On Wed, 16 Aug 2006, David Brownell wrote: > On Wednesday 16 August 2006 1:57 pm, Alan Stern wrote: > > + hcd->self.uses_dma = (dev->dma_mask && > > + dma_supported(dev, *dev->dma_mask)); > > That does look bizarre, given the variety of often-buggy implementations > of dma_supported() ... the ones looking at dev->dma_mask often ignore the > second parameter (assuming no devices have address range restrictions), > and many others ignore the first one (assuming all devices support DMA)!! > > But more to the point, dma_supported() is defined to test whether the > specified device supports a given range of addresses, 0..SecondParam. > And I'm unaware of a portable way to choose that parameter to indicate > to "all kernel memory defined as DMA-safe", which is what the USB stack > cares about.
Oh, okay. The description in DMA-API.txt wasn't very clear. > So I think it is far more correct to say > > hcd->self.uses_dma = (dev->dma_mask != NULL); > > Another way of putting it is that dma_supported() is badly misnamed; > it should be dma_supports_address_range() instead, accepting that the > range specifications are weak at best since they were designed only > to handle restrictions like "N-bit addresses" rather than real-world > restrictions like "only addresses from 0x01000000-0x02000000 work". I will revise the patch as you suggest. In any case, it seems like a good idea to encapsulate this test. Alan Stern ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
