05:18galimentor: I can clean up the dma code my own very likely, however i do not remember the the userspace drivers debate arguments, i think descriptive api design is great idea over syscalls function calls can be inlined and demangled into syscall that buffers the content, i could offer such binaries that adds syscall metadata, and there is address space randomization and cryptographically randomized
05:18galimentor: interrupt handler to behave as minor security tutor or guide, I finished all the modules and start to deal with security. My idea is to not allow userspace drivers to violate the API and abi of kernel, as such that my mode of execution can then reach to writing drivers offering special encoder for RMW ddr pci bar packet. I understand it's bit microkernelish or appleish, but instead of IPC
05:18galimentor: we do syscall/interrupt however.
05:20galimentor: other than that, linux has pushed through to prime time as monopoly holder in OS era.
05:31galimentor: So as for crypto i suppose an fpga fabric with user choosable cryptographic primitives or not at all if elected, to force consistency, i.e some type of usb key , or something integrated into future AMD and intel products, but i honestly never understood as to why AMD and intel discarded the mobile or so to speak smartphone market.
05:35galimentor: as to how i never understood why my dad borrowed million dollars to one of the biggest dickhead who tried to kill me, instead of investing the money into in house production of electronic equipment, with one year i would take over the mobile market with x86's arms and fpga's under my new fw umbrella.
05:35galimentor: and ontop of that manufacture my own chips in addition with my IP.
05:43galimentor: My propasal there is to now invest another million to cambodia to build a in house engineering complex to manufacture computer chips and eliminate the assholes from that area, like indrek raud to mention few, and collect the hotel from that idiot before, as Russia might be preparing to attack estonia.
11:11mripard: mlankhorst: did you end up sending the cgroup PR to airlied? Are we waiting on something?
11:57mlankhorst: mripard: Out for the week. :)
13:00mripard: mlankhorst: ack :) Do you want me to send it?
13:53stsquad: should mesa even attempt to access DRM with CONFIG_VKMS is enabled and there is no graphics hardware configured for a VM?
14:01emersion: what do you mean by "access DRM"?
15:15robclark: mareko: a630 runners are back
15:58alyssa: robclark: thanks, was about to ask
15:58stsquad: which mesa option builds /usr/lib/dri/vkms_dri.so?
16:17MrCooper: stsquad: FWIW, there's only really libdril_dri.so left, all the other names are just symlinks to that now
16:43alyssa: robclark: still having some hiccups https://gitlab.freedesktop.org/mesa/mesa/-/jobs/68924802
16:53robclark: not sure what is going on there.. restart the job?
17:08stsquad: MrCooper ahh ok - will check
17:08stsquad: the context is updating buildroot's mesa to the latest version and the switch from SWRAST to SOFTPIPE has broken a bunch of buildroots tests and I need to figure out why
17:28zmike: robclark: can you make it not back?
17:28zmike: cuz it seems to still not work
17:29zmike: https://gitlab.freedesktop.org/mesa/mesa/-/jobs/68927072
17:31robclark: looks like both those jobs were freedreno-cheza-19 so maybe something wrong with that one runner?
17:31zmike: 🤕
17:33robclark: hmm, it looks like it was fine before the NuC conked out over the weekend, but now all the jobs on freedreno-cheza-19 are hitting that error.. so I've paused the runner
22:29mlankhorst: mripard: yes please. :)
22:53mripard: mlankhorst: it's done. Enjoy your time off