#asahi-gpu · 2021-05-12 11:53
segher touch the voltage) people would be killing a lot of PC hardware#asahi-gpu · 2021-05-12 11:53
segher13:25 <@marcan> if overclocking damaged things permanently (when you don't#asahi-gpu · 2021-05-12 11:52
segheryeah, but that does not matter#asahi-gpu · 2021-05-12 11:52
marcanI can certainly figure out a PLL in under a month#asahi-gpu · 2021-05-12 11:52
segherunless you can do bringup activities in under a month?#asahi-gpu · 2021-05-12 11:52
segheryes, so it matters a lot#asahi-gpu · 2021-05-12 11:52
marcannobody's going to accidentally run something at 6GHz without noticing for that long :-)#asahi-gpu · 2021-05-12 11:52
marcan100x doesn't matter; 100x is 10 years to 36 days#asahi-gpu · 2021-05-12 11:51
segherbut yeah you do not mind much while just spraying the register space#asahi-gpu · 2021-05-12 11:50
segherand 100x makes it unsuitable during development even#asahi-gpu · 2021-05-12 11:50
seghermost things *will* run at 6GHz instead of 3GHz even (with good cooling). and if it reduces lifetime by 10x most people will care a LOT#asahi-gpu · 2021-05-12 11:25
marcanif overclocking damaged things permanently (when you don't touch the voltage) people would be killing a lot of PC hardware#asahi-gpu · 2021-05-12 11:24
marcanI've done the "let's crank up the clock to beyond design specs" thing too; what happened was the power circuitry couldn't handle it and the thing shut down :-)#asahi-gpu · 2021-05-12 11:24
marcanthat cleared up within a week or so though#asahi-gpu · 2021-05-12 11:24
marcanthe closest I've come to damaging hardware from this kind of thing was some temporary DC bias damage done to an LCD panel, but that was because I got the power sequencing wrong (which caused it not to init properly and bias the panel), so again, power stuff#asahi-gpu · 2021-05-12 11:23
marcansegher: most things won't *run* at 4GHz instead of 3.5GHz if 3.5GHz is nominal, and unless it reduces lifetime by a million times, nobody cares; these experiments last seconds to minutes.#asahi-gpu · 2021-05-12 10:44
seghermarcan: most PLLs don't allow crazily out-of-range settings... but running something at 4GHz instead of 3.5GHz will increase power consumpotion by 25% or so, reducing lifetime, maybe by a lot#asahi-gpu · 2021-05-12 03:23
marcanthat's usually where they are, though on SoCs with internal voltage control stuff that applies to internal MMIO too#asahi-gpu · 2021-05-12 03:03
phireyou mean external VRMs on i2c buses?#asahi-gpu · 2021-05-12 03:00
marcanVRM, yes#asahi-gpu · 2021-05-12 03:00
marcanyou'll just crash things#asahi-gpu · 2021-05-12 03:00
marcanit's pretty hard to break anything by messing with PLLs#asahi-gpu · 2021-05-11 18:15
seghermarcan: PLL and VRM stuff is something you may want to stay away from, and yup, anything external#asahi-gpu · 2021-05-11 18:13
segherJTL: on a machine where you were not supposed to be able to do that?#asahi-gpu · 2021-05-11 06:28
JTL(not M1, but something else I know)#asahi-gpu · 2021-05-11 06:28
JTLhehe#asahi-gpu · 2021-05-11 06:28
JTLremembers when doing remote kernel patching and debugging was hot shit#asahi-gpu · 2021-05-11 06:27
azonenbergNice#asahi-gpu · 2021-05-11 06:26
marcanI now have the same thing for the hypervisor; exceptions break to the same shell, I can inspect processor context, and as of yesterday get backtraces (with resolved symbols etc), decode panics, etc#asahi-gpu · 2021-05-11 06:26
marcanazonenberg: in case you haven't see it yet: https://github.com/AsahiLinux/docs/wiki/Developer-Quickstart#playground-shell#asahi-gpu · 2021-05-11 06:21
monochromaknows of some open source github repos where they goofed and left some internal debug features in, and deleted them from git, so they don't show up in the history, but obviously still in the repo if you know where to look#asahi-gpu · 2021-05-11 06:21
marcanand then that's just 8 seconds#asahi-gpu · 2021-05-11 06:21
marcanonly if I do something unforeseen to break execution, or actually crash the hardware, do I have to reboot#asahi-gpu · 2021-05-11 06:21
marcanI spend a lot of time making my tools efficient; my register poking is remotely controlled from python scripts running on my PC, and most exceptions are caught and recovered from#asahi-gpu · 2021-05-11 06:20
marcan26 seconds, a bunch of that is compression/checksumming that I want to speed up a bit#asahi-gpu · 2021-05-11 06:19
marcanlet me time an actual hypervisor run to the current crash point...#asahi-gpu · 2021-05-11 06:18
marcanalso my reboot-to-interactive-debug-shell cycle is ~8 seconds right now#asahi-gpu · 2021-05-11 06:18
lainthat's a nice goof#asahi-gpu · 2021-05-11 06:18
marcan:-)#asahi-gpu · 2021-05-11 06:18
marcanwell, you quickly build up a blocklist of registers you shouldn't touch#asahi-gpu · 2021-05-11 06:17
azonenbergmore like just causing lockups and panics by scribbling all over mmio regions#asahi-gpu · 2021-05-11 06:17
azonenbergand i wasn't talking about breaking hardware#asahi-gpu · 2021-05-11 06:15
marcan(they moved them elsewhere after that one botched release)#asahi-gpu · 2021-05-11 06:15
marcanso disassembling *those* blobs is perfecly kosher, which is cute#asahi-gpu · 2021-05-11 06:15
marcanI also mentioned in my stream yesterday that there is a *specific* goof where Apple, when they tried to close up M1 bring-up code (for some reason, ARM lawyers almost certainly), the first release was libary blobs included in their open source kernel release, which by all available documentation makes them openly licensed with no anti-RE clause#asahi-gpu · 2021-05-11 06:13
marcanand at least in the case of the M1, I think power management that deep down is handled by other cores running firmware we don't care about#asahi-gpu · 2021-05-11 06:13
marcanas long as you stay away from, like, regulator I²C spaces, it's very hard to actually blow modern hardware up#asahi-gpu · 2021-05-11 06:12
marcanjust a few crashes#asahi-gpu · 2021-05-11 06:12
marcannever destroyed any hardware#asahi-gpu · 2021-05-11 06:12
marcanazonenberg: re poking randomly, I've done literal "try flipping every single register bit" on a PS4 GPU (radeon) and on some register spaces on wii u