#
✕ nick:maz
#asahi · 2021-09-23 09:08
maz(\o/ UART and USB2 over a single USB-C connection -- I reached the 21st century!)
#asahi · 2021-09-23 09:06
mazbtw, the function you're looking for is irqd_get_trigger_type(), but you'll have to have use the corresponding setter in your set_type callback.
#asahi · 2021-09-23 09:04
mazmeh. if there is no other way, we'll have to live with it I guess.
#asahi · 2021-09-23 09:02
mazedge*
#asahi · 2021-09-23 09:02
mazdoes the SPI controller signal a level or dege interrupt?
#asahi · 2021-09-23 09:01
mazj_ey: how do you test it the first place?
#asahi · 2021-09-23 08:57
mazit is likely that setting it to OFF disables the latch, and you lose the edge that landed while masked.
#asahi · 2021-09-23 08:56
mazj_ey: I'm worried that this leads to lost interrupts. the expectations are that if an edge lands while the interrupt is masked, it fires on unmask.
#asahi · 2021-09-23 08:51
mazj_ey: but can't you read the state from the actual register?
#asahi · 2021-09-23 08:47
mazj_ey: you can get it indirectly, from irqdesc.
#asahi · 2021-09-23 08:43
mazj_ey: I guess that by doing that, you'd lose the pending state for edge interrupts. bad idea.
#asahi · 2021-09-23 08:00
mazgroup 0 meaning no interrupt I guess?
#asahi · 2021-09-23 07:57
mazin which case, I don't the the harm in having the driver evenly spreading the input pins over the 7 output lines, unless we want to play the priority game...
#asahi · 2021-09-23 07:55
maz_jannau_: OK. and you have a status per output interrupt to find out which interrupt has fired, I guess.
#asahi · 2021-09-23 07:53
mazis there any writeup about this thing?
#asahi · 2021-09-23 07:51
mazsure, but the question is whether you can have a a 32:1 mapping or not.
#asahi · 2021-09-23 07:50
mazif the GPIO block actually implements a 1:1 mapping between a single pin and an AIC interrupt, then the driver is completely wrong.
#asahi · 2021-09-23 07:49
mazbecause that's what the driver implements...
#asahi · 2021-09-23 07:48
mazsven: huh... doesn't the GPIO block expose a mux where any 32 pin can generate an interrupt that is then relayed by a single AIC interrupt?
#asahi · 2021-09-23 07:40
mazj_ey: I'm only trying to make sure that M1 is as clean as possible from an irqchip perspective so that I can point people at it and consider retiring.
#asahi · 2021-09-23 07:40
mazj_ey: fair enough. TBH, I don't think there is any major issue with this driver (we have far worse in the tree already).
#asahi · 2021-09-23 07:30
mazj_ey: nah, it's just some userspace candy for /proc/interrupts
#asahi · 2021-09-23 07:02
mazj_ey: this name is just a vague indication of where the interrupt is coming from, but is not meant to be a debug infrastructure. we have /sys/debug/irq for that.
#asahi · 2021-09-23 07:01
mazj_ey: that name doesn't have to be unique.
#asahi · 2021-09-22 16:23
mazj_ey: sure. I'll sketch something out for you.
#asahi · 2021-09-22 12:10
mazkettenis: I really hope it does. Otherwise, distinguishing between VFs is going to be impossible (once TB is up and that I can finally plug an SR-IOV device there).
#asahi · 2021-09-22 09:30
mazyup. even without TB, by the way. the radios are using two function from the same PCIe endpoint, and will share a single group with the current config.
#asahi · 2021-09-22 08:55
mazso either we keep your current approach of shadowing the config, or we move the sid2group release to apple_dart_release_device().
#asahi · 2021-09-22 08:47
mazsven: actually, you are right. the device goes away much earlier: https://paste.debian.net/1212807/
#asahi · 2021-09-22 08:38
mazlet me give KASAN a go.
#asahi · 2021-09-22 08:32
mazsven: ah, beers definitely have priority!
#asahi · 2021-09-22 08:32
mazsven: I reduced it to https://paste.debian.net/1212805/ which still works fine. am I missing something?
#asahi · 2021-09-22 08:31
mazsven: though I don't grok why you need to allocate a whole new master_cfg.
#asahi · 2021-09-22 08:03
mazsven: yup, that fixes it. thanks!
#asahi · 2021-09-21 16:39
mazsven: I'll have a stab at it.
#asahi · 2021-09-21 15:27
maz(though more gracefully)
#asahi · 2021-09-21 15:27
mazsven: in a similar vein, adding the removed device back fails as well: https://paste.debian.net/1212703/
#asahi · 2021-09-21 15:25
mazsven: yup, that fixes it.
#asahi · 2021-09-21 13:40
mazwell, I think there are about 4 of you now.
#asahi · 2021-09-21 13:39
mazsven: ah, gotcha.
#asahi · 2021-09-21 13:36
mazsven: don't we need this invalidation anyway? it would feel odd to keep TLBs around when we've dropped the PTs. agreed on the need for something to go upstream as a fix.
#asahi · 2021-09-21 11:13
mazkettenis: works for me.
#asahi · 2021-09-21 11:11
mazj_ey: OK. if you Cc me on your series, I'll try and pull all the strings together.
#asahi · 2021-09-21 11:10
maz(matrix invasion)
#asahi · 2021-09-21 11:09
mazkettenis: yup, that's the idea.
#asahi · 2021-09-21 11:04
mazj_ey: ah, cool. if you have a branch hanging somewhere, I could rebase the PCIe driver on top.
#asahi · 2021-09-21 11:03
mazs/blocker/blocked/
#asahi · 2021-09-21 11:03
mazI think the PCIe driver may be blocker by the pinctrl driver anyway (no idea what the status of that is).
#asahi · 2021-09-21 11:02
mazyeah, same here. I'll push out the PCIe driver v4 as well.
#asahi · 2021-09-21 11:00
mazHonestly, it looks good to me. so if you fix RobH's comments as well as the RC range I pointed out last week, we should be good. I'll ack that on the list.

plus ›