Re: port-sparc/56788 Rump kernel panic: kernel diagnostic assertion "old == LOCK_LOCKED" failed

[Nick Hudson <nick.hudson%gmx.co.uk@localhost>]

The following reply was made to PR port-sparc/56788; it has been noted by GNATS.

From: Nick Hudson <nick.hudson%gmx.co.uk@localhost>
To: Tom Lane <tgl%sss.pgh.pa.us@localhost>, gnats-bugs%NetBSD.org@localhost,
 port-hppa-maintainer%netbsd.org@localhost
Cc: 
Subject: Re: port-sparc/56788 Rump kernel panic: kernel diagnostic assertion
 "old == LOCK_LOCKED" failed
Date: Wed, 15 Jun 2022 07:22:40 +0100

 On 15/06/2022 02:18, Tom Lane wrote:
 >> I think that some (more) magic needs to be added so that the RAS sectio=
 n
 >> of rumpns__atomic_cas_32 needs the appropriate rasctl(2) call.
 >
 > Hmm.  I've not found what bit of magic prepends "rumpns_" to the symbol
 > visible in the source code, but I wonder whether the RAS_ADDR() macro
 > talks to it.  Maybe librump's copy of __libc_atomic_init is redundantly
 > re-registering the RAS address ranges in libc's atomic_init_testset.o,
 > rather than the ones in librump's copy?
 
 who knows? I'm not a member of the magic circle
 
 I think this is where the symbol renaming happens
 
 https://nxr.netbsd.org/xref/src/sys/rump/Makefile.rump#255
 
 Nick