03:39mwk: imirkin: it pops the control flow stack
03:40mwk: so yeah, it jumps to joinat, if that's what's on the stack
05:55jelly: hi, I have a GT720 (10de:1288) with passive heatsink running reasonably well with 3.19 kernel on debian testing. Compared to previous GT220, however, I can't find a way to look at the temperature reading with free drivers.
06:01jelly:looks at https://www.google.com/search?sitesearch=http://nouveau.freedesktop.org/wiki/&q=temperature but is not edified
06:01mupuf: jelly: run "sensors"
06:01mupuf: you may need to install lm_sensors
06:03jelly: mupuf: Ha! It works. sensors_detect used to find a separate chip on the old GT220, and did not find anything here so I never checked.
06:11mupuf: you're welcome
06:11mupuf: it must have been a long time ago!
06:18titan38: hi! what does nouveau require to have dri3 enabled. my xorg have both dri2 and dri3 as builtin
06:19titan38: when nouveau is loaded it tries to load dri2 module. why not dri3?
06:19titan38: [ 31.699] (II) [drm] nouveau interface version: 1.1.1
06:19titan38: [ 31.700] (II) Loading sub module "dri2"
06:19titan38: [ 31.700] (II) LoadModule: "dri2"
06:19titan38: [ 31.700] (II) Module "dri2" already built-in
07:11DL7LN_: hey everybody
07:12DL7LN: my hdmi-output is not working with X, i googled now for a few hours...now i need a little assistance to get on the right way...
07:13DL7LN: hdmi is the only connected...it delivers an output, as long as the kernel isn't loaded
07:13DL7LN: after that only the dvi works
07:14DL7LN: my xorg.conf says, that there's no monitor-section for HDMI-1
07:15DL7LN: but for DVI there isn't one neither
08:55Omar007: Say, does Nouveau currently support the GTX 970? The main site says it does but I'm receiving 2 errors and 2 warnings
08:56Omar007: Errors: failed to initialize sync subsystem -38 pointer to flat panel table invalid
08:56Omar007: Warnings: unknown connector type 70 Unknown-1 has no encoders, removing
08:58Omar007: FYI, I'm currently running Arch Linux with kernel version 3.19.0-1 as otherwise KMS would not work
08:58imirkin: Omar007: are you just getting these errors and it's annoying, or does it not work at all?
08:58Yoshimo: support is a bit far fetched, it is very very basic
08:59imirkin: Yoshimo: KMS is supposed to be fully functional
08:59imirkin: both the support and the GPUs are fairly new though, so there are probably things left presently unsupported
09:00Omar007: Since upgrading to 3.19 KMS is working fine for me (afaik anyway)
09:00Omar007: When I try to run gnome-session --session=gnome-wayland all I get is a brief flash of a gray background with a black mouse cursor before the screen enters stand-by mode
09:00Omar007: And these are the only oddities I can find on the system atm
09:04imirkin: DL7LN_: what gpu, what kernel? pastebin xorg log and dmesg
09:10DL7LN_: imirkin: i installed prop. nvidia-drivers now, now it's working...
09:27Omar007: Hmm weston also shows some finicky rendering issues; when starting weston, the screen shows flashes of colored artifacts (triangles, noise, w/e) for a sec or so, then the screen stays blank for a while (but does not enter stand-by) and after a while weston shows up
09:28Omar007: However, as soon as I move my mouse the cursor dissapears and it sometimes shows some more rendering artifacts now and then when I move my mouse
09:29Omar007: Haha as soon as I zoom in my cursor shows :D
09:29Omar007: It clearly shows rendering artifacts on the taskbar though
09:29Omar007: As I move
09:35imirkin: Omar007: weird, so the console renders fine? all the acceleration stuff should be disabled for you, so it's surprising that you'd see issues in X but not on the console
09:35Omar007: Pure console seems fine; the console inside weston has artifacts as soon as it moves
09:36imirkin: take a screenshot -- does the screenshot have artifacts? i assume not...
09:36Omar007: That just froze my whole system D:
09:37Omar007: wait it got unstuck :D
09:37Omar007: screenshot looks fine
09:38Omar007: But so do most things as long as nothing moves
09:38Omar007: except for EGL things like the smoke
09:38Omar007: those always show artifacts
09:39imirkin: well, i can only assume that skeggsb didn't see those, but he also had a testing sample size of 1...
09:39imirkin: [when doing the GM204 bringup]
09:39Omar007: Did he have a reference model?
09:39imirkin: i think so
09:40imirkin: pretty sure nvidia sent him something
09:41Omar007: In that case it probably is indeed
09:42Omar007: I have the G1 Gaming myself so I can imagine that it has differences when compared to the reference model
09:42Omar007: although I do believe it's currently the only 970 that actually has the same outputs as the reference model xD
09:58huehner: imirkin: related to my gm206 mail
09:58huehner: imirkin: is that patched kernel output + vbios already useful?
09:59huehner: as i had some trouble to get mmiotrace working
09:59huehner: also any easy way to test if the vbios file is extracted is fine before sending?
10:01imirkin: huehner: it's a little odd... based on the log, you should have a DVI and 3 DP ports
10:01imirkin: and a hypothetical "type 70" port, whatever that is
10:02imirkin: but then all your DP ports appear to be unhappy
10:02huehner: imirkin: the card does have 1*dvi+3*dp + 1*hdmi
10:02imirkin: could you give the DVI port a shot?
10:02huehner: + i have monitor connected to hdmi
10:02huehner: dvi-i with adapter to vga good enough?
10:02imirkin: hmmmmmm... i wonder if there aren't some shenanigans for that hdmi port
10:02huehner: as i have no monitor with native dvi
10:03Omar007: Heh in that case I wonder if then70 is also my HDMI ports :o
10:03huehner: Omar007: hehe ;)
10:03imirkin: Omar007: but your HDMI works fine right?
10:03Omar007: Idk I'm running on DP
10:04huehner: imirkin: for info: i do get hdmi image with efi simple but when it switched to nouveufb screen freezes (but can ssh in just fine)
10:04imirkin: it could be some revised hdmi port which was why they bumped the type
10:04Omar007: Want me to try? I may have an HDMI cable somewhere...
10:04imirkin: oooh, iirc they sent an updated dcb spec too
10:06huehner: imirkin: afaik all 9xx have hdmi 2.0 support first tiem ever for nvidia card
10:06imirkin: 0x70 = Virtual connector for Wifi Display (WFD)
10:06huehner: brb... testing dvi...
10:07imirkin: and i misread the connector table too...
10:07imirkin: 1 DVI, 1 HDMI, 3 DP
10:09imirkin: i also wonder whether "nouveau E[ DRM] failed to initialise sync subsystem, -38" is normal
10:09imirkin: Omar007 had a similar error message and his display seems (more) working
10:10Omar007: ^ until I turn on a second monitor on the same GPU
10:11Omar007: At that point nouveau starts spewing messages into the console and the second screen stays blank
10:11imirkin: Omar007: care to pastebin a log of those messages?
10:11imirkin: could also just be that hdmi is screwed on GM20x entirely. although i've gotta assume that skeggsb tested that... but who knows.
10:12imirkin: there was a thing where we were doing hdmi totally wrong on kepler, and that took like 1.5 years for us to find out, and very few people had actual issues.
10:15Omar007: imirkin: I believe it's the same one spammed infinitly. Give me a moment to boot the system again.
10:17huehner: imirkin: vga connected to that dvi does come up, even after modprobe nouveau i get a working console screen :)
10:18imirkin: huehner: ok cool. so sounds like just HDMI that's fubar'd
10:18huehner: imirkin: http://paste.debian.net/154118/ diff of dmesg
10:18imirkin: do you have a different hdmi monitor perchance?
10:19imirkin: erm... hopefully you get more than 1024x768 resolution though right?
10:19huehner: imirkin: didn't check much on that vga
10:20huehner: imirkin: other hdmi no now, could borrow one from work maybe during next week
10:20imirkin: nah don't worry about it
10:20imirkin: chances that it'd be a different outcome are quite low
10:20huehner: imirkin: but that could mean first gm206 patch for bring up -> split 124 case into 126 with GM206 label + otherwise same body?
10:20imirkin: but if it were easy, interesting to check -- occasionally screens can be very finicky about stuff
10:20huehner: imirkin: as that alone is alrady useful to get at least dvi up
10:21imirkin: huehner: yep, that sounds reasonable. send to skeggsb and cc nouveau@ + dri-devel@
10:22imirkin: and if you have a mmiotrace of the blob, send that email@example.com
10:22imirkin: (with HDMI in place)
10:22imirkin: and/or analyze it yourself
10:22imirkin: (that'd actually be better ;) if you can do it...)
10:22huehner: any pointers on how to do that?
10:23imirkin: once you have the trace, grab envytools
10:23huehner: i also got also a 1mb bios file from nvagetbios (entytools)
10:23imirkin: inside will be a tool called 'demmio'
10:23huehner: via prom method with a note (2nd bios present)
10:23imirkin: feed the trace to it, pipe output to less and.... enjoy :)
10:23huehner: ok ;)
10:24huehner: trace what?
10:24imirkin: even if you already have a trace, i'd recommend doing another one
10:24huehner: xinit 'sleep 10' as i that wiki page?
10:24imirkin: yeah, that's fine... as long as the outputs get turned on
10:24imirkin: but basically what i'd do instead is
10:24imirkin: xinit xterm or something
10:24imirkin: (i guess just xinit pops up an xterm)
10:24imirkin: then wait for things to settle a bit
10:24huehner: had some trouble not getting display lighting up last time need to re-try
10:25imirkin: then echo a mark (more on that later)
10:25imirkin: then unplug hdmi
10:25imirkin: then echo a mark
10:25imirkin: then plug hdmi
10:25imirkin: then echo a mark
10:25imirkin: and THEN you're good
10:25imirkin: well, make sure you wait a while... like at least a minute
10:25imirkin: maybe 3
10:26imirkin: to echo a mark, just echo "some useful message" > /sys/kernel/debug/tracing/trace_marker
10:26imirkin: which will show up in the mmiotrace log
10:26huehner: mark as something to find in the trace file i assume? to find those actions in there later?
10:26imirkin: that way it's easier to find your spot
10:26huehner: need to reorganize me a bit before that
10:27huehner: as the box with the gmx206 is my new main box
10:27huehner: so need to get some 2nd system up and running
10:27imirkin: yeah, it's miserable to do RE on your main box
10:27imirkin: i know, coz i do it...
10:27imirkin: except not anymore... mupuf has a remote RE box he lets me use :)
10:27huehner: where are the spare extra keyboards if you need them :(
10:28imirkin: in the trash can, where you threw them when you were like "man, extra keyboards are taking up too much space"
10:29huehner: that you get fro moving to another country so my pile of extra hardware if some miles away ;)
10:34Omar007:has 3 keyboards aside from his own
10:34Omar007: One of which still has a PS/2 connector :D
10:35imirkin: but no AT connector?
10:35huehner: Omar007: the one i am typing on is ps/2 also :D
10:35huehner: AT ones are at home
10:35huehner: that probably means we are getting old ...
10:36Omar007: You actually have something that you can plug that in to?
10:36Omar007: ok where the hell did my BIOS go?
10:38huehner: Q-Master: have yes, but not in use
10:40Omar007: That was probably directed towards me lol
10:41Omar007: huehner: ok good. If you didn't I really would've suggested to either send it to the museum or just dump it somewhere xD
10:43Omar007: Having said that, I still have a windows 3.11 laptop in here somewhere that people have suggested I should throw away already lol
10:43imirkin: for workgroups?
10:50Omar007: Not sure. I don't think so
10:50Omar007: Been to long since I've booted that thing lol
10:52Omar007: Sigh my experimenting seems to have killed all video output atm...
10:54huehner: hmm similar here
10:55huehner: plugging that vga once now all system bios/console output only vga, any idea how to get that back to hdmi?
10:55imirkin: bios controls that
10:58Omar007: Ok I'm gonna see if I have a DVI card
10:59Omar007: I mean cable
11:03huehner:waits 3min :)
11:09Omar007: Yay there is my video output!
11:10huehner: imirkin: looks like i need to add 126 chipset type to demmio first also
11:21huehner: imirkin: demmio can parse that trace file now
11:23Omar007: Ok I'm now unable to replicate my previous issue with 2 screens lol
11:23imirkin: huehner: cool, yeah. should be simple. feel free to send a pull req
11:23Omar007: it now clones and turning on/off works fine
11:23Omar007: Although I did also disconnect the 1 monitor that was in the other GPU..
11:24huehner: imirkin: yeah only 1.7million lines of text to understand now :D
11:24imirkin: huehner: well hopefully you used MARK's
11:24huehner: before unplug hdmi
11:24huehner: before replug
11:24imirkin: which should hopefully narrow it down a LOT
11:24huehner: after replug
11:24huehner: + 3min trailer
11:25imirkin: it came up fully before first unplug right?
11:25huehner: 1.4m lines until first mark (before unplug hdmi)
11:25huehner: came up with black screen + xterm without wm
11:26huehner: to next mark (before replug) so unplugged -> 36k lines only
11:26imirkin: yeah that's what xinit does by default
11:26imirkin: ok, well i'm more interested in the before plug -> after plug lines
11:26huehner: the replug sequence some 21k lines
11:26imirkin: cool, that's not so bad
11:27imirkin: you can probably ignore a ton of it
11:27imirkin: i'd start by focusing on the PDISP lines
11:28imirkin: i think your log made it seem like we don't detect that there's anything plugged in in the first place
11:29imirkin: (vs the other possibility of... we think everything's fine and monitor is happy, but it's not)
11:29imirkin: the former means that nouveau is either not getting the message about hdmi being plugged in
11:30imirkin: or when it goes to probe, it thinks nothing's there
11:30huehner: let me upload those files somewhere first
11:30imirkin: either way, a nouveau boot with nouveau.debug=debug drm.debug=0xe would be good to have on-hand (with hdmi plugged in)
11:30imirkin: or maybe even nouveau.debug=trace
11:34Omar007: Ok so it turns out weston uses the llvmpipe renderer
11:34imirkin: yeah that's expected
11:34imirkin: no accel on GM20x
11:35Omar007: On the plus side, gnome isn't crashing my system anymore
11:35Omar007: it's now killed by signal 11 xD
11:36Omar007: Its segfaultinh
11:37imirkin: afaik there's some heartache with weston or wayland or... something... with swrast on the latest mesas
11:37huehner: imirkin: http://www.huehner.org/temp/nouveau/
11:37huehner: imirkin: the replug.parsed.xz is decoded via demmio just the replug part cut out (by using the marks)
11:38imirkin: usually we don't distribute those
11:38imirkin: since we might want to change rnndb, which would then make demmio output diff things
11:38imirkin: huehner: do you have a commit to make demmio work with GM206 btw? if not, i'll make it happen
11:38huehner: i have a git diff here
11:39huehner: just not used how to convert that to nice git-format-patch whatever?
11:39imirkin: do you want to make a pull req? or mail the commit?
11:39huehner: mail git diff > my.patch acceptable?
11:39imirkin: git commit -a
11:39imirkin: enter in an appropriate commit description
11:39imirkin: git format-patch HEAD^..
11:40imirkin: git send-email --to 'firstname.lastname@example.org' *.patch
11:40imirkin: (or any diff way of mailing that email out... just please don't attach)
11:40imirkin: actually if you hastebin that patch file, that'd be good enough for me
11:40huehner: will do
11:41huehner: imirkin: http://paste.debian.net/154170/
11:41huehner: that 2nd nvhw not sure if needed for something but also had big switch of chipset types so added there as well
11:41imirkin: huehner: not the diff... the format-patch output... so i can git am it
11:41huehner: ok then let me prepare that properly first
11:41imirkin: huehner: so that i get your author info/etc
11:41imirkin: and don't have to come up with a commit log all by myself :)
11:41huehner: just not used to git yet
11:42imirkin: but familiar with source control presumably?
11:42huehner: brb.... reboot to debug for dmesg debug logs
11:42huehner: mercurial every day just not yet git
11:43imirkin: ah ok. so yeah, same basic deal
11:43imirkin: git is just like 100x faster
11:43imirkin: (and predates mercurial... but i digress)
11:51huehner: imirkin: http://www.huehner.org/temp/nouveau/startup-hdmi-nouveau-drm-debug & http://www.huehner.org/temp/nouveau/startup-hdmi-nouveau-drm-trace
11:51huehner: both from nouveau startup with only hdmi connected
11:53imirkin: seems like the logs are cut off
11:53imirkin: the trace one is, that is
11:54huehner: looks fine for me? both end in same mesage '[drm] Initialized nouveau 1.2.1 for me' ?
11:54huehner: [drm:drm_helper_probe_single_connector_modes_merge_bits] [CONNECTOR:27:HDMI-A-3] disconnected
11:54huehner: my gues would be that means is does not recognize the hdmi as connected
11:54imirkin: HDMI-A-3 -- why 3??
11:55imirkin: oh, coz you have an intel card
11:55huehner: also builtin intel but nothing connected to it
11:55imirkin: right, that's fine... just wanted to make sure it wasn't something more sinister
12:02huehner: imirkin: http://paste.debian.net/154189/
12:02huehner: do you need signed-off lines or similar also? or fine as it is?
12:03imirkin: no signoff is fine
12:04imirkin: you went a little overboard there... i've rephrased it a bit :)
12:05huehner: fine ;)
12:05huehner: for the linux kernel patch on top of what repo?
12:05imirkin: ben renamed everything
12:06imirkin: seems to happen about once a year... he gets bored or something :p
12:06huehner: that one i am watching a bit to see what's coming up in nouveau land
12:06imirkin: not a whole lot in there for 3.20 -- just the renames and a few smaller items
12:08huehner: new hw-bringup can still go into 3.20? as i think main drm pull request is already found?
12:08huehner: or for later?
12:09Eliasvan: imirkin: Hi, I've got a question, do you manage the mmiotraces sent to "mmio.dumps at gmail dot com"?
12:09imirkin: Eliasvan: manage is a strong word... we just keep them in the email account
12:09imirkin: we don't get enough of them to overflow the account limits
12:13Eliasvan: imirkin: Because I sent the following traces (chronological order) a while ago until now, (and I can't find them back in http://people.freedesktop.org/~imirkin/traces/ ) :
12:13Eliasvan: [NVE6-GTX765M]: reclocking trace
12:13imirkin: Eliasvan: yeah, those are just my traces... usually mmt traces, not mmiotraces
12:13Eliasvan: excuse me, typo, let's restart:
12:13Eliasvan: [NVD7-GT620M]: reclocking trace, take 2
12:13Eliasvan: [NVE4-GTX760]: reclocking trace
12:13Eliasvan: [NVE6-GTX765M]: reclocking trace
12:14Eliasvan: [NVC1-GT540M]: reclocking trace
12:14imirkin: although i wouldn't exclude the possibility that i have some mmiotraces in there
12:14Eliasvan: [NV108-GT740M]: reclocking trace
12:14imirkin: yeah, those are probably still in the email account
12:15Eliasvan: k, because I think if they would be public, more developers would be able to have access to them to do some rev eng
12:15imirkin: Eliasvan: mmmmmaybe. seems like a long shot. people who would be interested in doing RE usually come by here
12:16imirkin: if they express interest in that sort of thing, we'd happily give them access to the account
12:16imirkin: usually when someone expresses interest, i'll start them off with a simpler task
12:16imirkin: they usually don't come back
12:16Eliasvan: I remembered someone here having interest in doing so, but he didn't have access
12:17pmoreau: Tobias maybe?
12:17Eliasvan: yeah, I think so
12:18Eliasvan: "tobijk" was his name
12:18pmoreau: Should be him :)
12:18Eliasvan: probably :)
12:19Eliasvan: But I just need to be sure that the traces I'm making are in fact useful, I haven't got any feedback so far
12:21imirkin: Eliasvan: tbh, probably not very useful.
12:22imirkin: it is my opinion that to get reclocking working, it will take a developer with the hw to play with that hw a whole lot
12:22imirkin: a single trace from some hw that you can't play with will be of limited help
12:23imirkin: otoh, for something like brining hdmi up or whatever -- much more useful to have random traces
12:33imirkin: skeggsb: wtf is all this PDISPLAY.DAEMON stuff?? modesetting is handled by fuc in GM20x now?
12:39huehner: imirkin: i'll just try mapping that unknown connector 70 to hdmi in the code and see what happens
12:39huehner: maybe i get lucky
12:42imirkin: huehner: no
12:42Eliasvan: imirkin: ok, thanks, that was what I was afraid of ... Maybe I'll do something of RE myself on my HW, but the time that will happen will be far from now unfortunately. But let's hope it can still be used ;)
12:42imirkin: huehner: everything's fine on that front... we're just not detecting things correctly
12:42imirkin: huehner: i'm trying to find where in the code hpd's are serviced for SOR ports
12:45Eliasvan: imirkin: anyway, I also provided the vbioses of each card, so that on its own may be useful for other things
12:46imirkin: Eliasvan: cool, thanks!
12:46huehner: imirkin: ok then i'll better leave that to you, ping me if you need any more info or testing of patches
12:46imirkin: huehner: well... i'm about to give up
12:46imirkin: there are other things i need to do, and skeggsb is the one who's really familiar with this stuff
12:47Eliasvan: imirkin: no problem, I provided loads of other stuff too, full lspci -vvv, dmesg, frequency tables reported by the blob, and the like ;)
12:48imirkin: huehner: i'm insufficiently clear on how it's *supposed* to work, so working that out *and* working out what's wrong is a bit much
12:49imirkin: huehner: i'm pretty sure you need to focus on the SOR stuff - DAC is for VGA
12:49imirkin: and i see some SOR[0x2] communciations in your log, so perhaps that's it
12:50huehner: DAC and VGA i know
12:50huehner: what is SOR?
12:52imirkin: but that it writes to PDISPLAY+0x20/24 which is very odd
12:53imirkin: SOR = Something Output Register? who knows
12:53huehner: how can i test-run that code in the darktama repo?
12:53huehner: is there some better way that to copy it manually into a kernel tree?
12:53imirkin: it's set up to do out-of-tree module builds
12:53huehner: as had to redo patch there as the repo has all the renaming (compared to 3.19)
12:53imirkin: cd drm
12:56huehner: is nouveau.ko self-container? so just insmod of that? or need to copy bunch of .ko to /lib/modules after that?
12:56imirkin: not completely self-contained
12:56imirkin: it does have deps
12:56imirkin: modinfo nouveau
12:56imirkin: and manually modprobe all its deps
12:56imirkin: (or futz with depmod, but that's not worth it imo)
12:57huehner: thx... then running that .ko at least once before sending out patch ;)
12:58imirkin: seems like the classy thing to do
13:02huehner: still works
13:25huehner: imirkin: sent: http://lists.freedesktop.org/archives/nouveau/2015-February/020131.html
13:26imirkin: cool :)
13:26huehner: imirkin: in your initial gm206 mail you request vbios file also, where to send that to? in same mail to mmio.dumps as trace?
13:28huehner: how to check if that bios file is correct?
13:28huehner: gm206-asus-strix-gtx960-bios.bin: BIOS (ia32) ROM Ext. IBM comp. Video (122*512)
13:28imirkin: run it through nvbios
13:28huehner: from file looks good: size 1mb
13:29imirkin: actually that's way too big, but nvagetbios doesn't know any better
13:29huehner: nvbios seems to parse quite a bit but then segfaults :(
13:29huehner: let me check where
13:30huehner: probably some more reeing to do
13:30huehner: fails in parsing some power table or so
13:31huehner: 0x00000000004269e9 in envy_bios_print_power_cstep (bios=0xbd0010, out=0x7ffff7dd72a0 <_IO_2_1_stdout_>, mask=536870911) at /home/huehner/oss/x/envytools/nvbios/power.c:461
13:31huehner: 461 if (!cstep->ent2[i].valid)
13:31imirkin: don't worry about it
13:31imirkin: they changed the power table format, we need to redo it
13:33huehner: but i think we are a bit further away from reclocking maxwell anyway probably :D
13:34imirkin: well step 1 is figuring out the power table format
13:34imirkin: then there are 1000 more steps
13:36huehner: how to best sent the dmesg debug logs to ben? reply to my patch mail? or open an issue somewhere?
13:36imirkin: huehner: bugs.freedesktop.org Driver/nouveau (under xorg)
13:47huehner: imirkin: https://bugs.freedesktop.org/show_bug.cgi?id=89272
13:47imirkin: yep, i get email about those things :)
13:47imirkin: (well, the nouveau list does)
13:48huehner: so i'll get mail about my own bugs also, nice :)
13:49huehner: imirkin: thanks for your time and patience here :)
13:53imirkin: np. hopefully ben can make it work
17:24xormask: Hi guys, I'm at a hackathon with ~24 hours and a nvidia 750m, and I'm wondering if there is anything I could do to help nouveau out.
17:24xormask: I have contributed to the kernel before, but nothing in nouveau
19:48Tom^: fix pstates !
20:23imirkin: he's long gone... and it'd take more than 24h to learn about a lot of the relevant stuff... there are things you could do, but he left long ago.