On Wednesday 18 February 2009 17:56:24 Zhao, Yu wrote: > Matthew Wilcox wrote: > > On Wed, Feb 18, 2009 at 10:11:06AM -0800, Jesse Barnes wrote: > >> I wonder if it's worth adding better checks for bus vs CPU space mixups, > >> maybe with a sparse annotation or a new typedef... > > > > Part of the problem is that our PCI layer kludges all this around. > > First we read bus addresses into a struct resource, then we hand it > > off to arch code that frobs the struct resource. We also have struct > > pci_bus_region which is supposed to represent bus addresses, but we > > don't use it consistently. > > > > I've wanted to fix this for a while, but there always seems to > > be something more urgent to do. Plus it means fixing up a lot of > > architectures. > > Do you mean pcibios_bus_to_resource() doesn't work well, or something > else? Can you please advise what we should do now? No it should work; it's just that CPU addr vs. bus addr mixups are pretty common if you're used to dealing with x86 platforms, so we were talking about making the code a bit clearer about the distinction. Unfortunately things are a bit tangled, so it's kind of a chore. -- Jesse Barnes, Intel Open Source Technology Center -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
