RELENG_7_1: bce driver change generating too much interrupts ?

[Mike Jakubik]

On Sun, December 14, 2008 4:57 am, Xin LI wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Mike Jakubik wrote:
>> On Mon, December 8, 2008 5:22 pm, Mike Jakubik wrote:
>>> On Mon, December 8, 2008 5:12 pm, Xin LI wrote:
>>>
>>>> Which version are you currently using?  My previous commit only fixes
>>>> the excessive interrupt issue, I think this could be a different
>>>> problem, I'm taking a look at the code to see if I can have something
>>>> for you.
>>> I was running on the version just prior to the latest interrupt commit.
>>> I
>>> have now updated to the one with the interrupt fix. Will let you know
>>> if
>>> things change.
>>>
>>> Thank You.
>>
>> The interrupt rate has decreased significantly, however i am still
>> having
>> having problem with applications that hold stateful connections. The rx
>> errors are also still showing, i suspect this is related to the problem.
>> How can i roll back this driver to the last known good version?
>
> Hi, Mike,
>
> I think they are different problems.  Could you, please, give me
> feedback about whether:
>
>  - The old driver does not trigger the problem?
>
>  - The patched driver restore all the old driver behavior?

- Old driver.

I have been running the system for 4 days now with this driver. My
application has not stopped accepting connections, irq rate is low, and
there are no rx/tx errors reported. Everything looks good.

- Patched driver.

Your initial import plus the IRQ fix still shows rx errors, and my
application had stopped accepting connections. I have not tried the patch
in your last email, and im not sure when i will be able to to as these
systems are in production. Perhaps someone else could test it? As soon as
i get a chance i will let you know how it goes.

Thanks for the work on this.