#asahi-dev · 2024-06-09 19:04
jannauit has to show up since the firmware is loaded over pci but I remember that's fickle#asahi-dev · 2024-06-09 18:58
amarioguyjannau: i have the code for loading firmware in place, does that need to be done before the device can show up on the bridge?#asahi-dev · 2024-06-09 18:39
jannauamarioguy: are you loading the firmware? see https://github.com/AsahiLinux/linux/commits/bits/150-xhci-firmware/ and u-boot has similar changes#asahi-dev · 2024-06-09 18:20
amarioguy(yes i know it's been months since i've been on, irl stuff became a lot)#asahi-dev · 2024-06-09 18:19
amarioguyhey just had a quick question, i'm doing the PCIe port bringup sequence (as defined in U-Boot pcie-apple.c, setting up the PHY, REFCLK, etc), but the USB-A controller on the mac studio doesn't seem to be detected even after the entire pcie bus is being probed. is there something you need to do after the bringup sequence to make the USB-A controller itself visible on the PCIe bus?#asahi-dev · 2024-06-09 06:29
chadmedergh sdhci driver causing a serror after upgrading the studio to 14.5 sfw#asahi-dev · 2024-06-07 06:50
chadmedis anyone else testing the plasma 6.1 beta and getting random flashing? #asahi-dev · 2024-06-05 20:14
sveni looked through the tipd changes and didn't see anything suspicious fwiw#asahi-dev · 2024-06-05 20:00
jannausven: I suspect yesterday's ace issues were caused by a not properly plugged in power supply#asahi-dev · 2024-06-04 22:29
jannaudwc3, nvmem-efuses and gpu only only due to reverts of upstream changes. i.e. fine for now but we need other solutions. rust broke module parameters and the rawdevice constness hack, dwc3 phy init and nvmem broke offsets > 8#asahi-dev · 2024-06-04 22:20
jannauI've updated asahi-wip (and the bits/* branches) to rebase onto v6.9.3 but haven't tag anything yet, needs more testing#asahi-dev · 2024-06-04 20:56
jannauonly other change is one for tps25750_interrupt()#asahi-dev · 2024-06-04 20:54
jannauhttps://github.com/AsahiLinux/linux/commit/409c1cfb5a803f3cf2d17aeaf75c25c4be951b07#asahi-dev · 2024-06-04 20:51
jannauthe only conflicting changes were that the tps6598x irq handler now reads more than 64 bits since the event register on later tps6598x variants is larger#asahi-dev · 2024-06-04 20:49
jannauI'll push to asahi-wip without tag#asahi-dev · 2024-06-04 20:48
sven(unless we completely crash the firmware ofc)#asahi-dev · 2024-06-04 20:47
svenbut as long as we leave the secondary event registers alone we shouldn’t break anything #asahi-dev · 2024-06-04 20:47
sven*tipd#asahi-dev · 2024-06-04 20:47
svenI can take a look at tips changes tomorrow if there’s anything suspicious otherwise #asahi-dev · 2024-06-04 20:46
jannautesting linux 6.9 here#asahi-dev · 2024-06-04 20:46
svenwe don’t mess with its irqs anymore do we?#asahi-dev · 2024-06-04 20:46
svenSMC has a secondary i2c bus to that chip #asahi-dev · 2024-06-04 20:45
svenhuh no#asahi-dev · 2024-06-04 20:44
jannauseems to break the hypervisor usb port#asahi-dev · 2024-06-04 20:40
jannausven: did you ever see ace/tps6598x related smc syslog messages like https://gist.github.com/jannau/d321a9dcce1a53066843fc6a872c968a#asahi-dev · 2024-06-03 20:39
Lyudegotcha, I'll see if I can figure out where this got pulled in from then (I have a bunch of patches pulled in from asahi's tree so there's been a bit of confusion)#asahi-dev · 2024-06-03 20:38
jannauat least in the current tree#asahi-dev · 2024-06-03 20:37
jannaubut no rust/kernel/dma.rs in our tree#asahi-dev · 2024-06-03 20:37
jannauLyude: the asahi driver doesn't appear to interact with the DMA api directly. we only have https://github.com/AsahiLinux/linux/commit/b17553679437f79b5bf667aefe709dc9837fe5e4#asahi-dev · 2024-06-03 20:21
LyudeMainly trying to rebase the rvkms (rust virtual KMS driver) work I've been doing against the latest DRM bindings on the mailing list which is why I ask. I would link you to the repository for rvkms but unfortunately freedesktop's gitlab is temporarily in maintanence mode #asahi-dev · 2024-06-03 20:19
Lyudeah-: is this your patch and if so, do you have any idea what the purpose of Device::from_dev_no_reference was supposed to be? (or if there's a newer version of this patch that I should be using instead?) https://paste.centos.org/view/28e3678a . It gets used later in https://paste.centos.org/view/cb410144#asahi-dev · 2024-06-03 20:15
Lyudelooks like it also was actually a patch written by someone else, whoops lol#asahi-dev · 2024-06-03 20:13
j`eynot sure how to link it, but it's called fedora asahi dev#asahi-dev · 2024-06-03 20:11
Lyudej`ey: I didn't realize there was a matrix channel as well#asahi-dev · 2024-06-03 20:10
j`eycheck matrix!#asahi-dev · 2024-06-03 20:09
LyudeAny chance Lina's been around anywhere? Wondering if she could answer a quick question about of the rust DMA pool patches that she wrote up for asahi#asahi-dev · 2024-06-02 22:48
mariogripthis is with 14.1, so not sure if anything changed or?#asahi-dev · 2024-06-02 22:48
mariogripHello, I'm trying to get macvdmtool working, but im stuck trying to disable AppleSerialShim, when i do that i end up with a bootloop#asahi-dev · 2024-06-02 13:48
sventhat apparently sleeps up to 20ms #asahi-dev · 2024-06-02 13:47
svenpretty sure the timeout was much larger than 10ms in the polling code because it used msleep(1) fwiw#asahi-dev · 2024-06-02 13:47
jannauto my understanding / recollection 34edaaa716ba just implements error recovery. I think the driver / bus / chip was otherwise completely broken after running into the tiemout once#asahi-dev · 2024-06-02 13:46
sventhanks #asahi-dev · 2024-06-02 13:46
sven2022 🙃#asahi-dev · 2024-06-02 13:43
jannausven: that was a parallel development, see https://lore.kernel.org/asahi/a40dbff1-6aa1-dedf-7eb2-c75d75d808b2@marcan.st/#asahi-dev · 2024-06-02 11:23
svendo you remember if it's the improved pasemi_smb_clear and the error handling that actually fixed the tipd issues instead of an increased timeout?#asahi-dev · 2024-06-02 11:22
svenmarcan: so i'm looking at upstreaming misc stuff and came across https://github.com/AsahiLinux/linux/commit/34edaaa716ba55c9265dc55c8c006d7a0190c8a4 which mentions that "the driver uses a 10ms timeout" but afaict it already had a 100ms timeout before since https://github.com/AsahiLinux/linux/commit/e826192cc26bd69746bbf22e6bdf72b87cb3d97b#asahi-dev · 2024-06-01 12:11
jannaumaz: good to know#asahi-dev · 2024-06-01 11:59
chadmednot like the mac is trying to recharge them anyway right#asahi-dev · 2024-06-01 11:59
chadmedthe only difference is the chemistry so who cares#asahi-dev · 2024-06-01 11:45
maz2.7v