summaryrefslogtreecommitdiff
path: root/sys/arch/amd64/pci/iommu.c
AgeCommit message (Expand)Author
2008-12-03Remove the x86 and i386 prefixes to the bus_dma types. It's really quiteOwain Ainsworth
2008-04-28Sync PCI ids.Brad Smith
2007-09-17MALLOC/FREE -> malloc/free and M_ZERO changesCharles Longeau
2007-05-27be more agressive on invalidation; perform invalidates on load* and unloadJason Wright
2007-05-27remove two TODO items:Jason Wright
2007-02-09simplify the argument passing (use a pointer to bus_dma_segment_t insteadJason Wright
2007-02-09according to the errata, invalid pages shouldn't be used. Always makeJason Wright
2005-09-29Execute operations in the correct order. From jason@Marco Peereboom
2005-09-27ARGH! disable gart, I suck!Marco Peereboom
2005-09-27Clean up prints. From jason@Marco Peereboom
2005-06-17- remove old debugging codeMarco Peereboom
2005-06-08revert enabling iommu on amd64 as it breaks at least one MP host.Marco S Hyman
2005-06-07make it prettier; jason okTheo de Raadt
2005-06-06enable iommu on all systems where we support it. We'll support it on moreJason Wright
2005-06-02- wish there was a better way to do this... put splhigh() around theJason Wright
2005-05-27oops, need to get the boundary from the map when allocating gart addressesJason Wright
2005-05-27 - Use the direct map for mapping the PTEs.Artur Grabowski
2005-05-27use a fixed address for the base dva, also don't forget to setup theJason Wright
2005-05-27hook in iommu, but it's still disabled by default for nowJason Wright
2005-05-27fix printfJason Wright
2005-05-27handle multi-cpu GART allocation:Jason Wright
2005-05-26make all mappings valid/coherent. For pages not currently mapped, fillJason Wright
2005-05-26iommu meets amd64. This currently only works on machines where the GARTJason Wright
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-02 07:13:13 +0000