Am Freitag, den 08.06.2012, 16:03 +0300 schrieb Michael S. Tsirkin: > Why is this needed? > What's wrong with mapping resources through > /sys/bus/pci/devices/XXXXXresourceX > ? > Mmmh ok, the problem here is, that the UIO documentation states: "/dev/uioX is used to access the address space of the card. Just use mmap() to access registers or RAM locations of your card." and "From userspace, the different mappings are distinguished by adjusting the offset parameter of the mmap() call." This does not work and the "/sys/class/uio/uioX/maps/mapX/" directories do also not appear. I was very confused when I tried UIO the first time and it did not behave like it is described in the documentation. Anyway, the patch adds the described behaviour and I would more than happy, if it would be accepted for the mainline kernel. > > uio currently only supports devices which do not > do DMA. > > DMA from uio controlled devices is a no no unless > it's behind an IOMMU which can protect us from > random memory corruptions this could cause. > > In the later case it's OK but we need some code > to check this and program the IOMMU appropriately. OK, I see. Might it be OK if I just remove the DMA stuff from this patch and resubmit it? -- Gruß Dominic Frankfurt Institute for Advanced Studies (FIAS) Ruth-Moufang-Straße 1 D-60438 Frankfurt am Main Germany Phone: +49 69 79844114 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
