> FTR, I like this interface better since most drivers use ioremap() instead > of ioremap_nocache(). They generally should be using the latter. And pci_iomap is cleaner still. Plus pcim_iomap does resource tracking so all the masses of special case error code/cleanup can get cut down. > Is there any easy way to tell if the device driver should be using > uncached mappings vs cacheable mappings? > (Just from looking at the source code) There isn't one. > For storage/networking/audio cards I'm comfortable with the generalization > that they all should use uncacheable mappings (I'm sure there are some > exceptions.) I'm not with video devices. There are exceptions - I2O for example and there are other cards that use write merging when possible beyond video. Also btw vidoe depends on the chip - if you cache/write merge the framebuffer on a Voodoo 1/2 card you must be in 24/32bit video modes for example. Generally though PCI = controlled by hardware = uncached Alan -- 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