Howdy, I installed Nevada build 49 in a parallels VM on my intel mac mini, and when I run the following DTrace script: $ cat follow.d #pragma option D flowindent; pid$target:::entry, pid$target:::return {} $ cat format.cmds disk 0 exit $ dtrace -s follow.d -o follow.out -c "/usr/sbin/format -f format.cmds" dtrace: script ''follow.d'' matched 13641 probes The system hangs, and the following error is written to the console and /var/adm/messages: Dec 6 20:10:34 sparky unix: [ID 988528 kern.warning] WARNING: Unexpected INT 1 in user mode, dr6=ffff0ff0 Is anyone able to reproduce this? Any thoughts what might be causing it? Thanks, - Ryan -- UNIX Administrator http://prefetch.net
Adam Leventhal
2006-Dec-07 05:29 UTC
[dtrace-discuss] System hangs when invoking DTrace script
Hey Ryan, This might be an issue with the Parallels emulation environment. Do you have any more information? Can you boot into kmdb and get a dump while it''s hung? Adam On Wed, Dec 06, 2006 at 10:24:31PM -0500, Matty wrote:> Howdy, > > I installed Nevada build 49 in a parallels VM on my intel mac mini, > and when I run the following DTrace script: > > $ cat follow.d > #pragma option D flowindent; > > pid$target:::entry, > pid$target:::return > {} > > $ cat format.cmds > disk > 0 > exit > > $ dtrace -s follow.d -o follow.out -c "/usr/sbin/format -f format.cmds" > dtrace: script ''follow.d'' matched 13641 probes > > The system hangs, and the following error is written to the console > and /var/adm/messages: > > Dec 6 20:10:34 sparky unix: [ID 988528 kern.warning] WARNING: > Unexpected INT 1 in user mode, dr6=ffff0ff0 > > Is anyone able to reproduce this? Any thoughts what might be causing it? > > Thanks, > - Ryan > -- > UNIX Administrator > http://prefetch.net > _______________________________________________ > dtrace-discuss mailing list > dtrace-discuss at opensolaris.org-- Adam Leventhal, Solaris Kernel Development http://blogs.sun.com/ahl
I read something somewhere (helpful, I know) about Mac and/or something on Mac having "left debug interrupts in their BIOS". This could be another case of that; debug code left behind, harmless during normal execution, but triggers a breakpoint when running under a debugger or something else that traps breakpoints". I''ve thought that maybe there''s some opportunity for hardening the kernel and debuggers to simply ignore unexpected INT1/INT3 breakpoints, in an attempt to cope with such ill-advised released software. Adam Leventhal wrote:> Hey Ryan, > > This might be an issue with the Parallels emulation environment. Do you have > any more information? Can you boot into kmdb and get a dump while it''s hung? > > Adam > > On Wed, Dec 06, 2006 at 10:24:31PM -0500, Matty wrote: >> Howdy, >> >> I installed Nevada build 49 in a parallels VM on my intel mac mini, >> and when I run the following DTrace script: >> >> $ cat follow.d >> #pragma option D flowindent; >> >> pid$target:::entry, >> pid$target:::return >> {} >> >> $ cat format.cmds >> disk >> 0 >> exit >> >> $ dtrace -s follow.d -o follow.out -c "/usr/sbin/format -f format.cmds" >> dtrace: script ''follow.d'' matched 13641 probes >> >> The system hangs, and the following error is written to the console >> and /var/adm/messages: >> >> Dec 6 20:10:34 sparky unix: [ID 988528 kern.warning] WARNING: >> Unexpected INT 1 in user mode, dr6=ffff0ff0 >> >> Is anyone able to reproduce this? Any thoughts what might be causing it? >> >> Thanks, >> - Ryan >> -- >> UNIX Administrator >> http://prefetch.net >> _______________________________________________ >> dtrace-discuss mailing list >> dtrace-discuss at opensolaris.org >