On Fri, 2008-03-14 at 00:12 +0000, Sasha Khapyorsky wrote:
> On 11:13 Wed 12 Mar , Hal Rosenstock wrote:
> > >
> > > struct _ntc_144 {
> > > ib_net16_t pad1;
> > > - ib_net16_t lid; // lid where capability mask changed
> > > - ib_net16_t pad2;
> > > - ib_net32_t new_cap_mask; // new capability mask
> > > + ib_net16_t lid; // lid where change occured
> > > + uint8_t pad2; // reserved
> > > + uint8_t local_changes; // 7b reserved 1b local
> > > changes
> > > + ib_net32_t new_cap_mask; // new capability mask
> > > + ib_net16_t change_flgs; // 13b reserved 3b change
> > > flags
> >
> > Should this be padded out as in the 1.2.1 spec ?
>
> It is stated in the 1.2.1 that an upper bits in change_flgs are reserved
> for the same purpose - OtherLocalChanges mask. So it looks ok for me to
> have it as one field and not redo later.
Have what as one field ? change_flgs ?
I was referring to a pad3 at the end of the structure as in the spec.
-- Hal
>
> Sasha
> _______________________________________________
> general mailing list
> [email protected]
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
>
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
