Date: Wed, 09 Apr 2003 14:35:12 -0400 From: Mike Tancsa <mike@sentex.net> To: Tor.Egge@cvsup.no.freebsd.org Cc: stable@freebsd.org Subject: Re: 3ware 3dmd broken in STABLE ( due to cvs commit: src/sys/kern kern_descrip.c kern_fork.c ) Message-ID: <5.2.0.9.0.20030409142819.07c59c60@marble.sentex.ca> In-Reply-To: <20030409181835K.tegge@cvsup.no.freebsd.org> References: <5.2.0.9.0.20030409140404.07e83b18@marble.sentex.ca> <ZHob6D.A.T0J.JBcj-@coal.sentex.ca> <5.2.0.9.0.20030409121043.05c934a0@marble.sentex.ca> <5.2.0.9.0.20030409140404.07e83b18@marble.sentex.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
At 06:18 PM 09/04/2003 +0000, Tor.Egge@cvsup.no.freebsd.org wrote:
> >
> > OK, some more information. As I narrowed it down to one day worth of
> > commits, its easy enough to jump back and forth. Running ktrace -di
> > /usr/local/sbin/3dmd shows its barfing when it goes to fork.
>
> > 199 3dmd CALL rfork(0x10)
> > 199 3dmd RET rfork 200/0xc8
>
>This explains the problem. The program calls rfork(RFPROC) which now
>fails due to my commit to kern_fork.c. Reverting back to revision
>1.72.2.11 of kern_fork.c should allow 3dmd to run, but it will also
>allow for file lock leakage due to a shared file descriptor table
>being used without a shared p_leader.
Thanks. Is there a work around where the lock leakage can be fixed and the
3ware daemon can work ? I dont know who (if anyone) has the source code to
the 3dmd program, is this something that could be easily fixed in it ? Are
there any other programs that will break as a result of your commit as well ?
---Mike
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5.2.0.9.0.20030409142819.07c59c60>