[ofa-general] [PATCH/RFC 1/2] IB/core: Add support for "send with invalidate" work requests

[Steve Wise]

Tziporet Koren wrote:
> Roland Dreier wrote:
>> OK here's an updated series of the kernel side, with the invalidate
>> stuff moved to a new opcode.  I also decided after thinking about it
>> that I liked Eli's suggestion of putting the invalidate rkey in a union
>> with imm_data.  This won't work for libibverbs where we have to preserve
>> the API but I guess we can burn that bridge when we come to it...
>>   
>
> I think send w/invalidate is for kernel keys only (at least in IB) so 
> not clear we need it in libibverbs at all
>
For iWARP, its needed for user mode as well...

Steve.