01:14 <tpefreedom> Are mali bifrost and midgard gpus more common than the utgard series?

01:17 <HdkR> Utgard used to be more popular but that is likely swinging towards bifrost now

01:22 <tpefreedom> It seems that the community is having better luck with gpus for single board computers (aside from powervr) than the big dedicated gpus from amd and nvidia.

01:37 <HdkR> What do you mean? AMD open source stack is doing quite well

01:54 <tpefreedom> I've still heard that igpus work much better than dedicated gpus on x86.

01:56 <HdkR> No idea where that is coming from

01:56 <HdkR> Nouveau has problems obviously

01:56 <HdkR> But AMD with radeonsi and radv?

01:57 <HdkR> I have more display problems with my i965 laptop than that

01:59 <tpefreedom> huh

01:59 <tpefreedom> Modern x86 still has major issues.

01:59 <HdkR> Sure. Everyone has issues

02:05 <tpefreedom> Intel ME and AMD PSP.

02:05 <HdkR> Luckily that doesn't relate to those company's GPUs

02:07 <tpefreedom> Yeah.

02:07 <tpefreedom> That is more on the cpu side.

02:07 <tpefreedom> Are those company's GPUs compatible with ARM anyway?

02:08 <HdkR> AMD's is. Intel doesn't have a real dGPU yet.

02:11 <tpefreedom> Intel is making one. My hope is that it will be similar enough to their iGPUs to give GNU/Linux users better graphics cards with only free software.

02:12 <HdkR> Yes. They're making one. Can't make commentary if it'll work in ARM hardware until tested.

02:13 <HdkR> It'll likely work unless they heck up PCIe conformance like they did on the DG1 (or whatever the real problem is)

02:32 <alyssa> tpefreedom: it's my understanding the Intel dGPUs will just work on arm64.. whenever they're released

02:42 <icecream95> macc24: With a small kernel hack (removing the conditional where "failed to allocate beacon buffer" is printed in .../ath10k/mac.c) AP mode works for wifi on duet

02:43 <icecream95> The comment above the code mentions it could open a "possible security issue", but I'm not sure if that applies to the setup in duet or not...

05:20 <robclark> HdkR: it's been a while since I tried, but last time I tried I couldn't build intel drivers on arm since they depended on some x86 intrinsics.. although I guess that is fixable if anyone cared enough

05:21 <HdkR> I know there were some fixes on that front. Not sure if they were all handled though

05:21 <HdkR> streaming copy for the main one that was hit

05:21 <robclark> I guess w/ dGPU's there is some reason to actually care

05:22 <HdkR> Yea. I'm sure there will be plenty of people slapping an Intel Xe in to their Solidrun boards :D

05:23 <robclark> :-P

05:29 <HdkR> I'm looking forward to trying it

05:29 <HdkR> If I can get a new enough board with decent PCIe...

10:28 <HdkR> Just make sure to always expose PCIe atomics ;)

12:15 <macc24> icecream95: ack

15:11 <debiangamer> When I try to run vkcube with the Panfrost vulkan driver I get the following with latest Oibap ppa and Mali T820: ~$ VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/panfrost_icd.aarch64.json vkcube vkEnumerateInstanceExtensionProperties failed to find the VK_KHR_xcb_surface extension. Do you have a compatible Vulkan installable client driver (ICD) installed? Please look at the Getting Started guide for additional information.

15:11 <debiangamer> vulkaninfo WARNING: [Loader Message] Code 0 : terminator_CreateInstance: Failed to CreateInstance in ICD 4. Skipping ICD. WARNING: lavapipe is not a conformant vulkan implementation, testing use only. ERROR at /build/vulkan-tools-3Ouhix/vulkan-tools-1.2.162.0+dfsg1/vulkaninfo/vulkaninfo.h:248:vkEnumeratePhysicalDeviceGroupsKHR failed with ERROR_INITIALIZATION_FAILED

15:22 <macc24> debiangamer: consider not running away after 3 minutes