09:54lembeudu: well the filesystem specification in terms of addressing requires pretty bad ass brainpower still. I see you do not believe me that i actually have it.
09:57lembeudu: however it can be solved for addressing in many ways, but i suppose hardcoded indexes belong to divide and conquer hw addressing, or sampling, but it's not so much divide and conquer, cause it should be able to correct after first sample, it never is making a false address at second sample.
09:58lembeudu: so one cell is showing where the word is stored in the hdd.
10:24circleperimeter: at this point in the evolution, the code is not finished on linux, windows, osx and anyone working alone can allocate a big monopoly by doing things way better. And i have made already all the tests, but for safety am not storing those materials in my files and am developing this stuff on aircapped computer, however they planted another chip onto my neck which seems to be very capable
10:24circleperimeter: and they snatch all my code prolly one way or another, but force will be used against those people soon anyhow. There is a possibility to make manufacturing complex to asia for me, that has fences for thieves and guards for bully bisexuals and anal their thief anal buffalos, but those were junkies, liars and such more obnoxious rather than a threat. But new hardware is not required to
10:24circleperimeter: make very fast computation still, but i think it's still possible to make my own brand of hw too.
17:21dwlsalmeida[d]: does anybody know what this is?
17:21dwlsalmeida[d]: ERROR: [Loader Message] Code 0 : setup_loader_term_phys_devs: Failed to detect any valid GPUs in the current config
17:21dwlsalmeida[d]: ERROR at ./vulkaninfo/./vulkaninfo.h:247:vkEnumeratePhysicalDevices failed with ERROR_INITIALIZATION_FAILED
17:21dwlsalmeida[d]: again, trying to debug my colleague's setup...
17:21dwlsalmeida[d]: I can't reproduce
17:26mhenning[d]: Do they have the nouveau kernel driver loaded? Does VK_DRIVER_FILES point to the right place?
18:08dwlsalmeida[d]: mhenning[d]: Will try to get them to join the channel
18:08dwlsalmeida[d]: gfxstrand[d]: would you know anything about this?
18:08dwlsalmeida[d]: ffmpeg: ../src/vulkan/runtime/vk_object.h:120: vk_object_base_assert_valid: Assertion `base == NULL || base->type == obj_type' failed.
18:09dwlsalmeida[d]: -exec p *base
18:09dwlsalmeida[d]: $2 = {_loader_data = {loaderMagic = 140736817493152, loaderData = 0x7fffd80374a0}, type = VK_OBJECT_TYPE_UNKNOWN, client_visible = false, device = 0x200000, instance = 0x0, private_data = {elem_size = 8, node_size_log2 = 3, root = 0}, object_name = 0x0}
18:10dwlsalmeida[d]: this fence was crated successfully before being passed into vkWaitForFences
18:19dwlsalmeida[d]: airlied[d]: you didn't hit that in ffmpeg?
18:35dwlsalmeida[d]: using a watchpoint shows that the memory location holding the fence gets returned by nvk_device_memory_from_handle and is then overwritten by
18:35dwlsalmeida[d]: static void
18:35dwlsalmeida[d]: copy_bind(struct nvk_vid_mem *dst, const VkBindVideoSessionMemoryInfoKHR *src)
18:35dwlsalmeida[d]: {
18:35dwlsalmeida[d]: dst->mem = nvk_device_memory_from_handle(src->memory);
18:35dwlsalmeida[d]: dst->offset = src->memoryOffset;
18:35dwlsalmeida[d]: dst->size = src->memorySize;
18:35dwlsalmeida[d]: }
18:36dwlsalmeida[d]: so yeah, just me not being careful, sorry
19:11dwlsalmeida[d]: mhenning[d]: mhenning[d] btw, nouveau was blacklisted
19:57gfxstrand[d]: dwlsalmeida[d]: That's usually a use-after-free somewhere.
23:16gfxstrand[d]: gfxstrand[d]: It's alive
23:16gfxstrand[d]: https://tenor.com/view/its-alive-gene-w-ilder-gif-14865163
23:18gfxstrand[d]: Gotta remember at least 5GiB for /boot! 🤡
23:20orowith2os[d]: I don't think I've ever seen my /boot go above 500mb
23:20orowith2os[d]: Nor /boot/efi above 200mb
23:21redsheep[d]: Did you manage to order in a 96 GB kit that fast?
23:22mohamexiety[d]: gfxstrand[d]: I did the mistake of 1GB twice. do not recommend at all ðŸ˜
23:22mohamexiety[d]: orowith2os[d]: driver dev usually has you play with the kernel side and believe me it _grows A LOT_
23:23gfxstrand[d]: orowith2os[d]: You don't have enough initrams with GSP firmware blobs. ðŸ˜
23:23orowith2os[d]: mohamexiety[d]: Oh yeah, because you'd have several kernels installed, huh?
23:23mohamexiety[d]: yeah
23:23gfxstrand[d]: redsheep[d]: Yup. Amazon
23:23mohamexiety[d]: but yeah, congrats!
23:24redsheep[d]: Remember to update bios 🙂
23:43marysaka[d]: gfxstrand[d]: ... good reminder that I need to extend mine, currently have 500MB for it and uuum yeah
23:53gfxstrand[d]: redsheep[d]: Working on it now