09:35mwk: karolherbst: got a completely useless data leak
09:36mwk: turns out it's possible to learn the last value loaded from data memory by secure code
09:36mwk: unfortunately, it's most likely some useless return address from the stack :(
09:37karolherbst: yeah, this sounds useless
09:37mwk: completely useless.
09:37mwk: unless hmm
09:37mwk: unless it persists across reset
09:37mwk: but I doubt that
09:38karolherbst: the question is, even if it is not a return address and interesting data, would it help us?
09:39mwk: just throwing ideas out there.
09:39karolherbst: yeah, I know
13:16mwk: say, anyone looked at Falcon v6?
13:17karolherbst: mwk: are there pascals with v6?
13:18mwk: well, I just got a GP104
13:18mwk: and it has a lot of v6 Falcons
13:18karolherbst: I was under the impression, that those are v5
13:19mwk: I'm quite certain it says v6
13:19karolherbst: maybe it is v6, but v6 isn't that different?
13:19mwk: and I see a lot of funny new registers on these
13:19karolherbst: ohh okay
13:19mwk: *shrug* never seen it before
13:19karolherbst: the ISA looks basically the same though
13:19mwk: it definitely has a lot more MMIO regs
13:19mwk: we'll see, I'll hook it up to hwtest soon
13:20karolherbst: well I looked into the PMU on pascal a bit
13:20karolherbst: and since pascal they started to do all the reclocking bits on it except PCIe stuff
13:20karolherbst: even the volting regs are locked down now
14:30RSpliet: does that leave... no other overclocking methods than CoolBits?
14:48Tomin: hmm, nvidia drivers got somehow broken and I'm now using nouveau on my GTX 460. There is some static noise (or something, I'm not sure if that's correct word) in the picture, especially on dark gray background that my terminal has
14:48Tomin: it's not bad, but noticeable, I wonder if the low memory clocks have something to do with that
14:55RSpliet: Tomin: potentially... but I'd only expect that kind of noise on analogue displays
14:56RSpliet: (VGA, DVI-A)
14:56Tomin: this is DVI-D, and it's not really static noise like on analog TVs. (I just don't have the word for it)
14:57karolherbst: RSpliet: I guess so
14:58karolherbst: RSpliet: there are tools for windows though to do live overcocking. I have no idea how they are doing it though
15:03RSpliet: probably calling NVIDIA APIs
15:04RSpliet: there's not a lot else they _can_ do :-P
15:21pmoreau: I thought the pm runtime issue had gone away with 4.11, but apparently not. One more thing to look into.
18:38adsworth_: when submitting a bug is a journalctl --boot=-1 sufficient? The problem is that nouveau locks up and I can't get a kernel log with dmesg.
18:41RSpliet: adsworth_: that depends whether the log was saved (reboot with the magic sysrq?)
18:45adsworth_: RSpliet: hard shutdown pressing the power button for 5secs.
18:46adsworth_: there is a bunch of nouveau related messages at the end.
18:47pmoreau: adsworth_: If it is not enough, we will ask for more, don't worry ;-)
18:48pmoreau: What kind of issue are you having?
18:49adsworth_: system locks up directly after login to gnome. and it only happens sometimes.
18:49adsworth_: best type of bug.
18:49pmoreau: Yeah… :-/
18:49adsworth_: and it's been going on for at least the last 5 kernel versions or more.
18:49pmoreau: What GPU is that?
18:51adsworth_: one sec I'll check again. I always forget...
18:51adsworth_: if someone wants a peek at the last 150 lines of my journalctl. https://dpaste.de/chuZ/raw
18:53adsworth_: here is the lspci output for the card: 01:00.0 VGA compatible controller: NVIDIA Corporation GT218M [NVS 3100M]
18:55pmoreau: Is it reproducible by doing things in Skype? It seems to do something that makes Nouveau unhappy.
18:57adsworth_: not if it is always skype it always happens just after I login to GNOME.
18:58adsworth_: and I now run the new skype alpha, running the older version I had the same problem.
18:58pmoreau: That log should be sufficient as a first step. There may be a need for a new log with debug flags on, like "nouveau.debug=PGRAPH=trace", but I don't know which ones will be needed.
18:58RSpliet: that chip does have random hang issues with nouveau... I used to be all right just before my laptop died with X.org, but I think Wayland uncovers some new problems
18:59adsworth_: I'll just check my journal to see if I can find some more crashes. :)
18:59pmoreau: Is Skype automatically starting upon login?
18:59RSpliet: oh this looks like a proper bug, nost just a random hang
19:01pmoreau: adsworth_: If you want to continue using you card with Nouveau, you could try booting with `nouveau.noaccel=1` to disable hw acceleration.
19:02adsworth_: pmoreau: as it only happens on login I can live with it. I just reboot.
19:03adsworth_: RSpliet: I'm just checking older journals to see if it always is skype.
19:03adsworth_: and yes skype starts on login.
19:05RSpliet: adsworth_: try swapping to X.org when logging in and see if that makes a difference for you
19:06adsworth_: RSpliet: X.org instead of wayland?
19:06adsworth_: I don't think I'm using wayland.
19:06RSpliet: adsworth_: if that's Fedora 25, you might well be
19:07adsworth_: it's 24
19:07RSpliet: in that case you might not be :-D
19:07RSpliet: but that could imply older versions of software - I never keep up with old versions of Fedora
19:09adsworth_: the kernel is 4.10.14
19:10adsworth_: I have a second journal log where the culprit is thunderbird and not skype.
19:12adsworth_: so the big question is do I submit to the nouveau bug tracker or fedora bugtracker?
19:13adsworth_: RSpliet: does this https://dpaste.de/zhMF/raw mean I'm running under wayland?
19:15pmoreau: You seem to be indeed running Wayland.
19:15pmoreau: You should be able to switch to X on the login screen.
19:15adsworth_: OK, then I'll try to switch to X.org and see if that helps.
19:16pmoreau: And the bug report should be for the Nouveau bug tracker, as it is very likely to be a Nouveau issue.
19:17adsworth_: OK, will do that later. I'll attach both logs I have as it shows that it's not a skype related problem.
19:19pmoreau: I don't think you need to post both: I was just wondering, if Skype, or any application, was the culprit, would it be possible to find a reliable way to trigger the bug.
19:20adsworth_: I'm assuming it's a problem in the 2D driver i.e product “xorg”, component “Driver/nouveau”
19:22pmoreau: More like in the kernel module, but it is tracked by the same product as the 2D driver
19:23pmoreau: Wayland does not have an equivalent to the Nouveau DDX IIRC
19:25adsworth_: pmoreau: how can I find out which version I'm using?
19:26adsworth_: I'll be back in 30minutes or so, bedtime for the kids.
19:27pmoreau: adsworth_afk: Simply specify which kernel version you are using in the comment section, you can leave empty the version field.