root cause was something else  
again. Fiddling with the driver and the switch helped matters, but I  
still experienced random drops and warnings. Once I installed Gnome,  
the system began to bomb regularly with oversized frames.

Turns out that both the USB controller and the NIC were on the same  
IRQ. I'm not a hardware engineer (obviously) but it seems that  
software that probed the USB ports would cause problems for the  
Ethernet NIC.

I don't need USB at all, so I disabled it in the BIOS, and no problems  
for real this time. Just wanted to post this for the equally clueless.