#archlinux32 | Logs for 2018-10-06

Back
[00:16:53] -!- noctambulo has quit [Remote host closed the connection]
[00:17:22] -!- noctambulo has joined #archlinux32
[00:18:20] -!- noctambulo has quit [Remote host closed the connection]
[00:19:58] -!- noctambulo has joined #archlinux32
[00:20:39] -!- noctambulo has quit [Remote host closed the connection]
[00:21:26] -!- noctambulo has joined #archlinux32
[01:26:39] <thePiGrepper> Hi, is there any admin around?
[01:26:57] -!- thePiGrepper has quit [Quit: leaving]
[01:27:24] -!- thePiGrepper has joined #archlinux32
[02:19:12] -!- woshty has quit [Ping timeout: 252 seconds]
[02:31:09] -!- chromicant has quit [Read error: Connection reset by peer]
[02:32:06] -!- chromicant has joined #archlinux32
[03:17:38] <NoobAlice> thePiGrepper: I'm no admin, but if you ask your question, one of the admins may see it when they look at IRC again. :)
[03:34:15] noctambulo is now known as Chat-50038
[03:34:50] Chat-50038 is now known as Chat-50039
[03:46:02] -!- Chat-50039 has parted #archlinux32
[03:53:07] -!- noctambulo has joined #archlinux32
[04:00:20] -!- noctambulo has quit [Quit: WeeChat 2.2]
[04:01:49] -!- noctambulo has joined #archlinux32
[04:02:09] -!- noctambulo has quit [Client Quit]
[04:03:37] -!- noctambulo has joined #archlinux32
[04:05:00] -!- noctambulo has quit [Client Quit]
[04:06:47] -!- noctambulo has joined #archlinux32
[04:12:33] -!- noctambulo has quit [Quit: WeeChat 2.2]
[04:12:55] -!- noctambulo has joined #archlinux32
[04:17:45] -!- _noctambulo has joined #archlinux32
[04:18:05] -!- _noctambulo has quit [Remote host closed the connection]
[04:18:32] -!- _noctambulo has joined #archlinux32
[04:28:12] -!- noctambulo has quit [Quit: WeeChat 2.2]
[04:37:50] -!- bill-auger has quit [Quit: No Ping reply in 180 seconds.]
[04:39:14] -!- bill-auger has joined #archlinux32
[04:41:12] -!- _noctambulo has quit [Remote host closed the connection]
[04:43:46] -!- bill-auger has quit [Read error: Connection reset by peer]
[04:46:34] -!- bill-auger has joined #archlinux32
[04:48:46] -!- noctambulo has joined #archlinux32
[05:33:47] -!- chromicant has quit [Read error: Connection reset by peer]
[05:34:21] -!- chromicant has joined #archlinux32
[05:35:34] -!- chromicant has quit [Client Quit]
[05:38:32] -!- chromicant has joined #archlinux32
[05:54:08] -!- NoobAlice has quit [Quit: Leaving.]
[06:08:42] -!- chromicant has quit [Quit: ZNC 1.7.1 - https://znc.in]
[06:09:59] -!- noctambulo has quit [Remote host closed the connection]
[06:11:14] -!- chromicant has joined #archlinux32
[06:16:18] -!- noctambulo has joined #archlinux32
[06:28:44] -!- noctambulo has quit [Quit: WeeChat 2.2]
[06:29:16] -!- noctambulo has joined #archlinux32
[06:30:21] -!- noctambulo has quit [Client Quit]
[06:30:40] -!- noctambulo has joined #archlinux32
[06:34:52] -!- noctambulo has quit [Client Quit]
[06:35:25] -!- noctambulo has joined #archlinux32
[06:35:25] -!- noctambulo has quit [Client Quit]
[06:36:07] -!- noctambulo has joined #archlinux32
[06:36:35] -!- noctambulo has quit [Client Quit]
[06:37:05] -!- noctambulo has joined #archlinux32
[06:39:19] -!- noctambulo has quit [Client Quit]
[06:39:41] -!- noctambulo has joined #archlinux32
[06:40:02] -!- noctambulo has quit [Client Quit]
[06:40:28] -!- noctambulo has joined #archlinux32
[06:45:59] -!- noctambulo has quit [Quit: WeeChat 2.2]
[06:47:26] -!- noctambulo has joined #archlinux32
[07:39:47] noctambulo is now known as Chat-50001
[07:39:49] Chat-50001 is now known as Chat-50002
[07:39:51] Chat-50002 is now known as Chat-50003
[07:39:53] Chat-50003 is now known as Chat-50004
[07:39:55] Chat-50004 is now known as Chat-50005
[07:41:04] Chat-50005 is now known as Chat-50003
[07:41:23] Chat-50003 is now known as Chat-50004
[07:41:43] Chat-50004 is now known as Chat-50005
[07:42:03] Chat-50005 is now known as Chat-50006
[07:42:23] Chat-50006 is now known as Chat-50007
[07:43:23] Chat-50007 is now known as Chat-50010
[07:43:26] -!- Chat-50010 has parted #archlinux32
[07:56:19] -!- deep42thought has joined #archlinux32
[07:56:20] <buildmaster> Hi deep42thought!
[07:56:20] <buildmaster> !rq deep42thought
[07:56:21] <phrik> buildmaster: <deep42thought> brtln: you're "the arch linux", we're "the other arch linux" and alarm is "arches linux", then?
[07:57:37] <deep42thought> thePiGrepper: I'd consider tyzoid (US East Coast), abaumann and me (CET) an admin - however, how much the individual can help depends on what kind your problem is of
[08:09:01] -!- thePiGrepper has quit [Quit: leaving]
[08:09:19] -!- thePiGrepper has joined #archlinux32
[08:10:22] <thePiGrepper> hi deep42thought
[08:25:14] <bill-auger> thePiGrepper: if you have a specific question, do just ask it - you dont need to wait - it will be seen eventually
[08:39:50] -!- titus_livius has joined #archlinux32
[09:02:42] -!- noctambulo has joined #archlinux32
[09:06:23] noctambulo is now known as Chat-50001
[09:06:43] Chat-50001 is now known as Chat-50002
[09:09:43] Chat-50002 is now known as Chat-50001
[09:09:45] Chat-50001 is now known as Chat-50002
[09:10:00] Chat-50002 is now known as Chat-50001
[09:10:02] Chat-50001 is now known as Chat-50002
[09:10:03] Chat-50002 is now known as Chat-50003
[09:12:12] -!- abaumann has joined #archlinux32
[09:12:13] <buildmaster> Hi abaumann!
[09:12:13] <buildmaster> !rq abaumann
[09:12:14] <phrik> buildmaster: <abaumann> very soon only a machine learning algorithm will be able to devise a build plan for a Linux distribution..
[09:12:27] <deep42thought> Good morning, abaumann!
[09:12:33] <abaumann> hi deep42thought. :-)
[09:12:47] <abaumann> My new Pentium-II based archlinux32 mirror is up and running.
[09:12:51] <deep42thought> there is some circle in i486 around mesa, ghostscript and stuff
[09:12:57] <deep42thought> great!
[09:13:01] <abaumann> yeah. the mesa one I know.
[09:13:37] <abaumann> "/dev/sda6 has gone 3978 days without checking"
[09:13:45] <deep42thought> loool
[09:14:09] <abaumann> and the adaptec SCSI adapter says something about: I should make a backup and press any key, the hard disk is over the estimated life time.
[09:14:32] <abaumann> I always suspected, that this is just a stupid counter. The disk is in mint condition, has been lying around for at least 8 years
[09:15:07] <abaumann> SMART "Failure Prediction Threshold"
[09:15:15] <deep42thought> :-D
[09:15:16] <abaumann> I have to find a way to fiddle that one.
[09:16:14] <abaumann> I tried to build subversion yesterday (in order to resolve a cycle), and I failed in the Java-bindings which are so stupidly made you cannot switch them of.
[09:16:17] <abaumann> *off
[09:18:05] -!- Chat-50003 has quit [Quit: WeeChat 2.2]
[09:37:29] <rcf> ...another reason to be glad I moved from cvs straight to git....
[09:41:05] <deep42thought> bye
[09:41:07] -!- deep42thought has quit [Quit: Leaving.]
[09:41:32] <abaumann> cu :-)
[09:41:43] -!- abaumann has quit [Quit: leaving]
[09:47:47] * buildmaster goes insane.
[10:17:59] -!- woshty has joined #archlinux32
[11:29:00] -!- woshty has quit [Ping timeout: 252 seconds]
[11:49:53] -!- thePiGrepper has quit [Quit: leaving]
[11:57:56] -!- AndrevS has joined #archlinux32
[12:01:14] -!- thePiGrepper has joined #archlinux32
[12:02:12] -!- thePiGrepper has quit [Client Quit]
[12:32:11] -!- AndrevS has quit [Quit: umount /dev/irc]
[14:48:41] -!- bill-auger has quit [Remote host closed the connection]
[14:50:40] -!- bill-auger has joined #archlinux32
[14:59:38] -!- bill-auger has quit [Ping timeout: 245 seconds]
[16:13:10] -!- noctambulo has joined #archlinux32
[16:21:24] noctambulo is now known as Chat-50001
[16:21:44] Chat-50001 is now known as Chat-50002
[16:22:04] Chat-50002 is now known as Chat-50003
[16:22:14] -!- Chat-50003 has parted #archlinux32
[16:26:53] -!- noctambulo has joined #archlinux32
[16:27:31] -!- charims has joined #archlinux32
[16:28:45] -!- charims_ has quit [Ping timeout: 252 seconds]
[16:30:04] noctambulo is now known as Chat-50001
[16:40:04] Chat-50001 is now known as Chat-50002
[16:41:04] Chat-50002 is now known as Chat-50003
[16:42:04] Chat-50003 is now known as Chat-50004
[16:43:04] Chat-50004 is now known as Chat-50005
[16:44:04] Chat-50005 is now known as Chat-50006
[16:45:04] Chat-50006 is now known as Chat-50001
[16:47:04] Chat-50001 is now known as Chat-50003
[16:48:04] Chat-50003 is now known as Chat-50004
[16:49:04] Chat-50004 is now known as Chat-50001
[16:51:04] Chat-50001 is now known as Chat-50002
[16:52:04] Chat-50002 is now known as Chat-50003
[16:53:04] Chat-50003 is now known as Chat-50004
[16:54:04] Chat-50004 is now known as Chat-50001
[16:55:04] Chat-50001 is now known as Chat-50002
[16:56:04] Chat-50002 is now known as Chat-50003
[16:56:50] -!- thePiGrepper has joined #archlinux32
[16:57:04] Chat-50003 is now known as Chat-50004
[16:57:49] -!- Chat-50004 has parted #archlinux32
[17:37:24] -!- NoobAlice has joined #archlinux32
[17:43:45] -!- noctambulo has joined #archlinux32
[17:51:40] -!- deep42thought has joined #archlinux32
[17:51:41] <buildmaster> Hi deep42thought!
[17:51:41] <buildmaster> !rq deep42thought
[17:51:42] <phrik> buildmaster: <deep42thought> arch32 is on the bleeding edge of arch, which is on the bleeding edge of software?
[17:54:38] -!- thePiGrepper has quit [Ping timeout: 272 seconds]
[18:00:50] * buildmaster resumes sanity.
[18:26:27] -!- abaumann has joined #archlinux32
[18:26:28] <buildmaster> Hi abaumann!
[18:26:28] <buildmaster> !rq abaumann
[18:26:28] <phrik> buildmaster: <abaumann> the better the lint tools of a language, the crappier the language ;-)
[18:29:53] -!- ofara has quit [Quit: ofara]
[18:31:17] -!- ofara has joined #archlinux32
[18:38:43] -!- noctambulo has quit [Quit: WeeChat 2.2]
[18:38:51] <deep42thought> Hi abaumann!
[18:43:44] -!- thePiGrepper has joined #archlinux32
[18:47:49] thePiGrepper is now known as toto271
[18:53:58] toto271 is now known as thePiGrepper
[19:29:09] <abaumann> hi deep42thought
[19:29:13] <abaumann> sorry.. no see.
[19:29:17] <deep42thought> np
[19:31:48] <deep42thought> ah, a new toolchain is blocking other stuff :-)
[19:34:09] <abaumann> ou. :-)
[19:34:22] <abaumann> that's why the builds are stuck..
[19:36:37] <deep42thought> I have successfully compiled a version of ghostscript (with reduced dependencies), but I cannot build it with staging-i486-build and thus I cannot build it "the regular way" on a slave
[19:37:10] <deep42thought> I'd like to put this on your bootstrap staging, abaumann
[19:37:20] <abaumann> yep. in the meantime I'm ready
[19:37:41] <abaumann> there are two mirrors at the moment, so I need your ssh key.
[19:38:16] <abaumann> ssh -P 2223 httpupload@andreasbaumann.cc
[19:38:22] <abaumann> ssh -P 2227 httpupload@andreasbaumann.cc
[19:38:33] <abaumann> aeh. small -p :-)
[19:39:24] <deep42thought> gimme a moment, I need to put the children to bed, first ...
[19:39:39] <abaumann> .. sure :-)
[20:04:45] -!- djmoch has quit [Quit: ZNC 1.7.1 - https://znc.in]
[20:05:36] -!- djmoch has joined #archlinux32
[20:05:52] -!- rcf has quit [Ping timeout: 268 seconds]
[20:09:48] <deep42thought> thx, abaumann
[20:11:12] <abaumann> np.
[20:11:26] <abaumann> let's see how long this PII holds. :-)
[20:11:37] <abaumann> not what your administrator is recommending..
[20:18:06] <thePiGrepper> Hi, I have one questions regarding the PKGBUILDs used by arch32, are they just downstreamed from Arch and modified [if needed], or they can be customized to a greater degree? and if so, who/where that discussion takes place?
[20:19:44] <deep42thought> thePiGrepper: usually yes; yes; anywhere, but mostly here
[20:20:29] <deep42thought> anywhere = here, mailinglist, bug tracker, forums
[20:24:39] <thePiGrepper> I see. are there some kind of guidelines to follow before suggesting modifications to these PKGBUILDs, mainly to avoid wasting the time of anybody for something that's not acceptable. for example, If I would want to change a vim build option, that'd be a no right? without a doubt. but for example, if it were something like a compilation flag that mostly wont be used by the hardware targeted by
[20:24:45] <thePiGrepper> the distro, that might be discussed? something like that?
[20:26:35] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-10-06T20:17:01.L3uZys.stdin"?.
[20:27:45] <abaumann> Hi thePiGrepper. We usually try to stick closely to upstream in terms of what is in the packages and in terms of features. Ideally we would have close to zero patches on our own.
[20:28:20] -!- jarshvor has joined #archlinux32
[20:28:43] <deep42thought> I'd say maintainability it the top priority
[20:28:50] <deep42thought> which mostly implies what abaumann said
[20:28:56] <abaumann> :-)
[20:29:03] <abaumann> just shorter. :-)
[20:29:37] <deep42thought> but we (at least "I") are open to suggestions in like options to ./configure
[20:29:54] <abaumann> sure.
[20:30:27] <deep42thought> but before we continue our academic discussion: thePiGrepper what did you have in mind that should/could be changed?
[20:31:00] <deep42thought> ah, sry, you wrote it already
[20:31:11] * deep42thought is failing at doing multiple things in parallel
[20:31:20] <abaumann> np
[20:34:53] -!- rcf has joined #archlinux32
[20:39:39] <thePiGrepper> something specific doesn't come to mind right now, I'd need time to compile some examples, but there are those packages that are compiled assuming 'modern' hardware, for a general purpose distro like Arch. I just thought that, a nice idea would be to get something more than just maintainability out of the focus on *only* 32bit processors.
[20:40:16] <jarshvor> hi all, not sure if this is a good place to ask for some support or if its mainly for dev talk. having trouble with xorg (driver errors I believe).
[20:40:41] <jarshvor> and thanks for keeping arch32 alive btw =)
[20:41:04] <deep42thought> thePiGrepper: changing flags, that almost noone tangents is fine, I think. if it's a useful change with implications for many, a split package might be the way to go - but it sounds like something we would probably accept
[20:41:12] <deep42thought> jarshvor: it's the right place
[20:41:23] <deep42thought> abaumann: it's your turn ;-)
[20:41:29] <jarshvor> deep42thought: thanks ;)
[20:41:51] <jarshvor> xorg is spitting out this http://sprunge.us
[20:42:25] <jarshvor> im running on an old radeon, and its complaining with a bunch of errors at the end there.
[20:42:36] <Ingar> interesting
[20:42:39] <Ingar> the signal 4
[20:43:01] <Ingar> jarshvor: I have this in my virtual machines
[20:43:21] <jarshvor> im normally running headless but was just gonna set up a scanner so my dad could use it.
[20:43:49] <jarshvor> Ingar: virtual machines? I dont follow :/
[20:44:09] <jarshvor> what does the signal 4 signify?
[20:44:25] <Ingar> I have two archlinux32 virtual machines, where X worked fine. since a few weeks I get this error
[20:44:36] <Ingar> I suspect a mesa/kernel update
[20:44:36] <jarshvor> hmmm
[20:44:58] <jarshvor> alright.. so its not my hardware or something I fucked up and forgot about. that good then
[20:45:02] <jarshvor> ...somewhat.. =)
[20:45:12] <Ingar> Signal 4 is illegal instruction btw
[20:45:20] <Ingar> so I was thinking it was virtual machine related
[20:45:32] <jarshvor> well im running on real hardware... so..
[20:45:32] <Ingar> jarshvor: you have it on a real system ?
[20:45:35] <Ingar> k
[20:45:35] <jarshvor> yep
[20:45:45] <Ingar> what CPU, if I might ask
[20:45:50] <jarshvor> sure.. let me check
[20:45:58] <jarshvor> pentium 4 if I recall
[20:46:23] <jarshvor> http://sprunge.us
[20:47:21] <Ingar> I do have arch32 on real hardware, can't rmemeber if I actually updated that as well
[20:47:30] <Ingar> I think I was scared to break it
[20:48:16] <jarshvor> :S im must admit im updating a bit carelessly... just whenever I boot up that machine which is ocassionally. I guess I shoudnt be so trigger happy.
[20:48:50] <deep42thought> I update every morning - haven't broken too much so far ...
[20:49:09] <deep42thought> but it's a headless archlinux32 (the other archs are not all headless)
[20:49:40] <jarshvor> ohh thats much more frequently than me.. Its an old machine which I use only for emergencies? :S and its connected to a multi-printer which doesnt print but scans.. =) when it want to..
[20:49:58] <abaumann> P4, so it cannot be a SSE, SSE2 problem.
[20:50:03] <jarshvor> I might boot that machine maybe once every couple of months if at that
[20:50:12] <abaumann> but I'm sure the driver guys managed to smuggle in AVX or worse now :-)
[20:50:27] <deep42thought> for me it's the other way round: I reboot the machine whenever a new kernel or systemd ships in
[20:51:10] <jarshvor> deep42thought: heheh nice to know those little i686 are up and running and sweating the race ;)
[20:51:48] <abaumann> deep42thought: there is a sync mirror script changing permissions, that's why you had this repo-add lock error before.
[20:52:19] <deep42thought> umm, I'll use my usual approach: retry until it succeeds
[20:52:21] <jarshvor> (i do use arch32 much more routinely (on the tube/subway or on the go) on a small eePC laptop though.. =) to thanks for this)
[20:53:06] <Ingar> mm it my realbox has the same versions as my virtual machine
[20:55:12] <Ingar> oooh
[20:55:21] <jarshvor> Ingar: so you suspect this is a kernel or mesa issue? not ati drivers?
[20:55:37] <Ingar> because I have not ati
[20:55:40] * jarshvor quietly waits... listening to Ingar. =)
[20:55:41] <Ingar> but I jsut tested
[20:55:48] <Ingar> I changed the CPU configuration in my virtual machine and it works
[20:55:59] <Ingar> so it probably _is_ a signal 4
[20:56:01] <jarshvor> cpu configuration meaning..?
[20:56:04] * buildmaster resumes sanity.
[20:56:28] <Ingar> jarshvor: I can choose with CPU the virtual machine announces
[20:56:38] <abaumann> it's a mesa issue in the r300 driver..
[20:56:55] <abaumann> or radeon_drv.so
[20:57:06] <deep42thought> opcodesniffer!
[20:57:09] <abaumann> without a disassembly from gdb at the very place of the SIGILL hard to tell.
[20:57:09] <jarshvor> arent those ati drivers?
[20:57:18] <abaumann> yes. ./opcode in builder.bin
[20:57:46] <abaumann> /usr/lib/xorg/modules/drivers/radeon_drv.so looks like the free driver
[20:58:03] <jarshvor> but for the ati hardware yes?
[20:58:16] <deep42thought> something is wrong with the buildmaster: glibc has been rebuilt (successfully) the tenth time!
[20:58:25] <abaumann> congrats :-)
[20:58:39] <abaumann> like 'xxx failed with: 0 (Success)'
[20:58:42] <deep42thought> glibc-2.28-4.8 (ninth - sry, I can't count)
[20:59:02] -!- woshty has joined #archlinux32
[21:00:16] <abaumann> that's the problem: I don't have the right hardware around to test more modern ATI or Nvidia graphic cards. I mostly have Intel and MACH64, Elsa Gloria, Matrox (where I know X is no longer working)
[21:01:03] <jarshvor> mesa is for hardware acceleration trough opengl right?
[21:01:10] <abaumann> If you have a coredump, then you can use 'disassemble' at the location of SIGILL, this is the best way to debug the issue.
[21:01:27] <abaumann> mesa is a lot of things nowadays. also that, yes.
[21:01:34] <abaumann> but opengl is on it's way out, I understood..
[21:02:04] <jarshvor> if I were to uninstall the xf86-video-ati and leave a lower level driver like fb or vesa would that circumvent this issue?
[21:02:42] <jarshvor> does the problem lie in that its loading the ati driver along with mesa and mesa has some issue?
[21:03:07] <abaumann> I rarely had any success with that. on MACH64 I got black screens, on the AMDGNU Alix thingy I got green blocks and libvirt stopped to work for me alltogether
[21:03:42] <abaumann> I suspect, it's a small thing with the compilation flags for mesa..
[21:03:59] <jarshvor> hmmmm alrigh abaumann . Does my logic make sense though? Am I correct in suggesting it? even if it mgight not work? wanna make sure im understanding things correctly
[21:04:40] <abaumann> jarshvor: makes perfect sense, I would have tried the same things..
[21:05:12] <jarshvor> I probably have a bunch of old cards lying around. maybe some nvidia. I could try those. but I suspect if the problem is in mesa the same error might pop up.
[21:05:35] <jarshvor> unless I use the proprietary driver? :/
[21:05:43] <abaumann> well, it popped up in /usr/lib/xorg/modules/drivers/radeon_drv.so and /usr/lib/dri/r300_dri.so, this is ATI/Radeon
[21:05:55] <abaumann> Catalyst, good luck :-)
[21:06:00] <abaumann> never tried.
[21:06:07] <jarshvor> heheh no... I meant the nvdia
[21:06:12] <abaumann> ah. the same.
[21:06:31] <jarshvor> but yeah... I might just try swapping with anther card and trying nouveau
[21:06:36] <abaumann> The world is dominated by Intel/Nvidia/ATI and they all make quite some terrible binary blobs they call drivers..
[21:06:59] <abaumann> ..and X is basically supporting those 3 only. Or at least testing those 3 only.
[21:07:05] <jarshvor> abaumann: Im all for the open source drivers ;)
[21:07:14] <jarshvor> blobs and their makers should be burnt at the stake ;)
[21:07:25] <abaumann> It's understandable why they do it.
[21:07:37] <abaumann> But I prefer standards like VGA, VESA, etc.
[21:07:47] <abaumann> just nothing happened anymore in the last 10 years..
[21:09:00] <abaumann> deep42thought: I got a i486 glibc which build correctly it seems.
[21:09:23] <deep42thought> the problem is not, that it builds correctly
[21:09:35] <abaumann> ah. that it builds again. :-)
[21:09:38] <deep42thought> the problem is, that it gets rescheduled _although_ it built correctly
[21:09:42] <deep42thought> exactly
[21:09:43] <abaumann> we'll.. reproducability of build at work ;-)
[21:09:50] <deep42thought> lol
[21:10:05] * deep42thought disables the auto-rescheduler for now
[21:10:10] <abaumann> side node on ATI: the MAC works now, I had to hack the driver to load a VESA bios for KMS modesetting manually.
[21:10:17] <abaumann> And that thing is 64-bit.
[21:10:40] <abaumann> Now it goes black every 10 minutes and can not be woken up again..
[21:14:15] <Ingar> sry, I was away for a bit
[21:14:29] <abaumann> https://bbs.archlinux.org
[21:14:30] <phrik> Title:Load custom Radeon firmware for Macbook Pro / Kernel & Hardware / Arch Linux Forums (at bbs.archlinux.org)
[21:16:49] <jarshvor> fyi uninstalling xf86-video-ati and installing xf86-video-fbdev still gives me a Signal 4 error
[21:17:03] <jarshvor> http://sprunge.us
[21:17:55] <abaumann> Those "X drivers" you call -ati, -vesa, -fb are basically only for old X stuff. Mesa is always loaded for acceleration.
[21:18:15] <abaumann> Do you have a gdb ready and a coredump?
[21:18:40] <abaumann> Would really help me a lot to see on which instruction the thing dies this time..
[21:20:26] <jarshvor> oh.. I thought vesa (per the wiki) did not do any acceleration. hmmm so mesa is always loaded huh?
[21:20:54] <Ingar> http://sprunge.us
[21:20:59] <abaumann> mesa has software renders in libraries, but also those can have AVX or similar cod ein them.
[21:21:00] <Ingar> if I change CPU from SkyLake to CureDuo
[21:21:08] <Ingar> *CoreDuo
[21:21:27] <abaumann> sounds like AVX to me..
[21:22:26] <jarshvor> I dont have a gdb setup.. and Im not sure ive ever done any debugging like that.. so...
[21:22:32] <jarshvor> should I follow.. this? https://wiki.archlinux.org
[21:22:33] <phrik> Title:Step-by-step debugging guide - ArchWiki (at wiki.archlinux.org)
[21:23:01] <abaumann> mmh. this is more for developers.
[21:23:08] <abaumann> Do you have a shell?
[21:23:19] <abaumann> What is coredumpctl telling you?
[21:24:19] <abaumann> Those were my lastest fixes to get something out from mesa when running under libvirt: https://git.archlinux32.org
[21:24:20] <phrik> Title:archlinux32/packages: Package customizations and pure-i686 packages - Archlinux32 Gitea (at git.archlinux32.org)
[21:24:31] <abaumann> most likely this needs more thorough fixing..
[21:25:29] <abaumann> ah. SGML tools compile without any problem. Nobody is touching them and trying to squeeze some optimization code into them :-)
[21:27:20] <abaumann> jarshvor: never mind. If it happens regardless of the graphic card used or the drivers installed, I can reproduce it on my virtual machines (as soon as bridging decides to work again on 64-bit arch with new firewalling software)
[21:27:37] <abaumann> ah. and the 100% CPU usage problem on libvirt, I forgot.
[21:27:59] <abaumann> sorry, but my test enviroment is currently 100% borked..
[21:28:08] <jarshvor> abaumann: coredumpctl seems to have quite a few coredumps.. last ones about Xorg obviously
[21:28:21] <jarshvor> should I send them over┬┐┬┐
[21:28:21] <abaumann> cool :-)
[21:28:31] <abaumann> aeh.. no.
[21:28:43] <abaumann> coredumpctl debug
[21:29:13] <jarshvor> hehe alrigh =).. and out of curiousity... im guessing these coredumps take space on disk.. there are some from 2015.. where are they located?
[21:29:14] <abaumann> the use 'disass' to disassemle the code, you might have to go up in the call stack with 'bt' (for backtrace)
[21:29:39] <abaumann> they are either in the systemd journal (which was not so a wise decision) or they are stored externally somewhere on disk.
[21:29:59] <abaumann> there is no command to wipe them AFAIK
[21:30:00] <jarshvor> abaumann: gotcha
[21:30:23] <abaumann> me? hope not.. ;-)
[21:30:29] <jarshvor> :P
[21:30:55] <abaumann> depends on /etc/systemd/coredump.conf (Storage=external) and stuff where core dumps are.
[21:31:02] <abaumann> but coredumpctl can always find them.
[21:32:10] <jarshvor> http://sprunge.us
[21:32:39] <abaumann> yep. that's the call stack
[21:32:53] <jarshvor> wish I understoof what all that hex means.. =)
[21:33:02] <abaumann> simple
[21:33:03] <abaumann> #7 0x00000000b62b7167 n/a (r300_dri.so)
[21:33:03] <jarshvor> understood*
[21:33:32] <abaumann> this is a list of addresses of functions.
[21:33:36] -!- rcf has quit [Quit: WeeChat 2.0.1]
[21:33:45] <abaumann> _start is the entry point, then things get called.
[21:33:57] <abaumann> use 'bt' till you are at '#7'
[21:34:33] <abaumann> the use 'disass' to disassemble the code
[21:34:35] -!- rcf has joined #archlinux32
[21:34:49] <jarshvor> coredumpctl bt?
[21:34:54] <abaumann> you can paginate till you see a '=> ' on the call stack, there is the failing instruction
[21:34:54] <Ingar> it looks awfully similar http://sprunge.us
[21:35:05] <abaumann> yep.
[21:35:56] <abaumann> mmh. i'm searching a nice tutorial for this debugging thing.
[21:36:09] <jarshvor> :S im not quite getting this.. $coredumpctl bt unknown operation bt
[21:36:18] <abaumann> coredumpctl debug
[21:36:20] <abaumann> first
[21:36:28] <abaumann> then you get into a prompt (gdb)
[21:36:36] <jarshvor> ahh wait
[21:36:39] <jarshvor> missed something
[21:36:49] <jarshvor> failed to invoke gdb
[21:36:53] <jarshvor> let me install it?
[21:37:06] <Ingar> abaumann: how do I get back to #7 ?
[21:37:18] <abaumann> ah..
[21:37:21] <abaumann> sorry. my fault.
[21:37:27] <abaumann> bt shows the whole stack (backtrace)
[21:37:29] <abaumann> up
[21:37:32] <jarshvor> not at all. =) should have read throroughly
[21:37:32] <abaumann> and down
[21:37:37] <abaumann> are the commands to naviage the call stack
[21:37:39] <Ingar> yeah that I know ;)
[21:37:48] <Ingar> I just havent used gdb in a while ;)
[21:38:08] <abaumann> :-)
[21:38:38] <jarshvor> ok do im in gdb promt... did bt... for adresses? from 0 to 28
[21:38:58] <abaumann> yes, so you can navigate the stack with 'up' and 'down'
[21:39:10] <jarshvor> and where do you want it?
[21:39:26] <Ingar> mmm
[21:39:30] <Ingar> "No function contains program counter for selected frame."
[21:39:33] <deep42thought> abaumann: is your bootstrap mirror serving old/cached content? I uploaded glib2 to staging, but it doesn't pop up in my pacman
[21:39:34] <Ingar> for 7 and 8
[21:39:58] <abaumann> mmh. that's bad. so the code got inlined?
[21:40:15] <Ingar> that would explain why it would happen across drivers
[21:40:27] <abaumann> deep42thought: aeh, you uploaded to boostrap/staging, I hope.
[21:40:34] <deep42thought> yes
[21:40:38] <deep42thought> ah, found my error
[21:40:39] <abaumann> the others get overwritten from the master mirrors.
[21:40:49] <deep42thought> I repo-added to the wrong database file
[21:41:01] <deep42thought> e.g. staging instead of bootstrap-staging
[21:41:04] <abaumann> ah.
[21:41:09] <abaumann> yeah. forgot to tell you.
[21:41:51] <abaumann> mmh.. it's to much to run X now under gdb.. I fear that's the only way to get the code when SIGILL is happening.
[21:42:03] <abaumann> Well. I'll find a way to debug it..
[21:42:48] <abaumann> why is libvirt suddenly eating up all my CPU?
[21:44:54] <Ingar> that, I don't have
[21:45:47] <abaumann> mmhh. my virtual windows says 0x00005c, please press power button.. so something is completly weird with my setup (64-bit Archlinux and libvirt). Most likekely my 64-bit CPU now finally got too old for virtualization (Core Duo)
[21:46:19] <jarshvor> sorry guys.. I got a bit lost.. =) Is there anything I can assist in with my coredumps or do you guys have it under control? still in gdb
[21:46:37] <abaumann> sorry jarshvor, too many things in parallel :-)
[21:46:57] <abaumann> no, I think I have to reproduce the error here in order to find the cause.
[21:47:14] <Ingar> jarshvor: I guess the problem is mesa being compiled with funky CPU instructions that aren't supported on older systems
[21:47:25] <abaumann> exacltly.
[21:47:33] -!- turbokitty has joined #archlinux32
[21:47:42] <Ingar> debuggin it is a major PITA
[21:47:51] <abaumann> and I suspect SSSE3, SSSE4, AVX, AVX2 stuff, as the CPU is new enough to support SSE, SSE2, MMX
[21:49:05] <Ingar> maybe they are advertised by the buildhost ?
[21:49:25] <Ingar> I remember, back in the day, programs like mplayer determining CPU config at build time
[21:49:35] <abaumann> In my experience half of the time the autoconf helper are borked or just plainly wrong.
[21:49:49] <jarshvor> thats alright =) do let me know if I can assist.. (I understand though that this is above my understanding atm so by all means I dont want to distract you with baby tutoring..) Ill leave this in away and peruse it later for fun just to see what I can understand and not) ;)
[21:50:01] <jarshvor> thanks for your help and work guys! ;)
[21:50:06] <abaumann> nowadays they ask the compiler: can you generate SSE2 code, and the compiler will happilly do so.
[21:50:20] <abaumann> the only proper way is to have a switch --with-avx, --without-avx
[21:51:08] <abaumann> jarshvor: absolutely no problem :-)
[21:54:10] <abaumann> so, all my virtual machines either crash or use 100% after saying 'Loading initial ramdisk' (with debug enabled).. I'm lost..
[21:55:16] <abaumann> Oct 06 21:52:11 eurobuild3 systemd-machined[593]: New machine qemu-4-arch32-stable.
[21:55:26] <abaumann> oh. don't tell me systemd is fiddling there now too..
[21:55:56] <abaumann> tired and lost.. cu tomorrow. :-)
[21:56:05] <deep42thought> good night!
[21:56:13] <abaumann> thanks. :-)
[21:56:15] -!- abaumann has quit [Quit: leaving]
[21:57:08] <Ingar> computers... they suck! nn abaumann.
[22:19:01] -!- turbokitty has quit [Quit: Leaving]
[22:21:39] -!- turbokitty has joined #archlinux32
[22:33:27] -!- turbokitty has quit [Quit: Leaving]
[22:43:48] -!- bill-auger has joined #archlinux32
[23:29:03] -!- deep42thought has quit [Quit: Leaving.]