#
…

2026-06-14

00:00hdbngr has quit [Ping timeout: 480 seconds]
00:01ddxtanx has quit [Quit: Konversation terminated!]
00:01ddxtanx has joined #asahi-dev
00:15kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
01:16zumi has quit [Remote host closed the connection]
01:17zumi has joined #asahi-dev
01:22zumi has quit [Remote host closed the connection]
01:46kaleryeblueberryoboe has joined #asahi-dev
01:47hdbngr has joined #asahi-dev
01:56kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
01:56StanfordAppliedCyber[m]dfu restore, update to 26.5.1, smp still does not work
01:57StanfordAppliedCyber[m]known good m1n1 binary
01:58hdbngr has quit [Ping timeout: 480 seconds]
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:00readyee[m]Did you do software restore or firmware restore/“revive Mac”
02:04kaleryeblueberryoboe has joined #asahi-dev
02:09StanfordAppliedCyber[m]everything got wiped
02:11StanfordAppliedCyber[m]i'm really at a loss here, i'm not sure how to make progress nor how to debug this
02:11readyee[m]Try revive if something changed in the firmware going back to the older version might fix the issue
02:12StanfordAppliedCyber[m]well the issue is that 26.5 also broke for me for reasons I don't understand
02:12StanfordAppliedCyber[m]like i absolutely would have stayed on 26.5
02:12StanfordAppliedCyber[m]but i had to install the real macos kernelcache so i could boot and get some logs for a failing process
02:13StanfordAppliedCyber[m]that told me exactly what my error was, so my hypervisor should fully work and make it to a working userspace
02:13StanfordAppliedCyber[m]but i can't test it because smp doesn't work and setting cpus=1 just doesn't work, pmgr panics, if you set cpus=1 then toss the adt nodes, same deal
02:16StanfordAppliedCyber[m]and yep, debugusb doesn't work any any baud faster than 115200 doesn't work
02:18StanfordAppliedCyber[m]actually does anyone have ideas what underlying thing could be wrong with my machine? on 26.5, the moment i installed it then smp worked in raw mode, and it kept working for a while
02:18StanfordAppliedCyber[m]it only broke yesterday after i install the regular macos kernelcache
02:19StanfordAppliedCyber[m]i reboot into it, do a bunch of stuff, read logs, etc, then reflash m1n1
02:19StanfordAppliedCyber[m]then smp just.... doesn't work, the pstate is correct but the cores just never drop out of reset, same as 26.4.1
02:21readyee[m]By any chance did you enable CPU frequency before SMP?
02:24readyee[m]In the Shell.PY while the system booting up it tells you like what core you’re booted on what firmware you’re on did you double check that is the system and firmware is a 26.5.1
02:25StanfordAppliedCyber[m]everything looks totally normal as far as i can tell
02:33StanfordAppliedCyber[m]also for context all i'm doing is macvdmtool reboot serial, waiting a few seconds, then M1N1DEVICE=/dev/cu.debug-console python3 proxyclient/tools/shell.py, with the only patch being setting the baud to 115200 because anything higher breaks
02:33StanfordAppliedCyber[m]and from inside the shell running p.smp_start_secondaries() and proceed to see a wall of fail
02:34readyee[m]Do you think the chickens could be doing something with it?
02:35tobhe_ has joined #asahi-dev
02:39StanfordAppliedCyber[m]i mean, possibly? but why then does everyone else's device work, but mine doesn't
02:40StanfordAppliedCyber[m]i've done the most work on the neo but no one else hit as many problems as i did, everyone else i talked to had everything just work
02:42tobhe has quit [Ping timeout: 480 seconds]
02:48StanfordAppliedCyber[m]also apparently iboot is identical between 26.5 and 26.5.1
03:17kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
03:18StanfordAppliedCyber[m]can someone send me the full sysreg dump from a neo? specifically one where smp works
03:23kaleryeblueberryoboe has joined #asahi-dev
03:28codingkoopa3218 has quit []
03:28codingkoopa3218 has joined #asahi-dev
03:32codingkoopa3218 has quit []
03:34kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
03:37nora has joined #asahi-dev
03:39nora_ has quit [Ping timeout: 480 seconds]
03:42kaleryeblueberryoboe has joined #asahi-dev
03:50kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
03:57kaleryeblueberryoboe has joined #asahi-dev
04:06kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
04:11codingkoopa3218 has joined #asahi-dev
04:13kaleryeblueberryoboe has joined #asahi-dev
04:22kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
04:23pb17 has quit [Ping timeout: 480 seconds]
04:28hdbngr has joined #asahi-dev
04:29kaleryeblueberryoboe has joined #asahi-dev
04:30pb17 has joined #asahi-dev
04:37hdbngr has quit [Ping timeout: 480 seconds]
04:37kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
04:45kaleryeblueberryoboe has joined #asahi-dev
04:53kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
05:00kaleryeblueberryoboe has joined #asahi-dev
05:00hdbngr has joined #asahi-dev
05:09hdbngr has quit [Ping timeout: 480 seconds]
05:32StanfordAppliedCyber[m]yall think it's worth it to do a physical battery pull of this thing?
05:32StanfordAppliedCyber[m]it's the only debugging step i haven't done
05:32StanfordAppliedCyber[m]but i doubt it'll work
05:49StanfordAppliedCyber[m]yep, i did that thing where i load xnu, don't block the cpustart write, see if xnu can start its own cores, and no, xnu panics with the core didn't start message, nothing in my spin table
06:25kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
06:27hdbngr has joined #asahi-dev
06:32mattia013 has joined #asahi-dev
06:33kaleryeblueberryoboe has joined #asahi-dev
06:36hdbngr has quit [Ping timeout: 480 seconds]
06:41hdbngr has joined #asahi-dev
06:41yukaThere is something other than WFI losing state causing linux crashes with smp enabled on M4+
06:43yukaI now have a clean environment to test patches for WFI losing state in, by running it on M3 and setting CYC_OVRD_WFI_MODE(0) or CYC_OVRD_WFI_MODE(3)
06:44yukaThis CYC_OVRD setting makes the M3 cores intentionally lose state on WFI, and this way I was able to confirm that replacing wfi and wfit with nops works!
06:44yukaThere is some other race condition on M4 which causes the secondaries to crash with odd (seemingly random) stack traces
06:55hdbngr has quit [Ping timeout: 480 seconds]
07:02Tramtrist has quit [Remote host closed the connection]
07:11Tramtrist has joined #asahi-dev
07:11Tramtrist has quit []
07:11Tramtrist has joined #asahi-dev
07:22midou has quit [Read error: Connection reset by peer]
07:26midou has joined #asahi-dev
07:43kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
07:50kaleryeblueberryoboe has joined #asahi-dev
07:58kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
08:01kaleryeblueberryoboe has joined #asahi-dev
08:01hdbngr has joined #asahi-dev
08:11hdbngr has quit [Ping timeout: 480 seconds]
08:16kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
08:17kaleryeblueberryoboe has joined #asahi-dev
08:21hdbngr has joined #asahi-dev
08:22gargantua has joined #asahi-dev
08:30hdbngr has quit [Ping timeout: 480 seconds]
08:42hdbngr has joined #asahi-dev
08:47ddxtanx_ has joined #asahi-dev
08:47ddxtanx has quit [Read error: Connection reset by peer]
08:49kaleryeblueberryoboe has quit [Remote host closed the connection]
08:50kaleryeblueberryoboe has joined #asahi-dev
08:55hdbngr has quit [Ping timeout: 480 seconds]
09:08kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
09:14kaleryeblueberryoboe has joined #asahi-dev
09:28ddxtanx_ has quit [Remote host closed the connection]
09:30ddxtanx has joined #asahi-dev
09:34kaleryeblueberryoboe has quit [Remote host closed the connection]
09:36svenfwiw, we essentially got confirmation that wfi losing state on m4+ is expected because that cyc_ovrd thing we set just doesn’t exist anymore
09:36ddxtanx has quit [Remote host closed the connection]
09:36svenit’s effectively a cpu errata :/
09:43hdbngr has joined #asahi-dev
09:43chaos_princessput a wfe there
09:45yukaI mean, I can try it. It led to problems in the past.
09:45yukaAnd btw, this still does not fully make M4 boot reliably, it still crashes
09:45yukajust not because of losing state
09:48kaleryeblueberryoboe has joined #asahi-dev
09:48gargantu1 has joined #asahi-dev
09:50gargantua has quit [Ping timeout: 480 seconds]
09:51hdbngr has quit [Ping timeout: 480 seconds]
10:00yukaNope, wfe does not work, not even for booting a tiny kernel. It simply gets stuck and never wakes up again once the secondaries are initialized.
10:01kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
10:02kaleryeblueberryoboe has joined #asahi-dev
10:05hdbngr has joined #asahi-dev
10:10gargantu1 has quit [Ping timeout: 480 seconds]
10:12___nick___ has joined #asahi-dev
10:13yukaDoes anyone have an M4 Max and M5* to pull the MIDR part no from and run the WFI experiment?
10:15hdbngr has quit [Ping timeout: 480 seconds]
10:18yukaI'm strongly guessing M4* and M5* are all affected. Is there some other place I could find the MIDR part numbers?
10:22kaleryeblueberryoboe has quit [Remote host closed the connection]
10:23kaleryeblueberryoboe has joined #asahi-dev
10:25gargantua has joined #asahi-dev
10:29jannauonly the ones mentioned in Documentation/arch/arm64/cpu-feature-registers.rst. I can add the remaining ones but those are half 0 anyway
10:31jannauyuka: does the kernel boot with "nohlt" in the kernel command line
10:31NickChan[m]yuka: download xcode 27 or xcode 27 command line tools, nagivate to the MacOSX.sdk, then look at System/Library/Frameworks/Kernel.framework/Versions/A/Headers/arm/cpuid.h
10:31kaleryeblueberryoboe has quit [Remote host closed the connection]
10:33qyliss has quit [Quit: bye]
10:34NickChan[m]Note the "new performance cores" in M5 Pro/Max is m-Core in that header
10:34NickChan[m]and the "Super Cores" are p-Cores in the header
10:34qyliss has joined #asahi-dev
10:34yukaYes I already noticed that in the adts
10:34NickChan[m]"Efficiency cores" are still, well, e-Core
10:38bigmac has quit [Ping timeout: 480 seconds]
10:43kaleryeblueberryoboe has joined #asahi-dev
10:44hdbngr has joined #asahi-dev
10:52hdbngr has quit [Ping timeout: 480 seconds]
10:59___nick___ has quit [Remote host closed the connection]
11:06___nick___ has joined #asahi-dev
11:10yukaI will run the wfi experiment once more on my M1
11:16bigmac has joined #asahi-dev
11:23__nick__ has joined #asahi-dev
11:27___nick___ has quit [Ping timeout: 480 seconds]
11:28NickChan[m]technically applies to older socs too, but on them the deep wfi loses all state, so it's pointless to try to make force power down mode this way
11:28hdbngr has joined #asahi-dev
11:29NickChan[m]so M1+ appears to be the right call
11:35kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
11:37hdbngr has quit [Ping timeout: 480 seconds]
11:38nickchanah, I see it's replacing wfi with nop
11:39nickchanin which case that workaround would very much work on a7-a11
11:43kaleryeblueberryoboe has joined #asahi-dev
11:45hdbngr has joined #asahi-dev
11:53hdbngr has quit [Ping timeout: 480 seconds]
11:55yukaWFI experiment on my M1 does exactly the same: With CYC_OVRD_WFI_MODE(0) the boot core (core 0) retains register state, the others don't
11:58user890104what you're discussing reminds me of an issue I have on iPod nano 7th gen (S5L8740 with 1x Cortex-A5 core) - we use the nohlt kernel arg as a workaround
11:59nickchanthis also has complications for kvm but it's not like that would work on m4+ right now
12:00chaos_princesswfi in a vm just kicks you out to el2, no?
12:00hdbngr has joined #asahi-dev
12:01nickchanthere's HCR_EL2.TWI so I suppose so
12:03jannautrapping wfi isn't architecturally guaranteed but I can't see how a broken wfi may supported without trapping
12:09hdbngr has quit [Ping timeout: 480 seconds]
12:11__nick__ has quit []
12:12___nick___ has joined #asahi-dev
12:18kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
12:19yukaLike the plan is to re-introduce a way to properly sleep later in the boot process. This erratum is mostly in order to stop it from crashing earlier.
12:20yukasimilar to (out-of-tree) cpuidle-apple minus setting CYC_OVRD_WFI_MODE
12:20yukanot out of tree but not upstream, you know what I mean
12:21yukaWe also got flokli's M4 Pro to boot with smp with this patch and a lot of luck. It still crashes with seemingly random stacktraces most of the time though.
12:22kaleryeblueberryoboe has joined #asahi-dev
12:22nickchanif wfi loses state hw is not required to initialize to 0 again
12:22nickchanso random can still be lose state
12:23nickchanupdated soc codenames
12:25nickchanapple stopped stuffing secret codenames in ane fw though so only nevis and borneo
12:26nickchan...are added, then for some already known ones i added their other known names
12:26yukanickchan: the pattern is: something tries to write somewhere (can be tcp_init, some spinlock, allocator, whatever). the addresses it's trying to access (write to) look like a valid kernel address, but they are seemingly not mapped rw (they lost the 'w' bit? or they are the wrong kernel addresses in the first place, ...).
12:28nickchanthe only explaination i have for that is iboot setup the chicken bits wrong somehow
12:28nickchanwhich can lead to all sorts of botched behaviors as seen on previous socs
12:29yukaMy plan (not today) is to try to take one of the commonly occurring crash sites, and see which allocation should be accessed, then add logging to where it's allocated, and check if the page is rw at allocation and at access time.
12:30yukabotched chickens are not at all ruled out, but we need some more detailed information before we can file a bug report
12:43flokliAlso sent a PR for t6040 m1n1 bits. With maxcpus=1 it's booting linux with a minimal device tree (which I also intend to send), no need to only have this sitting on my machine.
12:46hdbngr has joined #asahi-dev
12:51sventhe aic sysreg will allegedly be fixed with the next beta fwiw
12:52nickchanprobably need a radar for the L2C registers as well
12:53nickchannot being able to handle or disable serror reporting is a weirdly broken state
12:53svendo we care? I don’t think we use that in Linux
12:54kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
12:54nickchansven: main impact is it's more easy to have unrecoverable errors by sending the wrong things via uartproxy
12:54hdbngr has quit [Ping timeout: 480 seconds]
12:55kaleryeblueberryoboe has joined #asahi-dev
12:56nickchanBut I think my opinion on this will be biased since I do not luxuries like "automatically booting to m1n1" in most of my m1n1 experiences
13:01yukait's quite annoying indeed when trying to scan mmio ranges
13:09gargantu1 has joined #asahi-dev
13:09kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
13:11gargantua has quit [Ping timeout: 480 seconds]
13:13kaleryeblueberryoboe has joined #asahi-dev
13:18yukawhere does the #define SYS_IMP_APL_CYC_OVRD ... belong in the kernel?
13:22kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
13:25maettu102 has quit [Quit: brb]
13:29kaleryeblueberryoboe has joined #asahi-dev
13:29maettu102 has joined #asahi-dev
13:35maettu102 has quit [Quit: brb]
13:37ralim8 has quit [Quit: The Lounge - https://thelounge.chat]
13:37ralim8 has joined #asahi-dev
13:39maettu102 has joined #asahi-dev
13:39kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
13:41kaleryeblueberryoboe has joined #asahi-dev
13:47hdbngr has joined #asahi-dev
13:49kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
13:55hdbngr has quit [Ping timeout: 480 seconds]
13:57kaleryeblueberryoboe has joined #asahi-dev
14:11Sobek[m]nickchan: And that chicken bit set up would be different from the one set up for macOS ?
14:12Sobek[m](Or macOS doesn't rely on WFI / WFI keeping state properly)
14:17gargantu1 has quit [Ping timeout: 480 seconds]
14:19berzrk_ has quit [Ping timeout: 480 seconds]
14:20chaos_princessmacos saves registers around wfi
14:21hdbngr has joined #asahi-dev
14:27Sobek[m]Hmm, and from the ARM spec perspective, is WFI required to save state or is the CPU current behaviour compliant ?
14:28chaos_princess"In some implementations, based on the configuration of system specific registers, WFI can be used as part of a
14:28chaos_princesspowerdown sequence where no interrupts will cause WFI wakeup events, and restoration of power involves resetting of
14:28chaos_princessthe PE. In those cases, the WFI is permitted to cause a loss of architectural state, as it is assumed that this state will have been saved by software as part of the powerdown sequence before the WFI."
14:28jannau"If the system is configured such that the WFI or WFIT instruction can be completed, then the architecture requires that the architectural state is not lost."
14:28chaos_princessand that's it
14:30hdbngr has quit [Ping timeout: 480 seconds]
14:31kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
14:32kaleryeblueberryoboe has joined #asahi-dev
14:44kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
14:44rensenwxre1 has quit []
14:47rensenwxre1 has joined #asahi-dev
14:48hdbngr has joined #asahi-dev
14:49kaleryeblueberryoboe has joined #asahi-dev
14:56hdbngr has quit [Ping timeout: 480 seconds]
14:59hdbngr has joined #asahi-dev
15:08hdbngr has quit [Ping timeout: 480 seconds]
15:09hdbngr has joined #asahi-dev
15:18hdbngr has quit [Ping timeout: 480 seconds]
15:18svenbut there’s not much Apple can do if that cyc_ovrd bit is gone at this point
15:20gargantua has joined #asahi-dev
15:27kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
15:34kaleryeblueberryoboe has joined #asahi-dev
15:36yukacan't make checkpatch.pl happy :(
15:49hdbngr has joined #asahi-dev
15:53svenwhat is it complaining about?
15:53yuka"ERROR: Macros with complex values should be enclosed in parentheses"
15:54jannaufor which line?
15:54yukaThen I put it in a do {} while (0) loop
15:54yukaand I get "WARNING: Single statement macros should not use a do {} while (0) loop"
15:54yukafor the wfi() macro with the ALTERNATIVE added
15:54svenyou also want at least two commits i think, one to add the MIDRs and another one for the wfi alternative thing
15:55yukaAh fair I looked at past commits which seemed to add the MIDRs at the same time as adding the errata, but maybe this one is a bit more controversial/complicated than those
15:55sven#define wfit(val) asm volatile(ALTERNATIVE("msr s0_3_c1_c0_1, %0", "msr s0_3_c1_c0_0, %0", ARM64_WORKAROUND_WFI_STATE) \ <-- what's that msr s0_3... thing? shouldn't that also be nop?
15:55yukait's wfit / wfet
15:55yukaI changed it to a nop already locally
15:56svenhm... i think the only SYS_IMP_APL regs we need in the kernel are for aic and those are just defined directly at the driver
15:56sveni guess the place can also just be yak shaved on the ML then :D
15:56jannaum1n1's i2c / smbus seems to busted. code to enter debugusb works when running under hv with i2c tracing
15:57svenah, pmu has some SYS_IMP_APL as well but there's also in a header file for it
15:57jannauapple's performance counters has sys_impl_apl as well but those are local as well
15:57yukathere doesn't seem to be a general apple sysregs header yet
15:58sveni'd also add "This workaround no longer exists since M4" to the commit message just to make it clear we're not doing this for fun
15:58jannaunot sure there should be one
16:00svenalso, fwiw, checkpatch is generally just a recommendation and does have false positives sometimes (especially for WARNINGS)
16:01kaleryeblueberryoboe has quit [Remote host closed the connection]
16:01kaleryeblueberryoboe has joined #asahi-dev
16:04hdbngr has quit [Ping timeout: 480 seconds]
16:04yukaI can count the number of patches I've sent to kernel mailing lists on one hand. Can I send my draft to some of you by mail to get a review before I send it to the list?
16:05svenfeel free to send them to sven@kernel.org
16:05hdbngr has joined #asahi-dev
16:10chaos_princesscover letter?
16:10svenyeah, i'd add a cover letter. and probablyt u
16:10sven*and probably use FIELD_GET for the CYC_OVRD check
16:11sveni.e. #define CYC_OVRD_WFI_MODE_MASK GENMASK(...) and then FIELD_GET(read_sysreg_s(...), CYC_OVRD_WFI_MODE_MASK) != 2
16:11yukanoted, thanks
16:11svenyou'll get complaints about the #define location but that's fine since we have no idea where to put them
16:13chaos_princessi also think other errata has kconfig symbols so people can compile it out
16:13svenwhatever you used to send the email works otherwise, it looks good
16:13gargantua has quit [Read error: Connection reset by peer]
16:14gargantua has joined #asahi-dev
16:14chaos_princessalso, recommendation - use b4
16:14sven^-- that
16:14svenespecially useful if you need a v2, v3, etc.
16:17kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
16:19hdbngr has quit [Ping timeout: 480 seconds]
16:19hdbngr has joined #asahi-dev
16:24kaleryeblueberryoboe has joined #asahi-dev
16:28hdbngr has quit [Ping timeout: 480 seconds]
16:29hdbngr has joined #asahi-dev
16:33gargantu1 has joined #asahi-dev
16:37jannaui2c fixed by switching from checking STATUS_XFER_BUSY to STATUS_XFER_ENDED
16:37gargantua has quit [Ping timeout: 480 seconds]
16:37hdbngr has quit [Ping timeout: 480 seconds]
16:38jannauenabling debugusb works now from within m1n1 on devices with i2c usb-c port controller
16:45hdbngr has joined #asahi-dev
16:47kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
16:54kaleryeblueberryoboe has joined #asahi-dev
17:04kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
17:05kaleryeblueberryoboe has joined #asahi-dev
17:08hdbngr has quit [Ping timeout: 480 seconds]
17:13nickchanI want to suggest some changes, but it looks like that git server isn't reachable right now :/
17:14nickchannamely pre-M1 should also be on the list
17:15nickchan(and i do not have th patches downloaded now)
17:18svenmeh, pre-m1 can always come later
17:18yuka has quit [Remote host closed the connection]
17:18svenand why do you need pre-m1 support?
17:18nickchansven: nah not really
17:18svendoesn't that already somehow work?
17:18yuka has joined #asahi-dev
17:18nickchanbased on previous interactions the only way to get cpu midr into the kernel is to lump them together with M-series...
17:18svenuh... where?
17:18nickchansven: uses same workaround as m1
17:18nickchanwfi force up
17:20svenwhich brings me back to "pre-m1 can always be a follow up patch"
17:21sveni half-expect only the workaround for m4+ will be accepted because really m1n1 is the only bootloader and we always set the correct bits in cyc_ovrd
17:21nickchanwell then that's fine it's not like having unused midr will do much
17:22yuka has quit [Remote host closed the connection]
17:22yuka has joined #asahi-dev
17:23nickchanI suppose I could summarize:
17:23yuka has quit [Remote host closed the connection]
17:24nickchanon A10/A10X/T2 the MPIDR_EL1 can appear to change while switching pstate because it's switching physical cores
17:24yuka has joined #asahi-dev
17:24nickchanas well as aarch32 can only be executed in performance core, attempting to do so on e-core or switching to p-core while executing results in illegal instruction
17:25nickchanand let's say people really do not like these kinds of behavior
17:25nickchanenough that the pmu patches that had little to do with that is affected
17:26nickchanwell, the server is accessible now
17:26yuka has quit [Remote host closed the connection]
17:26sveni'm very confused how any of this is related to the wfi errata thing
17:26yuka has joined #asahi-dev
17:27nickchanit is not
17:28kaleryeblueberryoboe has quit [Read error: Connection reset by peer]
17:28nickchani am looked at chance to upstream the patch that added all the pre-M1 midrs
17:28nickchan(i am using those midrs to ban executing aarch32 on a10 but that's besides the point)
17:29nickchanand by adding a workaround for all the apple silicon cpus which could all lose state on wfi, I can upstream that midr patch
17:29yuka has quit [Remote host closed the connection]
17:29yuka has joined #asahi-dev
17:30yukanickchan: do you have a link to the patch adding the other MIDRs?
17:31yukaAh, finally, my irc is also working again
17:31nickchanyuka: i am preparing it
17:32sveni still think that "the only way to get cpu midr into the kernel is to lump them together with M-series" conclusion is wrong at the very best. but i also don't see how adding them hurts *shrug*
17:32chaos_princesssomething something kernel is american congress where you need to stuff a year full of legislation^W patches into one bill :P
17:39yukaDo we know if there are _any_ devices with Apple MIDR which do _not_ have this behavior?
17:40yukaI mean, how far back did they use the Apple MIDR, and when did they introduce the deep wfi...
17:41chaos_princessi think all of them? at least all arm64 ones
17:43nickchani think everything with apple midr has the wfi lose state behavior
17:43nickchanthe only difference is on older socs the consequence of losing state is more severe
17:43nickchanas in, lose all state and restart from rvbar and highest EL on older
17:44nickchanbut on m1 the cpu really only looses gpr and simd state
17:44kaleryeblueberryoboe has joined #asahi-dev
17:46maettu102 has quit [Quit: brb]
17:46yukaThen we could just enable it for apple midr unconditionally? As I understand, there's no plan to change/fix this behavior anyways, and if they do change it we can always add an upper bound later.
17:50sventhat works I guess
17:51maettu102 has joined #asahi-dev
17:51hdbngr has joined #asahi-dev
17:51nickchani think that can work
17:51maettu102 has quit []
17:54nickchanenable it for cpu implementer 0x61 specifically
17:54jannaunickchan: if the older CPU cores go through reset that's not not architecturally unsound as the wfi never completes
17:55nickchani do know that but the workaround is still as valid as on m1,m2,m3
17:55jannaustill needs the fixup but someone might argue that those cores should boot with nohlt
17:56maettu102 has joined #asahi-dev
17:56nickchannohlt should work with m4 too shouldn't it
17:57yukawfit is used in delay
17:57yukaregardless of cpuidle
17:57jannaum4 has WFxT
17:59hdbngr has quit [Ping timeout: 480 seconds]
18:06steinsgate has joined #asahi-dev
18:19chaos_princessyuka: using b4 now?
18:19gargantu1 has quit [Ping timeout: 480 seconds]
18:19yukaWell b4 and then send the eml files using git send-email because I already configured that
18:20chaos_princessfor single-commit series, do not use the cover letter, just leave it as "EDITME:" or whatever it generates by default, and instead write the whole thing in the patch description
18:21yukaI wanted at least the comment about the defines location to go into the comment field
18:21yukabelow the --
18:22chaos_princessah. if it is intentional, then ok
18:22chaos_princessReviewed-by: Sasha Finkelstein <k@chaosmail.tech>
18:30___nick___ has quit [Remote host closed the connection]
18:36svenlooks good to me, will send r-b once it's on the list
18:51hdbngr has joined #asahi-dev
18:54kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:00yukaAnother question regarding upstream stuff (device tree bindings): I discovered some time ago that on A18 Pro we need to take into account the no-restore-gpio-pins adt property and tell the regmap to not read and write the regs of the gpios listed there. So essentially we would need a new apple,t????-pinctrl compatible to be able to express this, correct?
19:00hdbngr has quit [Ping timeout: 480 seconds]
19:01yukaor is adding an optional property possible in the existing compatible?
19:01chaos_princesshow bad it is if you do read/write it?
19:02yukabecause then the old driver which does not support the property yet will not attempt to operate on this device tree node?
19:03yukaI have not seen this on any other device, so I would name it after t8140?
19:03yukaI don't know for sure if some M4 / Pro / Max laptop doesn't have the same
19:04xal has quit []
19:04xal has joined #asahi-dev
19:04chaos_princesswe briefly had something similar in admac
19:05mattia013 has quit [Remote host closed the connection]
19:09mattia013 has joined #asahi-dev
19:10hdbngr has joined #asahi-dev
19:11sveni'd just use the first device where it was noticed
19:11svenif we had documentation i'd prefer the first device released where it happened but oh well
19:16kaleryeblueberryoboe has joined #asahi-dev
19:26hdbngr has quit [Ping timeout: 480 seconds]
19:27marvin24 has joined #asahi-dev
19:29kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:29nela has quit [Quit: Ping timeout (120 seconds)]
19:30nela has joined #asahi-dev
19:39chaos_princessright, it should check if we are actually at el2 and not el1
19:41svenoh right, el2 should just make sure to trap wfi from el1 and ensure it works
19:44chaos_princesswhat happens if wfi runs at el1 and isn't trapped? also state loss?
19:44kaleryeblueberryoboe has joined #asahi-dev
19:44svenno idea actually, but i'd guess so
19:45chaos_princessi don't think so
19:45sventhough actually.. who knows.
19:45chaos_princesscause wouldn't it break hvf completely then?
19:45svenyeah, that's what just came to my mind as well
19:45svenunless they somehow always trap it there i guess?
19:55kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:56crabbedhaloablut has quit []
19:57crabbedhaloablut has joined #asahi-dev
20:04yukaCan we run m1n1 in hvf? :D
20:07hightower2 has joined #asahi-dev
20:10chaos_princessit can run under m1n1 in el1
20:43kaleryeblueberryoboe has joined #asahi-dev
20:53hdbngr has joined #asahi-dev
20:53zenmaya[m] has joined #asahi-dev
20:55mattia013 has quit [Remote host closed the connection]
21:01hdbngr has quit [Ping timeout: 480 seconds]
21:06jannauI've tagged asahi-7.0.12-1 with the macOS 27 smc firmware battery fix
21:28hdbngr has joined #asahi-dev
21:34ddxtanx has joined #asahi-dev
21:36hdbngr has quit [Ping timeout: 480 seconds]
21:49hdbngr has joined #asahi-dev
22:03hdbngr has quit [Ping timeout: 480 seconds]
22:51heartburn has joined #asahi-dev
23:12hdbngr has joined #asahi-dev
23:21hdbngr has quit [Ping timeout: 480 seconds]
23:24ohtwoone has quit [Quit: Cya]
23:32ohtwoone has joined #asahi-dev
23:35hdbngr has joined #asahi-dev
23:44hdbngr has quit [Ping timeout: 480 seconds]
…