#asahi-dev · 2025-09-02 19:03
jannauI almost reported that the atc fuses m1n1 PR or dwc3/atcphy do not work on m2 until I remembered that your branch has only t8103 DT changes#asahi-dev · 2025-09-02 16:00
jannauj`ey: the problem doesn't apply to dcp as dcp limits fb flips, i.e. we have vblank hw, it's just inconvenient to use. if we used the timer we have deal with syncing it to dcp's swaps#asahi-dev · 2025-08-31 20:18
jannauI guess b4 export everything and then replace everything except the cover letter mail with `git format-patch` is easioer than editing mail files#asahi-dev · 2025-08-31 20:08
jannauI looked but it doesn't seem to support. git send-email supports it iirc#asahi-dev · 2025-08-31 20:02
jannauI've pushed asahi-6.16.4-1 to github, now to send the kasan fixes out#asahi-dev · 2025-08-31 20:01
jannausending patchsets with tons of CCs is annoying. I probably should use the web endpoint for those#asahi-dev · 2025-08-31 20:00
jannauI forgot to ask for acks but I assume it's ok to take the ack-ed commits through the asahi tree. it woould be nice if we can take those so I don't have to resend them for the v2#asahi-dev · 2025-08-31 19:58
jannausven, nickchan: or at least wait until the spmi changes are acked#asahi-dev · 2025-08-31 15:31
jannaunvmem_shift_read_buffer_in_place seems to be an harmless overread and was hiding a use after free via devres in simplefb.c#asahi-dev · 2025-08-31 14:47
jannauno, the while loop is ok, I just mentally already avoided the overread and forgot to advance `p` as well#asahi-dev · 2025-08-31 14:16
jannauthe overread is easily fixed but the `while (--extra >= 0) *p-- = 0;` puzzles me#asahi-dev · 2025-08-31 14:12
jannau"offset:1152 raw_len:4 bytes:4 bit_offset:23 nbits:2" specifically triggers kasan#asahi-dev · 2025-08-31 14:11
jannauas far as I understand it reads beyound the end of the buffer and writes before the start of it#asahi-dev · 2025-08-31 14:10
jannaugood that upstream atcphy gets rid of the fuses. nvmem_shift_read_buffer_in_place() triggers kasan and I have trouble wrapping my head around it#asahi-dev · 2025-08-28 16:44
jannaualready noted by nickchan and fixed locally#asahi-dev · 2025-08-28 15:22
jannaustrange, I would have expected it has aic3 like m3#asahi-dev · 2025-08-28 15:03
jannaui.e. it is what the generic compatible should have been#asahi-dev · 2025-08-28 15:02
jannausee the cover letter. I've used t8103 as base because that's what the drivers were written for#asahi-dev · 2025-08-28 14:59
jannaui2c has "apple,t8103-i2c" since the beginning#asahi-dev · 2025-08-28 14:57
jannausent those manually#asahi-dev · 2025-08-28 14:47
jannauyes, they failed due to "A message to this many recipients (64) would cause to exceed the per-10 minute message sending limit of 2000."#asahi-dev · 2025-08-28 14:01
jannaut6020 DT send#asahi-dev · 2025-08-27 17:49
jannaulibfdt doesn't do the translation, see dt_get_address() in m1n1's src/devicetree.c#asahi-dev · 2025-08-27 17:45
jannauthe addresses are translated via ranges in soc#asahi-dev · 2025-08-27 17:45
jannauthe fdt doesn't have the real mmio offsets#asahi-dev · 2025-08-27 16:57
jannauI'm annoyed but not annoyed enought to fight it either#asahi-dev · 2025-08-27 16:56
jannauI'm troubled by writing a justification beyond "we agreed to not use them" for the commit messages#asahi-dev · 2025-08-27 15:13
jannauthe samsung one wouldn't use "apple,spi" though#asahi-dev · 2025-08-27 15:03
jannau"apple,spi" wasn't a good idea and in our case having just the unrelieable information from the ADT is an argument for only using SoC specific ones#asahi-dev · 2025-08-27 15:02
jannauI'm actually not quite about what exactly Krzysztof complains. I don't think generic implementation / version compatibles like "nvme-ans2", "asc-mailbox-v4", "aic2" are uncommon nor did we start with that.#asahi-dev · 2025-08-27 14:09
jannaulet's see if there's a significant runtime impact of patching the whole DT. that might remind us of removing it#asahi-dev · 2025-08-27 14:03
jannauI'll use M1 as base compatible (if applicable) as that's what the driver was written for#asahi-dev · 2025-08-27 13:57
jannauI'll re-add the missing generic compatibles for t6020 in m1n1. for all I care we can leave the code in m1n1 forever#asahi-dev · 2025-08-27 13:52
jannaukettenis: https://lore.kernel.org/asahi/6365acb0-6e4c-4561-bc0f-d34637e8308e@kernel.org/#asahi-dev · 2025-08-27 13:51
jannauapparently not for new submissions even if they are compatible with generic one#asahi-dev · 2025-08-27 13:50
jannaumaybe we should ask for an exception for t6020 which already has downstream use outside of linux#asahi-dev · 2025-08-27 13:46
jannaufeels silly to do with "apple,pmgr" which isn't even used in a driver. it's relying on "syscon", "simple-mfd"#asahi-dev · 2025-08-27 13:25
jannauI'll add the generic compatibles for apple,t6020-* in m1n1#asahi-dev · 2025-08-27 13:23
jannauKrzysztof replied. 15 or so more patches, redoing the dt-bindings and updating the device trees#asahi-dev · 2025-08-27 13:18
jannauthe mac pro has this problem only for internal devices (which are still behind the pcie switch) so speaking about slots sounds equally wrong#asahi-dev · 2025-08-27 13:16
jannauit uses of_regulator_bulk_get_all() so we could add our own -supplies#asahi-dev · 2025-08-27 13:13
jannauah, the supplies mention slot#asahi-dev · 2025-08-27 13:11
jannauthe "slot" controller matches "pciclass,0604" which we have#asahi-dev · 2025-08-27 13:01
jannauthe pci slot power control driver uses the pcie supplies from https://github.com/devicetree-org/dt-schema/blob/v2024.11/dtschema/schemas/pci/pci-bus-common.yaml#L153 which we could back by gpio driven dummy regulator#asahi-dev · 2025-08-27 12:58
jannauI want to look if https://lore.kernel.org/linux-pci/20250819-pci-pwrctrl-perst-v1-0-4b74978d2007@oss.qualcomm.com/ works for the mac pro. if it does it should work for all other devices#asahi-dev · 2025-08-27 12:56
jannauthat alone is not enough since it needs to be synchronized with port probing#asahi-dev · 2025-08-27 12:55
jannaulinux has since a while a PCI slot power control driver.#asahi-dev · 2025-08-27 12:54
jannauwhy did I think we had the pwren-gpios in in the dt-bindings? I looked for mac pro support where it's more complicated due to the pcie switch#asahi-dev · 2025-08-27 12:51
jannauthe driver code is D3 handling. i.e. completely unrelated#asahi-dev · 2025-08-27 11:48
jannauno. looks like mmio addresses but the trailing 31 zeros are suspicios