wireguard-openbsd/sys/dev/pci/drm/amd/amdgpu, branch master WireGuard implementation for the OpenBSD kernel https://git.zx2c4.com/wireguard-openbsd/atom/sys/dev/pci/drm/amd/amdgpu?h=master 2021-02-20T21:50:04Z remove warning about amdgpu userptr ioctl being unimplemented 2021-02-20T21:50:04Z jsg jsg@openbsd.org 2021-02-20T21:50:04Z urn:sha1:ad2d38fa86fd2e27c36eb12497bcab3f0f6516fd matches radeon and i915 reported by Benjamin Baier The conversion of the IO BAR access did not completely mirror the 2020-11-19T08:53:14Z patrick patrick@openbsd.org 2020-11-19T08:53:14Z urn:sha1:9e1dc75e2daa85a654de4a3aeb461dca3f921917 behaviour of Linux' implementation: arm64's bus space operations have no barriers, so while Linux' iowrite32/ioread32 explicitly contain barriers, using bus space read/write is not enough on arm64. Add read barriers after a read to make sure that all reads have completed before the following reads. Add write barriers before a write to make sure all previous writes have completed before that write. This fixes panics on the HoneyComb LX2K with amdgpu(4). ok kettenis@ Switch from bus_space_read/write to linux io.h interfaces for all 2020-10-28T12:14:15Z jsg jsg@openbsd.org 2020-10-28T12:14:15Z urn:sha1:e54dbfe774d77b393515fd7292120986db1d1150 pci "memory space" io. Further reduces the diff to linux. ok kettenis@ Make mapping the "I/O Space" BAR optional. POWER9 systems don't support 2020-10-26T22:05:01Z kettenis kettenis@openbsd.org 2020-10-26T22:05:01Z urn:sha1:2c3ebbac5f6f3cd3a2ab5b778455a04feae2fabc "I/O Space" on their PCIe host bridges and for most (all) Radeon variants there is an alternative method to access the relevant registers. ok jsg@ handle capable(CAP_SYS_NICE) for allowing high priority contexts 2020-10-21T13:26:44Z jsg jsg@openbsd.org 2020-10-21T13:26:44Z urn:sha1:c8763d5d64aaf5f4188d12d31b457a5f488b5711 change drm_vma function arguments to take struct drm_file * 2020-10-21T02:16:53Z jsg jsg@openbsd.org 2020-10-21T02:16:53Z urn:sha1:0608e4cb21ccadfdade19dd76a78a2ab2aaa0e03 reduces the diff to linux 5.7.y ok kettenis@ consistently use dev_info() for sos fw version messages 2020-10-15T10:36:53Z jsg jsg@openbsd.org 2020-10-15T10:36:53Z urn:sha1:49078af9123567bc6a80b23df9d74266317e2c13 noticed in renoir dmesg don't print pci ids when amdgpu attaches 2020-10-15T10:04:39Z jsg jsg@openbsd.org 2020-10-15T10:04:39Z urn:sha1:46154dde19acbf082679ee40502de8f26e3f0402 Print asic name, rev and number of compute units. Suggested by and ok kettenis@ drm/amdgpu: move gpu_info parsing after common early init 2020-10-15T01:42:55Z jsg jsg@openbsd.org 2020-10-15T01:42:55Z urn:sha1:a89df194197bd58a2c08dcd0812b3400e1e52a6a From Alex Deucher 6e29c227a4976460ec6d4cc70b998e3a8c30c873 in mainline linux needed to load the correct gpu info firmware on raven2 (Ryzen 3 3200U/Ryzen 3 3250U/Athlon 300U/Athlon 3000G) drm/amdgpu: move discovery gfx config fetching 2020-10-15T01:37:29Z jsg jsg@openbsd.org 2020-10-15T01:37:29Z urn:sha1:7d292e99574b3ceb42838cf9ffad4c2bba189028 From Alex Deucher 6ba57b7a8f94fcf3df6883db362642a0075d185b in mainline linux