> > In particular, I think we need feedback on #3 from PCI and power > > management maintainers. > > I don't get your concern. > I think we can add device attribute file to let application to trigger > wakeup function, then NTB hardware will do the rest. NTB driver just > need to implement suspend/resume interface of PCI PM. > > Add one more thing, do you think NTB should support runtime power > management? >
I think it is good to make the power management functionality available. In other words, yes, to your last question. My concern is that I would like some degree of certainty that it is done right, in harmony with the rest of the kernel. I don't know what "done right" means in this case, which is why I would like someone else to review it. A smaller patch with only (and all of) the power management code will have a better chance of being reviewed. I'm also concerned about the waiting behavior in #2 and #3. I'm not saying it's wrong. At least now that behavior is noted in the api documentation; thanks for that. If a PCI or power management expert has no objection to the waiting behavior in #3, then I would be comfortable with that behavior in #2 as well. Allen -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/