#
✕ nick:sven
#asahi · 2021-04-10 20:47
svenhm, true. i do remember hitting something before the end of ram once after chainload m1n1 too often
#asahi · 2021-04-10 20:46
svenoh, the sep carveout area?
#asahi · 2021-04-10 20:41
svensure
#asahi · 2021-04-10 20:40
svenyou should probably do a closer review of the usb one though. that one's more critical
#asahi · 2021-04-10 20:38
svenno idea what me go into that direction when the new one is so much simpler
#asahi · 2021-04-10 20:38
svenheh, yeah. that framebuffer thing was somewhat strange
#asahi · 2021-04-10 20:37
svenhahaha :D
#asahi · 2021-04-10 20:36
svennot much changes for the usb one fwiw, i've just split the dart and pmgr stuff out because kettenis_ needed the first and i had updates for the second
#asahi · 2021-04-10 20:35
svenhttps://github.com/AsahiLinux/m1n1/pull/45 that one first, after that you'll get the usb one :)
#asahi · 2021-04-10 20:34
svenhah, looks like kettenis_ has some way of making marcan actually look at my PRs :P
#asahi · 2021-04-09 16:12
sveni guess anyone could create a wiki page like that :)
#asahi · 2021-04-09 16:02
svenbut i believe marcan also wants to do a monthly progress report
#asahi · 2021-04-09 16:02
sven"they're work in progress" ;)
#asahi · 2021-04-09 07:35
svenmarcan: is there anything in the pasemi i2c documentation about the base clock btw.? on the m1 the input is allegedly divided by 16 and I'm curious if something like that is already mentioned in there
#asahi · 2021-04-09 06:58
svenmarcan: great. i'll try to at least send the clock gating stuff (which is very dumb anyway) and possibly the i2c stuff this weekend then :)
#asahi · 2021-04-08 15:21
svennow... let's see what happens if i just run a kernel with 4k cpu/16k iommu pages and remove that bug_on
#asahi · 2021-04-08 15:16
svenkettenis_: alright, https://github.com/AsahiLinux/m1n1/pull/44 should be just the (slightly improved) pmgr code
#asahi · 2021-04-08 11:52
sven:-)
#asahi · 2021-04-08 11:51
svenso just add a udelay instead of a busy loop? :)
#asahi · 2021-04-08 11:50
svenI can split those from that MR this evening fwiw. looks like I need to resolve some conflicts anyway
#asahi · 2021-04-07 21:03
svenwill do, thanks!
#asahi · 2021-04-07 20:48
svenkettenis_: just pushed tunables_apply_local_addr(const char *path, const char *prop, uintptr_t base)
#asahi · 2021-04-07 20:47
sven+include
#asahi · 2021-04-07 20:47
svenbut after that i think i should include at least will in this discussion since he'll likely have a strong opinion how to deal with this situation as well
#asahi · 2021-04-07 20:47
svenyeah, i'll just remove that BUG_ON tomorrow and see what happens with 4k pages.
#asahi · 2021-04-07 20:16
svenstill not great though
#asahi · 2021-04-07 20:16
sven*things
#asahi · 2021-04-07 20:16
svenokay, so at least some thing will work with 4k pages then even if i don't find a way to switch the iommu to 4k pages as well
#asahi · 2021-04-07 20:11
svenkettenis_: it also sounds like it's the most flexible to me. the current function could just use that one as well internally so that we don't have to duplicate any code
#asahi · 2021-04-07 20:07
svenor just do that for now :)
#asahi · 2021-04-07 20:06
svensounds like we'll either need tunables_apply_local(.., base_addr) or tunables_apply_local(.., "parent-node", parent_reg_idx, addr_offset)
#asahi · 2021-04-07 20:04
svenmight be more flexible to just have the base address as an argument then
#asahi · 2021-04-07 20:04
svenah, okay. so it'll be somewhat ugly anyway.
#asahi · 2021-04-07 20:03
sven(i.e. it's nothing like regs[25] + some_offset?)
#asahi · 2021-04-07 20:03
sventrue. is that base address not based on a reg property at all?
#asahi · 2021-04-07 20:01
svendo you also need one that walks up the ADT or takes a parent node for the "regs" property?
#asahi · 2021-04-07 20:00
svenah, yeah. i've been meaning to get back to that
#asahi · 2021-04-07 19:49
svenlet's hope there's even more hw that's a variant of something known
#asahi · 2021-04-07 19:49
svengood news on the pcie root port though :)
#asahi · 2021-04-07 19:46
sventhere's also at least a BUG_ON((granule > PAGE_SIZE) in the linux iommu code which i should hit with 4k cpu pages
#asahi · 2021-04-07 19:46
sveni also couldn't really add a mapping from an iova/to a paddr that's not aligned to 16k easily
#asahi · 2021-04-07 19:16
svenlet's see if it's maybe just a bit somewhere that selects it
#asahi · 2021-04-07 19:16
svenhttps://gist.github.com/bazad/1faef1a6fe396b820a43170b43e38be1 i did find this though. there's some iphone that already has the t8020,dart with page-size (4): 0x1000
#asahi · 2021-04-07 19:15
sveni thought about that, but i'm not sure how. i think it's just a write protect but not 100% yet.
#asahi · 2021-04-07 19:00
svenwill pointed out on the ML that this will be trouble if the iommu only supports 16k pages
#asahi · 2021-04-07 18:59
svenugh.. the M1 support 4k pages as well, doesn't it?
#asahi · 2021-04-07 09:13
svenmarcan: ah, btw., can I just put the clock gate + i2c stuff under you in the MAINTAINERS file?
#asahi · 2021-04-07 09:06
svenworked fine for the second one though
#asahi · 2021-04-07 09:06
svenbut I didn't read any warnings etc. and overwrote one of my ssh keys on my yubikey
#asahi · 2021-04-07 09:05
sventhat's what I did a while ago :)

plus ›