Thanks all.  This is exactly what I observed in my program.

----- Original Message ----- 
From: "Tziporet Koren" <[EMAIL PROTECTED]>
To: "Roland Dreier" <[EMAIL PROTECTED]>
Cc: "Steve Wise" <[EMAIL PROTECTED]>; "openib-general" 
<[email protected]>
Sent: Sunday, July 23, 2006 5:56 AM
Subject: Re: [openib-general] libibverbs question


> Roland Dreier wrote:
>>     Steve> If a SEND is received on a QP and it generates a BOUNDS
>>     Steve> violation because the recv buffer posted is too small, 
>> will
>>     Steve> this generate an async error?
>>
>> No, I don't believe so.  The error can be returned synchronously as a
>> completion with status of "local length error."
>>
>> (I was actually too lazy to look at the IB spec, but you should be
>> able to confirm that by looking at the section of chapter 11 on
>> completion status)
>>
>>
>>
> Mellanox HCAs behaviour is:
> in RC - WQE completed with error
> in UD/UC - skip message and resume operation
>
> Tziporet
>
> 


_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Reply via email to