2.6.0 and Starfire NIC

2.6.0 and Starfire NIC

[Brian Macy]

When switching to 2.6.0 my Starfire NIC fails to function with an 
entertaining message:
Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02018101.
Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02010001.

I don't know if this is related but in 2.4 I get PCI bus congestion for 
the starfire adapter:
eth0: PCI bus congestion, increasing Tx FIFO threshold to 80 bytes
eth0: PCI bus congestion, increasing Tx FIFO threshold to 96 bytes
eth0: PCI bus congestion, increasing Tx FIFO threshold to 112 bytes
eth0: PCI bus congestion, increasing Tx FIFO threshold to 128 bytes
eth0: PCI bus congestion, increasing Tx FIFO threshold to 144 bytes
eth0: PCI bus congestion, increasing Tx FIFO threshold to 160 bytes


Brian Macy

Re: 2.6.0 and Starfire NIC

[William Lee Irwin III]

On Tue, Dec 30, 2003 at 08:48:13AM -0800, Brian Macy wrote:
> When switching to 2.6.0 my Starfire NIC fails to function with an 
> entertaining message:
> Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02018101.
> Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02010001.
> I don't know if this is related but in 2.4 I get PCI bus congestion for 
> the starfire adapter:
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 80 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 96 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 112 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 128 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 144 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 160 bytes

In 2.6, I get the message you got in 2.4.


-- wli

Re: 2.6.0 and Starfire NIC

[Ion Badulescu]

Hi Brian,

On Tue, 30 Dec 2003 08:48:13 -0800, Brian Macy <bmacy@macykids.net> wrote:
> When switching to 2.6.0 my Starfire NIC fails to function with an 
> entertaining message:
> Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02018101.
> Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02010001.

This says that the card's RX ring is empty. That could be because the 
driver fails to allocate any descriptors, though that's not very likely. 
Or it might be a bug...

Does it happen every time you try, or only sometimes?

> I don't know if this is related but in 2.4 I get PCI bus congestion for 
> the starfire adapter:
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 80 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 96 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 112 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 128 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 144 bytes
> eth0: PCI bus congestion, increasing Tx FIFO threshold to 160 bytes

Unrelated. These messages only tell you that the latency on your PCI bus 
is slightly higher than expected, and the driver is compensating for it.

Ion

-- 
  It is better to keep your mouth shut and be thought a fool,
            than to open it and remove all doubt.

Re: 2.6.0 and Starfire NIC

[Brian Macy]

This happened all 3 times I booted into the 2.6 kernel. If there is 
something you'd like me to try let me know.

Brian Macy


Ion Badulescu wrote:
> Hi Brian,
> 
> On Tue, 30 Dec 2003 08:48:13 -0800, Brian Macy <bmacy@macykids.net> wrote:
> 
>>When switching to 2.6.0 my Starfire NIC fails to function with an 
>>entertaining message:
>>Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02018101.
>>Dec 23 16:36:45 job kernel: eth0: Something Wicked happened! 0x02010001.
> 
> 
> This says that the card's RX ring is empty. That could be because the 
> driver fails to allocate any descriptors, though that's not very likely. 
> Or it might be a bug...
> 
> Does it happen every time you try, or only sometimes?
> 
> 
>>I don't know if this is related but in 2.4 I get PCI bus congestion for 
>>the starfire adapter:
>>eth0: PCI bus congestion, increasing Tx FIFO threshold to 80 bytes
>>eth0: PCI bus congestion, increasing Tx FIFO threshold to 96 bytes
>>eth0: PCI bus congestion, increasing Tx FIFO threshold to 112 bytes
>>eth0: PCI bus congestion, increasing Tx FIFO threshold to 128 bytes
>>eth0: PCI bus congestion, increasing Tx FIFO threshold to 144 bytes
>>eth0: PCI bus congestion, increasing Tx FIFO threshold to 160 bytes
> 
> 
> Unrelated. These messages only tell you that the latency on your PCI bus 
> is slightly higher than expected, and the driver is compensating for it.
> 
> Ion
>