Re: arc4random abort

[B Harder <brad.harder%gmail.com@localhost>]

> On Feb 28, 2025, at 14:52, Patrick Welche <prlw1%welche.eu@localhost> wrote:
> 
> On Fri, Feb 28, 2025 at 10:36:48PM +0000, Patrick Welche wrote:
>> Updating by building -current/amd64, I keep getting
>> 
>> (gdb) bt
>> #0  0x0000785bb12109fa in _lwp_kill () from /lib/libc.so.12
>> #1  0x0000785bb121be34 in abort () from /lib/libc.so.12
>> #2  0x0000785bb121bd9a in arc4random_prng_get[cold] () from /lib/libc.so.12
>> #3  0x0000785bb1202feb in arc4random_uniform () from /lib/libc.so.12
>> #4  0x0000785bb1202037 in __gettemp () from /lib/libc.so.12
>> #5  0x0000785bb1201f26 in mkstemp () from /lib/libc.so.12
>> #
>> 
>> with all sorts of binaries, like sshd, tar, make, etc.

mktemp, vi…

This seems like a serious tarpit…


>> 
>> Really puzzling is that I /rescue/untarred
>> pub/NetBSD-daily/HEAD/202502280540Z/amd64/binary/sets/
>> and am still seeing it (so that should be the daily libc).
>> 
>> Any guesses while I keep digging?
> 
> A second /rescue/untarring of base fixed things.
> 

I’m in the middle of a rescue attempt, but I’ll explore this too. I’d recommend -current users take care of their libc if they’re updating, or wait until this is understood.

-bch


> Very odd...
> 
> Patrick