#asahi · 2021-06-10 10:11
sven(and possibly another one to not try nvme_scan_ns_list which always seems to fail)#asahi · 2021-06-10 10:11
svenit needs NVME_QUIRK_SINGLE_VECTOR and NVME_QUIRK_SHARED_TAGS#asahi · 2021-06-10 10:05
svenneeds a different nvme_submit_cmd (which is not part of any ops) and a different nvme_common_complete_rq (which is part of the queue ops) and some additional mmio pokes#asahi · 2021-06-10 10:00
svennot sure yet :)#asahi · 2021-06-10 10:00
svenit might be enough to just use a different set of blk_mq_ops with two or so functions replaced#asahi · 2021-06-10 09:57
svenarnd: nice! i've been looking into how to do all the apple-specific stuff. not sure if just declaring a quirk and then patching random places like corellium did is the right way.#asahi · 2021-06-10 08:05
svenyeah, dwc3 is usb-c and pcie/xhci is usb-a#asahi · 2021-06-09 21:23
svenlike my code apparently! ;)#asahi · 2021-06-09 19:06
svenbloom: have you considered https://github.com/auchenberg/volkswagen as a solution to your test problem? :P#asahi · 2021-06-09 19:04
svenkettenis: https://github.com/AsahiLinux/m1n1/pull/73 that should work#asahi · 2021-06-09 19:02
svenah, okay#asahi · 2021-06-09 18:59
svenit won't put the dart back into bypass mode but if it never does usb_iodev_init it should still be left in bypass from iboot#asahi · 2021-06-09 18:58
svenhrm, isn't the dart in bypass mode if m1n1 doesn't initialize usb?#asahi · 2021-06-09 18:55
svenyup. i messed that up#asahi · 2021-06-09 18:54
svendid i mess that up. let me check#asahi · 2021-06-09 18:54
svenuh#asahi · 2021-06-09 18:48
svenlol, ok#asahi · 2021-06-09 18:47
svenwhat did i do now? :D#asahi · 2021-06-09 18:47
svenuh#asahi · 2021-06-09 16:20
sven(as soon as there's i2c and gpio-interrupt support)#asahi · 2021-06-09 16:19
svensince m1n1 sends the SPSS command now that driver should work without any patches already#asahi · 2021-06-09 16:19
svenyes, but i'm not sure if that's enough to trigger a phy/dwc reset#asahi · 2021-06-09 16:16
svenjannau: how that gpio driver coming along btw? i think we should be able to use i2c-gpio to talk to the pd chip once that one works and see if we can find a sane way to fix replugging#asahi · 2021-06-09 16:13
svenguess we might need that dwc glue node after all#asahi · 2021-06-09 16:13
svenso that means we'll find to figure out a way for the pd chip to somehow tell the phy driver to reinit everything#asahi · 2021-06-09 16:12
svenhm.. okay. so same behavior as linux right now#asahi · 2021-06-09 16:11
sven:>#asahi · 2021-06-09 16:10
sven+d#asahi · 2021-06-09 16:09
svenif their normal "new device connected" flow is pd innterrupt -> reinit phy -> kick xhci i'd expect it to never work if the PD interrupts are disable#asahi · 2021-06-09 16:08
sventhat kinda lines up with our hypothesis that it uses the pd interrupts to reinitialize the phy whenever a device is connected#asahi · 2021-06-09 16:07
sven:/#asahi · 2021-06-09 16:06
svendoes it not re-enable the interrupts because the chip is removed from the ADT or does it *never* re-renable interrupts?#asahi · 2021-06-09 15:42
svenah yes, that would be annoying#asahi · 2021-06-09 15:40
svenyeah, i'm just saying that apple abstracted away the power stuff for sink/source inside the PD chip. so i'm kinda hoping they also abstracted away a lot of crap into whatever ATC is#asahi · 2021-06-09 15:39
svenand yeah, ATC is where the magic happens#asahi · 2021-06-09 15:39
svenno idea. usually you apparently have to at least talk to the pd chip to switch it from source/sink/etc.#asahi · 2021-06-09 15:38
svenno idea. at least the whole power stuff is done automagically by that TI chip#asahi · 2021-06-09 15:37
svenmaybe usb3 just works by accident as well#asahi · 2021-06-09 15:37
svenyeah, let's see. i somehow managed to get the phy to work in both gadget and host mode for usb2 at least#asahi · 2021-06-09 15:37
svenabsolutely#asahi · 2021-06-09 15:36
svenhaven't looked into that at all because USB#asahi · 2021-06-09 15:36
sventhere's at least some co-processor involved in thunderbolt#asahi · 2021-06-09 15:35
svenoh, we'll figure it out. i'm just not looking forward to that#asahi · 2021-06-09 15:34
svenugh, the usb phy. that one's gonna be fun once we want usb 3 or thunderbolt support#asahi · 2021-06-08 20:25
sveneither way, very useful bot :-)#asahi · 2021-06-08 20:25
svenarnd: ah, yes. still some cleanup work left to be done. i think that build bot might take any git url it finds in MAINTAINERS #asahi · 2021-06-08 15:39
svenmy limited understanding is that a platform device is defined in a device tree (or possibly even in some code directly) as opposed to a device that can be enumerated on a bus#asahi · 2021-06-08 15:36
svenmakes sense, that probably made the rebase i had to do relatively painless#asahi · 2021-06-08 15:35
sventrue, probably also makes sense to split this out of pci.c as well#asahi · 2021-06-08 15:30
svenarnd: https://github.com/AsahiLinux/linux/commit/69142df77238252d5ab52d940bcff054c4dcd551 that's all i had to fix i your patches to make nvme as a platform device work