RE: New NTB API Issue
From: Allen Hubbe
Date: Thu Jun 22 2017 - 18:13:17 EST
From: Logan Gunthorpe
> On 6/22/2017 12:32 PM, Allen Hubbe wrote:
> > From: Logan Gunthorpe
> >> 2) The changes to the Intel and AMD driver for mw_get_align sets
> >> *max_size to the local pci resource size. (Thus making the assumption
> >> that the local is the same as the peer, which is wrong). max_size isn't
> >> actually used for anything so it's not _really_ an issue, but I do think
> >> it's confusing and incorrect. I'd suggest we remove max_size until
> >> something actually needs it, or at least set it to zero in cases where
> >> the hardware doesn't support returning the size of the peer's memory
> >> window (ie. in the Intel and AMD drivers).
> > You're right, and the b2b_split in the Intel driver even makes use of different primary/secondary
> bar sizes. For Intel and AMD, it would make more sense to use the secondary bar size here. The size
> of the secondary bar still not necessarily valid end-to-end, because in b2b the peer's primary bar
> size could be even smaller.
> > I'm not entirely convinced that this should represent the end-to-end size of local and peer memory
> window configurations. I think it should represent the largest side that would be valid to pass to
> ntb_mw_set_trans(). Then, the peers should communicate their respective max sizes (along with
> translation addresses, etc) before setting up the translations, and that exchange will ensure that the
> size finally used is valid end-to-end.
> But why would the client ever need to use the max_size instead of the
> actual size of the bar as retrieved and exchanged from peer_mw_get_addr?
The resource size given by peer_mw_get_addr might be different than the max_size given by ntb_mw_get_align.
I am most familiar with the ntb_hw_intel driver and that type of ntb hardware. The peer_mw_get_addr size is of the primary bar on the side to be the source of the translated writes (or reads). In b2b topology, at least, the first translation of that write lands it on the secondary bar of the peer ntb. That size of that bar could different than the first. The second translation lands the write in memory (eg). So, the end-to-end translation is limited by the first AND second sizes.
The first point is, the *max_size returned by intel_ntb_mw_get_align looks wrong. That should be the size of the secondary bar, not the resource size of the primary bar, of that device.
The second point is, because the sizes returned by peer_mw_get_addr, and ntb_mw_get_align, may be different, the two sides should communicate and reconcile the address and size information when setting up the translations.