19:32:03 FYI: Things with Gerrit ought to be better now. We were under attack from some putatively AI crawler spam bots 19:33:39 jclulow: thanks for fixing this! 19:35:33 Wow! Thanks for the diagnosis and cure. Fucking AI bots... 19:40:32 dumb question.. do we just completely disable the ability to omit frame pointers in our gcc, or is one of the -fno-XXXX-XXXX options implicitly including that -- -O implies -fomit-frame-pointer, so I'm guessing we must be doing something to undo that 19:40:45 but i'm not seeing the option for it 19:42:53 or i'm not seeing anything explicitly doing it 19:43:02 in our gcc 10, yes: https://github.com/illumos/gcc/commit/f924f6560cf80d524069f469449c9b3fd486f11b 19:43:18 ahh ok 19:44:02 our gcc 14 just got -fforce-omit-frame-pointer: https://github.com/illumos/gcc/commit/6c5ac80a6c7f68da03d1ba58cd9d686a46566828 19:49:19 i'm trying to dig into an ntp problem, and the binary wasn't built with frame pointers annoyingly 19:51:10 but as best as I can tell, it appears to be running yyparse() in a signal handler... 19:51:29 which I really want to confirm 20:00:51 since that seems a somewhat insane thing to do 20:02:38 which NTP versioN? 20:03:39 jbk: clobbered gcc hard to make it never omit them, and then added an option to force it to (because building firmware) 20:04:05 the whole purpose being to make it hard for build systems to helpfully turn them off 20:04:33 yeah, i think this ntp got built with a different compiler binary... 20:04:37 i need to dig into that... 20:04:40 so if they're off by surprise for you, you'd probably want to try to work out how 20:04:46 yeah, :\ 20:05:19 sommerfeld: 4.2.8P15 20:05:39 okay, the original, not one of the forks like ntpsec.. 20:05:42 there's a long standing assert that gets tripped 20:05:56 it's been seen on illumos, but also on several linux distros as well 20:06:04 so I don't think it's anything we're doing wrong.. 20:06:29 the fact that of the stack I can see, it appears to be in a signal handler 20:07:10 since I can see those bits from libc 20:07:55 and looking at all the callers... i just have to think i've missed something.., or there's more corruption that's makign things misleading... 20:13:44 My life has improved since switching to chrony haha 20:16:41 that's in the works.. 20:16:48 [illumos-gate] 16006 Need RFC8106 decode in snoop -- Bill Sommerfeld 20:58:22 is Gerrit slow for anyone else again? (did the AI spambots evade our blocks?) 20:59:31 seems quite ok 21:01:06 it made me log in again - after that it seems to be better. hmm. 21:12:53 It will make you log in again when network weather is bad 21:12:56 I do not understand why 21:13:25 or rather, it will ask you to. It will stop when the weather improves 21:30:52 I often reload when asked to login again.