#
✕ nick:marcan
#asahi · 2023-01-05 04:26
marcanwonders if all these crazy ideas about downgrades being necessary / upgrades breaking things come from people used to jailbreaking and hostile manufacturers...
#asahi · 2023-01-05 04:25
marcanSam___: you never have to downgrade anything, you only have to *upgrade* if you have an ancient macOS version (which is impossible on M2)
#asahi · 2023-01-04 16:41
marcansven: I reuse the same code ipsws use for the OS images, that's why I use zip
#asahi · 2023-01-03 15:49
marcanyes
#asahi · 2023-01-03 15:40
marcanchange the UUID to the original one, manually reinstall m1n1 stage 1 with an edited UUID at the tail of the binary, or just delete your APFS+ESP partitions again and reinstall in UEFI mode and then put back the ESP contents.
#asahi · 2023-01-03 14:23
marcanthen there's also piles of new WPA modes and such that are also missing
#asahi · 2023-01-03 14:23
marcansupport for the new PCIe core revisions, sleep state mailbox over the ring buffer, new scan interface stuff, etc are just the required ones to make Apple chips work properly
#asahi · 2023-01-03 14:23
marcansee our patchset :p
#asahi · 2023-01-03 14:22
marcannow it's just 8 years behind in feature support
#asahi · 2023-01-03 14:22
marcanyes, brcmfmac was probably a good Broadcom driver 8 years ago
#asahi · 2023-01-03 14:21
marcancomplete with #define-gated device configuration and all that lovely stuff
#asahi · 2023-01-03 14:21
marcanbcmdhd is what all the Android devices use
#asahi · 2023-01-03 14:21
marcanwhile they neglect brcmfmac
#asahi · 2023-01-03 14:21
marcanit is, and they are
#asahi · 2023-01-03 14:21
marcan(other than firmware selection, all the Apple platform support stuff is just implementing missing features already in bcmdhd)
#asahi · 2023-01-03 14:20
marcanmeanwhile my job now is to port features from bcmdhd to brcmfmac because Broadcom only have <20% of one employee working on it
#asahi · 2023-01-03 14:20
marcansigh
#asahi · 2023-01-03 14:19
marcanlol firmware under "cyfmac" with Broadcom device names
#asahi · 2023-01-03 14:18
marcanuhh no I haven't
#asahi · 2023-01-03 14:17
marcanthat commit branded the BCM4355 as CC89459, and a PCI ID that already has a Broadcom device sub-name as a Cypress one
#asahi · 2023-01-03 14:17
marcanyes
#asahi · 2023-01-03 14:16
marcanI sure hope that's not the case and this is just a stupid rebrand
#asahi · 2023-01-03 14:16
marcanbut if the Cypress chips now have the same chip IDs as Broadcom chips but are actually not the same chip, now we have a bigger problem
#asahi · 2023-01-03 14:15
marcan(but already upstream)
#asahi · 2023-01-03 14:15
marcanwhich I now have to fix because judging by what the Broadcom guy is saying, the Cypress patch is wrong
#asahi · 2023-01-03 14:15
marcanthis was a patch submitted by cypress, and the author never replied to my original query
#asahi · 2023-01-03 14:08
marcan*I* certainly can't divine out of a crystal ball what silicon revisions you've shipped with what revision ID mappings so we can actually do firmware requests properly
#asahi · 2023-01-03 14:07
marcanyou *work for the company*, you can *find out*
#asahi · 2023-01-03 14:07
marcanmeanwhile the Broadcom developer literally answered "who knows" to some of my questions about what devices are out in the wild
#asahi · 2023-01-03 14:07
marcanthis is just such a giant waste of time for everyone
#asahi · 2023-01-03 14:06
marcan*fight to
#asahi · 2023-01-03 14:06
marcan(after he ignored my patch long enough that another developer found and submitted a fix for the same issue, whom he is now also nitpicking - I guess he wants us to fix to see who fixes all the nitpicks first?)
#asahi · 2023-01-03 14:05
marcannow the nvmem maintainer is asking me *not* to fix a bug that is naturally fixed by the minor refactor involved in fixing the bug I was trying to fix
#asahi · 2023-01-03 14:05
marcanI had the SD driver maintainer complain that I aligned a block of #defines after adding one more that pushed the tabstop. ended up having to leave it unaligned.
#asahi · 2023-01-03 14:04
marcan(of the bad ones I mean, there are some good ones)
#asahi · 2023-01-03 14:04
marcanand they *don't get* that this is a problem. any of them.
#asahi · 2023-01-03 14:04
marcanexactly.
#asahi · 2023-01-03 14:03
marcanyeah.
#asahi · 2023-01-03 14:03
marcansigh.
#asahi · 2022-12-28 13:56
marcanthat's suspend though, not s2idle, and we don't have that OF property so it *should* be considering it ticking in suspend either way
#asahi · 2022-12-28 13:53
marcancc maz, ever seen that one before? (CLOCK_MONOTONIC advancing in s2idle)
#asahi · 2022-12-28 13:49
marcanso that's a generic bug; who is "officially" to blame, I have no idea, but it's not specific to Apple
#asahi · 2022-12-28 13:48
marcanthis is all generic ARM stuff, and the ARM timers are never going to stop in s2idle
#asahi · 2022-12-27 17:57
marcanasahi-<version>-<tagver>
#asahi · 2022-12-27 17:55
marcanasahi tagged releases
#asahi · 2022-12-27 17:54
marcanI should test that...
#asahi · 2022-12-27 17:54
marcanthat would be a fun bug to file
#asahi · 2022-12-27 17:54
marcanbut then we should be able to cause epic filesystem corruption on macOS too
#asahi · 2022-12-27 17:54
marcanif we can't make it work with downstream hubs it's a lost cause anyway
#asahi · 2022-12-27 17:53
marcanI'm going to sleeep.

plus ›