On (11/08/07 13:27), Garrett D'Amore wrote: > > This is weird. > > Out of curiosity, why is sad.h needed? I don't see anything in it that I > can think of dld.h as implicitly needing. > > Its unfortunate that sad.h pollutes the namespace this way, but I think it > was designed with the idea that only the sad driver itself, along with > whatever userland component is associated with it (autopush?) would include > it. > > I also note that those portions of the namespace that it pollutes with do > not appear to be documented anywhere. Maybe they could be changed? (I > guess this would impact userland components like autopush?) >
don't know- Brussels inherited this part from clearview-uv- maybe someone there can clarify? <..discussion about private property for checksum offload snipped..> > I don't think QE has a problem with bge busted hardware, do they? I know > that *Sun* NICs have an implementation flaw, but I wasn't aware of any > relating to bge. > > The EBUSY idea has merit, if you're going to leave this in. > > It seems like you could probably activate this in the descriptor rings, so > you don't have to check on it at run time. Of course, that means a full > device reset (well at least reprogramming the rx rings) is probably > required. Ok. As I said, let me confer with the Beijing team and get back on this one. >> >>> bge_ndd.c: line 33. Surely this warning should come from the MAC layer >>> rather than the driver? Nemo could intercept the ndd ioctl, and issue >>> the warning. (It can also look to see if the driver already supports >>> brussels at that point, before issuing the warning.) >>> >> >> That's an interesting suggestion. You are proposing that mac_ioctl >> examine the mc_callbacks for (MC_SETPROP|MC_GETPROP), and issue >> the warning if the ((struct iocblk *)mp->b_rptr)->ioc_cmd is ND_SET >> or ND_GET, right? yes, that's a good idea- I'll implement it. >> > > Thanks, that was exactly what I meant. :-) Consider it accepted! (will spin out a new webrev after I coalesce the review comments) Thanks very much for the review! --Sowmini
