00:29:24 thaks illumos im mad 00:29:38 and i like it 00:51:43 ? 02:55:58 Ellenor waht do you mind about a HELL. Do you know where it placed in this world? 02:56:49 That's absolutely incoherent, Guest25. 02:58:16 Name it or I will name this cruel words 02:58:44 I do not understand. 02:58:57 MISCONFIGURATION 02:59:04 mmm ENTROPY 02:59:21 Try saying it in Finnish or Swedish and translating it through https://deepl.com/en/translator 03:02:31 This make our world bleeding in pain. Many names Door to Hell have: linux, k8s. BUT true delight can be explained in as right configuration of reality! proper OS to use, proper file system, proper networking stack. This is what to think about 03:04:45 jclulow, rmustacc, richlowe, I think we have a troll 03:08:14 You ask me to ynderstanding. You got what you got. Not any weight of meanins can be light as a feather. Sorry if I disturbed your patience too much Ellenor 03:11:08 You make no sense. 03:11:20 You also make no sense through the translator. 06:33:55 usually best not to interact with bots 06:35:15 Didn't know kiwiirc made bots. 09:51:24 Ellenor wise mind make sense throught heart. Like illumos developers. No man speak C code. But heart do. Let's speak from my heart to your heart, sir: "thruth exists, let's find it" 09:52:41 Please talk to a psychiatrist. 09:53:58 Ellenor don't tell 'em I have phone there. 09:54:55 Ok? 09:55:07 Are you supposed to be in locked prison? 09:57:39 Modern psychatry is full of tricks. No man can solve them all to leave this institution. But I'm trying founding illumos church 09:59:10 faith is a key to soul insides 10:02:56 It makes your perception far more wide than inprinment place's walls, it's building. I can fly over horison finding love taking a walk across my heaven garden full of beatiful flowers 10:03:44 illumos does not want a church. thank you 10:05:32 your answer made a new dialectical pair to discuss. Sometime later. Thank you Ellenor 17:49:04 I can report that jbk's https://code.illumos.org/c/illumos-gate/+/4804 is working for me on a system with the necessary feature.. 17:49:05 → CODE REVIEW 4804: 13962 Support using CPUID instruction for TSC frequency when available (NEW) | https://www.illumos.org/issues/13962 18:22:31 apparently there's something similar for new(er?) AMD CPUs using some MSRs, but I don't have access to any systems for testing 18:26:57 What MSR? 18:27:45 Or is this documented somewhere in the APM jbk? 18:28:00 i believe so, let me see if I can find it again... 18:28:39 Okay. I can always go digging in the PPR, just there's a lot of that so if there's a hint somewhere, that'd be speed things up. 18:39:54 heh yeah... i remember seeing it, but didn't know if there was enough additional information that'd let us know if the MSR would be implemented (and with those semantics) 18:44:11 ahh.. (unverified) it looks like it might be from the P0 speed (at least on supported models) in 0xC001_0064 18:44:31 (though might require a bit of decoding) 19:03:05 jbk: Is someone else already consuming that or just looking at a specific part of the APM? 19:03:11 But yes, that's the PStateDef MSR. 20:10:29 i'm not sure -- i think there might be some linux patches floatign around, but it's unclear what the state of them are... 20:13:37 i did notice that both freebsd and linux will also use leaf 0x16 on Intel (despite Intel describing it as 'informational only'...even though they appear to be the ones that contributed the code to Linux)... but I figured that could be a different discussion since the leaf 0x15 bits are explicit about being usable (when present on Intel) 20:13:55 so i stopped w/ 0x15 20:14:18 (and could always add later if the consensus is we feel comfortable enough with them) 20:23:34 jbk: There's language in the ppr that describes this. The biggest challenge will be working through VMs in practice. 20:24:31 I guess we have to assume that someone that defines the P0 frequency in a pstate MSR and defines PState support then will use the TscRateMsr if we move systems to help us deal with the change. 20:25:24 We'd probalby have to go through each PPR to figure out where this applies since its not architectural it seems. 20:25:51 yeah -- since I had no systems to even test with, it was just something I noticed in passing 20:25:58 We do. 20:26:07 I'll talk to some folks and see if it makes sense for us. 20:26:32 we have 1 amd system in the lab now, but waiting for Toomas's loader fixes (he's made a lot of progress, so fingers crossed) 20:34:44 [illumos-gate] 13962 Support using CPUID instruction for TSC frequency when available -- Jason King 20:35:10 (and yes, that was started that far back :P) 20:37:13 Until I settle in to $NEW_JOB in a 2-3 weeks, if anyone is looking for `danmcd` you're going to have to settle for `kebe` (my owned-by-self "work laptop" IRC client). 20:57:24 jbk: looks like a linux patch to use the PStateDef MSR is in patch limbo. https://patchew.org/linux/20251105064231.220065-1-s.horvath⊙oca/ 21:05:44 that patch looks at the ~12 low order bits of that MSR; on my zen4c the contents of that MSR are consistent with what the patch does -- I get 2550 MHz which is consistent with what it calibrated at (2549997167 Hz) 21:46:45 on an unrelated note, finally made some progress in my (many years) saga of 'SATA hot plug not working' in my home server 21:47:13 we're not enabling the cold presence detect interrupt (despite there being a handler for it) in the ahci driver 21:48:10 with that, i can cfgadm -x sata_port_activate and see an inserted drive (this is with sata_auto_online already set to 1) 21:48:31 though i feel like maybe that step shouldn't be needed