Masa Murayama wrote:
> ----- Original Message -----
>> Date: Mon, 14 Sep 2009 00:16:45 -0500
>> From: Steven Stallion <[email protected]>
>> To: Masa Murayama <[email protected]>
>> CC: [email protected]
>> Subject: Re: [driver-discuss] Hard hang during interrupt processing
>>
>>
>>> Steve,
>>>
>>> Oh, I'm sorry. I tested netperf for about 1 minute. I'll test for longer
>>> duration.
>> Hello Masa,
>>
>> Have you had any luck yet with a longer netperf run ?
>>
>> Cheers,
>>
>> Steve
> 
> Steve,
> 
> I think I have progress with following changes. Now re works for over one hour
>  with bi-directional netperfs.
> 
>  I changed re_intr() to handle rx overflow interrupt according with the way do
> cumented in dp8390 series datasheet.
> 
>  I changed re_intr() to handle one or more interrupt reasons at a call.
> 
>  I added clearing rx interrupt reason on each packet reception to suppress
>  rx spurious interrupt errors.
> 
>  I added CR_RD_ABRT bit when a tx packet is sent.
> 
> The modified source code is here. Please try it.
> 
> http://homepage2.nifty.com/mrym3/taiyodo/re.c.new3

Hello Masa,

I am still seeing the hang. I am able to duplicate this fairly easily by
ftping a large file between machines. Have you tried using ftp rather
than netperf?

Regards,

Steve

_______________________________________________
driver-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/driver-discuss

Reply via email to