<< | Thread Index | >> ]    [ << | Date Index | >> ]

Subject: Re: cipe-win32-pre9 fix
From: "Damion K. Wilson" <dwilson,AT,ibl,DOT,bm>
Date: Fri, 11 May 2001 19:46:08 +0200
In-reply-to: <31276.988416041@www32.gmx.net>

Hi, Erik

I know exactly how you feel. In fact, It took me a long time to remove the 
spinlock stuff because I wasn't convinced that they could even be the cause. 
Then, after removing them, the stability increased. I still don't know why, 
which is why I'm being so careful. Right now, the driver is the most stable 
it's ever been, but it's still less stable than the (really) old version.

I'd prefer if you could tell me why the spinlocks are working for you. I know 
what spinlocks are for, but I would like to have proof that there is a 
pointer math overrun or something occurring to cause the crash. An easy way 
of doing this is to add a reference counter to the list routines.

I'll try this stuff out when I get a chance, but right now I'm getting my 
consulting company off the ground. Wish me luck :-)

BTW: Use the NT4.0 DDK for building so that the driver will run unmodified on 
both platforms.


*********** REPLY SEPARATOR  ***********

On 5/11/01 at 12:27 AM Erik Wallin wrote:

>Damion Wilson wrote:
>> That doesn;t make sense to me either. I tell you what, I'll try
>> judiciously adding the spinlocks to just the linked list handling. It
>> shouldn't speed up the code, though. What is your speedup ? Furthermore, my
>> setup has been stable for months (I'm running three instances) with ping
>> times (round trip) about .8 ms slower than "on the wire" speed.
>Sorry, I'm mixing ideas about speedup with my quest for stability. The
>spinlocks are not giving any speedup. All I care about right now is a stable
>driver. Then we can play around with speedups. I'm sorry to say it, but on 
>the driver (Pre-9) is not stable. Maybe it is on NT?
>It does require *a lot* of packets before it crashes, but it does. So far 
>managed to crash it on every machine I tried by sending a few hundred Mb of
>data. I've not been able to crash the one with spinlocks in so far.
>Is really no one else experiencing these problems on W2K? It would be very
>interesting to get reports from people who have tried Pre-9 and if they found
>it to be stable or not. Right now I feel like I'm the only one whining.
>Adding the spinlocks would be great. I really hope that it was the bug biting
>me. I've rolled out a few tests to other poeple and instructed them to keep
>their crash dumps if they get any blue screens. We should see if there are 
>IRQL problems.
>I've been running this machine for over two weeks now with the modified cipe
>driver. The service keeps crashing occasionally, but not the driver.
>Message sent by the cipe-l,AT,inka,DOT,de mailing list.
>Unsubscribe: mail majordomo,AT,inka,DOT,de, "unsubscribe cipe-l" in body
>Other commands available with "help" in body to the same address.
>CIPE info and list archive: 

<< | Thread Index | >> ]    [ << | Date Index | >> ]