#
✕ nick:kettenis2
#asahi · 2021-03-22 13:21
kettenis2and with option I mean property
#asahi · 2021-03-22 13:21
kettenis2and with option I mean property
#asahi · 2021-03-22 13:07
kettenis2correllium has a "sid-remap" option; the ADT has a "remap" option
#asahi · 2021-03-22 13:07
kettenis2correllium has a "sid-remap" option; the ADT has a "remap" option
#asahi · 2021-03-22 12:39
kettenis2and the dart driver would look at "regs" and notice that there are two sets of registers and decide to mirror the page tables?
#asahi · 2021-03-22 12:39
kettenis2and the dart driver would look at "regs" and notice that there are two sets of registers and decide to mirror the page tables?
#asahi · 2021-03-22 12:38
kettenis2you'd have a single node for the usb dart and a singly iommus reference in the usb controller node
#asahi · 2021-03-22 12:38
kettenis2you'd have a single node for the usb dart and a singly iommus reference in the usb controller node
#asahi · 2021-03-22 12:37
kettenis2that would simplify the binding isn't it?
#asahi · 2021-03-22 12:37
kettenis2that would simplify the binding isn't it?
#asahi · 2021-03-21 15:22
kettenis2and the "global" part doesn't do dma does it?
#asahi · 2021-03-21 15:22
kettenis2and the "global" part doesn't do dma does it?
#asahi · 2021-03-21 15:22
kettenis2well, xhci is what is used for host mode
#asahi · 2021-03-21 15:22
kettenis2well, xhci is what is used for host mode
#asahi · 2021-03-21 15:20
kettenis2is there a separate dart for device mode and host mode?
#asahi · 2021-03-21 15:20
kettenis2is there a separate dart for device mode and host mode?
#asahi · 2021-03-20 11:37
kettenis2will take a closer look tomorrow
#asahi · 2021-03-20 11:37
kettenis2will take a closer look tomorrow
#asahi · 2021-03-20 11:37
kettenis2sven: I made a comment on the binding diff
#asahi · 2021-03-20 11:37
kettenis2sven: I made a comment on the binding diff
#asahi · 2021-03-19 21:12
kettenis2it usually gets fixed (or fixes itself) pretty quickly
#asahi · 2021-03-19 15:14
kettenis2sven's code uses the ADT, so we don't need to add bindings for them in the FDT
#asahi · 2021-03-19 15:10
kettenis2magic register initialization sequences
#asahi · 2021-03-19 15:09
kettenis2sven wrote some m1n1 code to do that
#asahi · 2021-03-19 15:09
kettenis2the plan there was for m1n1 to apply them
#asahi · 2021-03-19 15:09
kettenis2biggest issiue is the tunables
#asahi · 2021-03-19 15:08
kettenis2yup
#asahi · 2021-03-19 15:07
kettenis2so we need to come up with proper bindings
#asahi · 2021-03-19 15:06
kettenis2but since we don't have a proper binding yet, I added a few nodes to the devicetree using whatever corellium came upwith
#asahi · 2021-03-19 15:05
kettenis2it pretends to be a linux kernel ;)
#asahi · 2021-03-19 15:05
kettenis2m1n1 even
#asahi · 2021-03-19 15:05
kettenis2it boots from mini
#asahi · 2021-03-19 15:01
kettenis2it seems to be fairly straightforward though; basically a two-level page table with 16k pages
#asahi · 2021-03-19 15:00
kettenis2maz: not as far as I can tell
#asahi · 2021-03-19 14:59
kettenis2but a dart iommu driver will also be needed
#asahi · 2021-03-19 14:59
kettenis2if you use my u-boot fork you should be able to get away with a fairly minimal pcie host bridge driver
#asahi · 2021-03-19 14:58
kettenis2maz: pcie support will give you network (both wired and wireless) and working type-A ports
#asahi · 2021-03-19 14:52
kettenis2damnit; mine even
#asahi · 2021-03-19 14:52
kettenis2minr even
#asahi · 2021-03-19 14:52
kettenis2that is minr ;)
#asahi · 2021-03-19 14:51
kettenis2https://gist.github.com/kettenis/8a810aa333e08a95222f80cea2a57d5d