Date: 28 Nov 2007 17:41:12 +0100 From: "Arno J. Klaassen" <arno@heho.snv.jussieu.fr> To: Norberto Meijome <freebsd@meijome.net> Cc: "Cc: \"Andrew Tulloch\"" <andrew.tulloch@complinet.com>, freebsd-java@freebsd.org Subject: Re: FreeBSD 7.0-BETA3 JDK16 libthr crash Message-ID: <wp3auqqq53.fsf@heho.snv.jussieu.fr> In-Reply-To: <20071128165010.1160cb7b@meijome.net> References: <366BFCC95F75344DA1E0C4598B3CE19D886987@gboff01-exbe1.complinet.local> <wpve7nrbdh.fsf@heho.snv.jussieu.fr> <20071128165010.1160cb7b@meijome.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Norberto Meijome <freebsd@meijome.net> writes: > On 27 Nov 2007 15:50:18 +0100 > "Arno J. Klaassen" <arno@heho.snv.jussieu.fr> wrote: > > > this looks like what I get with 7.0-BETA3 with SCHED_ULE on > > a 2x Opteron 285 (Dual core) with my own java program : > > > > Fatal error 'mutex is on list' at line 432 in file /files/bsd/src7/lib/libthr/thread/thr_mutex.c (errno = 0) > > maybe a long shot, but I was getting errors like that in some programs ( I don't think they were java progs) after doing an upgrade from 6 to 7, and some programs were using a mix of 6.x and 7 libraries.... I eventually fixed it by rebuilding all libraries under 7, otherwise u can try installing the compat6x port. nope, this is a completely clean 7 install, starting from cvsup-sources as of Oct 11 (were still -current at that time I guess). Anyway, the good news is that the MFC of rev 1.35 of intr_machdep.c (amd64) seems to solve, or at least alleviate, the problem : no more never seen before errors while running test for about a day now. Before they showed up rather quickly. I as well use hand-mfced version 1.287-289 for kern_fork.c, but without rev 1.35 of intr_machdep.c that didn't matter. Could releng_6 need as well this "Extend critical section coverage" since I see the same problems on (my one and only) 2x2way releng_6 with jdk15? Best, Arno
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?wp3auqqq53.fsf>