04:09:40 <gitomat> [illumos-gate] 17210 meoi expectations broken by 17136 -- Patrick Mooney <pmooney⊙pc>
14:19:36 <gitomat> [illumos-gate] 17167 dld: unsigned is never less than zero -- Toomas Soome <tsoome⊙mc>
15:47:50 <jbk> hrm... there isn't a way with USCSI to direct a CDB down a specific path of a multipath device, is there?
15:56:31 <tsoome_> you can disable other path
15:58:00 <tsoome_> but otherwise it depends if you have A/A or A/PG
16:04:14 <jbk> hm.. i see an undocumented USCSI_PATH_INSTANCE and uscsi_path_instance in struct uscsi_cmd (ulong)..
16:06:40 <jbk> it looks like the instance#s are maybe just assigned from a global int that's just incremented every time a new path is created
17:26:02 <tsoome_> hm... stupid spam detectors somehow illumos-advocates mail ended in junk.:(
17:30:01 <jbk> i've had that happen a few times
17:31:43 <gitomat> [illumos-gate] 17161 mxfe: variable dereferenced before check -- Toomas Soome <tsoome⊙mc>
18:18:13 <Guest7> What would it take to compile illumos-gate with Clang?
18:27:21 <tsoome_> a lot.
18:28:41 <alanc> presumably the first step would be to take the illumos-specific gcc patches and figure out how to implement them in an illumos-specific clang
18:30:39 <tsoome_> it is not hopeless, it has been improving, but it still needs effort. the compiler patches are to provide some specific features we want to have, those are not quite needed if you want to just build the code. 
18:34:29 <tsoome_> once in a time we had assembler course in uni and to get to complete it one guy was tasked with chess related problem. when time was up, he had really beautiful chessboard drawn on screen, but problem was still not solved;)
18:41:57 <Woodstock> they obviously hadn't mastered the tao of programming.
20:54:45 <richlowe> alanc: the illumos-specific patch that matters is msave-args, which is also the hard one, but is only relevant to running.
20:55:06 <richlowe> I have marginal faith it is more readily implemented as a clang plugin than a gcc one
20:55:15 <richlowe> well, an llvm plugin really
21:25:53 <jbk> it looks like openbsd added it to clang/llvm, though not as a plugin
21:26:28 <richlowe> then perhaps you can't do it there, either.
21:26:44 <richlowe> but that's good to know.  I hope tsoome knows too
21:27:03 <tsoome_> havent checked that much:)
21:27:06 <jbk> https://github.com/openbsd/src/blob/0fbcf26120002658371aac08d58f0dc72325daa0/gnu/llvm/llvm/lib/Target/X86/X86.td#L494-L496
21:27:14 <tsoome_> but here we are;)
21:27:19 <jbk> obviously not all of the work, but at least the hint suggesting it's there
21:27:38 <richlowe> hadfl: You might be interested in this too
21:29:35 <hadfl> oh :)
21:35:23 <jbk> obviously i've also not used it, so caveat developer :P
21:41:53 <richlowe> conditor?
21:52:45 <jbk> *shrug* no idea :)