#
#asahi · 2025-10-30 14:45
svenyup, and i'd rather make some people happy with basic m3 support than waiting until we have full gpu just to make the peanut gallery happy
#asahi · 2025-10-30 14:45
j`eysven: expert mode doesnt really mean much as soon as peopel with M3 know it's behind expert mode
#asahi · 2025-10-30 14:44
chaos_princessyou don't need to read it to shitpost and troll in the comment section
#asahi · 2025-10-30 14:44
lm_The internet peanut gallery is always going to fling shit at ambitious projects
#asahi · 2025-10-30 14:44
sveni just don't read phoronix because it's a waste of time tbh
#asahi · 2025-10-30 14:44
sveni'd honestly be fine with enabling it inside expert mode or whatever once the basics work
#asahi · 2025-10-30 14:43
fl0_idhaha sure
#asahi · 2025-10-30 14:43
chaos_princessright, but i'd rather not give the moronix comment section extra ammo :P
#asahi · 2025-10-30 14:43
fl0_idchaos_princess yeah i know. which is also why I get why you probably wouldn't want to enable sth very basic in the installer
#asahi · 2025-10-30 14:43
svensure, but they're already claiming that for whatever reason right now and they'll never be happy *shrug*
#asahi · 2025-10-30 14:42
chaos_princessit is going to be fine for people who understand what goes into it, but unfortunately there is a certain crowd who will just instantly go "waaah, no gpu, asahi sux"
#asahi · 2025-10-30 14:42
svenyeah
#asahi · 2025-10-30 14:42
fl0_idyeah i get that
#asahi · 2025-10-30 14:42
fl0_idand me f.e. would much prefer to hack on it from a machine that somewhat works
#asahi · 2025-10-30 14:42
svennow that being said, gpu acceleration is going to be the everest-sized one :D
#asahi · 2025-10-30 14:42
fl0_idchaos_princess well yeah. I didn't mean to say they are all the same
#asahi · 2025-10-30 14:41
lm_Yeah, enough to get the ball rolling
#asahi · 2025-10-30 14:41
fl0_idsven same
#asahi · 2025-10-30 14:41
sven+be
#asahi · 2025-10-30 14:41
svenit's not gonna be great but i think it's gonna fine for a start
#asahi · 2025-10-30 14:40
sveni also don't think that m3 with software acceleration will be all that bad tbh
#asahi · 2025-10-30 14:39
svenbut the only way to learn is to get through that
#asahi · 2025-10-30 14:39
svenyeah, and if you've never done those baby steps before they will feel quite huge
#asahi · 2025-10-30 14:34
chaos_princessit is a pyramid made of a bunch of baby steps and one everest-sized one
#asahi · 2025-10-30 14:32
fl0_idbtw, kind of interesting that that one reddit m3 thread was mentioned in the update. but then in that thread marcan for some reason was active and kind of putting people down. I mean technically he's probably right that device trees are the easy part or whatever else he said, but maybe it can be a pyramind made from baby steps...
#asahi · 2025-10-30 14:08
svenand then there's truly weird stuff like dwc3 on these machines: that sits behind *two* separate iommus and gadget mode and one half of xhci uses the first one and the other half of xhci the second one 🙃
#asahi · 2025-10-30 13:45
chaos_princessyea, i thought of mentioning the whole part where iommus can be multiple-in-one, but decided not to
#asahi · 2025-10-30 13:45
mazchaos_princess: it's more subtle than that. IOMMUs usually have multiple page-tables, one per set of devices that share the same "source-id" (stream-id, requester-id...).
#asahi · 2025-10-30 12:38
monthebiffappreciate the knowledge
#asahi · 2025-10-30 12:37
monthebiffexcellent cheers
#asahi · 2025-10-30 12:37
chaos_princessusually platform firmware tells the os in some way what the topology is
#asahi · 2025-10-30 12:36
chaos_princessif multiple devices are behind one iommu, they must be in the same group as it is physically impossible to have different page tables for them
#asahi · 2025-10-30 12:36
chaos_princessdepending on hardware topology
#asahi · 2025-10-30 12:35
monthebiffok cool. and iommu groups are defined by the OS? based on rules/ conventions defined... where?
#asahi · 2025-10-30 12:33
chaos_princessthere is a set of hardware registers for each iommu, and the os either has them system-global, or also does context switching depends on what device it is
#asahi · 2025-10-30 12:33
monthebiffso theres a hardware register, and every time a process starts, the OS finds some empty location for some new page tables and sets the register as part of context switching? when we say 'iommu have their own page tables'... how does that fit that model?
#asahi · 2025-10-30 12:31
chaos_princessthe major exception is gpus where it is common for the userspace to have more control (mediated by the os ofc)
#asahi · 2025-10-30 12:30
chaos_princessand less as a part of a process and more as a part of a driver
#asahi · 2025-10-30 12:30
monthebiffthought so
#asahi · 2025-10-30 12:30
chaos_princessos
#asahi · 2025-10-30 12:30
monthebiffbut to be clear *you* being the operating system as part of spinning up a process? or *you* being the program itself running?
#asahi · 2025-10-30 12:30
chaos_princess(that ofc assumes you have a sane iommu, which we do, and i hope whatever you are reading models a sane iommu)
#asahi · 2025-10-30 12:29
monthebiffyup just learned that
#asahi · 2025-10-30 12:28
j`eyyou allocate some memory for the page tables, and like with the CPU where you use TTBR*_EL1, there's some iommu specific 'page_tables_here' register
#asahi · 2025-10-30 12:28
j`eyno special memory location really
#asahi · 2025-10-30 12:28
monthebiffright ok... ill need to read more on that
#asahi · 2025-10-30 12:27
chaos_princessthey do sit in memory, but there is no such thing as "the page tables"
#asahi · 2025-10-30 12:27
monthebiffright ok
#asahi · 2025-10-30 12:27
chaos_princessok, ignore the "no"
#asahi · 2025-10-30 12:26
chaos_princessyou just give the iommu a memory address and say "page tables live here"

plus ›