#
✕ nick:jannau
#asahi-dev · 2026-08-25 18:06
jannaushouldn't device_links ensure that
#asahi-dev · 2026-08-25 18:02
jannauah, yes thunderbolt3 hubs
#asahi-dev · 2026-08-25 18:01
jannaumy hp dock has a (power) button but I think that's a usb2 device
#asahi-dev · 2026-08-25 18:00
jannauI'd expect that even keyboards with usb3 hub still use usb2 for the keyboard itself
#asahi-dev · 2026-08-25 17:59
jannauwouldn't a keyboard use usb2?
#asahi-dev · 2026-08-25 17:58
jannaudoes thunderbolt expect that the things are exactly the same after wakeup? I guess it can't really since it has to deal with hotplug during suspend anyway
#asahi-dev · 2026-08-25 15:48
jannausven: https://github.com/AsahiLinux/linux/commit/305b77680a33700995dfdc05e8c56cdfe128b0d0 blocks suspend while acio is up
#asahi-dev · 2026-08-25 15:20
jannaumaybe aquantia like the 10G one? Broadcom has afaik none
#asahi-dev · 2026-08-25 14:37
jannauoh, the acio driver has no PM handling at all
#asahi-dev · 2026-08-25 14:29
jannausven: I wonder why it's a problem though. afaik children should suspend before it parent and resume after
#asahi-dev · 2026-08-25 13:52
jannauthe mini has a new ethernet nic, 2.5 GBps
#asahi-dev · 2026-08-25 13:51
jannauso it still doubles a Max which resulted in a different chip id in the past
#asahi-dev · 2026-08-25 13:50
jannauyes, m5 pro and max probably share the CPU die, m5 ultra has 2 CPU (and GPU) dies
#asahi-dev · 2026-08-25 13:49
jannauerr, the Max variant always had that. DCP and 4 DCPext
#asahi-dev · 2026-08-25 13:48
jannaunickchan: T6051 or T6052 would be consistent with M1-M3. will have the same MIDR as T6050
#asahi-dev · 2026-08-25 13:46
jannaudid we notice that M5 Max has 5 DCPs?
#asahi-dev · 2026-08-25 13:43
jannaustill tb4 / DP 1.4 on M6
#asahi-dev · 2026-08-25 13:41
jannaum6 mac mini has only from 24GB RAM the full memory bandwidth
#asahi-dev · 2026-08-25 13:37
jannauitr kinda had. when it was introduced the super cores were still called perfomance cores
#asahi-dev · 2026-08-25 13:36
jannauthere are other arm socs with 3 core types though
#asahi-dev · 2026-08-25 13:35
jannauit is if I'm not mistaken. m5 had super and efficiency and m5 pro/max super and new performance core
#asahi-dev · 2026-08-25 13:34
jannauno, that's just iphones, laptops are usually october/november
#asahi-dev · 2026-08-25 13:32
jannauonly mac mini and studio refreshed
#asahi-dev · 2026-08-25 13:32
jannauyeah!, 3 core types on M6 (2, super, 4 performance, 6 efficiency): https://www.apple.com/newsroom/2026/08/apple-introduces-m6-and-m5-ultra-for-a-big-leap-in-performance-and-ai-compute/
#asahi-dev · 2026-08-25 13:15
jannauany complications in calling apple_cio_tbt_switch_set(TYPEC_THUNDERBOLT_SWITCH_OFF) during suspend? either directly (or through tipd, probably not as it shouldn't know that acio needs switched off)
#asahi-dev · 2026-08-25 12:56
jannauacio appears to be shutdown on suspend. last rtkit syslog message "yslog message: clx.c:776: CLX CFG: clx_en changed 0x0 --> 0x3" and before that two messages with tear_down
#asahi-dev · 2026-08-25 12:54
jannausven: SError on wakeup in DART (t6020: 0xb01a80080)
#asahi-dev · 2026-08-25 12:47
jannauno SError if I'm not mistaken
#asahi-dev · 2026-08-25 12:47
jannauI did accidentally at some point with tbt-reset-wip, iirc no wakeup but I'm not sure macsmc-input was enabled
#asahi-dev · 2026-08-25 09:02
jannaureported but not sure if Apple still cares about Sequoia
#asahi-dev · 2026-08-25 08:50
jannauinvalid PTE on dart-acio1 sid 1
#asahi-dev · 2026-08-25 08:47
jannauwow, I've panicked macos with iperf3 over usb4net
#asahi-dev · 2026-08-25 08:45
jannauso remap works there as well
#asahi-dev · 2026-08-25 08:44
jannausolved, it was fedora's firewall
#asahi-dev · 2026-08-25 08:35
jannautcp connect fails immediately with "Failed to connect to" with the correct ip but on not assigned ip addresses it takes normal timeout of ~3s
#asahi-dev · 2026-08-25 08:33
jannauless than 1ms and the connected device should not have any other network interface up
#asahi-dev · 2026-08-24 17:20
jannauah, I did the dma-range to #iommu-cells conversion before merging the m3 dts. I was already wondering why we don't need 5 iommu-cells
#asahi-dev · 2026-08-24 16:38
jannauiirc port numbers should be without gaps (even if there is a reason for the gap)
#asahi-dev · 2026-08-24 16:36
jannauit's 24
#asahi-dev · 2026-08-24 16:26
jannauisn't it "just" 23 IOMMU entries on M3*?
#asahi-dev · 2026-08-24 15:50
jannauI guess they never found the time/justification to implement this in the driver
#asahi-dev · 2026-08-24 10:31
jannausven: the crazy tbt setup broke (different selection of DUTs though). only 2 host on DFPs are detected without apparent cause (without debug logs). 3rd host is running macos
#asahi-dev · 2026-08-24 10:28
jannauchadmed: I did not. I merged it because v7.2 did rename drm_atomic_state to drm_atomic_commit and I wanted to avoid causing conflicts. it's impossible for anyone to use that "accidentally"
#asahi-dev · 2026-08-24 08:28
jannauI seem to have compared the wrong kernel builds. remap helps in all cases but the effect seems to be negligible when using DMA-FQ and dropping the tlb flush after mapping
#asahi-dev · 2026-08-24 08:17
jannauso it has to be smarter than re-routing it to all remapped streams
#asahi-dev · 2026-08-24 08:16
jannauremap helps if I keep the flush tlb in apple_dart_iotlb_sync_map. 1.86 Gbits/sec -> 10.7 Gbits/sec
#asahi-dev · 2026-08-24 08:06
jannaumost if not all of the improvement comes from DMA-FQ / map_pages non-invalidation
#asahi-dev · 2026-08-24 07:57
jannauacceptable '14.7 / 22.6 Gbits/sec' now but I did multiple changes: DMA-FQ and no invalidate after map_pages. I'll test the impact of remap now
#asahi-dev · 2026-08-24 07:52
jannaujust done that. apple_dart_hw_enable_translation() only sets TTBR for the first sid and all other enabled sids just remap to the first. invalidate_tlb then breaks after invalidating the first sid
#asahi-dev · 2026-08-24 07:21
jannaunot yet on m3*

plus ›