#
✕ nick:stintel
#asahi-alt · 2024-03-27 13:25
stintelgoing to reboot in 6.8 and try again to see what it does in a hopefully clean state
#asahi-alt · 2024-03-27 13:23
stintelyeah me neither
#asahi-alt · 2024-03-27 13:21
stintel01:00.1 Network controller: Broadcom Inc. and subsidiaries BRCM4387 Bluetooth Controller (rev 07)
#asahi-alt · 2024-03-27 13:21
stintel01:00.0 Network controller: Broadcom Inc. and subsidiaries BCM4387 802.11ax Dual Band Wireless LAN Controller (rev 07)
#asahi-alt · 2024-03-27 13:21
stintel15" air m2 here
#asahi-alt · 2024-03-27 13:21
stintelnyaomixyz: what hardware are you on?
#asahi-alt · 2024-03-27 13:18
stintel:)
#asahi-alt · 2024-03-27 13:12
stintelchanges between 14 and 16 seem mostly drm related, so I'd say unlikely this would introduce a regression in bt
#asahi-alt · 2024-03-27 13:09
stintelthat's actually how I ran bcachefs.git for years :)
#asahi-alt · 2024-03-27 13:08
stintelv6.6.5 no longer merged cleanly and conflicts were too many so that's why I stuck with 6.6.4
#asahi-alt · 2024-03-27 13:08
stintelI switched to 6.8 before 6.6-15 came out, iirc
#asahi-alt · 2024-03-27 13:08
stintelwhat I usually do is: git checkout asahi-6.6-14; git checkout -b build-6.6.x; git merge linux-stable/v6.6.x
#asahi-alt · 2024-03-27 13:07
stintelthose are two tags, so you should be able to exactly replicate that
#asahi-alt · 2024-03-27 13:07
stintellike I said, asahi-6.6-14 + linux-stable/v6.6.4
#asahi-alt · 2024-03-27 13:06
stintelGentoo
#asahi-alt · 2024-03-27 13:05
stintelI guess it depends on how corrupt things are :)
#asahi-alt · 2024-03-27 13:03
stintelscrub is missing, but fsck is there, and can repair lots of things
#asahi-alt · 2024-03-27 13:02
stinteltooling and repairability is actually already quite good in bcachefs
#asahi-alt · 2024-03-27 09:42
stintelso there's not many options :)
#asahi-alt · 2024-03-27 09:41
stinteland I consider any other fs without checksumming and snapshots not a proper fs for this age
#asahi-alt · 2024-03-27 09:40
stintelbeen bitten too often by btrfs bugs
#asahi-alt · 2024-03-27 09:40
stintelI've been using it since 2018, and now that it's finally in mainline I plan to move all my non-embedded Linux machines (physical and virtual) to bcachefs rootfs
#asahi-alt · 2024-03-27 09:27
stintelright, that means I should hold off on reinstall with bcachefs rootfs until this is sorted, as I won't be able to boot that with 6.6
#asahi-alt · 2024-03-27 09:21
stintelI've been running it fine without other issues for ~2 weeks
#asahi-alt · 2024-03-27 09:21
stintel"completely unusable" --> bluetooth audio is completely unusable, to be clear
#asahi-alt · 2024-03-27 09:20
stintelI'll make sure to keep this kernel around for when I'm traveling lol
#asahi-alt · 2024-03-27 09:20
stintelbooted into 6.6.4 (asahi-6.6-14 + linux-stable/v6.6.4) works like charm
#asahi-alt · 2024-03-27 09:19
stintel6.8.0 here is completely unusable, built from asahi-wip-6.8 342f90828c4167c94ee184ca05eeacb22255ac9b
#asahi-alt · 2024-03-27 09:16
stintelwhich kernel version you on?
#asahi-alt · 2024-03-27 09:15
stintelhaha, testing it now and it is completely unusable
#asahi-alt · 2024-03-27 09:10
stintelused*
#asahi-alt · 2024-03-27 09:10
stintelI've use my m2 air on a plane once or twice with B&W PX7 and afair it worked fine, but that was ~3 months ago, so might be a regression somewhere
#asahi-alt · 2024-03-27 09:09
stintelmaybe try completely unloading the brcmfmac* modules too
#asahi-alt · 2024-03-27 09:05
stintel2.4GHz is hit-or-miss, if you live in a house with no close neighbors, it can work very well, if you live in an area with apartment buildings ... good luck! I scan > 100 2.4 SSIDs, I've had times where toggling a zigbee light took ~10s
#asahi-alt · 2024-03-24 11:48
stintelsys-kernel/asahi-sources-6.8.0
#asahi-alt · 2024-03-24 11:48
stintel➜ cat /etc/portage/profile/package.provided
#asahi-alt · 2024-03-24 11:47
stintelyou can always use package.provided
#asahi-alt · 2024-03-05 13:59
stintelemerge: there are no ebuilds to satisfy "=media-libs/mesa-24.1.0_pre20240228[video_cards_asahi(+)]".
#asahi-alt · 2023-12-29 15:09
stintelsam_: here's the full build.log https://gist.github.com/stintel/e1e373ec7b47bea759b237a1eb624b8a
#asahi-alt · 2023-12-29 02:39
stintel8' that is compiled without support for 'fullfp16'
#asahi-alt · 2023-12-29 02:39
stintelsam_: so now it fails later during build: ../../third_party/xnnpack/src/src/amalgam/gen/neonfp16arith.c:114:36: error: always_inline function 'vaddq_f16' requires target feature 'fullfp16', but would be inlined into function 'xnn_f16_avgpool_minmax_ukernel_9p8x__neonfp16arith_c
#asahi-alt · 2023-12-29 01:34
stintellet me confirm tomorrow to be 100% sure
#asahi-alt · 2023-12-29 01:33
stintelsam_: oh wait, it got past the initial early error but because I had to leave before the build would finish, I never actually finished the build
#asahi-alt · 2023-12-29 01:32
stintelwhich probably explans why the person in the ticket was having trouble with enabling lld
#asahi-alt · 2023-12-29 01:31
stinteladding that in both CFLAGS and CXXFLAGS did not work
#asahi-alt · 2023-12-29 01:31
stintelsam_: don't remember if I got back to you about chromium with lld: it builds. but requires LDFLAGS="-fuse-ld=lld"
#asahi-alt · 2023-12-22 09:58
stintelso maybe I was still > 75 and didn't notice the echo 100 > charge_control_end_threshold not having worked, and somehow the plasma widget decided this meant "charger not powerful enough"
#asahi-alt · 2023-12-22 09:57
stinteliirc
#asahi-alt · 2023-12-22 09:57
stintelalso I read in the code something that charge_control_start_threshold is writeable to not confuse userspace but it's either 100 if charge_control_end_threshold is 100, otherwise it's always 75
#asahi-alt · 2023-12-22 09:56
stintelI guess it's just a bit finicky

plus ›