[ofa-general] [PATCH/RFC] Add support for "send with invalidate" to libibverbs

[Talpey, Thomas]

At 03:39 AM 4/2/2008, Dotan Barak wrote:
>If the value of device_attributes.num_mw is more than zero => the device 
>supports memory windows, so i think this flag
>can be safely removed.

I agree with removing the flag, but if you mean "max_mw", looking at the
tree, there are a few problems with the > zero assertion. :-)

drivers/infiniband/hw/ehca/ehca_hca.c 376:
	props->max_mw = min_t(unsigned, rblock->max_mw, INT_MAX);

drivers/infiniband/hw/nes/nes_verbs.c 3915:
	props->max_mw = nesibdev->max_mr;

Note, ehca may set it to huge negative values, and nes puts the wrong
value in the attribute field! (typo?)

The good news is, the AMSO1100 seems to get it right. ;-)

I'm still looking to be able to test the NFS/RDMA client over memory
windows. The code's all there in the RPC layer, just not in the
providers.

Tom.