14:05:56 <danmcd> @Woodstock --> Can you set the RTI wheels in motion for fenix illumos#17781  ?
14:05:57 <fenix> BUG 17781: swprintf(3c) does not work properly (New)
14:05:57 <fenix> ↳ https://www.illumos.org/issues/17781 | https://code.illumos.org/c/illumos-gate/+/4729
14:09:06 <Woodstock> danmcd: will do
20:31:08 <richlowe> rmustacc: I broke fmd's ses in the merge, do you know where logs about those things live?
20:48:25 <heat> hi, so I'm trying to boot illumos (i've tried latest omniosce stable, lts, and bloody; also openindiana) on qemu 11.0.0 and I keep getting this page fault, consistently: https://usercontent.irccloud-cdn.com/file/cmBY9w4H/image.png  I've tried playing with KVM vs no KVM, cpu types, no dice. any ideas? command line is "qemu-system-x86_64 -cdrom ~/Downloads/omnios-bloody-20260319.iso -m 8G --no-shutdown --no-reboot"
20:52:52 <heat> FTR i'm on an amd ryzen 9 9900X but I can repro this with KVM and _without_ KVM, so I don't think that's the problem
20:53:21 <danmcd> @heat it's at CPU bringup time... and you just confirmed it fails this way on the bare-metal, too?
20:54:06 <heat> i didn't try this on bare-metal, closest to that was qemu + kvm
20:57:21 <danmcd> Okay... it MAY be some smbios thing.  real_mode_stop_cpu_stage2_end+0x7171 is a call to smbios_open().
20:57:32 <danmcd> Lots of missing stack between that and the bcopy.
20:58:13 <richlowe> we had an arm bug on a freebsd host where the smbios just wasn't there, and we did something like that.
20:58:26 <heat> ok! so I tried OVMF to rule out BIOS vs UEFI, and OVMF seems to boot
20:58:29 <danmcd> smbios_open() calls bcopy 7 different times, and THAT's assuming bcopy isn't getting called further deep into smbios.
20:58:33 <danmcd> Thanks @richlowe.  :)
20:59:02 <heat> so whatever seabios is giving illumos, it's not enjoying it too much
20:59:11 <danmcd> Yeah... probably not.  :)
21:02:08 <richlowe> we failed earlier on arm most recently I guess, but the problem was we took the x86 code to find the SMBIOS tables in the old mpbios way where you scan low memory for a magic string.
21:02:46 <richlowe> and that's bad news both in general in a machdep way, and in specific if you don't find one I think?
21:05:47 <richlowe> I sure am glad someone made fmd's option stuff as baroque as possible
21:11:22 <LXGHTNXNG> xD
21:31:58 <richlowe> for the IRC log searchers: the answer is `(cd /; /usr/lib/fm/fmd/fmd -o fg=true -odebug=all -odbout=stderr)` you need to `cd`, fmd doesn't.