From: David W. <dw...@in...> - 2011-01-06 16:28:23
|
On Tue, 2010-12-14 at 11:51 -0800, Jason Lunz wrote: > > > Instead, you should solve this problem in UML code. I do not know how, > > but may be you can add readb/writeb there which actually do nothing or > > print a scary warning, or do BUG(), and let things which use them just > > fail run-time. > > Something like this could work, but it would be error-prone for anyone > else who attempts using iomem-requiring drivers on uml. Instead of > getting obvious compile failures we'd have broken drivers that BUG() or > emit scary warnings. That doesn't seem to me like an improvement. Drivers should *never* BUG() or crash, or busy-loop, on getting 0xFF when they read from hardware. That can happen anyway in some circumstances. Doesn't iSeries take this approach? I don't much like the patch that Artem took into his l2-mtd tree; it doesn't even let you build mtdchar, which really *ought* to be permitted. It also didn't allow the nandsim or mtdram devices, which are purely virtual. I think I'd prefer something similar to your original, Jason. I don't think the HAS_IOMEM dependencies have to be *so* complex to maintain. If anything we're just going to err on the side of inclusion and you'll occasionally have to send us patches to "hide" things from you again. -- David Woodhouse Open Source Technology Centre Dav...@in... Intel Corporation |