Re: NetBSD 10.0 - configure test process hang
Martin Husemann <[email protected]>
| Newsgroups | gmane.os.netbsd.ports.sparc |
|---|---|
| Message-ID | <[email protected]> |
On Fri, Nov 08, 2024 at 04:54:39PM +0100, Riccardo Mottola wrote: > > That is the reason why we don't have COMPAT_NETBSD32 test runs for > > sparc on sparc64 hardware - the tests basically never complete. > > > > The official sparc testbed runs single CPU and doesn't get into this kind > > of trouble. > > faster CPUs? in terms of MHz of processing power? faster CPUs trigger the race more easily. > Is this a new thing? I did not experience it in 9.x series, appears to > be a 10.0 novelty. It has been this way ~forever. > SO this is for both SPARC 32 & 64 bit? Anywhere else? It only happens with sparc (32bit) userland. It probably will go away if you compile 32bit userland for ultrasparc cpus only (but AFAICT noone has exactly identified the failing piece of code). Martin