08:23 <milkylainen> basic kernel allocation question. The sizes of max allocation is restricted by normal order? Even for DMA allocs? What restricts the max size of a normal dma alloc, beside congestion inside the typical zone?

09:10 <milkylainen> arnd: are the ti81xx also considered for some soon-to-be cleanup?

09:12 <milkylainen> or are cortex-a8 designs safe for now? :)

09:13 <maz> ARMv7 should be safe until we retire the whole of the 32bit ARM port.

09:14 <ukleinek> ..ooOO(next week?!)

09:14 <Xogium> noooo please never do that

09:14 <Xogium> ^^

09:14 <maz> I wouldn't mind, but some may object.

09:15 <Xogium> couldn't use my nice armv7 board to talk with y'all anymore :(

09:16 <ardb> maz: shouldn't we drop alpha first? :-)

09:17 <milkylainen> ok. I was thinking that the ti81xx was never that well supported anyway. With the nice accelerators and such. TI never bothered to do mainline support for them afaiu?

09:18 <Xogium> I'd ask since when does TI does bother with mainline support, but it'd just me being snarky

09:20 <maz> ardb: probably.

09:41 <arnd> my estimate is that arch/arm/ will be removed in about 15 (+/- 5) years, along with support for any other 32-bit hardware that may hang around that long. There is a good chance we'll kill armv4/v5/v6 support before v7, but then again they might all hang around until 32-bit kernels become unmaintainable

09:42 <Xogium> very good chance for... Wait wait I hadn't realized v4 was still a thing

09:43 <arnd> we still have five armv4 platforms,plus another handful of armv4t

09:44 <arnd> mach-rpc is somewhere betweeen armv3 and armv4 actually and will have to die once gcc-8 becomes the minimum version

09:44 <arnd> (or gcc-9 perhaps)

09:44 <Xogium> holy

09:44 <bjdooks> :sadacornnoises:

09:44 <bjdooks> I still have two

09:44 <Xogium> :D

09:45 <bjdooks> I even have a Hydra for one

09:46 <Xogium> I like old tech

09:59 <milkylainen> Is max zone order 15 ok for a 24G machine. I do have use for a larger max. Consequences?