Switch to DuckDuckGo Search
   April 16, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >


NOTICE: This channel is no longer actively logged.

Toggle Join/Part | bottom
[00:04:31] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[00:09:35] <papertigers> neirac: ping
[00:09:53] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Read error: Connection reset by peer)
[00:10:11] <papertigers> when you added the epoll stuff to rust's libc crate where did you grab the definitions from? I was curious if oracle solaris had epoll as well now
[00:14:17] <jbk> i'd still like to see us use event ports instead :)
[00:14:55] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has quit IRC (Quit: jellydonut)
[00:16:08] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[00:16:37] <papertigers> was just talking about that in another window
[00:16:58] <papertigers> I have an inital branch from awhile back. I would like to take another stab at it. But it looks like theres another user doing work as well
[00:17:22] *** phyre <phyre!~phyre___@78.30.16.107> has quit IRC (Remote host closed the connection)
[00:18:08] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has joined #illumos
[00:34:59] *** andy_js <andy_js!~andy@90.218.209.121> has quit IRC (Quit: andy_js)
[01:03:56] *** wacki <wacki!~wacki@i577A59B5.versanet.de> has joined #illumos
[01:08:35] *** wacki <wacki!~wacki@i577A59B5.versanet.de> has quit IRC (Ping timeout: 260 seconds)
[01:15:57] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has quit IRC (Remote host closed the connection)
[01:25:38] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has quit IRC (Ping timeout: 265 seconds)
[01:38:37] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[01:54:11] <rmustacc> jbk: Given that the number of signals has increased you definitely want to make sure that's not encoded there. Any reason to start with the behavior of supplying a buffer and count? Or is the idea there for reuse?
[01:54:19] <rmustacc> You'll want a free function of libproc allocated the memory too.
[01:55:19] <rmustacc> A size_t feels a little weird for a count of entries. A uint_t may make more sense. *shrug*
[01:57:52] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[01:59:30] *** hemi770_ is now known as hemi770
[02:01:02] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[02:04:42] <jbk> mostly since the other proc_get_xxx functions tend to take a pointer to the buffer to populate.. though that doesn't work too well here since the size needed could change in the future (and actually it looks like lx branded processes on smartos might have a different amount already)
[02:05:46] <jbk> though one other wrinkle.. struct sigaction does vary between 32/64 bit as well..
[02:05:47] <rmustacc> Take a look at proc_get_fdinfo(3PROC).
[02:06:15] <rmustacc> OK, then a single wrapper like this probably isn't appropriate.
[02:07:01] <jbk> (because it includes a field for the address of the signal handler)
[02:07:35] <rmustacc> What's the consumer here?
[02:09:16] <jbk> the immediate one was to convert w & whodat to use libproc (instead of walking /proc itself).. when it's calculating JCPU and PCPU it ignores processes for that session that ignore SIGINT and SIGQUIT
[02:09:31] <jbk> though, having psig also use it as well would be nice
[02:09:37] <rmustacc> I would fix the security thing with the minimal bit necessary.
[02:09:46] <rmustacc> And come back to clean it up as another pass (imho).
[02:10:04] <rmustacc> Depends how gross it is.
[02:10:15] <rmustacc> But what are the /proc semantics on reaidng the sigaction struct?
[02:10:51] <rmustacc> I would assume this is something where it might unhelpfully say the bitness needs to match the target, but no idea, to be honest.
[02:11:13] <jbk> well i think in this case, it largely still works post-64 bit world by accident :)
[02:11:25] <jbk> as i looked at it
[02:11:26] <rmustacc> It appears the structs are the same size, but the offsets aren't.
[02:14:56] <rmustacc> So seems like having in libproc which smooths this over is useful. But it maybe can't be a stock struct sigaction.
[02:21:30] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has quit IRC (Ping timeout: 258 seconds)
[02:34:31] <richlowe> tsoome_: what's up?
[03:01:57] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has joined #illumos
[03:04:55] *** wacki <wacki!~wacki@i577A59B5.versanet.de> has joined #illumos
[03:09:39] *** wacki <wacki!~wacki@i577A59B5.versanet.de> has quit IRC (Ping timeout: 256 seconds)
[03:16:22] *** wiedi_ <wiedi_!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has quit IRC (Read error: Connection reset by peer)
[03:16:31] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[03:24:00] *** varna <varna!~varna@60.176.146.221> has quit IRC (Read error: Connection reset by peer)
[03:27:49] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has quit IRC (Ping timeout: 258 seconds)
[03:27:57] *** varna <varna!~varna@125.120.227.3> has joined #illumos
[03:29:57] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has quit IRC (Quit: Lost terminal)
[03:52:41] *** xzilla_ is now known as xzilla
[04:08:05] *** ke5c2fin <ke5c2fin!~ke5c2fin@c-67-176-119-108.hsd1.co.comcast.net> has joined #illumos
[04:21:12] *** AlecTaylor <AlecTaylor!~AlecTaylo@unaffiliated/alectaylor> has joined #illumos
[04:21:15] <AlecTaylor> hi
[04:21:48] <idodeclare> hello
[04:22:23] <AlecTaylor> Is there a CI solution for illumos? - I wrote a little C library to bootstrap different OSs and I've got GitHub Actions (Windows, Linux, macOS) and Cirrus CI (FreeBSD), but now need something for illumos
[04:58:11] *** ke5c2fin <ke5c2fin!~ke5c2fin@c-67-176-119-108.hsd1.co.comcast.net> has quit IRC (Ping timeout: 260 seconds)
[05:05:55] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has joined #illumos
[05:08:00] *** AlecTaylor <AlecTaylor!~AlecTaylo@unaffiliated/alectaylor> has quit IRC (Quit: This computer has gone to sleep)
[05:10:46] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has quit IRC (Ping timeout: 256 seconds)
[05:23:59] *** arnoldoree <arnoldoree!~arnoldore@2001:d08:2183:afdb:c1cd:b139:c4b1:918f> has joined #illumos
[06:25:31] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has joined #illumos
[06:47:40] *** Teknix <Teknix!~pds@69.41.134.110> has quit IRC (Ping timeout: 256 seconds)
[06:50:59] *** Teknix <Teknix!~pds@69.41.134.110> has joined #illumos
[06:58:55] <gitomat> [illumos-gate] 12424 openeepr: variable may be used uninitialized -- Toomas Soome <tsoome at me dot com>
[07:01:08] <gitomat> [illumos-gate] 12516 bignum: File does not contain CTF data -- Toomas Soome <tsoome at me dot com>
[07:01:12] *** BOKALDO <BOKALDO!~BOKALDO@81.198.20.220> has joined #illumos
[07:03:43] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has quit IRC (Quit: leaving)
[07:03:57] *** tsoome_ <tsoome_!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: Leaving)
[07:04:06] *** tsoome <tsoome!~tsoome@7129-0555-394f-640e-2f80-4a40-07d0-2001.sta.estpak.ee> has joined #illumos
[07:06:59] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has joined #illumos
[07:11:41] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has quit IRC (Ping timeout: 258 seconds)
[07:25:42] *** phyre <phyre!~phyre___@78.30.16.107> has joined #illumos
[07:28:09] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has joined #illumos
[07:31:41] *** joltman <joltman!znc@gateway/vpn/privateinternetaccess/joltman> has quit IRC (Ping timeout: 250 seconds)
[07:33:57] *** joltman <joltman!~znc@unaffiliated/joltman> has joined #illumos
[08:14:00] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[08:20:19] <jbk> hrm..
[08:28:22] <Agnar> moin
[08:29:41] <jbk> # nm /lib/libc.so | grep itoa | wc -l
[08:29:42] <jbk> 3
[08:29:47] <jbk> seems like one would be sufficient..
[08:34:06] *** neuroserve <neuroserve!~toens@ip-178-202-216-85.hsi09.unitymediagroup.de> has quit IRC (Ping timeout: 258 seconds)
[08:40:41] *** neuroserve <neuroserve!~toens@ip-84-118-130-67.unity-media.net> has joined #illumos
[08:50:34] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has quit IRC (Remote host closed the connection)
[08:51:01] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has joined #illumos
[08:58:59] *** ibenn <ibenn!~benn@HSI-KBW-095-208-237-062.hsi5.kabel-badenwuerttemberg.de> has joined #illumos
[09:00:40] *** ibenn <ibenn!~benn@HSI-KBW-095-208-237-062.hsi5.kabel-badenwuerttemberg.de> has quit IRC (Client Quit)
[09:03:07] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 250 seconds)
[09:08:34] *** ibenn <ibenn!~benn@HSI-KBW-095-208-237-062.hsi5.kabel-badenwuerttemberg.de> has joined #illumos
[09:10:30] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[09:15:33] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has quit IRC (Quit: leaving)
[09:18:38] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has joined #illumos
[09:29:23] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has joined #illumos
[09:29:49] *** kayront <kayront!~kayront@unaffiliated/kayront> has quit IRC (Read error: Connection reset by peer)
[09:30:21] *** kayront <kayront!~kayront@zbase.xen.prgmr.com> has joined #illumos
[09:30:44] *** kayront is now known as Guest34625
[09:31:28] *** man_u <man_u!~manu@89-92-19-81.hfc.dyn.abo.bbox.fr> has joined #illumos
[09:42:31] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[09:43:20] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has joined #illumos
[09:55:40] *** andy_js <andy_js!~andy@90.218.209.121> has joined #illumos
[09:56:59] <sjorge> hmm I don't think we have a CI infra somewhere for illumos right? I guess it would be useful in the long run with rust, go, (corecrl *crosses fingers*) slowling getting support
[09:57:08] <sjorge> And well openzfs if we add illumos back to the new upstream
[09:57:28] <sjorge> jbk: I reworked it based on round 2 of feedback, still works - https://github.com/joyent/illumos-joyent/compare/master...sjorge:pam_list?expand=1
[09:58:10] <sjorge> Spend an hour trying to figure out why I was not getting it the first go, I had swapped my check for realloc to a B_FALSE one... then fix it but I forgot to reboot the zone (I lofs mount the files in for testing)
[10:06:28] <sjorge> also saw your reply on the openzfs issue logged yesterday (?), how far behind are we? Jerry did a lot of catch up work for crypto right?
[10:17:39] *** nde <nde!uid414739@gateway/web/irccloud.com/x-wuhedspkfwknirwv> has quit IRC (Quit: Connection closed for inactivity)
[10:44:59] <Agnar> alanc: regarding https://github.com/0xdea/advisories/blob/master/2020-07-solaris-whodo-w.txt, it does not Seg Fault on my OI hipster
[11:00:54] *** Guest34625 is now known as kayront
[11:09:07] <sjorge> ooh https://www.illumos.org/issues/12509 this is a good idea, we keep hitting that in pkgsrc too where we need to create a wrapepr script
[11:09:21] <sjorge> Which works, but sucks if we want to run the service as non root, for things that don't drop privilages
[11:09:27] <sjorge> Because then that approuch doesn't work
[11:13:53] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[11:14:07] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Client Quit)
[11:29:52] *** AlecTaylor <AlecTaylor!~AlecTaylo@203.166.234.35> has joined #illumos
[11:54:51] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has quit IRC (Quit: nbhauke)
[11:57:56] <v_a_b> sjorge Yes. Shouldn't be that hard to implement. Not sure if alp is going to tackle it.
[12:02:53] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[12:03:11] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Client Quit)
[12:05:08] <tsoome__> igork do you still have the link to sparc prom debugging hints?
[12:34:44] <igork> tsoome__: one moment
[12:35:58] <igork> one is: https://docs.oracle.com/cd/E19457-01/801-7042/801-7042.pdf
[12:36:21] <igork> https://docs.oracle.com/cd/E36784_01/pdf/E36858.pdf
[12:36:26] <igork> another one
[12:37:41] <igork> another one: https://blogs.oracle.com/cwb/entry/debugging_sparc_reall_and_i
[12:37:51] <igork> and interest one: https://arcb.csc.ncsu.edu/~mueller/codeopt/codeopt00/notes/sparc.html
[12:38:45] <igork> http://docs.oracle.com/cd/E19253-01/816-1177-10/debug.html#pgfId-17069
[12:39:52] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Read error: Connection reset by peer)
[12:40:04] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[12:44:52] <tsoome__> thanks:)
[12:45:43] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[13:28:08] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has joined #illumos
[13:28:42] *** AlecTaylor <AlecTaylor!~AlecTaylo@unaffiliated/alectaylor> has quit IRC (Quit: This computer has gone to sleep)
[14:01:47] *** jemershaw <jemershaw!~jemershaw@c-68-83-252-28.hsd1.pa.comcast.net> has quit IRC (Remote host closed the connection)
[14:01:59] *** jemershaw <jemershaw!~jemershaw@c-68-83-252-28.hsd1.pa.comcast.net> has joined #illumos
[14:16:58] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has joined #illumos
[14:23:25] *** AlecTaylor <AlecTaylor!~AlecTaylo@unaffiliated/alectaylor> has joined #illumos
[14:33:07] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has quit IRC (Quit: Leaving.)
[14:35:21] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has joined #illumos
[14:48:51] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[15:28:47] *** AlecTaylor <AlecTaylor!~AlecTaylo@unaffiliated/alectaylor> has quit IRC (Quit: Leaving)
[15:32:32] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[15:48:26] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has joined #illumos
[15:59:02] *** nde <nde!uid414739@gateway/web/irccloud.com/x-zyekkbryknugxuar> has joined #illumos
[16:18:02] *** phyre <phyre!~phyre___@78.30.16.107> has quit IRC (Remote host closed the connection)
[16:23:35] *** phyre <phyre!~phyre___@78.30.16.107> has joined #illumos
[16:29:53] <jbk> Agnar: then that's probably just luck then.. i was able to reproduce it with a pretty recent-ish illumos-gate, and stop it w/ the fix i wrote yesterday
[16:31:49] <Agnar> jbk: funny.
[16:36:22] <Agnar> ah, right, now I see, sorry my fault
[16:37:08] <Agnar> got it faulting
[16:38:33] <jbk> i'm doing a bit more testing just to make sure nothing broke w/ my fixes (I did add some additional hardening beyond the immediate buffer overflow)
[16:39:38] <jbk> and since there's no stability attributes associated with either utility (if anything should be 'not an interface'), the implicit assumption is that we're ok with truncation of output
[16:40:00] <jbk> (i think it's very reasonable, but just to be explicit)
[16:40:44] <Agnar> well, ps does truncate...so why not w?
[16:46:29] <am11> under linux ucontext.h defines REG_CSGSFS as: `REG_CSGSFS, /* Actually short cs, gs, fs, __pad0. */`, is there way to define it on Solaris x64?
[16:49:21] <jbk> am11: what is the context?
[16:51:15] *** ibenn <ibenn!~benn@HSI-KBW-095-208-237-062.hsi5.kabel-badenwuerttemberg.de> has quit IRC (Quit: Leaving)
[16:51:21] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has joined #illumos
[16:51:34] <KungFuJesus> jbk: So did
[16:51:51] <KungFuJesus> Saso Kiselkov contact you directly. Or did he just expect me to relay that message to you?
[16:52:22] <KungFuJesus> do you have enough to (re)work his patches to improve the situation with the context switching with AES-NI?
[16:52:46] <am11> jbk: dotnet runtime requires it in one of its header `context.h`: `Header file for thread context utility functions.`
[16:52:50] <jbk> the ucontext_t.uc_mcontext.gregs[] stores CS, GS, and FS
[16:53:10] <jbk> the defines REG_CS, REG_GS, REG_FS contain the indicies for them
[16:53:32] <jbk> KungFuJesus: i have not seen an email from him yet
[16:53:43] <am11> https://github.com/dotnet/runtime/blob/c2f47d9/src/coreclr/src/pal/src/include/pal/context.h#L124
[16:54:04] <jbk> i mean, the code itself is fine -- i just wanted to get confirmation on the copyright notices (or lack thereof in places)
[16:55:14] <jbk> heh.. every bit is sacred on linux i guess :) (packing all of those registers into one entry)
[16:56:20] <am11> jbk: i can define it as `REG_CS` to pass by, that's how it is defined for other platforms in that file.
[16:56:31] <jbk> am11: you probably want something like `#ifdef __sun #define MCREG_SegCs(mc) ((mc).gregs[REG_CS]) #else (existing line) #endif
[16:57:07] <am11> yup, exactly that :)
[17:03:34] <rmustacc> sjorge, v_a_b: I'm actually looking at that in my spare time (12509)
[17:12:13] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has quit IRC (Remote host closed the connection)
[17:13:08] *** varna <varna!~varna@125.120.227.3> has quit IRC (Ping timeout: 258 seconds)
[17:14:53] *** varna <varna!~varna@125.120.227.3> has joined #illumos
[17:15:23] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[17:17:42] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has quit IRC (Quit: nbhauke)
[17:47:42] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has joined #illumos
[17:54:24] <KungFuJesus> jbk: you mean whodo?
[17:54:37] <KungFuJesus> whodat would be an awesome name for that binary, though
[17:54:52] <jbk> err yeah
[17:55:06] <jbk> i think i've been down here too long :)
[17:55:20] <KungFuJesus> that's it, I'm writing whodat
[17:55:33] <KungFuJesus> like finger on steroids
[17:58:22] <v_a_b> rmustacc Nice! Looking forward to testing it. Toyed with the idea of doing it myself, but given my previous track records it would be finished in 2032.
[17:59:34] <am11> noticed another difference, in linux' `context.h` header, `mcontext_t.fpregs` is a pointer, however, in solaris mcontext.h, `mcontext_t.fpregs` is not a pointer. will add introspection for this case.
[18:03:29] <jbk> sorry about that
[18:06:11] <jbk> it's a new south/new orleans/cajun locale :)
[18:07:06] <toastersonerson1> why does that remind me of the movie Airplane again
[18:08:23] *** neuroserve <neuroserve!~toens@ip-84-118-130-67.unity-media.net> has quit IRC (Ping timeout: 256 seconds)
[18:10:33] <KungFuJesus> it will compliment the command 'whodis'
[18:10:56] <KungFuJesus> as in, 'new phone, whodis?'
[18:11:08] <toastersonerson1> sigh: -bash: whodis: command not found
[18:11:14] <toastersonerson1> i hoped...
[18:12:19] <alanc> ln -s finger whodat
[18:12:31] <toastersonerson1> but it could be a programm that asks the sender of a text for details.
[18:12:56] <andyf> Anything's better than "pinky"...
[18:13:04] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has quit IRC (Quit: nbhauke)
[18:27:07] *** jimklimov2 <jimklimov2!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has quit IRC (Ping timeout: 258 seconds)
[18:27:26] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has quit IRC (Ping timeout: 256 seconds)
[18:29:44] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[18:44:04] *** jimklimov1 <jimklimov1!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[18:47:07] *** neuroserve <neuroserve!~toens@ip-178-202-216-85.hsi09.unitymediagroup.de> has joined #illumos
[19:00:24] <sjorge> rmustacc: nice!
[19:14:18] <gitomat> [illumos-gate] 12475 want cyclic_move_here() -- Patrick Mooney <pmooney at pfmooney dot com>
[19:14:41] <LeftWing> The first bhyve transport is away!
[19:14:50] <LeftWing> *cheers*
[19:22:01] <toastersonerson1> Upstream?
[19:27:38] <danmcd> Yes. You can't have bhyve without cyclic_move_here().
[19:31:48] <am11> is this `cwd` https://code.woboq.org/userspace/glibc/sysdeps/unix/sysv/linux/x86/sys/ucontext.h.html#_libc_fpstate::cwd, same thing as `struct _fpchip_state.cw` in Solaris' mcontext.h?
[19:33:19] <toastersonerson1> Ah makes sense once you know. My brain still keeps reading cycle as in bycicle
[19:33:33] <KungFuJesus> Has there been any upstreamed effort for a OI SPARC port again? I just tried to upgrade Solaris 11.3 on my SPARC machine and...had some pretty miserable results. Technically it's not a supported upgrade path
[19:34:05] <danmcd> KungFuJesus: Best sparc results, AIUI, are from Tribblix or Dilos.
[19:35:38] <KungFuJesus> I kind of like IPS, though. I mean there are things I dislike about the implementation (like the 20 minute process of transactional package system thing), but I like the built in BE support and the simple interface
[19:36:11] <KungFuJesus> oh and the rolling updates via Hipster's IPS
[19:37:59] <wacki> I think tsoome is working on OI for SPARC, maybe he will chime in...
[19:40:51] <KungFuJesus> tsoome__: where can I subscribe?
[19:40:54] <toastersonerson1> tsoome what hotlines does a server do?
[19:42:58] *** tsoome_ <tsoome_!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[19:43:14] <tsoome> toastersonerson1, on x86, it does set up smap entries.
[19:43:24] *** tsoome_ <tsoome_!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Client Quit)
[19:43:29] <jbk> am11: i think so
[19:43:48] <jbk> though i'm not super-familiar with the fp stuff.. but it seems very likely
[19:57:53] <jbk> tsoome: not to be confused with hothlines, which could be very cold :P
[19:59:16] <tsoome> :D
[20:05:36] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has quit IRC (Quit: leaving)
[20:11:08] <jbk> though maybe a couple weeks too early for that joke..
[20:13:14] *** ypankov <ypankov!~ypankov@85.175.139.251> has joined #illumos
[20:27:51] <am11> jbk: at the callsite it is `FPREG_ControlWord(native) = lpContext->FltSave.ControlWord` and macro on linux expands to cwd, so the chances of ambiguity are bleak. :)
[20:28:24] <jbk> also, still unresolved about the crt1.o bits..
[20:28:52] <jbk> it may be easier to do a second PR for libunwind that fixes it there
[20:30:05] <jbk> basically it can switch from using the eh_frame data to using frame pointers (which should always be present for stuff compiled on illumos).. except when it does that, it pops out a frame with ip=0, when it should really have just stopped there
[20:31:08] *** neirac <neirac!~cneira@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[20:31:18] *** tsoome_ <tsoome_!~tsoome@91.209.240.229> has joined #illumos
[20:31:34] *** tsoome__ <tsoome__!~tsoome@91.209.240.229> has quit IRC (Read error: Connection reset by peer)
[20:31:50] <jbk> it is arguable that compliance with the amd64 sysv ABI requires that there be eh_frame data that explicitly terminates a stack walk
[20:32:33] <jbk> arguable since the spec says 'see DWARF'
[20:32:50] <jbk> and DWARF just mentioned offhand 'yeah, you do this until you see X'
[20:32:56] <jbk> but
[20:33:17] <jbk> fixing it means having to compile on a system with the fix
[20:34:36] <jbk> if libunwind would accept a PR to terminate at ip=0 when using frame pointers (or rather don't treat that as a separate frame).. that would avoid having to worry about a sufficiently updated system when building
[20:36:04] <jbk> (since we try to be careful about not breaking ABIs as much as we can, it's not unusual to build something on the oldest release you want to use, knowing that it should work on future releases w/o issues
[20:36:47] <jbk> i know it's not always the case elsewhere, but we strive to try to avoid forcing people to recompile just because you updated the OS
[20:44:35] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has quit IRC (Ping timeout: 260 seconds)
[20:45:03] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has joined #illumos
[20:51:43] *** nbhauke <nbhauke!~hauke@55d44f26.access.ecotel.net> has joined #illumos
[21:18:51] <danmcd> Fired the ion cannon for the 2nd transport (cc LeftWing, pmooney)
[21:19:03] <LeftWing> pew pew!
[21:19:23] <danmcd> (good thing this isn't iMessage...)
[21:23:17] <gitomat> [illumos-gate] 12477 ctxops should use stack ordering for save/restore -- Patrick Mooney <pmooney at pfmooney dot com>
[21:23:28] <pmooney> danmcd: thanks
[21:28:08] *** BOKALDO <BOKALDO!~BOKALDO@81.198.20.220> has quit IRC (Quit: Leaving)
[21:29:31] <danmcd> YW, and thank you for tying the two bugs together.
[21:33:18] <pmooney> np
[21:47:38] *** c4rc4s_ <c4rc4s_!~c4rc4s@unaffiliated/c4rc4s> has joined #illumos
[21:48:49] *** c4rc4s <c4rc4s!~c4rc4s@unaffiliated/c4rc4s> has quit IRC (Ping timeout: 264 seconds)
[21:50:04] *** c4rc4s_ is now known as c4rc4s
[21:58:54] *** lpsmith_ is now known as lpsmith
[21:59:16] *** wacki <wacki!~wacki@i577B8B6D.versanet.de> has quit IRC (Quit: Lingo: www.lingoirc.com)
[22:01:46] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has quit IRC (Remote host closed the connection)
[22:02:26] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has joined #illumos
[22:30:35] *** man_u <man_u!~manu@89-92-19-81.hfc.dyn.abo.bbox.fr> has quit IRC (Quit: man_u)
[22:59:40] <Smithx10> LeftWing: It looks like that go-kstat lib.... isn't being maintained
[22:59:54] <LeftWing> Who owns it?
[22:59:56] <Smithx10> What's the plan for libraries and such
[23:00:16] <Smithx10> https://github.com/siebenmann/go-kstat
[23:00:24] <LeftWing> Ah, that guy
[23:00:30] <LeftWing> You could ask him
[23:00:39] <LeftWing> Maybe he would transfer ownership to the illumos org
[23:00:54] <LeftWing> And you an neirac could maintain it! :D
[23:00:56] <Smithx10> Yeah, that's what I was curious about
[23:01:02] <Smithx10> hahahahah LeftWing I like your style!
[23:01:29] <neirac> LeftWing I would need to learn more go then, haha.
[23:01:47] <LeftWing> It is never a bad time for more learning, I always say
[23:02:01] <Smithx10> ./mutley laugh
[23:02:12] <LeftWing> I am currently learning to type again, for instance, on this ortholinear keyboard
[23:02:41] <LeftWing> It is remarkably frustrating not being able to type haha
[23:02:54] *** ypankov <ypankov!~ypankov@85.175.139.251> has quit IRC (Ping timeout: 265 seconds)
[23:03:37] <neirac> LeftWing that's true, but when I bought a kinesis advantage I forced mysefl to type in that until it was comfortable is the only way
[23:03:50] <jbk> heh.. that's like when I got my work mbp :)
[23:04:31] <jbk> the almost non-existent travel + relearning where to rest your fingers when not on the keys so as to not trigger 'possessed laptop' mode (hitting the touchbar) took a bi
[23:04:34] <jbk> t
[23:04:44] <LeftWing> neirac: I am currently on Day 2 of that process haha
[23:05:15] <LeftWing> I like having my arms further apart, I just wish I could type a c instead of a v
[23:06:31] <neirac> LeftWing is the only way, work on it until is natural.
[23:06:42] <LeftWing> Indeed
[23:07:05] <LeftWing> Going from 100+ WPM to 20 is just jarring is all haha
[23:12:17] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[23:22:37] <am11> jbk: regarding https://github.com/libunwind/libunwind/pull/171, is it so that we would need to wait for the crt1.o fix to get the matching result?
[23:28:16] <jbk> yeah.. you'll get fewer tests passing without the other bit
[23:28:47] <jbk> i may just go ahead and maybe update that PR w/ the bits i mentioned earlier that might allow it to not require that...
[23:29:45] *** KungFuJesus <KungFuJesus!~adam@72.49.160.91> has quit IRC (Quit: leaving)
[23:32:22] <am11> thanks. that would be nice if it could be made self contained to elide the crt patch requirement. :)
[23:34:15] <nde> If i'm compiling with gcc, should something different be used over -Bsymbolic-functions ?
[23:44:38] *** andy_js <andy_js!~andy@90.218.209.121> has quit IRC (Quit: andy_js)
[23:51:25] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[23:57:11] *** phyre <phyre!~phyre___@78.30.16.107> has quit IRC (Remote host closed the connection)
top

   April 16, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >