#
✕ nick:i509VCB
#asahi-gpu · 2023-08-16 15:42
i509vcbHowever here's the kicker
#asahi-gpu · 2023-08-16 15:42
i509vcbradv and nvk will return an error correctly
#asahi-gpu · 2023-08-16 15:42
i509vcbthat little pseudocode snippet should be returning VK_ERROR_INVALID_EXTERNAL_HANDLE since a file is definitely not a prime fd
#asahi-gpu · 2023-08-16 15:41
i509vcbwhich seems bizzare since those cannot be prime fds which are what vulkan drivers typically consume with the dma buf handle type
#asahi-gpu · 2023-08-16 15:41
i509vcbvkGetMemoryFdPropertiesKHR(device, fd)
#asahi-gpu · 2023-08-16 15:41
i509vcbfd = open("~/Pictures/ferris.txt")
#asahi-gpu · 2023-08-16 15:40
i509vcbAlso this means you can do something like
#asahi-gpu · 2023-08-16 15:40
i509vcbYou could pass -1 (invalid fd) and it technically is ok with that function
#asahi-gpu · 2023-08-16 15:40
i509vcbhowever a lot of implementations (v3dv, tu, pvr, anv, hasvk) return VK_SUCCESS regardless of the file descriptor you pass into that function
#asahi-gpu · 2023-08-16 15:39
i509vcbSo vkGetMemoryFdPropertiesKHR is used primarily used for determining the memory type of a dmabuf you are importing.
#asahi-gpu · 2023-08-16 15:38
i509vcbBefore I go over to #dri-devel to try to figure out which implementations are wrong, I'll try to explain what I think is wrong with v3dv and co in that function
#asahi-gpu · 2023-08-16 15:37
i509vcbSometimes I don't read enough if the last message
#asahi-gpu · 2023-08-16 15:37
i509vcberr yeah agxv rebase with new kernel release
#asahi-gpu · 2023-08-16 15:33
i509vcb /s/to/get
#asahi-gpu · 2023-08-16 15:33
i509vcbwell assuming memory didn't just to carpet bombed
#asahi-gpu · 2023-08-16 15:32
i509vcbI'm running the asahi distro kernels, my stuff is mainly in memory land so feel free to bump if ella is okay with it
#asahi-gpu · 2023-08-16 15:14
i509vcbalyssa: I looked at other drivers and now I realize I need to lay out a case why every driver except for nvk and radv is implementing vkGetMemoryFdPropertiesKHR wrong
#asahi-gpu · 2023-08-16 01:23
i509vcbnothing of value will be lost at least
#asahi-gpu · 2023-08-16 01:13
i509vcbpulls out 2nd laptop
#asahi-gpu · 2023-08-16 01:09
i509vcbBut did not enable the swapfile yet
#asahi-gpu · 2023-08-16 01:09
i509vcbI *might* have accidentally ran mkfs.swap on RecoveryOSContainer
#asahi-gpu · 2023-08-16 01:07
i509vcbI think I fucked up a lot harder
#asahi-gpu · 2023-08-16 01:07
i509vcbnever mind the missing swapfile
#asahi-gpu · 2023-08-16 00:58
i509vcbnarrator: in a moment of stupidity i5 realizes his swapfile isn't even setup
#asahi-gpu · 2023-08-16 00:54
i509vcbshould I be getting stuff oom killed running dEQP-VK.memory.*?
#asahi-gpu · 2023-08-15 21:05
i509vcbagxv currently unsets VK_SYNC_FEATURE_TIMELINE, I'm guessing the kernel driver work for timeline semaphores is not yet done?
#asahi-gpu · 2023-08-15 19:36
i509vcbI've gotten api.external.memory.* to pass
#asahi-gpu · 2023-08-15 19:36
i509vcbalyssa: expect VK_EXT_external_memory_dma_buf soon in agxv, just a bunch of other stuff I need to implement from dependent extensions.
#asahi-gpu · 2023-08-15 14:12
i509vcbif page size is an issue we can always just have another patch
#asahi-gpu · 2023-08-15 14:11
i509vcbHence the question around vm_page_size
#asahi-gpu · 2023-08-15 14:10
i509vcbGL_EXT_external_objects is less specific but it should match vulkan
#asahi-gpu · 2023-08-15 14:08
i509vcb> deviceUUID must be immutable for a given device across instances, processes, driver APIs, driver versions, and system reboots.
#asahi-gpu · 2023-08-15 14:08
i509vcbThe vk spec states
#asahi-gpu · 2023-08-15 13:49
i509vcbOn that topic what is vm_page_size supposed to be?
#asahi-gpu · 2023-08-15 13:46
i509vcbAlso I decided to throw the abi version in the hash for good measure
#asahi-gpu · 2023-08-15 13:46
i509vcbLast I recall Linux can't really do dynamic, per process page sizes
#asahi-gpu · 2023-08-15 13:45
i509vcbalyssa: do we want to consider the vm_page_size in the uuid as well?
#asahi-gpu · 2023-08-15 05:07
i509vcbSHA1("agx" + gpu_generation + gpu_variant + gpu_revision) if you were wondering
#asahi-gpu · 2023-08-15 05:03
i509vcbI implemented the uuid stuff, heavily inspired by freedreno, do I need to throw igalia in the copyright header for that?
#asahi-gpu · 2023-08-15 04:05
i509vcbdevice and driver uuids, generally systems don't have more than one physical device per system so I can just do something like freedreno and hash a constant for that?
#asahi-gpu · 2023-08-14 23:24
i509vcb(Answer to my question, run with deqp instead of deqp-runner)
#asahi-gpu · 2023-08-14 22:55
i509vcbIs there some argument I can pass to see the logs so I know why some deqp test is skipped?
#asahi-gpu · 2023-08-14 19:12
i509vcbI guess I did something stupid
#asahi-gpu · 2023-08-14 19:06
i509vcbI suspect I did something weird earlier
#asahi-gpu · 2023-08-14 19:05
i509vcbI nuked my clone and am trying again
#asahi-gpu · 2023-08-14 18:48
i509vcbWhy am I getting stuck on amd64 binaries trying to be run?
#asahi-gpu · 2023-08-14 18:16
i509vcbarch
#asahi-gpu · 2023-08-14 18:16
i509vcbErr not OS
#asahi-gpu · 2023-08-14 18:15
i509vcbnot sure why it defaults to x86_64
#asahi-gpu · 2023-08-14 18:15
i509vcbFigured it out, for some reason I had to set -DDE_OS=DE_CPU_ARM_64

plus ›